• hierarchical clustering;
  • data-mining;
  • high-throughput experimentation;
  • flow cytometry;
  • cell-surface antigens;
  • cellular phenotypes;
  • cellular profiling


  1. Top of page
  2. Abstract
  6. Acknowledgements


This study examined whether hierarchical clustering could be used to detect cell states induced by treatment combinations that were generated through automation and high-throughput (HT) technology. Data-mining techniques were used to analyze the large experimental data sets to determine whether nonlinear, non-obvious responses could be extracted from the data.


Unary, binary, and ternary combinations of pharmacological factors (examples of stimuli) were used to induce differentiation of HL-60 cells using a HT automated approach. Cell profiles were analyzed by incorporating hierarchical clustering methods on data collected by flow cytometry. Data-mining techniques were used to explore the combinatorial space for nonlinear, unexpected events. Additional small-scale, follow-up experiments were performed on cellular profiles of interest.


Multiple, distinct cellular profiles were detected using hierarchical clustering of expressed cell-surface antigens. Data-mining of this large, complex data set retrieved cases of both factor dominance and cooperativity, as well as atypical cellular profiles. Follow-up experiments found that treatment combinations producing “atypical cell types” made those cells more susceptible to apoptosis.


Hierarchical clustering and other data-mining techniques were applied to analyze large data sets from HT flow cytometry. From each sample, the data set was filtered and used to define discrete, usable states that were then related back to their original formulations. Analysis of resultant cell populations induced by a multitude of treatments identified unexpected phenotypes and nonlinear response profiles. © 2007 International Society for Analytical Cytology

Three individually successful techniques in the fields of biotechnology and drug discovery are high-throughput (HT) screening platforms (1), flow cytometry (2–4), and data-mining tools, such as hierarchical clustering (5–8). In this study, these three techniques have been combined and used successfully to detect cell states induced by multifactor combinations. The creation of multifactor combinations and analysis of the results of cell-based in vitro assays can better elucidate how cells behave in a complex microenvironment (9–11). Cells behave differently when treated with single factors as compared to treatment with multiple factors either in sequence or in combination (12, 13). Thus, by limiting the evaluation of treatment combinations, traditional experimentation does not approach a cell's natural environment and thus in vitro results may not reflect the in vivo situation (14–16).

For cell-based studies, HT, combinatorial platforms, aided by automation are more useful than traditional, manual experimentation because of three advantages (7, 17–19). First, HT-automated platforms have enabled the creation of large numbers of multifactor combinations that can elucidate how cells behave in a complex microenvironment and, consequently, reduce the limitations of traditional experimentation, in which only one or two variables are manipulated at a time and improve correlations between in vitro and in vivo results. Second, automated platforms have provided a hands-off method for accurately measuring and dispensing minute, and often viscous samples with increased accuracy and precision. Third, HT platform discovery and use requires sample miniaturization, resulting in a lower cost per sample, thus allowing analysis of more samples, and the exploration of a larger experimental space.

High-throughput platforms have not been used together with flow cytometry until recently (20). A well-known advantage of flow cytometry is its ability to evaluate multiple parameters for each cell, such as cell size, cell granularity, DNA content, cell viability, differentiation, proliferation, production of soluble antigen, and expression of cell-surface antigens (which can also indicate activation) (3, 4, 21–23). This advantage is particularly useful in the analysis of hematopoietic cell populations, because cell populations can be potentially analyzed using over 200 CD (cluster of differentiation) cell-surface markers. With the introduction of HT samplers to flow cytometry, a large diverse set of treatment conditions can be analyzed for a large number of read-outs simultaneously (20). By expanding both the number of input variables and the number of readouts, a larger experimental space can be explored, increasing the probability of finding conditions that can produce diverse phenotypic profiles.

In spite of these potential benefits, flow cytometry has disadvantages that have likely contributed to its non-use in HT platform-based approaches. Flow cytometry generates large, bulky data sets that are typically analyzed using two dimensional scatter plots. This makes it difficult, or impossible, to identify and compare trends in the data efficiently. Thus, traditional data analysis, consisting of generating graphs in 2-D format, may not be useful for analysis of data from HT flow cytometry. More commonly, traditional data analysis is being replaced by high-powered data-mining techniques in order to analyze the large amounts of data generated using HT, automated platforms.

One example of a data-mining technique is hierarchical clustering (24, 25). Hierarchical clustering is a common method used to determine clusters of similar data points in multidimensional spaces and allows for easy handling of large bulky data sets. Rather than merely confirming already known relationships, hierarchical clustering facilitates the probability of finding rare, unexpected, or nonlinear events in complex high-dimensional data sets. The incorporation of HT experimental platforms should greatly increase the likelihood of finding improved correlations between in vitro and in vivo systems. Data mining tools, such as hierarchical clustering, can help to identify novel responses and alternate formulations that can produce a particular cellular profile.

Here, HT platforms, flow cytometry, and data-mining tools, such as hierarchical clustering, are all used to analyze the effects of complex treatments from combinations of factors on cell states, by analyzing both large data sets from multiple read-outs and experiment-wide responses to many different treatments in one plot. Through the use of data-mining techniques, rare, unexpected response profiles and nonlinear interactions were found in complex high-dimensional data sets. Finally, flow cytometry is shown to be an effective tool for defining discrete phenotypic states that can be related through informatics to the formulations that produced them.


  1. Top of page
  2. Abstract
  6. Acknowledgements

Cell Culture

HL-60 cells, a generous gift from A. Means, (Duke University Medical Center, Durham, NC) were maintained in Dulbecco's Modified Eagle Medium (DMEM) (VWR, West Chester, PA) supplemented with 10% fetal bovine serum (FBS) (Hyclone, Logan, UT) and 1% penicillin/streptomycin (VWR) and grown in a humidified atmosphere at 37°C under 5% CO2. Each experiment was performed using cells of approximately the same passage number (p 10–20).

Chemicals and Antibodies

To induce differentiation, cells in logarithmic growth were seeded at 4.0 × 104 cells/well in 96-well, flat-bottom plates (VWR) and treated with single, binary, or ternary combinations of the following differentiation factors: Phorbol 12-myristate 13-acetate, (PMA) (VWR): 81 nM, 16 nM, 0.81 nM [100 nM, 1nM]; 1-α, 25-dihydroxycholecalciferol, (Vitamin D3) (Sigma Aldrich, St. Louis, MO): 100 nM, 1 nM, 10 pM [100 nM, 1 nM]; dimethylsulfoxide, (DMSO) (VWR): 0.26 M, 0.18 M, 0.13 M [0.19 M, 0.15 M]; all-trans Retinoic acid, (ATRA) (VWR): 10 μM, 10 nM, 10 pM [50 μM, 500 nM]; Media (pH 7.8) sodium butyrate, (VWR): 600 μM, 300 μM, 100 μM [500 μM, 200 μM] (Table 1).

Table 1. Summary of Differentiation Factors
Differentiation factorResulting cell lineage
  1. Differentiation factors selected for binary and ternary experiments and the known cell lineage that HL-60 cells differentiate into when treated with those factors.

Vitamin D3Monocyte
Sodium butyrate (pH 7.8)Eosinophil/basophil

Following differentiation, the cells were stained with some combination of the following antibodies to cell-surface antigens and vital stain, (Pharmingen, San Jose, CA): FITC-mouse antihuman CD3, R-PE-mouse antihuman CD3, PE-CY7 mouse antihuman CD11b/Mac-1, APC-mouse antihuman CD13, R-PE mouse antihuman CD13, APC-mouse antihuman CD14, R-PE mouse antihuman CD33, FITC mouse antihuman CD34, APC-mouse antihuman CD34, PE-CY5-mouse antihuman CD 42b, PE-CY7-rat antimouse/antihuman CD45R/B220, FITC-rat antimouse/antihuman CD45R/B220, AlexaFluor 488-mouse antihuman CD56, FITC-mouse antihuman CD57, Cy-Chrome-mouse antihuman CD62P, R-PE-mouse antihuman CD66, FITC-mouse antihuman CD66b, FITC-mouse antihuman CD83, APC-mouse antihuman CD86, R-PE mouse antihuman CDw125, APC-mouse antihuman CD235a, PE-Annexin V, and 7-AAD (Table 2).

Table 2. Summary of Cell Markers
Cell markerCell lineage
  1. The cell surface markers for phenotypic profiling were chosen to span specific lineages of the hematopoietic tree. In addition, cell markers representing early and late stage apoptosis were also selected.

B220B lymphocyte
CD3T lymphocyte
CD11b/Mac1Myeloid cell
CD13Granulocyte/monocyte Progenitor (GMP)
CD34Hematopoietic stem cell (HSC)
CD56Natural killer cell
CD57Natural killer cell
CD83Dendritic cell
CD86Dendritic cell
AnnexinEarly apoptosis
7-AADLate apoptosis

Process Flow

Each experiment consisted of three phases: (a) design of experiment (DOE), (b) execution of experimental protocols, and (c) data analysis. The DOE was facilitated by a web-based interface that allows the generation of sample-plate layout schemes and the choice of input parameters (variables) such as biological or pharmacological factors, concentration, experimental controls, time, and temperature. These parameters were assembled into a set of instructions for driving automated preparation and for tracking samples. Once experiments were performed, data-mining tools were developed that enabled exploration of that combinatorial space. The combinatorial space can be defined as the multidimensional space of possible formulations made by independently varying the individual input parameters—in other words, the range of possible formulations containing some combination of many possible treatments.

High-Throughput Combinatorial Process

Five combinations of factors in either binary (i.e. combinations of 2 chosen factors at 3 concentrations) or tertiary (i.e. combinations of 3 factors at 2 concentrations) were created using the design of experiment (DOE) software. Stock solutions of each differentiation factor were prepared manually and used for automated dispensing. A robotic sample processor (Tecan Genesis™, Durham, NC) was used to dispense the specified reagent combinations within a sterile enclosure suitable for cell-culture work. Binary combination treatments of cells were created by dispensing 100 μl of a 2× stock solution into each well. Ternary combination treatments of cells were created by dispensing 70 μl of a 3× stock solution.

The plate layout was generated for a 96-well plate by TransForm Pharmaceutical's Process Informatics System (Lexington, MA). Each combination of factors was performed in triplicate. Once the layout was generated, a worklist was created and submitted to a database. From this worklist, a set of instructions was created for dispensing differentiation factors by the Tecan Genesis to control and direct robotic operations. For quality assurance, precision, and accuracy of the dispensing syringes were tested and validated prior to experimental execution. Each plate contained experimental and process controls: one positive control (HL-60 cells treated with 100 μM vitamin D3 and stained with CD14-APC), two negative controls (untreated, stained with plate antibody cocktail), and four internal process controls (untreated cells stained with CD45 FITC, CD45 PE, CD45 PE-Cy5, or either CD45 PE-CY7 or CD45 APC).

The worklist was executed by Tecan Gemini software to direct robotics operations. Combinatorial sample dispensing into prebarcoded, 96-well plates was carried out under sterile conditions by the Tecan Genesis™. After sample dispensing, a robotic arm transferred plates to an integrated cell-culture incubator. The cells (4 × 104 cells in 250 μl media) were seeded into the prepared plates using a semi-automated Multidrop system (Day 0) from Titertek™. The plates were incubated for 5 days in a standard cell-culture incubator. On Day 2, media was exchanged by dispensing a duplicate set of plates. Resulting cell states were assessed by flow cytometry using a panel of antibodies directed toward cell-surface markers.

Analysis of Cell-Surface Marker Expression

On Day 5, cells were washed and incubated with γ-globulin from rat, mouse, or both (Jackson Immunoresearch, West Grove, PA) for 15 min at room temperature. Plates were stained with antibody cocktails according to manufacturer's instructions (adjusted for approximate cell number/well) for 25 min at room temperature on a rocker in the dark.

Cells were washed with PBS and resuspended in 1% Ultrapure methanol-free formaldehyde (Polysciences, Warrington, PA) and analyzed on a BD FACS Calibur™ flow cytometer with an up-front HT sampler. A gate was created using a forward versus side scatter plot to exclude cellular debris. Cells located within this gate were analyzed for expression of cell surface markers using fluorescently labeled antibodies. 7-AAD was not used in a gate to exclude dead cells. Instead, 7-AAD was included as part of the analysis to determine whether particular factor treatment combinations caused increased levels of apoptosis in the cell populations compared to control. At least 5,000 cells were collected inside the gate for each sample.

Assessment of Apoptosis

HL-60 cells were incubated with PMA alone, DMSO alone, or PMA + DMSO, and then washed and incubated with either 5, 10, or 50 ng/ml idarubicin (Sigma) for 24 h. Control wells contained either undifferentiated, untreated cells or undifferentiated, idarubicin-treated cells.

Apoptosis was assessed by flow cytometry using Annexin V-PE + 7-AAD labeling. Experiments were carried out according to manufacturer's instructions, which were adjusted for approximate cell number.

Data Processing

The labeling of any sample well with antibodies is limited by the number of fluorochromes with distinct emission spectra that can be detected at one time by the flow cytometer. In this study, the FACS Calibur™ flow cytometer (BD Biosciences, San Diego, CA) was used because it can collect samples in a 96-well plate format using a high-throughput (HT) sampler. The FACS Calibur is limited to detection of only four colors at one time; therefore, the total number of antibodies was divided into cocktails containing four fluorescently labeled antibodies per cocktail (Fig. 1a). Replicate plates contained treatment combinations to which one antibody cocktail was added (Fig. 1b). Data from each sample well on a plate was stored by an acquisition template in the Flow Cytometry Standard 2.0 (FCS2.0) data format. The BD CellQuest Pro™ instrument software was used to set markers to define the percent expression for any cell-surface marker based on isotype-negative controls. The collected data were extracted and stored in a database (Fig. 1c).

thumbnail image

Figure 1. Data Collection Strategy. (a) The FACS Calibur™ can detect only four colors at one time; therefore, the total number of fluorescently labeled antibodies was divided into cocktails containing four antibodies per cocktail. (b) Replicate plates of treatment combinations were made for each cocktail. An acquisition template was created to store data from each sample well on a plate in the Flow Cytometry Standard 2.0 (FCS2.0) data format. (c) Markers were set to define the percent expression for any cell-surface marker based on isotype-negative controls. The data was collected using HT flow cytometry, and the data were extracted and stored in a database. (d) For each treatment, informatics was used to create a feature vector that summarizes statistics for each cell-surface marker. The statistics included the percentage of cells positive for a particular cell-surface marker, mean fluorescence of positive cells, median fluorescence of positive cells, and the coefficient of variation. (e) By data reduction, a composite signature of expression for all cell-surface markers in each treatment was created and is characterized by a two-dimensional array of 12 numerical values.

Download figure to PowerPoint

For each treatment, a feature vector was created that summarizes statistics for each cell-surface marker. The statistics included the percentage of cells positive for a particular cell-surface marker, mean fluorescence of positive cells, median fluorescence of positive cells, and the coefficient of variation. These statistics were collected to summarize the response of a cell to a combinatorial treatment and used to compose a profile of the cell's state. A composite signature for each treatment was defined as a collation of related output statistics. The individual markers in the experiment were chosen to identify characteristic expression of differentiated cells as uniquely as possible. Hence, for the purpose of creating signature profiles and exploring broad patterns in the output space, the distributions of marker expression were assumed to be independent of each other. In the case where a set of closely related markers are used in an experiment, co-expression and principal factors can be used to study the variance in the data set.

The collected data were validated with the help of positive, negative, and process controls on each plate. The treatment response in each well is normalized by comparing the fluorescence intensity distribution with untreated, unstained wells (negative controls) on the same plate. Replicate measurements for a treatment are combined by evaluating the median result for each marker.

The histograms for individual markers were assumed to be mutually exclusive. The collected data were validated with the help of positive, negative, and process controls on each plate. Day-to-day variation in fluorescence readings was assessed and corrected for by using cumulative distribution plots of control wells across all plates. The histograms were inspected manually to identify and handle insufficient or corrupted data from particular markers in a treatment.

Preprocessing of raw flow-cytometry data generates three views of experimental data. (1) The well data view corresponds to four-marker data from physical sample wells. (2) The formulation data view collates information from all markers used in the experiment and associates it with each combinatorial treatment. (3) The normalized data view handles variability in sample readings across plates by comparing the treatment response of a sample with that of untreated, stained wells (negative controls) on each plate. Replicates were evaluated by applying the median results for each marker.

The feature vectors are based on a particular statistic for selected markers to summarize the profile of a cell's state and the Euclidean distance between feature vectors for any two cells that defines the similarity between those two cells. In this study, informatics tools generated a feature vector to contain the percentage of cells positive for a certain cell-surface marker or stain (Fig. 1d). Finally, a composite signature of expression for all cell-surface markers in each treatment is characterized by a one dimensional array of 12 numerical values (Fig. 1e). An acquisition template was created to store data from each sample well on a plate in the Flow Cytometry Standard 2.0 (FCS2.0) data format.

Generation of Clustering Algorithms and Tartan Plots

Unsupervised clustering with the average-linkage hierarchical clustering algorithm (24, 25) was used to identify interesting partitions in composite-signature profiles. Because summary statistics were used in the feature vector, Euclidean distance and the Pearson correlation coefficient were used to define similarity between the profiles. Cluster analysis was performed to identify sets of co-expressed markers. Saved cluster information was used to interrogate the corresponding combinatorial treatments and to guide further experimentation in an iterative manner.

The data sets can be divided into several partitions by agglomerative hierarchical clustering, which groups highly similar cellular responses into a cluster. Similar clusters are progressively merged in a tree-like manner using a hierarchical clustering algorithm to form larger clusters. Finally, larger clusters are merged to form one tree in which each cellular response (a leaf node of the tree) is ordered so that similar responses are close to each other. This ordering can be observed in a hierarchical clustering Tartan plot (24, 25).

For example, Figure 2a is a Tartan plot that represents all cell-surface markers examined. In this plot, both the x-axis and y-axis show individual treatment combinations, so each point in the plot represents the comparison of two composite signatures from two treatment combinations. These points are color-coded according to the similarity between the two composite signatures at that point: similar treatment combination pairs are colored red (hot color), and dissimilar treatment combination pairs are colored blue (cold color). The hierarchical clustering algorithm, which orders the composite signatures in the plot, positions treatment combinations that produce the same result next to each other. This ordering results in the red points falling into squares along the diagonal (white dots) of the Tartan plot. Each red square located along the diagonal can be considered to represent a distinct cell state (outlined by black squares). The size of the red square represents how many treatment combinations cause differentiation of the cells into that particular cell state [Fig. 2a, compare squares 3, (large) and 4, (small)].

thumbnail image

Figure 2. Hierarchical clustering using all cell-surface markers to visualize changes in cell state. (a) A hierarchical clustering “Tartan” plot is based on the similarity between cell states. Each point in the plot represents the comparison of two composite signatures. The points are color-coded according to the distance between the two composite signatures at a given point, with similar composite signatures being colored red (hot) and dissimilar composite signatures being colored blue (cold). The composite signatures in these plots have been ordered by a hierarchical clustering algorithm, which puts similar composite signatures next to each other. This ordering results in the red points falling into squares along the diagonal of the Tartan plot. (b) The heat map plot (a profile of cells positive for marker expression for each of the indicated markers) is associated with the clustering Tartan plot in Figure 2a. For example, the cluster of composite signatures numbered 73–83 (upper white box) is characterized by high levels of CD11b and CD13 and CD33, where as the cluster of composite signatures numbered 65–72 (lower white box) is characterized by high levels of CD33 and CD13, but a low level of CD 11b.

Download figure to PowerPoint

In a Tartan plot, the colors surrounding the squares show how similar or dissimilar one cell state is from its neighbor. A square that has warm colors (yellow, orange) separating it from another square indicates that the cell states are similar. In Figure 2a, squares 1 and 2 are separated mostly by yellow. Conversely, a square that has cold colors (blue, green) separating it from another square indicates dissimilar cell states. Squares 1 and 3 are separated by blue.

The Tartan plot in Figure 2a is associated with a heat map (Fig. 2b), which is a kind of adjacent column chart that profiles the expression of cell-surface antigens. In Figure 2b, each number on the y axis represents a treatment. Each column on the x axis represents a cell-surface marker. The heat map is color-coded from dark green, indicating a cold region or low expression of the marker, to red, indicating a hot region or high expression of that marker. Heat maps are excellent for viewing flow cytometry data because the expression of one marker in all the experimental treatments can be easily observed by looking up one column from the x axis, and the expression of all markers in one treatment can be easily compared by looking across a row from the y axis.


  1. Top of page
  2. Abstract
  6. Acknowledgements

The acute promyelocytic leukemia cell line HL-60 was selected as the biological model for several reasons (26, 27). First, it is a very well-studied cell line and reproducing previously published data was necessary to validate the system. Second, the HL-60 cell line can respond to multiple biological and pharmacological factors and can differentiate along multiple pathways within the myeloid lineage (28–32). This multiplicity is necessary to assess multiple inputs and multiple endpoints in parallel. Third, this cell line shows evidence of nonlinear, non-obvious interactions, essential to prove TransForm's ability to explore the combinatorial space (33–40).

The HL-60 cells were exposed to five well-investigated pharmacological agents known to initiate differentiation in these cells: DMSO, vitamin D3, PMA, sodium butyrate, (pH 7.8), and ATRA. These agents, alone or in limited combinations, have been shown repeatedly to promote differentiation of HL-60 cells along three distinct pathways within the myeloid lineage: neutrophil, monocyte, and eosinophil/basophil (9, 29–32, 36, 41).

Tartan plots and heat maps can be generated by clustering either all cell-surface markers or a constrained subset of markers. Figure 2a shows a tartan plot that clusters all cell-surface markers. All individual red squares outlined in black are clusters of treatments that yielded a similar cellular profile but have a distinctly different phenotypic profile from the next red square outlined in black (compare squares 1–7). The squares are separated from each other by areas of yellow, green, or blue. The associated heat map in Figure 2b shows the relative intensities of marker expression for the treatment combinations in each square. For example, treatments 68–72 (lower white box) can be compared with square 4 in Figure 2a and treatments 73–82 (upper white box) can be compared with square 5 in Figure 2a.

If the analysis is constrained to cell-surface markers that represent a particular cell lineage (for example, markers representing the myeloid lineage), further stratification of phenotypes is observed. Tartan plots and heat maps were generated from clustering cell-surface markers that are restricted to myeloid cells (CD66b, CD11b/Mac1, CD13, CD14) (Figs. 3a and 3b) (13, 42–45). The Tartan plot shows a series of red squares along the diagonal separated by some yellow and a lot of blue, meaning that distinct cellular profiles arise based on experimental treatments (Fig. 3a). In fact, a pattern that is most representative of monocytes (CD66b, CD11b/Mac-1+, CD 13+, CD14+) (13, 42–45) can be viewed in the heat map (Fig. 3a, compare formulations 29–45 black box with Fig. 3b, formulations 29–45 white box).

thumbnail image

Figure 3. Hierarchical clustering using cell-surface markers for a specific cell lineage. (a) Tartan visualization. To validate the expected cell states, the data was clustered using markers specific for cells of the myeloid lineage (CD66b, CD11b/Mac-1, CD13, and CD 14). Red boxes along the diagonal line indicate formulations producing states similar to one another. (b) A heat map shows the percentage of positive cells for each of the indicated markers associated with the clustering Tartan plot in Figure 3a. For example formulations 30–45 show low expression of CD66b and high expression of CD 11b, CD13, and CD14, indicative of a monocyte cell. The black box around formulations 30–45 in Figure 3a corresponds to the white box around the same formulation in Figure 3b.

Download figure to PowerPoint

One convenient feature developed in this software is the ability to retrieve the treatment combinations giving rise to cells exhibiting a profile of interest (Fig. 4a, blue highlighted area). In this case, treatments 29 through 45, which produced the “monocyte-like” pattern in the heat map, contained vitamin D3, PMA, or both. These results are consistent with literature findings demonstrating that monocytes are produced following treatment of HL-60 cells with vitamin D3, PMA, or both. The difference in expression patterns between treatment wells can be examined by viewing the composite signatures for each of the wells that produced a particular profile (Fig. 4b).

thumbnail image

Figure 4. Retrieval of formulations producing a particular cell type. (a) Clicking on a square cluster in a Tartan plot opens the formulation viewer, which shows the formulations that give rise to that particular cell type (blue highlighted area). (b) A composite signature viewer for that particular formulation and marker set can also be displayed, so that each individual composite signatures for a particular treatment in a cluster can be quantitatively viewed. Each line is a composite signature for an individual treatment. All the signatures resulting from treatments 30–45 are very similar, which means that the cellular profiles are similar.

Download figure to PowerPoint

An advantage of data-mining is the ability to discover nonlinear or unexpected events, such as factor dominance, unexpected cell cooperativity, or synergies and anergies (data not shown). Factor dominance occurs when the combination of two treatment factors produces a profile that resembles a profile produced by one of the factors and not the other. This was observed when PMA, 0.81 nM (PMAlo) (Fig. 5a) was combined with sodium butyrate, 100 μM, (sodium butyratelo) (Fig. 5b). The resultant profile was most similar to that of PMAlo (Fig. 5c). Dominance by one factor in one treatment did not, however, predict the outcome when that factor was used in a different treatment. When DMSO, 0.18 M (DMSOmed) (Fig. 5d) was combined with PMAlo (Fig. 5a), the resultant profile was most similar to that of DMSO (Fig. 5e), indicating that, for this treatment paradigm, DMSO and not PMA was the dominant factor.

thumbnail image

Figure 5. Combinations reveal nonlinear, non-obvious interactions between factors. (ae) Binary combinations can demonstrate factor dominance. Signature profiles show the percentage of cells positive for the cell-surface markers CD66b, CD11b/Mac-1, CD 13, and CD 14. The signature profile of cells treated with a combination of PMAlo + sodium butyratelo (c) is more similar to that of cells treated with PMAlo alone (a) than to that of cells treated with sodium butyratelo alone (b). PMA does not always act as the dominant factor, however. When PMAlo (a) is combined with DMSOmed (d), DMSO behaves as the dominant factor (e). (fh) Binary combinations result in non-obvious interactions between factors. A “bin” is a grouping of antibodies with the same fluorescent label due to the low probability for expression in these populations. Here CD34, CD3, B220, and CD56 were all labeled with FITC. Neither sodium butyratemed (f) nor DMSOmed (g) alone produce cells positive for CD 125w, whereas cells treated with both sodium butyratemed and DMSOmed (h) produce cells expressing high levels of CD125w. Higher order combinations result in unexpected interactions between differentiation factors. (i, j) The ternary combination of DMSOhi+NaButlo+RAlo produced a phenotype distinctive from those resulting from respective unary and binary inputs. DMSOhi+NaButlo+RAlo treatment results in down regulation of CD83 and CD235 (black arrows) suggesting negative synergies between these factors. These synergies would not have been predicted from the unary and binary data. (j) Changing concentrations of the differentiation factors in a ternary mix produces distinct phenotypes, illustrating complexity of the DFs interactions.

Download figure to PowerPoint

Unexpected cell cooperativity occurs when a combination treatment produces a unique profile different from the profiles of the individual treatments. This is illustrated by sodium butyratemed 300 μM, (sodium butyratemed) (Fig. 5f) and DMSOmed (Fig. 5g). When present alone, each produced cells negative for cell-surface marker CD 125w. However, when used in combination these factors produced a population expressing high levels of the cell-surface marker CD125w (Fig. 5h).

Another example of unexpected cooperativity is the ternary combination of DMSO, 0.15 M, (DMSOlo), sodium butyrate, 500 μM (sodium butyratehi), and ATRA, 500 nM (ATRAhi), which produced a profile that had different levels (much lower) of expression of CD83 and CD 235a when compared to the profiles of the individual factors and of the binary combinations (black arrows, Fig. 5i). Furthermore, the combination of DMSO, sodium butyrate, and ATRA at different concentrations resulted in different signature profiles (Fig. 5j). These results show that factor type, combination, and concentration are all important for determining cell phenotype.

Data mining allowed us to identify and retrieve atypical cellular profiles that are not obvious on visual inspection alone. For example, some treatment combinations produced cell populations exhibiting high expression of lymphocyte markers (CD3, B220), hematopoietic stem cell (HSC) markers (CD 34), and erythrocytic markers (CD235) (Fig. 6a). Moderate to high expression of myeloid lineage markers (CD66b, CD11b, CD33, CD13, CD14) were also observed for these profiles (Fig. 6a). At the time of publication, no known cell population expressing high levels of multiple lineage markers had been reported, and such cell populations are considered atypical. Interestingly, all treatments producing this atypical cellular profile contained both PMA and DMSO, whereas treatments containing either PMA or DMSO alone did not produce this atypical cell profile (Fig. 6b, highlighted in blue).

thumbnail image

Figure 6. Hierarchical clustering reveals functional cell states. (a) HL-60 cells treated with binary combinations of PMA and DMSO produce abnormal composite signatures. (b) The highlighted blue box shows the formulations that produced the abnormal composite signatures in Figure 6a. In this case, all formulations contained both PMA and DMSO. (c) Flow cytometric analysis of HL60s treated with 2 ng/ml idarubicin. Extent of apoptosis is indicated by 7-AAD and Annexin 5 markers. HL-60s treated with both PMA and DMSO are more sensitive to apoptosis following treatment with idarubicin than those treated with PMA or DMSO alone. (i) Undifferentiated cells, no idarubicin; (ii) Undifferentiated cells, 2 ng/ml idarubicin; (iii) Cells treated to differentiate with DMSOhi + 2 ng/nl idarubicin; (iv) Cells treated to differentiate with PMAlo + 2 ng/ml idarubicin; (v) Cells treated to differentiate with PMAlo + DMSOhi, no idarubicin; (vi) Cells treated to differentiate with PMAlo+ DMSOhi + 2 ng/ml idarubicin. (d) HL-60 cells treated to undergo differentiation with both PMA and DMSO show a synergistic response at low doses of idarubicin compared to cells treated with PMA or DMSO alone and control cells that have been not been stimulated to differentiate.

Download figure to PowerPoint

The identification of unusual cellular phenotypes provides a launching point to explore the biology of those populations. We proposed that unusual phenotypes may be associated with altered sensitivity to chemotherapeutic drugs (39, 46, 47). To test this hypothesis, differentiation was induced in the presence of PMA alone, DMSO alone, or mixtures of PMA and DMSO over 5 days (Fig. 6). The differentiated cell populations were subsequently treated with idarubicin, an anthracyclin derived antibiotic, commonly used to promote apoptosis in leukemic cells (37) Treatments containing both PMA and DMSO accelerated cell apoptosis, compared with treatments of either PMA or DMSO alone (Fig. 6c). This acceleration was most evident at idarubicin doses of less than or equal to 10 ng/ml, whereas higher idarubicin doses induced all cells to undergo apoptosis (Fig. 6c). Recent studies by Marekova et al. have shown that PMA and DMSO individually render HL-60 cells resistant to idarubicin-induced apoptosis (47), indicating that the combination of PMA and DMSO changes the susceptibility of cells to idarubicin (9, 37). Furthermore, similar increases in idarubicin-induced apoptosis were observed in combinations of PMA and other factors (ATRA, sodium butyrate, or vitamin D3). No such increases occurred with DMSO or combinations of DMSO with other factors (data not shown).


  1. Top of page
  2. Abstract
  6. Acknowledgements

The three techniques of HT experimentation, flow cytometry, and hierarchical clustering have been combined and used successfully to detect different cell phenotypes resulting from various combinations of multifactor treatments. Incorporating HT-automated dispensing platforms and informatic data mining tools into traditional assays has allowed the execution and analysis of more complex experiments which are thus becoming more mainstream. In this study, HT experimentation was extended to another application: flow cytometry.

Despite the pronounced advantages of HT experiments for cell-based assays, the use of flow cytometry in HT experimentation has not been widely used until recently (20). The introduction of HT samplers allows experiments to be performed at higher throughput and on a smaller scale, thus using less material and becoming more economical. The adoption of high throughput applications to flow cytometry will result in the generation of larger and more complex data sets. Conventional flow cytometry data mining applications are inadequate to handle this increased data complexity and therefore data sets can quickly become too difficult to manage and the identification of trends within the data set can often be missed. Here, the application of hierarchical clustering and other data-mining applications, now common place (e.g., in genomics research) apply to these large, complex data sets.

Hierarchical clustering not only identified trends in the data sets but also revealed nonlinear, non-obvious responses in the data. In this case, experiment-wide patterns of response under different treatment scenarios were examined. The expression of all cell-surface markers for a particular treatment and the expression of any one cell-surface marker over all treatments can be assessed from one plot. To confirm the validity of this approach, both the positive controls and the treatments that produced them were correctly identified and correlated to previously published results (31, 32, 36, 40, 42–45, 48, 49).

Data mining allowed us to identify and retrieve atypical cellular profiles that are not obvious on visual inspection alone. The ability to discover atypical cell profiles and explore their physiological relevance is facilitated by hierarchical clustering and provides a launching point to explore the biology of a particular cell population. Without the added ability to mine large data sets, opportunities to discover and provide insight into interesting biological phenomena may be missed. In this paper, this principle was illustrated by identifying a subpopulation of HL-60 cells that had enhanced susceptibility to chemotherapy-induced apoptosis in follow-up manual experimentation. Therefore, a potential application of these discovery methods may be that it is a useful tool for producing treatments of future clinical value. This could be accomplished either by identifying new combination treatments or if the mechanisms of the factors producing the desired effect are known, other agents which utilize the same pathways or molecules known to interact with these pathways may be used in subsequent combinatorial screens.

The incorporation of flow cytometry into a high-throughput platform has advantages over other common techniques that incorporate high-throughput dispensing platforms or advanced data-mining techniques. One such technique is molecular profiling, in which discrete phenotypes are identified by changes in gene expression rather than protein expression. Therefore, their identification is susceptible to effects from post-translational modification; so the expression profiles may not solely determine the final phenotype. Flow cytometry has the advantage that phenotypes are captured on the basis of protein expression; a more physiologically relevant parameter.

Molecular profiling of individual cells is also cumbersome. Most cell populations are heterogeneous mixtures of subpopulations. This heterogeneity is not compensated for in RNA and DNA microarray analyses, so the results reflect a mixture, rather than a single population of cells. Consequently, absolutely pure cell populations must be analyzed. Flow cytometry has the advantage that it does not require pure cell populations because it can quickly sort and examine large numbers of cells individually and separate them according to specific antigenic profiles.

Another advantage of incorporating flow cytometry is that it allows for multiparametric data capture of cell populations exposed to multifactor combinations. Multifactor combinations are rarely tested in traditional experiments and are often limited to testing only one or two components at a time. Thus the in vivo microenvironment can not be fully captured in vitro, resulting in poor correlation between in vitro and in vivo experiments. For example, in vivo, cells are targeted with multiple stimuli that occur either in combination or sequentially and include soluble and insoluble factors (50, 51), mechanical and electrical signals (50), osmotic pressure (52), hydrophobicity (53), and oxygen tension (52). In these experiments, the experimental design included both multiparameter inputs and multiparameter outputs. Performing experiments using multiple inputs and outputs is important, particularly in cell-based assays and provides rich data sets for the evaluation of trends.

While attempting to improve the overall sensitivity of the algorithms for detecting phenotypic profiles, some limitations of the technique were detected. One limitation was that not all formulations known to produce monocytes were discovered. For example, treatments containing vitamin D3lo alone were not observed in the monocyte cluster, although they do, in fact, produce monocytes. The most persistent limitation was that some profiles fell into multiple clusters. This limitation occurred because some differences between pairs of 16-dimensional histograms were lost in data reduction. For instance, the measure works well when using cell-surface markers that shift from either zero to a positive or a positive to zero surface marker expression. However, more subtle shifts in the mean fluorescence intensity are not captured as well by the measure. Alternatively, a distribution-free similarity measure such as the Kolmogorov-Smirnov measure that evaluates the difference between cumulative distributions from the profiles would also be able to handle cases like these (54). In future work, algorithms will be designed that better assess small shifts in mean fluorescent intensity for a particular marker.

Another limitation of hierarchical clustering was the selection of cell biomarkers. The proper selection of cell biomarkers is crucial for accurate experimental analysis. The markers should be selected only if they are found exclusively on the phenotype being sought. For instance, a cell-surface marker found on multiple cellular phenotypes is not as helpful as one found on a cell induced to differentiate into only one cell type. Finally, the proper labeling of cells by the antibodies must be verified. If an antibody is defective, then it provides unreliable information, and data may be misinterpreted.

Calibration and normalization of antibodies were not a concern for this application in flow cytometry. The responses of the antibodies in each well were normalized by comparing the fluorescence intensity distribution with untreated, unstained wells (negative controls) on the same plate. Day-to-day variation in fluorescence readings was assessed and corrected for by using cumulative distribution plots of control wells across all plates. A composite signature profile was generated for each unique treatment combination and compared against every other treatment combination in the experiment. Therefore, relative rather than absolute values of antibody expression were most important for our purposes.

The experimental goal for this paper was to validate our approach using the HL-60 cell line. The benefit of using a cell line for this purpose included easier use and handling, less sensitivity to environmental manipulation, result predictability, and a generous literature landscape for comparison of results. However, there are exciting potential applications for using HT combinatorial platforms together with HT flow cytometry and informatic data mining in primary isolated cells. For example, finding the combination of factors necessary to sustain embryonic stem cells in their undifferentiated state without the need for a feeder layer and/or finding the combinations of factors that provide more directed cell differentiation (55). Another application that could benefit from this type of approach is autologous cell therapy, a process that incubates specific cytokines in a medium containing an autograft, allowing for selective amplification of particular cell populations, which can then be returned to the donor (56, 57). This technique has often been hindered by the difficulty optimizing the many conditions necessary to produce the large, optimal cell number of the desired population necessary (58). The common challenge of these two applications is to develop and define methods that promote directed differentiation or expand the desired populations of cells without becoming contaminated with undesired cell populations. By implementing the approach described here, perhaps some unknown parameters concerning these fields of research can be revealed.

New automated ways to manipulate the complex input variables, along with data-mining tools that can analyze the large amounts of data produced from these experiments have proven successful. They can be used to model the in-vivo situation more accurately than possible with other methods. Therefore, HT flow cytometry coupled with hierarchical clustering has an important place in the investigation of biological systems in which multiple experimental parameters may be manipulated.


  1. Top of page
  2. Abstract
  6. Acknowledgements

The authors would like to thank Randy Forsyth, Colin Gardner, Sherry Morissette, and Evie Sun for their critical review of this paper, as well as Joe Bielitski and Laura Keating.


  1. Top of page
  2. Abstract
  6. Acknowledgements
  • 1
    Kariv I,Rourick R,Kassel D,Chung T. Improvement of “hit to lead” optimization by integration of in vitro HTS experimental models for early determination of pharmokinetic properties. Comb Chem High Throughput Screen 2002; 5: 459472.
  • 2
    Cox B,Denyer JC,Binnie A,Donnelly MC,Evans B,Green DV,Lewis JA,Mander TH,Merritt AT,Valler MJ,Watson SP. Application of high-throughput screening techniques to drug discovery. Prog Med Chem 2000; 37: 83133.
  • 3
    Davey HM,Kell DB. Flow cytometry and cell sorting of heterogeneous microbial populations: The importance of single-cell analyses. Microbiol Rev 1996; 60: 641696.
  • 4
    Herzenberg LA,Parks D,Sahaf B,Perez O,Roederer M,Herzenberg LA. The history and future of the fluorescence activated cell sorter and flow cytometry: A view from Stanford. Clin Chem 2002; 48: 18191827.
  • 5
    Li L,Darden D,Weinberg C,Levine A,Pederson L. Gene assessment and sample classification for gene expression data using a genetic algorithm/k-nearest neighbor method. Comb Chem High Throughput Screen 2001; 4: 727739.
  • 6
    Nagorsen D,Wang E,Panelli M. High-throughput genomics: An emerging tool in biology and immunology. Am Soc Histocompatibility Immunogenet 2003; 9698.
  • 7
    Chang C. The current novel high-throughput technology landscape. Drug Discov Dev 2004; 7: S14S17.
  • 8
    Stegmaier K,Ross KN,Colavito SA,O'Malley S,Stockwell BR,Golub TR. Gene expression-based high-throughput screening(GE-HTS) and application to leukemia differentiation. Nat Genet 2004; 36: 257263.
  • 9
    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: 24582462.
  • 10
    Spradling A,Drummond-Barbosa D,Kai T. Stem cells find their niche. Nature 2001; 414: 98104.
  • 11
    Streuli C. Extracellular matrix remodelling and cellular differentiation. Curr Opin Cell Biol 1999; 11: 634640.
  • 12
    Behonick DJ,Werb Z. A bit of give and take: The relationship between the extracellular matrix and the developing chondrocyte. Mech Dev 2003; 120: 13271336.
  • 13
    Kitano H. Systems biology: A brief overview. Science 2002; 295: 16621664.
  • 14
    Abbott A. Cell culture: Biology's new dimension. Nature 2003; 424: 870872.
  • 15
    Davidson EH,Rast JP,Oliveri P,Ransick A,Calestani C,Yuh CH,Minokawa T,Amore G,Hinman V,Arenas-Mena C,Otim O,Brown CT,Livi CB,Lee PY,Revilla R,Rust AG,Pan Z,Schilstra MJ,Clarke PJ,Arnone MI,Rowen L,Cameron RA,McClay DR,Hood L,Bolouri H. A genomic regulatory network for development. Science 2002; 295: 16691678.
  • 16
    Wolf K,Mazo I,Leung H,Engelke K,von Andrian UH,Deryugina EI,Strongin AY,Brocker EB,Friedl P. Compensation mechanism in tumor cell migration: Mesenchymal-amoeboid transition after blocking of pericellular proteolysis. J Cell Biol 2003; 160: 267277.
  • 17
    Borisy AA,Elliott PJ,Hurst NW,Lee MS,Lehar J,Price ER,Serbedzija G,Zimmermann GR,Foley MA,Stockwell BR,Keith CT. Systematic discovery of multicomponent therapeutics. Proc Natl Acad Sci USA 2003; 100: 79777982.
  • 18
    van der Pouw Kraan TC,Kasperkovitz PV,Verbeet N,Verweij CL. Genomics in the immune system. Clin Immunol 2004; 111: 175185.
  • 19
    Hertzberg RP,Pope AJ. High-throughput screening: New technology for the 21st century. Curr Opin Chem Biol 2000; 4: 445451.
  • 20
    Edwards BS,Oprea T,Prossnitz ER,Sklar LA. Flow cytometry for high-throughput, high-content screening. Curr Opin Chem Biol 2004; 8: 392398.
  • 21
    Irish JM,Hovland R,Krutzik PO,Perez OD,Bruserud O,Gjertsen BT,Nolan GP. Single cell profiling of potentiated phospho-protein networks in cancer cells. Cell 2004; 118: 217228.
  • 22
    Perez OD,Nolan GP. Simultaneous measurement of multiple active kinase states using polychromatic flow cytometry. Nat Biotechnol 2002; 20: 155162.
  • 23
    Sachs K,Perez O,Pe'er D,Lauffenburger DA,Nolan GP. Causal protein-signaling networks derived from multiparameter single-cell data. Science 2005; 308: 523529.
  • 24
    Duda R,Hart P. Pattern Classification and Scene Analysis. New York: Wiley; 1973.
  • 25
    Eisen MB,Spellman PT,Brown PO,Botstein D. Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA 1998; 95: 1486314868.
  • 26
    Collins SJ. The HL-60 promyelocytic leukemia cell line: Proliferation, differentiation, and cellular oncogene expression. Blood 1987; 70: 12331244.
  • 27
    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, α1,25-dihydroxyvitamin D3 or 12-O-tetradecanoyl phorbol-13-acetate. Leuk Res 1998; 22: 537547.
  • 28
    Collins SJ,Ruscetti FW,Gallagher RE,Gallo RC. Normal functional characteristics of cultured human promyelocytic leukemia cells (HL-60) after induction of differentiation by dimethylsulfoxide. J Exp Med 1979; 149: 969974.
  • 29
    Fischkoff SA,Pollak A,Gleich GJ,Testa JR,Misawa S,Reber TJ. Eosinophilic differentiation of the human promyelocytic leukemia cell line, HL-60. J Exp Med 1984; 160: 179196.
  • 30
    Hutt-Taylor SR,Harnish D,Richardson M,Ishizaka T,Denburg JA. Sodium butyrate and a T lymphocyte cell line-derived differentiation factor induce basophilic differentiation of the human promyelocytic leukemia cell line HL-60. Blood 1988; 71: 209215.
  • 31
    McCarthy DM,San Miguel JF,Freake HC,Green PM,Zola H,Catovsky D,Goldman JM. 1,25-dihydroxyvitamin D3 inhibits proliferation of human promyelocytic leukaemia (HL60) cells and induces monocyte-macrophage differentiation in HL60 and normal human bone marrow cells. Leuk Res 1983; 7: 5155.
  • 32
    Rovera G,Santoli D,Damsky C. Human promyelocytic leukemia cells in culture differentiate into macrophage-like cells when treated with a phorbol diester. Proc Natl Acad Sci USA 1979; 76: 27792783.
  • 33
    Brown G,Bunce CM,Rowlands DC,Williams GR. All-trans retinoic acid and 1 α,25-dihydroxyvitamin D3 co-operate to promote differentiation of the human promyeloid leukemia cell line HL60 to monocytes. Leukemia 1994; 8: 806815.
  • 34
    Bunce CM,Wallington LA,Harrison P,Williams GR,Brown G. Treatment of HL60 cells with various combinations of retinoids and 1 α,25 dihydroxyvitamin D3 results in differentiation towards neutrophils or monocytes or a failure to differentiate and apoptosis. Leukemia 1995; 9: 410418.
  • 35
    Kang SN,Chung SW,Kim TS. Capsaicin potentiates 1,25-dihydoxyvitamin D3- and all-trans retinoic acid-induced differentiation of human promyelocytic leukemia HL-60 cells. Eur J Pharmacol 2001; 420: 8390.
  • 36
    Tanaka H,Abe E,Miyaura C,Kuribayashi T,Konno K,Nishii Y,Suda T. 1 α,25-Dihydroxycholecalciferol and a human myeloid leukaemia cell line (HL-60). Biochem J 1982; 204: 713719.
  • 37
    Tsuda T,Wong D,Dolovich J,Bienenstock J,Marshall J,Denburg JA. Synergistic effects of nerve growth factor and granulocyte-macrophage colony-stimulating factor on human basophilic cell differentiation. Blood 1991; 77: 971979.
  • 38
    Verlinden L,Verstuyf A,Mathieu C,Tan BK,Bouillon R. Differentiation induction of HL60 cells by 1,25(OH)2D3, all trans retinoic acid, rTGF-β2 and their combinations. J Steroid Biochem Mol Biol 1997; 60: 8797.
  • 39
    Wallington LA,Bunce CM,Durham J,Brown G. Particular combinations of signals, by retinoic acid and 1 α, 25 dihydroxyvitamin D3, promote apoptosis of HL60 cells. Leukemia 1995; 9: 11851190.
  • 40
    Zheng X,Chang RL,Cui XX,Kelly KA,Shih WJ,Lin Y,Strair R,Suh J,Han ZT,Rabson A,Conney AH. Synergistic effects of clinically achievable concentrations of 12-O-tetradecanoylphorbol-13-acetate in combination with all-trans retinoic acid, 1α,25-dihydroxyvitamin D3, and sodium butyrate on differentiation in HL-60 cells. Oncol Res 2000; 12: 419427.
  • 41
    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: 29362940.
  • 42
    Altieri DC,Edgington TS. A monoclonal antibody reacting with distinct adhesion molecules defines a transition in the functional state of the receptor CD11b/CD18 (Mac-1). J Immunol 1988; 141: 26562660.
  • 43
    Goyert SM,Ferrero E,Rettig WJ,Yenamandra AK,Obata F,Le Beau MM. The CD14 monocyte differentiation antigen maps to a region encoding growth factors and receptors. Science 1988; 239: 497500.
  • 44
    Sakai K,Hattori T,Sagawa K,Yokoyama M,Takatsuki K. Biochemical and functional characterization of MCS-2 antigen (CD13) on myeloid leukemic cells and polymorphonuclear leukocytes. Cancer Res 1987; 47: 55725576.
  • 45
    Watt SM,Sala-Newby G,Hoang T,Gilmore DJ,Grunert F,Nagel G,Murdoch SJ,Tchilian E,Lennox ES,Waldmann H. CD66 identifies a neutrophil-specific epitope within the hematopoietic system that is expressed by members of the carcinoembryonic antigen family of adhesion molecules. Blood 1991; 78: 6374.
  • 46
    Ketley NJ,Allen PD,Kelsey SM,Newland AC. Mechanisms of resistance to apoptosis in human AML blasts: The role of differentiation-induced perturbations of cell-cycle checkpoints. Leukemia 2000; 14: 620628.
  • 47
    Marekova M,Vavrova J,Vokurkova D. Dose dependent biological effects of idarubicin in HL-60 cells: Alterations of the cell-cycle and apoptosis. Acta Medica (Hradec Kralove) 2000; 43: 6973.
  • 48
    Matsuhisa T,Mori Y. Mode of differentiation of human promyelocytic leukemia cell line, HL-60, by 1 α,25-dihydroxyvitamin D3. Blood Cells Mol Dis 1995; 21: 4248.
  • 49
    Miyaura C,Abe E,Kuribayashi T,Tanaka H,Konno K,Nishii Y,Suda T. 1 α,25-Dihydroxyvitamin D3 induces differentiation of human myeloid leukemia cells. Biochem Biophys Res Commun 1981; 102: 937943.
  • 50
    Ingber DE. Mechanical signaling and the cellular response to extracellular matrix in angiogenesis and cardiovascular physiology. Circ Res 2002; 91: 877887.
  • 51
    Lelievre S,Weaver VM,Bissell MJ. Extracellular matrix signaling from the cellular membrane skeleton to the nuclear skeleton: A model of gene regulation. Recent Prog Horm Res 1996; 51: 417432.
  • 52
    Garcia AM,Black AC,Gray ML. Effects of physicochemical factors on the growth of mandibular condyles in vitro. Calcif Tissue Int 1994; 54: 499504.
  • 53
    Saltzman WM. Principles in Tissue Engineering. San Diego: Academic Press; 2000 pp 221235.
  • 54
    Knuth E. The Art of Computer Programming. Reading, MA: Addison-Wesley; 1998. pp 4552.
  • 55
    Semb H. Human embryonic stem cells: Origin, properties and applications. APMIS 2005; 113: 743750.
  • 56
    Lundell BI,Vredenburgh JJ,Tyer C,DeSombre K,Smith AK. Ex vivo expansion of bone marrow from breast cancer patients: Reduction in tumor cell content through passive purging. Bone Marrow Transplant 1998; 22: 153159.
  • 57
    Williams SF,Lee WJ,Bender JG,Zimmerman T,Swinney P,Blake M,Carreon J,Schilling M,Smith S,Williams DE,Oldham F,Van Epps D. Selection and expansion of peripheral blood CD34+ cells in autologous stem cell transplantation for breast cancer. Blood 1996; 87: 16871691.
  • 58
    Zanstra P,Eaves C,Piret J. Environmental requirements of hematopoietic progeneitor cells in ex vivo expansion systems. In: SchindhelmK, NordanR, editors. Ex Vivo Cell Therapy. San Diego: Academic Press; 1999. pp 246264.