Our results demonstrate that rat ADSC could be converted into neurospheres using a procedure similar to the one used for propagation of genuine neural stem cells. In addition to generating neuronal- and glial-like cells, neurosphere cells from rat ADSC could differentiate into SC-like cells. We showed further that SC-like cells were functional since these cells could secrete soluble factors and form myelin structures with neuronal neurites. Functional properties, especially formation of myelin structures with neuronal neurites, further indicated that SC-like cells from rat ADSC were closely similar to genuine SC.
Primary cultures of adipose tissue are heterogeneous, containing hepatopoietic cells, endothelial cells, smooth muscle cells and pericytes [9
]. However, the number of these other cells is small, and the frequency of these other cells will diminish quickly through serial passages [21
]. Also, ADSC can differentiate into several mesenchymal tissue lineages, including adipocyte and osteoblast. Rat ADSC within 3–5 passages we used in our experiment were CD29 and CD44 positive, CD31, CD106, CD184, CD34 and CD45 negative, and could undergo osteogenic and adipogenic differentiation. All these characteristics of rat ADSC in our experiments are consistent with previous reports [10
Although ADSC and MSCs share many common biological characteristics, the two populations are not identical [12
]. Immunocytochemical analysis shows that surface epitope profiles of the two populations are different [10
]; although it is well established that both MSCs and ADSC can undergo chondrogenic differentiation [22
], Kang et al show that ADSC but not MSCs could undergo chondrogenic differentiation under the conditions used in their study; ADSC may have significantly higher neural differentiation capacities than those of MSCs [12
]; the distinctions between the two populations may also extend down to the gene level [10
]. Although MSCs can be induced to differentiate along SC lineage [7
], the differences between the two populations mentioned above suggest that whether ADSC could be induced to differentiate along SC lineage needs to be confirmed.
Some recent studies show that neural crest stem cells can be harvested by means of neurosphere method from various seemingly "mesodermal" tissues of adult animals, such as heart [23
] and hair follicular dermal papilla [24
]. Kang et al. show that ADSC can be converted into neurospheres [12
], and a preliminary report further suggests that neurosphere cells derived from ADSC may have neural crest-like properties [13
]. In our experiment, since neurospheres converted from rat ADSC could differentiate into SC-like cells which belong to peripheral nervous system, these neurospheres should have the characteristics of peripheral nervous system.
Recently, nestin expression has also been observed in myogenic cells, hepatic cells and endothelial cells, which indicates that nestin may not be used as a specific marker for neural stem cells. However, in our experiment, neurosphere cells from rat ADSC can be induced into neuronal- and glial-like cells, which strongly indicates that neurosphere cells derived from rat ADSC have neural stem cell-like properties. Hermann et al suggest that neural stem cell-like cells converted from MSCs are real neural stem cells [25
]. The immature neural stem cells would be more suitable for the treatment of neurodegenerative diseases than fully differentiated neural cells, because fully differentiated neurons can not survive detachment and subsequent transplantation procedures [26
Neural stem cells from CNS can be maintained in an undifferentiated status by bFGF and EGF [27
]. When exposed to RA, neural stem cells will exit from cell cycle and differentiate into nerve cells [29
]. In our experiment, neurosphere cells will differentiate in the presence of RA and in the absence of EGF and bFGF.
FSK can elevate the level of intracellular cyclic adenosine monophosphate (cAMP) and cAMP signal may be an intracellular signal during several different stages of SC development. In cultured SC, cAMP elevation can mimick SC responses in the presence of axons during myelination in vivo [30
]. In addititon, FSK can enhance the responsiveness of SC to SC mitogens, such as PDGF-BB and glial growth factor [31
]. PDGF-BB can induce SC proliferation in the presence of serum and FSK [32
]. Heregulin is a subtype of neuregulin-1 and neuregulin-1 is now regarded as the pivotal signal that controls SC at every stage of the lineage [33
]. Neuregulin-1 type II, also known as glial growth factor, can induce instructively cultured neural crest cells into SC [34
]. In the presence of Heregulin (neuregulin-1 type I), MSCs can be induced into SC-like cells [35
]. A mixture of cytokines mentioned above may synergize to induce neurosphere cells into SC-like cells.
SC can produce a number of neurotrophic factors, and a combination of these and other SC-derived soluble factors have been referred to as 'anti-neuroblastoma' agents [36
]. Pigment epithelium-derived factor is now regarded as the key factor responsible for SC's ability to induce the differentiation of SH-SY5Y cells [19
]. It is likely that SC-like cells from rat ADSC in our experiment produced at least some 'anti-neuroblastoma' agents produced by genuine SC since SC-like cells could induce the differentiation of SH-SY5Y cells efficiently. SC-like cells induced from MSCs can cause neurite growth of dorsal root ganglion neurons in vitro [7
], which supports that SC-like cells from rat ADSC may produce some soluble factors. These SC-derived factors, such as pigment epithelium-derived factor, can promote survival and neurite outgrowth of neurons. SC-like cells from ADSC may be useful for the treatment of diseases in peripheral nervous system (e.g., nerve injuries) and CNS (e.g., multiple sclerosis).