We analyzed whether DLK PREF that was firstly identified as

We analyzed whether DLK1(PREF1), that was firstly identified as a putative marker for murine preadipocytes (Smas and Sul, 1993), could be used as additional marker to characterize human ASC. We found that DLK1(PREF1) is highly expressed in human ASC. The DLK1(PREF1) protein levels were similar in human ASC proliferating in PM4 medium and in density-arrested G0/G1-phase ASC. In the course of adipogenic differentiation the DLK1(PREF1) protein level declined. DLK1(PREF1) was almost undetectable in adipocytes 14days after the induction of adipocyte differentiation. Thus DLK1(PREF1) is expressed in proliferating ASC, in density-arrested ASC, and most likely at early stages of adipogenic differentiation but completely downregulated in human adipocytes. During adipogenic differentiation of MEFs (mouse embryonic fibroblasts) the DLK1(PREF1) protein levels are relatively low in density arrest before differentiation, transiently increased at days 1 and 2, and then decreased and were undetectable at day 5 after conversion into adipocytes (Wang and Sul, 2009). Similar in adipogenesis of the preadipocyte cell line NIH 3 T3 L1 the DLK1(PREF1) protein levels are low in density arrest before differentiation and transiently increase until day 2–3 (Garces et al., 1999). When these ion channels cease clonal expansion and enter terminal adipogenic differentiation (Farmer, 2006), the DLK1(PREF1) protein levels decrease but membranous forms of DLK1(PREF1) are detectable until day 5 (Garces et al., 1999). Thus certain forms of DLK1(PREF1) protein are still present at early stages of terminal adipogenic differentiation and may contribute to adipocyte differentiation (Garces et al., 1999). DLK1(PREF1) has been identified as a negative regulator of adipocyte differentiation in mouse preadipocytes (Smas and Sul, 1993). Moreover, in its active form, DLK1(PREF1) was shown to function in an autocrine and paracrine manner as soluble factor to inhibit differentiation of murine multipotent MSC into adipocytes, osteoblasts and chondrocytes (Wang and Sul, 2009; Wang et al., 2010). This suggests that DLK1(PREF1) has a general role in keeping mouse precursor cells in an undifferentiated state. In the present study we found that DLK1(PREF1) functions as negative regulator of adipogenesis in human ASC. Previous studies have identified DLK1(PREF1) as regulator of differentiation in human BM MSC (Abdallah et al., 2004; Jing et al., 2009) and CB MSC (Kluth et al., 2010). DLK1(PREF1) was used as a marker to distinguish human unrestricted somatic stem cells and CB MSC (Kluth et al., 2010). Moreover, DLK1(PREF1) was employed as surface marker for the isolation and differentiation of chondrogenic cells derived from human embryonic stem cells (Harkness et al., 2009). DLK1(PREF1) is also involved in additional differentiation processes in rodent cells, including haematopoiesis (Moore et al., 1997), pancreatic islet cell differentiation (Carlsson et al., 1997), Schwann cell differentiation (Costaglioli et al., 2001), hepatic cell differentiation (Tanimizu et al., 2004), and differentiation of muscle satellite cells. Moreover, DLK1(PREF1) is expressed in thymocytes and neuroblastoma cells [reviewed in Sul, 2009]. Thus DLK1(PREF1) should be used in combination with additional markers to distinguish ASC from MSC and other cell types, especially in combination with the CD34+ and CD31− immunophenotype, as considered earlier (Gesta et al., 2007). Studies in mice suggest that the age, depot site and sex of the donor can influence the features and functionality of the derived ASC (Cartwright et al., 2010; Gesta et al., 2006; Majka et al., 2010). Clinical studies found a correlation between donor age and differentiation capacity of the given ASC (Madonna et al., 2011; Schipper et al., 2008; Zhu et al., 2009). In humans also an influence of the BMI (van Harmelen et al., 2003) and fat depot origin (Tchkonia et al., 2002, 2007) on ASC features was shown. In the present study, we isolated ASC from subcutaneous adipose tissue pads of four different women with an age between 30 and 47years and a BMI between 22 and 27kg/m2. No differences in the analyzed cell surface marker patterns were found between the ASC from the abdominal subcutaneous depots of the different donors. Moreover, all ASC populations had a high capacity to differentiate into adipocytes in vitro, as demonstrated by the formation of high levels of four terminal adipogenic differentiation products, FABP4, adiponectin, leptin and triglycerides. This suggests that the relatively small differences in age and BMI between our four donors did not account for major differences in these ASC features. We detected however differences in the strength of the adipogenic maturation at the level of single ASC isolated from the same depot. This was reflected by the difference in the quantity and size of lipid droplets formed 20days after the induction of adipogenesis (Fig. 2H, compare white and yellow arrows). These findings are in keeping with the differences in the single cell FABP4 protein level (Fig. 2A, compare white and yellow arrows). These data suggest that there may be heterogeneity regarding the adipogenic capacity of ASC within the same depots. Almost all passage 5 ASC stained strongly positive for the marker composition DLK1(PREF1)+/CD105+/CD90+/CD34+/CD31−/FABP4−. However, this marker combination cannot distinguish ASC with high or low adipogenic capacity. Evidence for differences between ASC populations cloned from a single human subcutaneous fat depot regarding the differentiation capacity was previously shown (Kirkland et al., 1993; Tchkonia et al., 2005; Zimmerlin et al., 2010). Common observations in our study were islets of strongly differentiated ASC surrounded by areas of weaker differentiated ASC. This could, for example, result from differences in the autocrine/paracrine micromilieu caused by a different cell density. Such findings could mirror the existence of different subtypes of ASC in subcutaneous fat depots in vivo, as discussed previously (Gesta et al., 2007). More studies are necessary to better understand these observations.