Although the detailed mechanisms underlying the increase
Although the detailed mechanisms underlying the increase in PDX1+NKX6.1+ cell induction in aggregation cultures remain unknown, this finding raises the possibility that pancreatic bud formation may be regulated by the molecules associated with contact-dependent cellular interactions, such as extracellular matrices and adhesion molecules, and other membrane-bound receptors (Brafman et al., 2013; Crisera et al., 2000; Ekblom et al., 1998). Possible mechanisms include the involvement of laminin–integrin signaling. During pancreatic development, epithelial cell delamination and migration into mesenchyme is the early event in the differentiation process into endocrine buy CMX001 (Villasenor et al., 2012). However, we found that the expression levels of pan-laminin were similar between NKX6.1+ cells and NKX6.1− cells in both aggregation and monolayer cultures (data not shown). Although we could not exclude the possibility that laminin isoforms are different between the cell populations, we assumed that typical laminin–integrin signals are not likely to be directly involved. In addition, based on the findings obtained with the other extracellular matrices (Fig. S1), at a minimum, laminin-111, vitronectin (main component of Synthemax), fibronectin and collagen I are not likely to be involved. Instead, other cell surface-derived signals, such as those associated with the apicobasal polarity (Villasenor et al., 2010; Kesavan et al., 2009) and planar cell polarity pathways (Cortijo et al., 2012), or unknown autocrine/paracrine molecules, may be involved. Further studies with comparative analyses of the gene expression profiles between 2D and aggregation cultures would be helpful to elucidate the factors responsible for the regulation of NKX6.1 induction in developing pancreatic buds. It is possible that the signals elicited by a high cell density increased the ratio of PDX1+NKX6.1+ cells via the direct induction of differentiation into PDX1+NKX6.1+ cells, while the selective proliferation of progenitor cells and/or selective apoptosis of uncommitted cells are also possible mechanisms. The number of proliferating cells was low on days 1–2 in Stage 4, during which time the PDX1+NKX6.1+ cells emerged in the cell aggregation cultures (Figs. S5A and C). In addition, the total cell number in the aggregation cultures was not increased at Stage 4, days 1–2, suggesting the involvement of selective proliferation to be less likely (Fig. S3). On the other hand, the number of apoptotic cells was increased in the cellular aggregates, consistent with the idea that aggregation cultures selectively delete cells uncommitted to the PDX1+NKX6.1+ cell lineage (Figs. S5B and D). In fact, more apoptotic cells were found within the PDX1− cell population (Fig. S5D), which suggests that selective apoptosis is a possible mechanism. In the future, the identification of marker genes for pancreatic cells uncommitted to NKX6.1+ cells and application of subsequent live cell imaging with the reporter cell line of the marker would help to clearly demonstrate the involvement of selective apoptosis in the increased PDX1+NKX6.1+ cell ratio observed in high-density cultures. It was previously reported that the activation of BMP signaling facilitates the proliferation of NKX6.1+ cells (Sui et al., 2013). In contrast, our data indicate that the inhibition by adding NOGGIN increased the NKX6.1 expression (Fig. 5A). Consistent with our findings, other reports have demonstrated that NKX6.1+ pancreatic progenitors were generated under continuous suppression of BMP signaling in hESC differentiation cultures (Rezania et al., 2012, 2013; Schulz et al., 2012; Bruin et al., 2013), and that the phosphorylation of smad1/5/8, downstream effectors of BMP signaling, was suppressed in Nkx6.1+ cells from developing chick pancreatic buds (Ahnfelt-Ronne et al., 2010). In addition, the BMP inhibition by in ovo gene transfer of Noggin into the presumptive pancreatic region of chick embryos resulted in pancreatic bud formation without further branching and epithelialization (Ahnfelt-Ronne et al., 2010). We reasoned that BMP inhibition is involved in the commitment to NKX6.1+ cells, but that BMP signaling is necessary for the subsequent proliferation and differentiation. In agreement with this, we found that the induction rate of PDX1+NKX6.1+ cells was continuously increased with the removal of NOGGIN after day 4 of Stage 4 (Fig. 5B). Although BMP activation may expand NKX6.1+ cells, we assume that the first four days of aggregation cultures are critical for the commitment of PDX1+NKX6.1+ cells, and that all three factors are beneficial for the PDX1+NKX6.1+ cell induction.