For the first time we showed that a secreted

For the first time, we showed that a secreted form of human ICAM-4 (sICAM-4) is made from human erythroblasts. Similarly, erythroblasts derived from murine Friend virus anemia (FVA) tachykinin receptor are known to secrete ICAM-4 (Ihanus et al., 2007; Lee et al., 2003). As the erythroblasts mature, they reside in areas peripheral to macrophages within the erythroblastic island (Bain et al., 1996), and the direct effects of macrophages decrease and the effects of adjacent erythroid cells increase. We therefore speculated that the addition of the sICAM-4 could mimic co-stimulation among erythroblasts. The addition of recombinant ICAM-4 enhanced cell differentiation to RBCs by decreasing abnormal cytokinesis in vitro. Such abnormal cytokinesis is known as nuclear dysplasia, and occurs when the nucleus does not properly divide to form daughter cells. These dysplastic cells show multi-nucleation states, fragmented nuclei, or irregular nuclear contours. Erythroid dysplasia leads to inefficient enucleation and cell death in vivo and in vitro. The role of DLC-1 in cytokinesis and co-binding with ICAM-4 suggests that the effects induced by ICAM-4 are related to DLC-1. This relationship should be further evaluated in culture systems and in pathologic diseases in vivo. The elucidation of the role of sICAM-4 is also important for practical purposes. The progressive loss of viable cells during successive stages of cell culture is well known, especially in terminally matured erythroblasts, resulting in inefficient generation of RBCs. Rigorous attempts have been made to create a platform for efficient erythrocyte production, but these attempts have met with failure. In our culture system, with its absence of supporting stromal cells and serum/plasma, a supraoptimal density led to significantly enhanced cell survival, enucleation, and finally, to a higher amount of generated RBCs, indicating that erythroblast-erythroblast contact may mimic in vivo conditions (Chasis, 2006). Even though these erythroblast contacts do not equal the most powerful effects of macrophages on erythropoiesis, the availability of sICAM-4 for in vitro terminal erythropoiesis is valuable. This clearly suggests that our culture system, with its addition of sICAM-4, could be used for mass RBC production using a bioreactor by reducing culture area and enhancing culture efficiency. Culturing precursor cells of hematopoietic origin at a high density induces a “crowding” effect, which results in a general decline in metabolic activity (Sand et al., 1977). Additional evidence demonstrates that inhibitory cytokines can accumulate in cultures through lactate production (Sand et al., 1977). Therefore, we analyzed conditioned media to investigate the effects of culture density on cell metabolism, and did not observe any variations in metabolic activity or cellular homeostasis.
Contributors
Conflict of interest statements
Acknowledgments This study was supported by a grant from the Korean Health Technology R&D Project, Ministry of Health and Welfare, Republic of Korea (A101782-200).
Introduction The 3-dimensional structure of the brain is integral in development (Georgieva et al., 2009). It has been shown that neural networks display optimal information processing when they achieve a complexity threshold (Song and Abbott, 2001). The development of mature and functional chemical synapses in the CNS requires the presence of gap-junctions, which are highly expressed in complex CNS tissue (Todd et al., 2010). In addition, the formation of functionally mature synapses is highly dependent on glutamate release from functional presynaptic terminals, and readily occurs in slice culture preparations, which display a 3-dimensional structure and electrical activity (Cai et al., 2004). Although glutamate-induced synapse formation has been observed in monolayer preparations, these synapses are not as electrically robust, and glutamate release has to be induced by non-physiological chemical agents. Three-dimensional systems, such as SFEBs have been shown to successfully recapitulate cortical development (Eiraku et al., 2008), but there are some drawbacks to their use that parallel those observed in acute slices taken from brain tissue. For example, SFEBs grown using typical methods form thick, dense structures that are difficult to live image or obtain electrophysiological recordings from, and require dissociation for efficient analysis. This undermines many of the advantages of a 3-D system in terms of cell–cell contact and network formation.