To determine tissue specificity in CESC differentiation we p

To determine tissue specificity in CESC differentiation, we performed chondrogenic and osteogenic induction under normoxia and hypoxia, respectively. Our study revealed tissue specificity indicating that hypoxia promoted chondrogenesis but inhibited osteogenesis in CESCs. This tissue specificity of CESC differentiation in response to physiological hypoxia contributed to our understanding of the mechanism of CEP degeneration in the context of CESCs. Due to the avascular nature, in CEP the oxygen tension was as low as 1% (Lee et al., 2007). As previous studies described, the CEP was free of blood vessels in groups of healthy juveniles and adolescents without a history of LBP. In adults and seniors with DDD, there were defects in the CEP whereby blood vessel invasion could be observed (Nerlich et al., 2007). Blood vessel invasion was also noted in painful human IVD (Freemont et al., 2002). In osteoarthritis, osteochondral angiogenesis within subchondral spaces was often caused by the invasion of neovessels that increased the oxygen level (Walsh et al., 2010). It is reasonable to assert that, in degenerated CEP, blood vessel invasion through small fissures accompanied by an increased oxygen tension in the microenvironment could disrupt the physiological hypoxic microenvironment of CESCs. This destruction could inhibit CESCs chondrogenic differentiation and facilitate osteogenic differentiation, which could initiate the loss of chondrification and the acquisition of ossification in CEP. The ossification of CEP resulted in a poor capacity to resist mechanical stress and poor nutrient exchange; complete IVD degeneration was thus initiated. In a study of chondro-osteogenic differentiation in CESCs, we observed a critical role of HIF1A in regulating the direction of differentiation. HIF1A was a key cellular regulator in responding to a low oxygen level (Coimbra et al., 2004). A number of cartilage-related and osteogenesis-related Calcium Ionophore I manufacturer were under transcriptional control by HIF1A (Robins et al., 2005; Wagegg et al., 2012). Given this background, we mimicked “chemical hypoxia” under normoxia via DMOG upregulating the expression of HIF1A and “chemical normoxia” under hypoxia by YC1 downregulating the expression of HIF1A. We observed that the promotion of chondrogenesis and the inhibition of osteogenesis were directly correlated with the expression of HIF1A in CESCs. The mode of action of MIF in previous studies can be encapsulated in five aspects: (1) MIF receptor, (2) activation of ERK1/ERK2, (3) upregulation of TLR4 expression, (4) suppression of p53 activity, and (5) inhibition of JAB1 activity (Calandra and Roger, 2003; Lue et al., 2002). We next investigated the presence of other pathways. Unexpectedly, in CESCs an increased nuclear MIF expression under hypoxia was observed. As previous studies described, a high nuclear expression of MIF could also be observed in the tissues of lung adenocarcinoma, glioblastoma multiforme, pituitary adenoma, and aggressive bladder cancer (Kamimura et al., 2000; Markert et al., 2001; Pyle et al., 2003; Taylor et al., 2007); the same was observed in our study using CESCs. Tumor tissue is very hypoxic, and we speculated that increased nuclear MIF expression may occur in this hypoxic microenvironment. MIF may not act in the traditional manner but may play a biological role in the nucleus. After the analysis of the three-dimensional structure of MIF, we considered that MIF might lack a typical DNA-binding motif. In this study, ChIP showed the potential interaction between MIF and the promoter regions of SOX9 and RUNX2, leading us to conclude that MIF could bind to DNA indirectly via intermediate proteins. From the result of the luciferase report, we observed that MIF could participate in the regulation of transcription. Based on the above information, we think that MIF could act as a “transcriptional regulator” that participates in the regulation of transcription indirectly, but not as a “transcription factor” that functions via direct binding to DNA sequences. A limitation of this study is that the intermediate mechanism has not been fully investigated. Further studies are required to illuminate this mechanism.