PDB-MSCs also share some properties of pluripotent embryonic stem cells as well as other properties of multipotent stem cells [ 16 ].
Recently, urine-derived stem cells USCs which are isolated from urine have been studied as a promising candidate for many tissue engineering therapies due to their multilineage differentiation properties into osteocytes, chondrocytes, adipocytes, neurocyte, myocytes, and endothelial cells and sufficient proliferation activities [ 13 , 25 , 26 ]. Advantages to the use of USCs include noninvasive and low-cost harvesting as well as being considered for ethical use.
Additionally, USCs have been isolated from autologous urine which do not induce immune responses or rejection [ 25 ]. Therefore, USCs are considered to be an attractive alternative source of multipotent stem cells that have been appropriated for a large variety of uses. In this study, we only focus on the differences in proliferation and differentiation potentials of USCs, PDB-MSCs, and BMSCs by comparing their morphologies, immune-phenotypes, proliferation capacities, and differentiation potentials osteogenic, adipogenic, chondrogenic, and endothelial.
Human bone marrow samples were obtained from six patients age from 45 to 65 years old who underwent a total hip replacement at the orthopedic department of the West China Hospital after providing written informed consent. BMSCs were isolated using the method outlined in our previous report [ 27 ]. The 4th passage and 10th passage cells were used in the morphologic analysis, and remaining cells from the 4th passage were used in other assays.
Human placenta samples were obtained from three healthy donor mothers age 28 to 33 years after providing written informed consent. Briefly, decidua basalis was collected and washed in PBS to remove residual blood. The samples were then mechanically minced into small particles and digested with 0. The 4th passage and 10th passage cells were used in the morphologic analysis and remaining cells from the 4th passage were used in other assays.
We obtained human urine samples from five healthy male adult donors age from 24 to 30 years old after receiving written informed consent.
USCs were isolated following the protocol laid out in our previous report [ 25 ]. To determine the total number of living cells shed into the urine, cells were stained with trypan blue and counted. Colony-forming ability analysis was modified from our previous study [ 15 ]. After 8 days, cells were fixed with methanol and stained with 0. Cell proliferation was assessed according to our previous study method [ 15 ].
Cell viability was monitored on days 0, 1, 3, 5, 7, and 9, respectively. CD31 expression after endothelial differentiation was tested under the same procedure. In order to analyze the multipotency of PB-MSCs, osteogenic, adipogenic, chondrogenic, and endothelial differentiation was performed in specific induction media. Osteogenic and adipogenic differentiation methods were optimized from our previous study [ 15 ].
Red staining was detected by filtering by color thresholds L for Alizarin Red staining in the Lab color space. Fixed thresholds were used for each set of images. The ratio of positive pixels to the total number of pixels per image was quantified. The medium was replaced every 3 days. Red staining was detected using color thresholds L for Oil Red O staining in the Lab color space.
The ratio of positive pixels to the total number of pixels per image was calculated. Chondrogenic differentiation was performed following the protocol described in our previous study [ 25 ]. The medium was replaced every 2—3 days. The medium was then replaced every 3 days. After 8 days, cells were collected for CD31 expression analysis, tube formation assay, and real-time PCR testing. In vitro tube formation assay was conducted in order to further investigate endothelial differentiation function.
A fluorescence microscope Olympus IX50, Japan was used for the analysis of network formation. The number of tubes was counted using the Image-Pro Plus 6. RT-PCR was performed to analyze the endothelium-related gene expression after endothelial induction for 8 days and chondrogenic cell-related gene expression after chondrogenic induction for 21 days.
Each target gene expression was analyzed and compared to the housekeeping gene glyceraldehydephosphate dehydrogenase GAPDH. All data are expressed as. Statistical analysis was performed to use SPSS After culturing for 7 to 10 days, adherent cells from bone marrow, urine, and placenta digestion began to form cell clones. USCs had the highest growth rate from day 7 to day 9. Phenotypic analysis was performed using flow cytometry Figure 3. Under conditioned culture medium for certain days, all the cells were positively stained for Alizarin Red, Oil Red O, and Toluidine Blue, which demonstrated their multilineage differentiation potential Figure 4.
Their differentiation capabilities, however, were different with respect to the positive staining area as evaluated by ImageJ and RT-PCR results. In differentiation analysis of endothelial cells, nontreated groups exhibited negative results with CD31 expression under 0.
CD31 expression of endothelial-induced USCs is the highest among three endothelial-induced groups. In addition, we performed tube formation assay to assess stem cell vascularization potential. Quantitative results demonstrated that USCs had the highest tube number per millimeter square in both treated and nontreated groups Figures 7 a and 7 b.
Lastly, gene expression levels of two endothelial cell markers, CD31 and vWF, were all enhanced in three groups after induction and the expression levels of differentiated USCs were significantly higher than in the other two groups.
Human stem cells harvested from different tissues have distinct features, including cell proliferation, colony formation ability, and differentiation capability.
Thus, it is essential to obtain identifying biophysical information about each cell type in order to optimize experimental and clinical selection of seed cells. Here for the first time, the bio-characteristics of stem cells derived from human bone marrow, placenta decidua basalis, and urine were investigated, specifically cell morphology, proliferation, phenotype, and multilineage differentiation properties.
Although this study did not include other common stem cell candidates, comparison of the most promising or most widely used seed cells including USCs, BMSCs, and PDB-MSCs might provide valuable evidence for stem cell transplantation research. This urothelial-like shape was consistent with recent evidence from other groups [ 13 , 25 ].
The efficiency to form colonies still remains an important assay for the quality of cell preparations [ 31 ]. Despite no previous direct comparison study of these three cell types, previous research predicted our findings as BMSCs seemed to be inferior to other multipotent stem cells regarding the growth kinetics [ 14 ]. Therefore, CFU and proliferation assays allowed comparing stem cell proliferation and stemness capabilities using optimized cell culture conditions.
In other words, the P—D partition is not an artifact of network topology. In fact, it does not depend on any particular network at all. Probability of Obtaining P—D Partition. Loosely aggregated coexpression clusters can be also derived from all genes on the microarray to significantly overlap with P nodes and D 1, nodes modules Experiment 3 in Table 2.
The P—D partitions can actually be detected in almost any network of the same number of nodes Experiment 4 in Table 2 or the same number of edges Experiment 5 in Table 2 as the HPRD network but randomly sampled from an extended PPI network. The extended PPI network contains the updated HPRD plus two yeast two-hybrid interactome maps [ 19 , 20 ] and covers 7, proteins, 3, of which have expression profiles on the Affymetrix U95 array.
Although they correspond to smaller and smaller fractions of the total genes available on the microarray when the PCC cutoff increases, the fraction of genes corresponding to P and D modules maximizes around PCC cutoffs of 0. Furthermore, relationships among the P, D, and N clusters did not change; only the number of genes and interactions varied to some extent without altering most of the enriched functions in each clusters unpublished data. The answer lies in the difference in the stability of detecting these network modules.
Only after we examined the distribution of the anti-correlated interactions, of which a large number between the D and the N modules are evident see below , we decided to further cluster only the genes in the D and N clusters. These suggest that by concentrating on only the correlated and anti-correlated interactions, we enriched for the genes of the P, D, and N clusters. To more rigorously test the stability of finding the P—D partition, we compared the chance of finding anti-correlated network modules with or more nodes in each module and their overlap to the P and D modules with or without extracting NP networks prior to clustering the genes in the network.
These reductions in the probability of finding anti-correlated modules and further reductions in the identification of P and D modules among the input genes or all modules point to a role of using the appropriate PCC cutoff and integrating true PPIs on the stability of P and D module identification.
However, the contribution of integrating the PPI network is not limited to the module identification, but more importantly is linked to the identification of the large PPI interface between the P and D modules that potentially coordinate the cellular proliferation and differentiation processes see below. These controls demonstrate that integrating the interactome, extracting the NP network, and applying an appropriate PCC cutoff ensured a high probability of stably detecting the P and D modules and improved their homogeneity, probably by filtering out most gene pairs that function in irrelevant tissue or cell types or under irrelevant physiological conditions.
The P—D partitions and their transcriptional anti-correlation can also be seen in the fruit fly. We used the adult whole-fly expression profiles to probe the dynamic gene relationships in the network. Here, we used these profiles to extract the network modules based on anti-correlated and correlated interactions across different fly populations using the automated analysis pipeline described in Figure 1.
While the composition of the P module is largely conserved between the human brain and the fly, that of the D module is quite different between the two species. In particular, the apoptosis pathways are only enriched in the human brain D module. The enriched differentiation markers in D modules are also different; in the human brain, there are the neuronal markers; in the fly, there are genes involved in eye development unpublished data , which is consistent with their tissue- and organism-specific requirements for differentiation.
We examined the percentages of human gene orthologs that can be found in yeast, worm, fly, and mouse. Similar evolutionary patterns can be seen for fly P and D modules Figure 2 D.
The above observations indicate that the P module is more conserved from the single-cellular organism yeast to the multicellular organisms C. The conservation of the P—D partition, their relationship, and the similar evolutionary profiles between fly and human also indicate that these observations cannot be due to sample variations introduced by sample preparations or other technical factors, but instead reflect true biological features of the gene networks of different multicellular organisms.
A switch between differentiation and proliferation has been demonstrated in myoblast C2C12 cells [ 1 ]. Inhibition of the P—D interface protein HDAC4 has been shown to promote differentiation and inhibit proliferation, whereas inhibition to another interface protein, SRF, does the reverse [ 1 ]. These findings and many others collectively point to the existence of the switch between proliferation and differentiation at the cellular level.
Indeed, we found a decrease of P expression and an increase of D expression when fly, rat, mouse, or human cells of various cell types are switched from the proliferation to the differentiation state upon induction by various external stimuli. In this analysis, we used previously published data on human endometrial stromal cell differentiation induced by cyclic AMP, mouse C2C12 myoblast differentiation upon shifting to differentiation medium, mouse smooth muscle cell differentiation induced by retinoic acid, the inhibition of proliferation and induction of differentiation by the FGF of rat chondrocytes, and fruit fly neural progenitor cell differentiation detailed sample information is available in Table S2.
Consistent with the conservativeness of the P module, P is more uniformly suppressed upon differentiation of various different tissues in various different organisms. For example, the expression of fly P genes, especially of those derived under diet restrictions or low-food conditions, is suppressed in all cell types Figure 3 A, and middle and bottom rows in Figure 3 B. In contrast, the expression of the human brain D is strongly induced during human endometrial stromal cell differentiation, and less so during mouse and fly cell differentiation; the expression of fly D genes is only most strongly induced in fly cells, but less so in cells of other organisms Figure 3.
A The suppression of P and induction of D expression upon cellular differentiation. The expression levels of genes in the human brain P, D, N, and S modules and fly P and D modules under high-food HF or low-food LF conditions listed in column headers are compared between the undifferentiated and differentiated samples listed in row headers by paired Student t -test.
Red and green colors indicate an increase and a decrease in the differentiated samples, respectively. The color intensity represents the —log p- value of the Student t test between the undifferentiated and differentiated samples. See text and Table S2 for details about the cell lines and differentiation conditions. B The average expression level of P and D genes during the time course of the cell differentiation process of the human endometrial stroma cell left column plots , the mouse myoblast middle column plots , and the rat chondrocyte right column plots.
The plots in the top row are the average expression of human brain P and D genes left column or their mouse middle column or rat right column homologs upon induction of differentiation; plots in the middle and bottom rows are those of the human, mouse, and rat homologs of the fly genes in P and D modules derived under high- and low-food conditions, respectively.
The expression levels of P and D modules are indicated by the green and lavender lines, respectively. The suppression of P and induction of D can occur during a short window. The other experiments presented in panel A but not here include only single timepoints. One way the systems level controls are achieved might be through circulating hormones and growth factors, as many of them and their downstream regulation molecules are present at the P—D interface.
In addition to facilitating the module detection, integrating the interactome and the transcriptome also revealed a large number of PPIs between a limited number of proteins forming a PPI interface between P and D modules. As expected, anti-correlated interactions preferentially bridge between the transcriptionally anti-correlated P and D modules.
The probability of the anti-correlated interactions bridging the P and D modules compared with bridging any modules is 1. From the GO terms overrepresented at the interfaces, the P interface is enriched for transcription factors and the D interface is enriched in cell-cycle checkpoint, DNA repair, and receptor signaling genes Table 3.
We therefore compared the protein interaction degrees, the percentage of oncogenes and tumor suppressor genes, and the percentage of genes in the feedback loops between the interface and the non-interface, or core genes. Even though all the feedback loops are still of very limited coverage, it is already evident that most of these feedback controls are between the P and D modules and mediated by anti-correlated interactions Figure 4 D. Nearly all the proteins involved in these feedback loops are transcription regulators, and many of the loops are formed by both PPI and transcriptional regulations Figure 4 D.
A The average degree k of the nodes at the human brain and fly P—D protein interaction interfaces and inside the modules cores. B The percentage of proto-oncogenes and tumor suppressor genes at the human brain P—D protein interaction interfaces and inside the modules cores.
C The percentage of genes located in feedback loops at the human brain P—D protein interaction interfaces and inside the modules cores. D Network consisting of only the feedback loops traversing protein interaction interfaces and inside the modules cores.
Solid edges represent directional protein interactions; dashed edges, transcriptional regulations. Red and green edges represent transcriptional correlations and anti-correlations, respectively. Feedback loops of three or more nodes that traverse the P—D interface are listed.
Pathways 5, 8, 9, and 10 are related; 6 and 7 are also related, and they are between all three modules. The font colors for D, P, and N genes are lavender, green, and light green, respectively. Here, hematopoietic stem cells in the bone marrow are a type of multipotent stem cells that can differentiate into any type of blood cells.
However, myeloid and lymphoid progenitor cells differentiated from the hematopoietic stem cells are oligopotent stem cells. That means; the myeloid progenitor cells can only differentiate into red blood cells, mast cells, granulocytes, and platelets while the lymphoid progenitor cells differentiate into lymphocytes and natural killer cells.
Also, the unipotent cells that occur in different tissues and organs can only differentiate into a specific type of cells in that tissue.
For example, hepatoblasts in the liver can only differentiate into hepatocytes. Cell proliferation refers to the process which results in an increase of the number of cells and is defined by the balance between cell divisions and cell loss through cell death or differentiation. Cell differentiation refers to the process by which a less specialized cell becomes a more specialized cell type.
Thus, this explains the main difference between cell proliferation and cell differentiation. The cell number increases due to cell proliferation while cells become functionally specialized due to cell differentiation. Another difference between cell proliferation and cell differentiation is that the cell proliferation occurs first in stem cells while the proliferated cells then differentiate into functional cell types. Moreover, cell proliferation occurs through cell growth and cell division while cell differentiation occurs through the regulation of gene expression.
While cell proliferation is important for both replenishing and replacing cells, cell differentiation is important for replacing the dead or damaged cells in a tissue. Interestingly, there appears to be a more consistent role for the unique pro-proliferative functions of cyclin-D2, whereas cyclin-D1 has been associated with unique pro-differentiation functions.
By gain of function and loss of function studies in the spinal cord, Lukaszewicz and Anderson have revealed that cyclin-D1 has a cell-cycle-independent role in the differentiation of motor neurons. Moreover, cyclin-D1 can have a direct role in the control of gene expression in the developing retina, apparently by promoter binding and recruitment of epigenetic modifiers in a cdk-independent manner Bienvenu et al.
Not all cdks promote cell cycle progression; cdk5 is an unusual member of the cdk family in that it is not activated by cyclins but instead by the binding of an unrelated protein p Moreover, the expression of cdk5 is not associated with cycling cells but, instead, is highly expressed in differentiating and mature neurons in which it plays diverse roles such as promoting neuronal migration, neurite extension and synaptogenesis, as reviewed in Dhariwala and Rajadhyaksha and Dhavan and Tsai An intriguing finding is that cyclin-E expression is retained in terminally differentiated neurons in which it has a cell-cycle-independent function to facilitate synapse formation, through the binding and sequestering of cdk5 in a kinase-inactive complex Odajima et al.
Studies such as these demonstrate the benefit of further examination of cell cycle regulators that are expressed in a manner inconsistent with their known roles, as this may well reveal new and unexpected functions.
These cdkis are also ideally placed both spatially and temporally to coordinate both cell cycle exit and neuronal differentiation during development Nguyen et al. Overexpression of p27Xic1 at a high level in the developing Xenopus embryo leads to cell cycle arrest and massive cell death. However, the expression of lower levels of p27Xic1 results in the precocious differentiation of neural plate progenitors into primary neurons Vernon et al.
Furthermore, the depletion of p27Xic1 impairs the formation of endogenous primary neurons and results in the accumulation of progenitor cells that are unable to transition to differentiation Vernon et al. Functions of mammalian p27Kip1 in regulating proliferation are readily demonstrated by the pknock-out mouse model, which exhibits systemic hyperplasia and increased cellularity in many tissues and organs Fero et al.
Loss of function approaches also identify roles of p27Kip1 specifically in the developing CNS. For example, p27Kip1 functions in regulating the proliferation of transit amplifying progenitors in the developing SVZ Doetsch et al.
The thickened cerebral cortex in p27Kip1-null mice results from the expansion of projection neurons in layers II-IV and GABAergic interneurons in layers V and VI; based on birth date, this indicates altered neuron production during mid- to late-term neurogenesis, which is consistent with the time at which p27Kip1 mRNA peaks in these progenitor cells Goto et al. Interestingly, however, cell cycle length and G1 phase are not found to be significantly altered by the loss of p27Kip1 function Goto et al.
Intriguingly, the overexpression of Xenopus p27Xic1 in the mammalian retina produces similar effects with increased numbers of Muller glial cells, but overexpression of mammalian p27Kip1 does not directly alter retinal cell fate Dyer and Cepko indicating a mechanistic difference between these two orthologues.
In support of this idea, expression patterns and genetic manipulation strategies have revealed specific roles for p27Kip1 and p57Kip2 in regulating neurogenesis in various subsets of progenitor cells. In the cortex, p57Kip2 is more abundantly expressed during early corticogenesis Tury et al. P27Kip1 expression is highest at later stages Tury et al. Furthermore, although both p27Kip1 and p57Kip2 act as modular proteins, the domains required for specific non-cell-cycle functions differ between the two, possibly indicating a different mechanism of action Tury et al.
Overexpression experiments indicate that p57Kip2 is more effective than p27Kip1 in inducing neuronal differentiation Tury et al. An increased understanding of the mechanistic action of these two cdkis will help to elucidate their various shared or non-redundant functions see below. Both p27Kip1 Nguyen et al. Furthermore, cdkis have been implicated in the differentiation of glial cells in the nervous system, and p27Kip1 and p21Cip1 might serve functionally separate and non-redundant roles during oligodendrocyte differentiation.
Classic studies show that p27Kip1 gradually accumulates in oligodendrocyte progenitors and forms a component of both the timer and effector mechanisms that determine a limited number of cell divisions before terminal differentiation for a review, see Durand and Raff However, both p27Kip1 and p21Cip1 are required for oligodendrocyte differentiation; whereas p27Kip1 is required for proper cell cycle withdrawal, p21Cip1 is instead required for the onset of differentiation, independently of its function as a cdki Zezula et al.
Similarly, p57Kip2 levels increase over time in oligodendrocyte precursors and form part of the intrinsic timer mechanism to regulate the number of divisions before differentiation Dugas et al. Whereas p27Xic1 overexpression in Xenopus clearly slows the cell cycle, its ability to induce ectopic neuronal differentiation notably localises to an N-terminal domain of the molecule and is independent of its role as a cdki, being possibly related to an ability to stabilise proneural protein Neurogenin2 Vernon et al.
The separation of functions to distinct structural domains is also conserved in mammalian p27Kip1, together with an ability to interact with and stabilise proneural protein Ngn2 to promote neuronal differentiation in the mammalian brain Nguyen et al.
Consistent with this, p27Kip1 knockout cells from the adult mouse SVZ region have reduced neuronal output attributable to the enhanced degradation of proneural proteins via the proteasome Gil-Perotin et al. Cdkis can regulate gene expression either directly or by regulating transcription factor function. P27Kip1 has also recently been shown to be part of a repressive complex on the Sox2 promoter Li et al.
Similarly, p57Kip2 can interact with nuclear receptor Nurr1 in a cell-cycle-independent manner to promote a dopaminergic fate of midbrain neurons Joseph et al.
However, p57Kip2 can also repress the transcriptional activity of proneural proteins such as Ascl1 and NeuroD, which might be important to enable proper glial cell differentiation Joseph et al. Thus, p57Kip2 might regulate neurogenesis and gliogenesis in a context-dependent manner Tury et al. These studies illustrate nicely the multi-functionality of cell cycle regulators that allows the precise coordination of the many parameters that accompany the transition from progenitor to differentiating neuron.
We undoubtedly have some way to go to identify all the ways that this crucial class of cell cycle regulators is able to influence the differentiation and maturation process.
The retinoblastoma Rb protein is another key regulator of G1, traditionally recognised for its central role in the G1 restriction point to control the commitment of the cell to a further round of replication for a recent review, see Dick and Rubin The role of Rb in aspects of neuronal differentiation and survival has been an evolving story since the early s. Generation of the Rb-null mouse revealed an embryonic lethal phenotype; death occurs between E14 and E15, with embryos displaying severe developmental defects that notably affect the nervous and haematopoietic systems Clarke et al.
Ectopic mitoses are observed throughout the CNS and peripheral nervous system, and this is accompanied by massive cell death, particularly in the hindbrain and sensory ganglia Lee et al. Mechanistically, this has been associated with abnormal S phase entry attributable to elevated E2F DNA-binding activity and with activation of pmediated apoptosis in the CNS Macleod et al.
Subsequent work has demonstrated, however, that this severe null-phenotype results indirectly from placental insufficiency Wu et al. Loss of Rb leads to extensive proliferation of the trophoblast cells, which compromises placental vasculature; sustenance of Rb-null embryos with a wild-type placenta can prevent most of the neurological and haematopoietic abnormalities otherwise observed Wu et al.
Thus, mass apoptosis in the null model may be triggered indirectly, for example, by hypoxia, rather than as a direct effect of Rb loss in neuronal cells, and this is supported by chimeric studies in mice expressing both wild-type and Rb-null cells. Similar to Rb-null embryos, the brains of mid-gestation chimeras show extensive ectopic S phase entry but, in contrast to the Rb-null model, Rb-deficient cells in chimeras are still able to survive and differentiate into neurons, albeit arrested at the G2 phase of the cell cycle.
Additionally, adult brains show an overall normal architecture Lipinski et al. Tissue-specific knock-out of Rb in the developing telencephalon also results in ectopic cell divisions without widespread apoptosis, and the ectopically dividing cells are able to express early neuronal markers Ferguson et al. Furthermore, conditional mutant cortices are still able to generate the full repertoire of cortical projection neurons and interneurons, despite the abnormal terminal mitosis Ferguson et al.
This is consistent with observations during Xenopus development in which Rb is not absolutely required for neuronal differentiation and, indeed, Rb remains hyper-phosphorylated, and therefore presumably inactive, well into tadpole stages, even though extensive neuronal differentiation has occurred by then Cosgrove and Philpott A refined model can therefore be presented whereby Rb functions cell-autonomously to regulate the cell cycle, but largely indirectly in neuronal differentiation and survival, with only a few specific cell types displaying a selective Rb requirement Lipinski et al.
For example, within the nervous system, a selective cell-autonomous role for Rb is described for the survival of cerebellar Purkinje neurons Lipinski et al. Consistent with this model, even in the presence of a wild-type placenta, the rescued Rb-null mice still die perinatally with defective skeletal myogenesis and excessive apoptosis in the lens of the eye de Bruin et al. Additionally, Rb might function subsequent to the neuronal specification and early differentiation stage, instead influencing aspects of neuronal maturation and migration.
Loss of Rb function in the developing mouse cortex results in impaired radial migration of early born dorsal telencephalon neurons and defective tangential migration of GABAergic interneurons from the ventral telecephalon Ferguson et al. Mechanistically, Rb represses the E2F-mediated transcription of a chemotropic ligand receptor, neogenin, that otherwise leads to aberrant migration and adhesion Andrusiak et al.
More recent work has focused on the mechanistic basis of Rb function in specific cell types. In addition to its established function of regulating the E2F transcription factor family, which is crucial for driving the expression of a number of vital cell cycle progression factors such as cyclin-E and cdc2 , Rb has been shown to regulate several aspects of neurogenesis directly and might interact directly with bHLH or HLH proteins, as reviewed in Ferguson and Slack Interestingly, it is in the context of pituitary tumorigenesis that the interaction between Rb and the bHLH inhibitor ID inhibitor of differentiation proteins has been described Lasorella et al.
This interaction prevents ID activity and promotes differentiation, further supporting a tumour suppressor activity of Rb that goes beyond its classic role in the inhibition of cell cycle progression. Even within the Rb-E2F pathway, our comprehension of both cell-cycle- and non-cell-cycle-associated functions is evolving.
For example, within the retina, Rb limits proliferation through the inhibition of E2F1, and yet Rb independently regulates the differentiation of cholinergic starburst amacrine cells SACs through E2F3a, without influence on cell cycle kinetics Chen et al. Although our understanding has progressed from the early Rb-null models, we still do not know whether the phosphorylation status of Rb affects its neurogenic activity and, thus, to what extent the roles of Rb in neuronal differentiation are cell-cycle-dependent.
Geminin is a further example of an important factor with separable and conserved roles in the cell cycle and neurogenesis, as reviewed in Seo and Kroll and Luo and Kessel Geminin has previously been associated with the regulation of DNA replication licensing; accumulation during S phase enables Geminin to inhibit the re-initiation of DNA synthesis and thereby to prevent two rounds of DNA replication within each cell cycle.
Concurrent with the characterisation of its role in replication, Geminin has also independently been identified as a neuralising protein in Xenopus embryo ectoderm and is highly expressed at the onset of gastrulation in the area that later forms the neural plate Kroll et al. Geminin promotes early neural lineage specification from pluripotent progenitor cells but then keeps these progenitors in a proliferative and neural-primed state prior to subsequent differentiation Seo and Kroll ; Luo and Kessel In this respect, Geminin participates in the dynamic equilibrium between proliferation and differentiation in neuronal progenitors, a function that appears to involve competitive interactions with various transcription factors and chromatin remodelling complexes Luo and Kessel ; Pitulescu et al.
In Xenopus embryos, this role in early neural specification is correlated with an ability to suppress BMP4 expression within the presumptive region of the neuroectoderm and with the up-regulation of the expression of proneural genes Kroll et al.
More recent work has revealed that Geminin additionally establishes an epigenetic state that favours the adoption of neuroectodermal fate, resisting sub-threshold stimuli for alternative lineage fates Lim et al. This is dependent upon Polycomb repressor function Lim et al. Despite its key role in promoting neuronal lineage specification, Geminin subsequently maintains the neural progenitor state and resists premature neuronal differentiation in both Xenopus and mammalian cells Seo et al.
Consistent with a role in epigenetic regulation, Geminin promotes a bivalent chromatin state in mammalian cells, at key transcription factor genes that promote neurogenesis; the presence of both activating and repressive histone modifications enables these genes to be repressed but poised for activation Yellajoshyula et al. In this way, Geminin prevents the premature activation of the neurogenic cascade downstream of Ngn2 and NeuroD and regulates the timing of neurogenesis; activation of bHLH target genes occurs as Geminin levels decrease at the onset of neurogenesis Seo et al.
An emerging and recurrent theme is the ability of Geminin to competitively inhibit transcription factor activity. For example, Geminin has a bidirectional inhibitory interaction with Six3, a homeodomain protein involved in eye formation, and a similar relationship is observed with patterning factor Hox homeodomain proteins Pitulescu et al.
Whereas Geminin can inhibit Hox protein function by associating with the Hox-Polycomb multi-protein complex, this interaction additionally prevents Geminin from regulating Cdt1 during DNA replication licensing Luo et al. The ability of Geminin to participate in many independent functions is facilitated by the distinct structural domains within the protein for a review, see Pitulescu et al.
For example, cell-cycle-associated roles reside in the C terminus, whereas neuralisation functions require the N terminus Kroll et al. Further characterisation of these multiple protein interactions will no doubt improve our understanding of the mechanisms by which this key protein influences the balance between proliferation and differentiation.
0コメント