- Research article
- Open Access
Multistable and multistep dynamics in neutrophil differentiation
© Chang et al; licensee BioMed Central Ltd. 2006
Received: 05 October 2005
Accepted: 28 February 2006
Published: 28 February 2006
Cell differentiation has long been theorized to represent a switch in a bistable system, and recent experimental work in micro-organisms has revealed bistable dynamics in small gene regulatory circuits. However, the dynamics of mammalian cell differentiation has not been analyzed with respect to bistability.
Here we studied how HL60 promyelocytic precursor cells transition to the neutrophil cell lineage after stimulation with the differentiation inducer, dimethyl sulfoxide (DMSO). Single cell analysis of the expression kinetics of the differentiation marker CD11b (Mac-1) revealed all-or-none switch-like behavior, in contrast to the seemingly graduated change of expression when measured as a population average. Progression from the precursor to the differentiated state was detected as a discrete transition between low (CD11bLow) and high (CD11bHigh) expressor subpopulations distinguishable in a bimodal distribution. Hysteresis in the dependence of CD11b expression on DMSO dose suggests that this bimodality may reflect a bistable dynamic. But when an "unswitched" (CD11bLow) subpopulation of cells in the bistable/bimodal regime was isolated and cultured, these cells were found to differ from undifferentiated precursor cells in that they were "primed" to differentiate.
These findings indicate that differentiation of human HL60 cells into neutrophils does not result from a simple state transition of a bistable switch as traditionally modeled. Instead, mammalian differentiation appears to be a multi-step process in a high-dimensional system, a result which is consistent with the high connectivity of the cells' complex underlying gene regulatory network.
It is commonly postulated that bistability governs cellular differentiation in mammalian cells [14–16] athough the underlying genetic regulatory networks there are much more complex, but this has never been demonstrated experimentally. Instead of constructing artificial networks to exhibit bistability [6–8, 15], we examined the validity of the bistable model in the context of mammalian differentiation by carrying out single cell analysis of human HL60 promyelocytic cells that are chemically induced to differentiate into neutrophils by treatment with dimethyl sulfoxide (DMSO). These studies show that a surface marker for differentiated neutrophils, CD11b (Mac-I) , is expressed in an all-or-none manner within individual cells, whether analyzed over time or in response to different levels of stimulus. However, detailed kinetic studies of the transition rate suggest that mammalian cell-fate switching may not simply be a bistable transition. Instead, differentiation appears to be a more complicated multistep process, a result which is consistent with the complexity of the underlying gene regulatory network which extends beyond the two-gene circuits used to model bistability.
Results and discussion
Bistability and bimodality
The human promyelocytic HL60 cells robustly differentiate into neutrophils within 6 days in the presence of 1.25% (v/v) DMSO , reaching stationarity with 50–70% of cells in the differentiated state as evaluated by morphological, biochemical, and molecular markers (see Material and Methods). We studied differentiation in HL60 cells by monitoring the expression of CD11b (Mac-1), a well-established surface marker for differentiated neutrophils . Western blotting was used to measure the expression of CD11b as a population average, and in parallel flow cytometry was used to resolve expression of CD11b at the level of individual cells in the same population. Although the latter measurements are common place, a detailed kinetic analysis has not been reported previously.
Mean and inter-quartile ratio (IQR) of fluorescence intensities for Fig. 3.
Moreover, when we carried out similar studies in which we stimulated HL60 precursor cells with different concentrations of DMSO (0% to 1.1%) and analyzed CD11b expression at the stationary state (7d), we also observed bimodality (Fig. 3C, D). Again, analysis of the whole population using Western blots showed a gradual increase in CD11b expression with increasing DMSO concentration (Fig. 3C), whereas flow cytometry histograms of single cells demonstrated bimodality of the CD11b signal at 0.9% and 1.1% DMSO (Fig. 3D). Again, if the bell-shaped histogram of CD11 expression level shifted from low to high intensity values while maintaining its overall shape, it would indicate a gradual switching kinetics at the single cell level (Fig. 1B); however this was not observed (Fig. 3B, D). As in the time-course experiment, the non-monotonic evolution of the spread (Table 1) excludes the possibility of graded differentiation kinetics in individual cells.
The "blurring" of the two peaks that we observed in the bimodality (Fig. 3B,D) may be due to inherent population heterogeneity (e.g. due to stochasticity of gene expression ) which would lead to partial overlapping of the CD11bLow and CD11bHigh peaks, but does not invalidate the underlying switch-like kinetics.
To confirm that the observed bimodal response in CD11b expression was due to switch-like dynamics, rather than an ultrasensitive transition (i.e., steep step in the dose-response curve that acts a threshold ), we explored whether HL60 cells exhibit hysteresis in their CD11b expression response. Hysteresis implies a history-dependence of the response to the same stimulus and is a unique characteristic of a bistable system . Here, hysteresis would manifest in the shape of the dose-response curve measured at stationary states for each DMSO dose, such that a stepwise reduction of the stimulus strength would produce a "lagging" of the corresponding decrease in response strength when compared to the dose-response curve obtained by increasing the strength of the stimulus.
Cells were first stimulated with increasing concentrations of DMSO for 7 days, reaching steady-state ("forward reaction"), and the proportion of cells that became CD11bHigh was recorded. For the "backward reaction", maximally differentiated cells (treated with 1.25% DMSO for 7 days) were resuspended with various concentrations of DMSO for another 7 days to arrive at new stationary states, and the fraction of CD11bHigh cells was similarly noted. Care was taken to ensure that the cells were never exposed to normal medium (except for the 0% DMSO backward reaction data point), because it cannot be excluded that such a short pulse of DMSO-free treatment may create unexpected and lasting effects interfering with hysteresis.
The demonstration of bimodality and hysteresis in CD11b expression of differentiating HL60 cells is consistent with an underlying gene regulatory network of the structure shown in Fig. 2 that can give rise to bistable behavior. Such an architecture has been found in the regulatory circuit of transcription factors implicated in neutrophil differentiation involving the transcription factors GATA2 (= X) and PU.1 (= Y) which mutually inhibit each other [23–25]. For this well-studied system, a relatively broad range of interaction strength and stability of these factors gives rise to two equilibrium states [7, 12, 26, 27]: state a in which GATA2 expression is high and PU.1 is low, and conversely, state b, in which PU.1 is high and GATA2 is suppressed. Biologically, state b may represent the differentiated neutrophil state because endogenous or enforced PU.1 upregulation activates the expression of many neutrophil specific genes, including the surface marker CD11b [28, 29]. In contrast, state a represents the progenitor cell with low PU.1 and, in the case of the HL60 cells, higher levels of GATA2 . Although artificially isolated as a module from a larger network, this small two-gene circuit captures the observed discreteness of a cell fate "switch" from the progenitor to the differentiated state. However, because the molecular targets of DMSO remain unknown, we cannot formally demonstrate how the hysteresis with respect to varying doses of DMSO arose; it only provides phenomenological support for bistability.
Bistability has been proposed as a generic principle that governs differentiation in higher metazoans. However, since mammalian cell differentiation is controlled by regulatory interactions between hundreds, perhaps thousands of genes, and not by isolated one- or two-gene modules as widely assumed in bistability models , the multi-dimensionality of the switch-dynamics may be concealed by measuring a single variable. Using computational models, it has been previously shown that high-dimensional equivalents of bistability can exist in large genetic networks [32–38]. In these 'multi-stability' models, given some architectural constraints of the network, multiple equilibrium states, or "attractor states" may co-exist in high-dimensional state space. Experimental evidence for the existence of such high-dimensional attractor states that represent differentiated phenotypes has recently been shown in populations of HL60 cells using DNA microarray-based dynamic gene expression profiling . Thus, we next examined whether the dynamics of switching in individual HL60 cells harbor evidence of multiple dimensions that could be revealed by monitoring a single variable (CD11b).
Three possible outcomes could be expected. First, the sorted CD11bLow subpopulation could display a decreased rate of generating CD11bHigh cells compared to the native mock-sorted cells. This outcome would indicate that the CD11bLow subpopulation consisted of cells that were inherently more resistant to DMSO induction than native cells. In this case, the coexistence of both states in the bimodal culture would be due to heterogeneity of intrinsic responsiveness to the differentiation stimulus and selection, rather than from bistability as proposed in the bistable switch model of differentiation [6–9]. Second, the sorted CD11bLow subpopulation could have the same differentiation kinetics upon restimulation with DMSO as the native cells. This would indicate that the CD11bLow subpopulation of the bimodal distribution consisted of cells that "by chance" had not yet differentiated . This possibility would not only support a simple bistable switch but also indicate that the state transition is a purely stochastic process. Stochasticity is often observed in cell fate choice and transitions in multipotent progenitor cells or stem cells [41–43], and could be related to the probabilistic manner by which cell type-specific genes are regulated by cis-regulatory elements [44, 45]. The third and last possibility is that the sorted CD11bLow subpopulation exhibits an increased rate of producing CD11bHigh cells. This result would point to some "additive" effect of the two rounds of stimulation with DMSO wherein the first stimulation leads to progress in differentiation that is "stored" in state space dimensions other than CD11b.
Interestingly, even after restimulating these cells (i.e., two rounds of DMSO in total), a CD11bLow subpopulation was still observable. We thus asked the same question for the priming process as we originally did for the differentiation process: does the heterogeneity in the priming process result from the existence of other metastable "pre-primed" states or do the unprimed cells represent a resistant sub-population? To address this question, the CD11bLow subpopulation that appeared after one round of re-stimulation with DMSO was FACS sorted and restimulated for a second time with 0.8% DMSO for another seven days (three rounds of DMSO treatment in total) (Fig. 6). Surprisingly, this subpopulation again showed accelerated CD11b expression kinetics upon restimulation when compared to the native control population (Fig. 7A), but had a rate of generating CD11bHigh cells comparable to that exhibited by the subpopulation that was only exposed to DMSO for two rounds of stimulation (Fig. 7A). These results rule out the preexistence of a resistant subpopulation, and suggest that no additional intermediate steps between the CD11bLow and the CD11bHigh states can be discerned with the stimulation scheme used here.
Taken together, these results indicate that human HL60 cell differentiation is a multi-step process, consisting of at least two steps: (1) an initial transition step from the native CD11bLow state to the "primed" CD11bLow state and (2) a second step from the "primed" state to the CD11bHigh state. The observation that a second round of sorting and restimulation did not alter the rate of CD11bHigh cell production indicates that the process of "priming" (step 1) went through to completion (e.g. all "primed" cells are in the same state) at the perturbation strength conferred by 0.8% DMSO for 7 days. In contrast, the second step leading to the high expression of CD11b appeared to be a switch that only partially ran to completion in 0.8% DMSO, hence exposing not only the existence of a "primed" undifferentiated state, but also the stochastic nature of its transition to the differentiated CD11bHigh state . Given the design of our experiments, however, it was not possible to determine whether HL60 cells that are in the CD11bHigh state must also pass through multiple sequential states to be fully differentiated with respect to all state space dimensions. Nevertheless, the existence of two discernible states (native and "primed") among the CD11bLow subpopulation supports the existence of "deterministic heterogeneity" within this population as a result of multistability within the genome-wide regulatory network. This heterogeneity does not measurably contribute to additional population dispersion of CD11b expression levels in the CD11bLow population. Instead, it represents an additional state in which the cells exhibit an increased readiness to express CD11b upon restimulation.
Our results may also explain why cellular differentiation processes often take as long as several days to weeks to complete, although molecularly, they essentially consist of a change in gene expression profile which could be completed in a day at the level of individual genes. Specifically, the results also explain why, despite hysteresis, prolonged exposure (> 3 days) to the stimulating agents DMSO and all-trans-retinoic acid is necessary to achieve maximal neutrophil differentiation in HL60 cells ([18, 48]; H.H.C. and S.H., unpublished observations). This is in contrast to other HL60 differentiation processes, such as macrophage differentiation, for which brief exposure to TPA of a few hours is sufficient .
At the moment, the molecular mechanisms that establish the primed state are not known. It is likely that in order to switch the transcriptome to that of the differentiated state, multiple waves of transcriptional activation must occur in which newly synthesized transcription factors regulate other (transcription) factors. Thus, the primed state may reflect a state in which such intermediate regulatory proteins have become available. Moreover, remodeling of chromatin to make loci of differentiation-specific genes accessible for transcription may contribute to the multi-step characteristics of differentiation with kinetically identifiable intermediate states .
In this study, we examined the dynamics of mammalian cell differentiation by studying the expression kinetics of the differentiation marker CD11b during neutrophil differentiation in DMSO-treated HL60 cells. Although the behavior of this marker is in agreement with simple bistability models [10, 11], our detailed analysis revealed that differentiation is actually a multi-step process consistent with a model in which multiple coupled switches along various state space dimensions give rise to multistable states that represent high-dimensional attractors in the genome-wide cell regulatory network [33, 39]. Based on a purely dynamical and phenomenological analysis, we were able to identify a "primed" state in HL60 differentiation characteristic of cells that had not made the all-or-none phenotypic switch, but had proceeded partially along the path of differentiation. Although we were unable to take a "bottom-up" approach as in the studies of well-characterized microorganisms  or engineered gene networks , our treatment strategy and sorting scheme allowed us to study mammalian cell differentiation without knowing the underlying gene regulatory network beyond the GATA2/PU.1 switch. This approach opens a new way of dissecting the multi-step process of cellular differentiation into a sequence of discrete metastable intermediate states that evade conventional time-course analysis of entire populations. Specifically, our results suggest that the regulation of differentiation may involve "unanticipated" gene dimensions which do not directly affect the expression of a measured marker. The existence of multi-dimensional, multistable behavior during cell fate switching in mammalian cells has important implications in the way differentiation is viewed and ultimately, in how processes such as lineage commitment of stem cells during tissue development can be explained and controlled.
Cell culture and differentiation
HL60 cells (ATCC) were cultured in IMDM medium (ATCC) supplemented with 10% fetal bovine serum and 1% glutamine plus penicillin and streptomycin. Cells of passage 7 (after receipt from ATCC) at a density of 1.0 × 106 cells/ml and growing at a basal rate of 1.3–1.7 day -1 were treated with variable concentrations of DMSO (Sigma) ranging from 0.3% to 1.25% (v/v) to induce differentiation. At each time point, cells were harvested from the suspension culture, pelleted, and processed for either Western blot and/or flow cytometry analysis (see below). Differentiation was monitored primarily with CD11b expression by flow cytometry, but morphology by Giemsa stain and nitroblue tetrazolium-reducing activities were also utilized. A stationary state was reached at day 6 since the fraction of differentiated cells and the level of expression of CD11b did not further increase when cells were monitored up to day 12 after induction of differentiation.
Western blot analysis
1 × 106 cells were pelleted and directly lysed with 20% sample-loading buffer for SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and immediately boiled for 5 min at 95°C. 30–50 μl of total cell lysate were fractionated on SDS-PAGE gel and transferred to nitrocellulose membranes. Following blocking with 5% milk/PBST (phosphate buffered saline with Tween 20), the membrane was probed with a 1:500 dilution of CD11b/Mac-1 antibody (BD Pharmingen). Antibody binding was detected with a 1:5000 dilution of peroxidase labeled anti-mouse IgG (Vector) and luminescence was detected with Supersignal West Dura Signal reagents (Pierce).
Immunofluorescence staining of live cells for flow cytometry
For the Guava- PCA system (see below) 200,000 cells were pelleted and incubated in 7 μl of CD11b/MAC-I R-PE conjugated fluorescence antibody (BD Pharmingen) on ice for 30 min, washed with ice-cold 1% fetal calf serum/PBS/0.01% NaN3 (NaN3 is left out in sorting experiments), and resuspended in the same buffer at 106 cells/ml density for analysis. Intracellular phosphorylated-Erk levels were detected using the BD PhosFlow kit (BD Pharmingen) and the protocol provided. Briefly, 200,000 cells were fixed with BD PhosFlow Fix Buffer (BD Pharmingen) at 37°C for 10 min, pelleted, washed with BD PhosFlow Perm/Wash Buffer (BD Pharmingen) twice, incubated with 5 μl of a 1:5 dilution of Anti-Phospho-ERK1/2:PE conjugated fluorescence antibody (BD Pharmingen) in the dark at room temperature for 1 hour, washed again with Perm/Wash Buffer, and resuspended in the same buffer at 106 cells/ml density for analysis. For fluorescence-activated cell sorting, staining was scaled up 10-fold to 50 μl of CD11b/MAC-I R-PE conjugated fluorescence antibody (BD Pharmingen) per 106 cells and cells were resuspended at 8–10 × 106 cells/ml. Pilot antibody titration experiments were performed to ensure that staining occurred at least at 2-fold saturation. Ice-cold 1% fetal calf serum/PBS/0.01% NaN3 was used to establish background signal with unstained cells.
Flow cytometry and Fluorescence Activated Cell Sorting (FACS)
Flow cytometry was performed on a Guava-PCA microfluidic-based flow cytometer (GuavaTechnologies, Inc). Fluorescence activated cell sorting was performed with either a Becton Dickinson FACSVantage (Becton Dickinson) or a Becton Dickinson FACSAria (Becton Dickinson) flow cytometer. Data analysis was done with either CytoSoft™ 2.1.1. (GuavaTechnologies, Inc) or WinMDI software. For cell sorting, starting cell number ranged between 40–80 × 106 cells, and cells were sorted into ice-cold medium for a maximum of 3 hours. Gates for sorting the CD11bLow subpopulation in the 0.8% DMSO-treated samples were set relative to an untreated, native population. The latter was also mock sorted and processed in exactly the same way as the former to control for the effects of FACS sorting on cellular expression of CD11b. To remove the staining antibody before reculturing, pelleted cells were suspended in pH.2.25 MES (morpholinoethanesulfonic acid)/Tris buffer for 30 s. A 10-fold volume of pH 7.4 PBS was immediately added for neutralization and the cells were pelleted and resuspended in culture medium. After antibody removal the cells had fluorescence signal intensities on par with unstained HL60 cells and exhibited normal viability for future immunofluorescence staining.
This work was funded by grants to S.H. from the Air Force Office of Scientific Research (F49550-05-1-0078) and the Patterson Trust, and to D.E.I. from the National Institutes of Health (CA55833) and the Army Research Office (W911NF-04-1-0273). H.H.C. is supported by the Presidential Scholarship and the Ashford Fellowship of Harvard University. H.H.C. would like to thank K. Farh for critical reading of the manuscript.
- Rubin H: Mechanisms for enduring biological change. Am J Physiol. 1992, 262 (1 Pt 1): L111-3.PubMedGoogle Scholar
- Krutzik PO, Irish JM, Nolan GP, Perez OD: Analysis of protein phosphorylation and cellular signaling events by flow cytometry: techniques and clinical applications. Clinical Immunology. 2004, 110 (3): 206-221. 10.1016/j.clim.2003.11.009.View ArticlePubMedGoogle Scholar
- Novick A, Weiner M: Enzyme Induction as an All-or-None Phenomenon. Proc Natl Acad Sci U S A. 1957, 43: 553-566.PubMed CentralView ArticlePubMedGoogle Scholar
- Waddington CH: Principles of Embryology. 1956, London , Allen & Unwin LtdGoogle Scholar
- Bagowski CP, Ferrell JE: Bistabillity in the JNK cascade. Current Biology. 2001, 11 (15): 1176-1182. 10.1016/S0960-9822(01)00330-X.View ArticlePubMedGoogle Scholar
- Becskei A, Seraphin B, Serrano L: Positive feedback in eukaryotic gene networks: cell differentiation by graded to binary response conversion. EMBO J. 2001, 20 (10): 2528-2535. 10.1093/emboj/20.10.2528.PubMed CentralView ArticlePubMedGoogle Scholar
- Gardner TS, Cantor CR, Collins JJ: Construction of a genetic toggle switch in Escherichia coli. Nature. 2000, 403 (6767): 339-342. 10.1038/35002131.View ArticlePubMedGoogle Scholar
- Ozbudak EM, Thattai M, Lim HN, Shraiman BI, Van Oudenaarden A: Multistability in the lactose utilization network of Escherichia coli. Nature. 2004, 427 (6976): 737-740. 10.1038/nature02298.View ArticlePubMedGoogle Scholar
- Xiong W, Ferrell JEJ: A positive-feedback-based bistable 'memory module' that governs a cell fate decision. Nature. 2003, 426 (6965): 460-465. 10.1038/nature02089.View ArticlePubMedGoogle Scholar
- Delbrück M: Discussion. Unités biologiques douées de continuité génétique Colloques Internationaux du Centre National de la Recherche Scientifique. 1949, CNRS, ParisGoogle Scholar
- Monod J, Jacob F: Teleonomic mechanisms in cellular metabolism, growth, and differentiation. Cold Spring Harb Symp Quant Biol. 1961, 26: 389-401.View ArticlePubMedGoogle Scholar
- Kaplan DGL: Understanding Nonlinear Dynamics . 1995, New York , Springer, 1View ArticleGoogle Scholar
- Tyson JJ, Chen KC, Novak B: Sniffers, buzzers, toggles and blinkers: dynamics of regulatory and signaling pathways in the cell. Curr Opin Cell Biol. 2003, 15 (2): 221-231. 10.1016/S0955-0674(03)00017-6.View ArticlePubMedGoogle Scholar
- Thomas R: Laws for the dynamics of regulatory networks. International Journal Of Developmental Biology. 1998, 42 (3): 479-485.PubMedGoogle Scholar
- Kramer BP, Fussenegger M: Hysteresis in a synthetic mammalian gene network. Proceedings Of The National Academy Of Sciences Of The United States Of America. 2005, 102 (27): 9517-9522. 10.1073/pnas.0500345102.PubMed CentralView ArticlePubMedGoogle Scholar
- Mariani L, Lohning M, Radbruch A, Hofer T: Transcriptional control networks of cell differentiation: insights from helper T lymphocytes. Progress In Biophysics & Molecular Biology. 2004, 86 (1): 45-76. 10.1016/j.pbiomolbio.2004.02.007.View ArticleGoogle Scholar
- Trayner ID, Bustorff T, Etches AE, Mufti GJ, Foss Y, Farzaneh F: Changes in antigen expression on differentiating HL60 cells treated with dimethylsulphoxide, all-trans retinoic acid, alpha1,25-dihydroxyvitamin D3 or 12-O-tetradecanoyl phorbol-13-acetate. Leuk Res. 1998, 22 (6): 537-547. 10.1016/S0145-2126(98)00041-1.View ArticlePubMedGoogle Scholar
- Collins SJ, Ruscetti FW, Gallagher RE, Gallo RC: Terminal differentiation of human promyelocytic leukemia cells induced by dimethyl sulfoxide and other polar compounds. Proc Natl Acad Sci USA. 1978, 75 (5): 2458-2462.PubMed CentralView ArticlePubMedGoogle Scholar
- Kaern M, Elston TC, Blake WJ, Collins JJ: Stochasticity in gene expression: from theories to phenotypes. Nat Rev Genet. 2005, 6 (6): 451-464. 10.1038/nrg1615.View ArticlePubMedGoogle Scholar
- Koshland DEJ, Goldbeter A, Stock JB: Amplification and adaptation in regulatory and sensory systems. Science. 1982, 217 (4556): 220-225.View ArticlePubMedGoogle Scholar
- Laurent M, Kellershohn N: Multistability: a major means of differentiation and evolution in biological systems. Trends Biochem Sci. 1999, 24 (11): 418-422. 10.1016/S0968-0004(99)01473-5.View ArticlePubMedGoogle Scholar
- Studzinski GP, Brelvi ZS: Changes in proto-oncogene expression associated with reversal of macrophage-like differentiation of HL 60 cells. J Natl Cancer Inst. 1987, 79 (1): 67-76.PubMedGoogle Scholar
- Graf T: Differentiation plasticity of hematopoietic cells. Blood. 2002, 99 (9): 3089-3101. 10.1182/blood.V99.9.3089.View ArticlePubMedGoogle Scholar
- Zhang P, Behre G, Pan J, Iwama A, Wara-Aswapati N, Radomska HS, Auron PE, Tenen DG, Sun Z: Negative cross-talk between hematopoietic regulators: GATA proteins repress PU.1. Proc Natl Acad Sci U S A. 1999, 96 (15): 8705-8710. 10.1073/pnas.96.15.8705.PubMed CentralView ArticlePubMedGoogle Scholar
- Zhang P, Zhang X, Iwama A, Yu C, Smith KA, Mueller BU, Narravula S, Torbett BE, Orkin SH, Tenen DG: PU.1 inhibits GATA-1 function and erythroid differentiation by blocking GATA-1 DNA binding. Blood. 2000, 96 (8): 2641-2648.PubMedGoogle Scholar
- Huang S: Multistability and Multicellularity: Cell Fates as High-dimensional Attractors of Gene Regulatory Networks. Computational Systems Biology. Edited by: A. K, R. E. 2005, Elsevier Acadmic PressGoogle Scholar
- Angeli D, Ferrell JE, Sontag ED: Detection of multistability, bifurcations, and hysteresis in a large class of biological positive-feed back systems. Proceedings Of The National Academy Of Sciences Of The United States Of America. 2004, 101 (7): 1822-1827. 10.1073/pnas.0308265100.PubMed CentralView ArticlePubMedGoogle Scholar
- Pahl HL, Scheibe RJ, Zhang DE, Chen HM, Galson DL, Maki RA, Tenen DG: The proto-oncogene PU.1 regulates expression of the myeloid-specific CD11b promoter. J Biol Chem. 1993, 268 (7): 5014-5020.PubMedGoogle Scholar
- Gangenahalli GU, Gupta P, Saluja D, Verma YK, Kishore V, Chandra R, Sharma RK, Ravindranath T: Stem cell fate specification: role of master regulatory switch transcription factor PU.1 in differential hematopoiesis. Stem Cells Dev. 2005, 14 (2): 140-152. 10.1089/scd.2005.14.140.View ArticlePubMedGoogle Scholar
- Walsh JC, DeKoter RP, Lee HJ, Smith ED, Lancki DW, Gurish MF, Friend DS, Stevens RL, Anastasi J, Singh H: Cooperative and antagonistic interplay between PU.1 and GATA-2 in the specification of myeloid cell fates. Immunity. 2002, 17 (5): 665-676. 10.1016/S1074-7613(02)00452-1.View ArticlePubMedGoogle Scholar
- Huang S: Back to the biology in systems biology: what can we learn from biomolecular networks. Brief Funct Genomics Proteomics. 2004, 2 (4): 279-297. 10.1093/bfgp/2.4.279.View ArticleGoogle Scholar
- Kauffman SA: Metabolic stability and epigenesis in randomly constructed genetic nets. J Theor Biol. 1969, 22 (3): 437-467. 10.1016/0022-5193(69)90015-0.View ArticlePubMedGoogle Scholar
- Kauffman SA: The Origins of Order. 1993, New York , Oxford Univ. PressGoogle Scholar
- Bilke S, Sjunnesson F: Stability of the Kauffman model. Physical Review E. 2002, 65 (1):Google Scholar
- Guckenheimer JHP: Nonlinear Oscillations, Dynamical Systems, and Bifurcations of Vector Fields. Applied Mathematical Sciences. 1983, New York , Springer, 42:Google Scholar
- Kappler K, Edwards R, Glass L: Dynamics in high-dimensional model gene networks. Signal Processing. 2003, 83 (4): 789-798. 10.1016/S0165-1684(02)00479-6.View ArticleGoogle Scholar
- Mochizuki A: An analytical study of the number of steady states in gene regulatory networks. Journal Of Theoretical Biology. 2005, 236 (3): 291-310. 10.1016/j.jtbi.2005.03.015.View ArticlePubMedGoogle Scholar
- Aldana M, Cluzel P: A natural class of robust networks. Proc Natl Acad Sci U S A. 2003, 100 (15): 8710-8714. 10.1073/pnas.1536783100.PubMed CentralView ArticlePubMedGoogle Scholar
- Huang S, Eichler G, Bar-Yam Y, Ingber DE: Cell fates as high-dimensional attractor states of a complex gene regulatory network. Phys Rev Lett. 2005, 94 (12): 128701-10.1103/PhysRevLett.94.128701.View ArticlePubMedGoogle Scholar
- Acar M, Becskei A, van Oudenaarden A: Enhancement of cellular memory by reducing stochastic transitions. Nature. 2005, 435 (7039): 228-232. 10.1038/nature03524.View ArticlePubMedGoogle Scholar
- Ogawa M: Stochastic model revisited. Int J Hematol. 1999, 69 (1): 2-5.PubMedGoogle Scholar
- Hume DA: Probability in transcriptional regulation and its implications for leukocyte differentiation and inducible gene expression. Blood. 2000, 96 (7): 2323-2328.PubMedGoogle Scholar
- Enver T, Heyworth CM, Dexter TM: Do stem cells play dice?. Blood. 1998, 92 (2): 348-51; discussion 352.PubMedGoogle Scholar
- Walters MC, Fiering S, Eidemiller J, Magis W, Groudine M, Martin DI: Enhancers increase the probability but not the level of gene expression. Proc Natl Acad Sci U S A. 1995, 92 (15): 7125-7129.PubMed CentralView ArticlePubMedGoogle Scholar
- de Krom M, van de Corput M, von Lindern M, Grosveld F, Strouboulis J: Stochastic patterns in globin gene expression are established prior to transcriptional activation and are clonally inherited. Mol Cell. 2002, 9 (6): 1319-1326. 10.1016/S1097-2765(02)00558-0.View ArticlePubMedGoogle Scholar
- Wang X, Studzinski GP: Activation of extracellular signal-regulated kinases (ERKs) defines the first phase of 1,25-dihydroxyvitamin D3-induced differentiation of HL60 cells. J Cell Biochem. 2001, 80 (4): 471-482. 10.1002/1097-4644(20010315)80:4<471::AID-JCB1001>3.0.CO;2-J.View ArticlePubMedGoogle Scholar
- Hasty J, Pradines J, Dolnik M, Collins JJ: Noise-based switches and amplifiers for gene expression. Proc Natl Acad Sci U S A. 2000, 97 (5): 2075-2080. 10.1073/pnas.040411297.PubMed CentralView ArticlePubMedGoogle Scholar
- Breitman TR, Selonick SE, Collins SJ: Induction of differentiation of the human promyelocytic leukemia cell line (HL-60) by retinoic acid. Proc Natl Acad Sci USA. 1980, 77 (5): 2936-2940.PubMed CentralView ArticlePubMedGoogle Scholar
- Rovera G, Santoli D, Damsky C: Human Promyelocytic Leukemia-Cells In Culture Differentiate Into Macrophage-Like Cells When Treated With A Phorbol Diester. Proceedings Of The National Academy Of Sciences Of The United States Of America. 1979, 76 (6): 2779-2783.PubMed CentralView ArticlePubMedGoogle Scholar
- Georgopoulos K: Haematopoietic cell-fate decisions, chromatin regulation and ikaros. Nat Rev Immunol. 2002, 2 (3): 162-174. 10.1038/nri747.View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.