Loss of the osteogenic differentiation potential during senescence is limited to bone progenitor cells and is dependent on p53

Abstract

DNA damage can lead to the induction of cellular senescence. In particular, we showed that exposure to ionizing radiation (IR) leads to the senescence of bone marrow-derived multipotent stromal cells (MSC) and osteoblast-like stromal cells (OB-SC), a phenotype associated with bone loss. The mechanism by which IR leads to bone dysfunction is not fully understood. One possibility involves that DNA damage-induced senescence limits the regeneration of bone progenitor cells. Another possibility entails that bone dysfunction arises from the inability of accumulating senescent cells to fulfill their physiological function. Indeed, we show here that exposure to IR prevented the differentiation and mineralization functions of MSC, an effect we found was limited to this population as more differentiated OB-SC could still form mineralize nodules. This is in contrast to adipogenesis, which was inhibited in both IR-induced senescent MSC and 3T3-L1 pre-adipocytes. Furthermore, we demonstrate that IR-induced loss of osteogenic potential in MSC was p53-dependent, a phenotype that correlates with the inability to upregulate key osteogenic transcription factors. These results are the first to demonstrate that senescence impacts osteogenesis in a cell type dependent manner and suggest that the accumulation of senescent osteoblasts is unlikely to significantly contribute to bone dysfunction in a cell autonomous manner.

Figure 1. Senescence of multipotent and committed stromal lineages following exposure to IR.

(A) Murine bone marrow-derived multipotent stromal cells (MSC), osteoblasts (OB–SC) and pre-adiopocytes (3T3-L1) were exposed (IR) or not (CTRL) to 10 Gy IR and 7 days later stained for the expression of the senescence-associated β-galactosidase (SAβ-gal). (B) Quantification of the proportion of SAβ-gal positive cells in each population. (C) Sustained activation of the DNA damage response in stromal populations was measured by staining for the presence of 53BP1 DNA damage foci (in red) one week post exposure to IR. Nuclei were counterstained with DAPI. (D) The proliferation capacity of MSC, OB–SC and 3T3-L1 cell population was determined using a CFU assay one week post-exposure or not to IR. Mean ± standard error of at least three individual experiments is shown. p values were obtained by performing a Student’s t-test.

Figure 2. Exposure to IR abrogates adipogenesis independently of the stromal lineage potential.

(A) MSC and 3T3-L1 cell populations were exposed (IR) or not (CTRL) to 10 Gy IR and one week later placed in adipogenic differentiation media. Representative photographs showing lipid accumulation stained with Oil Red O is shown for each population. (B) Quantification of lipid accumulation was determined by the extraction of Oil Red O staining and detection by spectrophotometry. (C–D) Expression of PPARγ was determined by quantitative real-time PCR using RNA extracted from control and IR-induced senescent MSC and 3T3-L1 populations cultured or not in adipogenic differentiation media. Mean ± standard error of at least three individual experiments is shown. p values were obtained by performing a Student’s t-test. *: p value < 0.05.

Figure 3. Abrogation of osteogenic differentiation potential following irradiation is limited to stromal progenitor cells.

(A) MSC and osteoblasts (OB–SC) were exposed (IR) or not (CTRL) to 10 Gy IR and one week later placed in osteogenic differentiation media for 14 to 21 days. Representative photographs showing mineralization nodules accumulation stained with Alizarin Red S is shown for each population. Scale bar: 2mm. Phase contrast photograph showing the presence of senescent MSC in absence of mineralization is also shown. (B) Quantification of mineralization was determined by the extraction of Alizarin Red S and detection by spectrophotometry. (C and D) Expression of Runx2 and Osx was determined by quantitative real-time PCR using RNA extracted from control and IR-induced senescent MSC and OB–SC populations cultured or not in osteogenic differentiation media. Mean ± standard error; *: p value < 0.05.

Figure 4. IR-induced senescent MSC failed to generate bone in vivo.

(A) Schematic of the experiment. Control or IR-induced senescent MSC were mixed with HA/TCP particles along with collagen and injected subcutaneously to the flank of mice. 10 weeks post injection, implants were retrieved from the animals, embedded in plastic, sectioned and stained with Goldner’s trichrome to detect bone formation. (B) Representative images from n= 6 implants per group showing mineralization (Goldner’s trichrome in green) from control or IR-induced senescent MSC. Implants were counterstained with hematoxylin eosin. Scale bar: 300µm.

Figure 5. Loss of osteogenic but not adipogenic potential in senescent MSC is p53 dependent.

(A) MSC derived from p53 knockout mice (MSC-p53KO) were exposed (IR) or not (CTRL) to 10 Gy IR and one week later stained for the expression of SAβ-gal activity. (B) The proliferation capacity of MSC-p53KO was determined using a CFU assay one week post-exposure or not to IR. (C) One week post exposure or not to IR, MSC-p53KO were placed in adipogenic differentiation media for 7 to 14 days. Representative photographs showing lipid accumulation stained with Oil Red O is shown. Scale bar: 200µm. (D) Quantification of lipid accumulation has determined by the extraction of Oil Red O staining and detection by spectrophotometry. (E) Expression of PPARγ was determined by quantitative real-time PCR using RNA extracted from control and IR-induced senescent MSC-p53KO cultured in adipogenic differentiation media. (F) One week post exposure or not to IR, MSC-p53KO were placed in osteogenic differentiation media for 14 to 21 days. Representative photographs showing mineralization nodules accumulation stained with Alizarin Red S is shown. (G) Quantification of mineralization was determined by the extraction of Alizarin Red S staining and detection by spectrophotometry. (H) Expression of Runx2 and Osx was determined by quantitative real-time PCR using RNA extracted from control and IR-induced senescent MSC-p53KO populations placed in osteogenic differentiation media. Mean ± standard error of at least 3 individual experiments is shown; *: p value < 0.05.