This protocol describes the use of centrifugal elutriation to separate primary acute lymphoblastic leukemia cells into different cell cycle phases.
The ability to synchronize cells has been central to advancing our understanding of cell cycle regulation. Common techniques employed include serum deprivation; chemicals which arrest cells at different cell cycle phases; or the use of mitotic shake-off which exploits their reduced adherence. However, all of these have disadvantages. For example, serum starvation works well for normal cells but less well for tumor cells with compromised cell cycle checkpoints due to oncogene activation or tumor suppressor loss. Similarly, chemically-treated cell populations can harbor drug-induced damage and show stress-related alterations. A technique which circumvents these problems is counterflow centrifugal elutriation (CCE), where cells are subjected to two opposing forces, centrifugal force and fluid velocity, which results in the separation of cells on the basis of size and density. Since cells advancing through the cycle typically enlarge, CCE can be used to separate cells into different cell cycle phases. Here we apply this technique to primary acute lymphoblastic leukemia cells. Under optimal conditions, an essentially pure population of cells in G1 phase and a highly enriched population of cells in G2/M phases can be obtained in excellent yield. These cell populations are ideally suited for studying cell cycle-dependent mechanisms of action of anticancer drugs and for other applications. We also show how modifications to the standard procedure can result in suboptimal performance and discuss the limitations of the technique. The detailed methodology presented should facilitate application and exploration of the technique to other types of cells.
Cultured cells typically grow asynchronously and individual cells are present in different phases of the cell cycle. Some organisms exhibit naturally synchronous cell cycles or can be synchronized by specific physiological stimuli. For example, nuclei within giant plasmodia of the slime mold Physarum polycephalum divide in a highly synchronous fashion1, and cells of the green algae Desmodesmus quadricauda can be synchronized by alternating light and dark periods2. While such organisms offer unique experimental attributes, they inadequately model the complexities of mammalian cells. The ability to artificially synchronize mammalian cell populations has been central to advancing our understanding of cell cycle regulation and the molecular basis of cell cycle checkpoints. Common methods include serum deprivation, chemical block and release, or exploiting physical characteristics3,4. Withdrawal of serum often causes cells to enter quiescence, and the re-addition of serum promotes cell cycle reentry into the G1 phase5. Chemical inhibitors include agents such as excess thymidine or hydroxyurea which block cells at the G1/S boundary, or microtubule inhibitors which typically arrest cells in the M phase3,4. Approaches that exploit physical characteristics include mitotic shake-off which can be used to enrich for mitotic cells since they are less adherent than interphase cells6. However, all of these techniques have potential drawbacks. For example, not all cell types maintain viability in the absence of serum or the presence of chemical inhibitors, and the yield of cells after mitotic shake-off is limited without prior synchronization with mitotic inhibitors.
With the exception of early embryonic cell cycles, where cells progressively decrease in size as they divide7, most cells advancing through the cycle undergo growth phases and become larger. This property is exploited in the technique of counterflow centrifugal elutriation (CCE), which can be used to separate cells differing in size and hence in cell cycle phase8,9. During CCE, cells are under the influence of two opposing forces: centrifugal force, which drives the cells away from the axis of rotation, and fluid velocity (counterflow), which drives the cells towards the axis of rotation (Figure 1). Cells in the elutriation chamber reach an equilibrium position where these forces are equal. The key factors dictating the equilibrium position are cell diameter and density. As the flow rate of the buffer solution is increased and the counterflow drag force outweighs the centrifugal force, a new equilibrium is established, causing a change in position of cells inside the chamber. All cells are shifted toward the chamber exit, resulting in smaller ones leaving the chamber first, whereas the larger cells stay within the chamber until the counterflow rate is increased sufficiently to promote their exit. The cells escaping the elutriation chamber with consecutive increases in counterflow rate can be collected in specific fractions and each fraction contains cells of sequentially increasing sizes. Instead of conducting elutriation with incremental increases in counterflow rate, sequential decreases in centrifugal force by decreasing centrifugation speed will accomplish the same result. The process of elutriation requires optimization depending on cell type, elutriation buffer employed, and the specific apparatus used. The rate of sedimentation of cells under these conditions is best described by Stokes Law: SV = [d2(ρp-ρm)/18η].ω2r, where SV = sedimentation velocity; d = diameter of the particle; ρp = density of the particle; ρm = density of the buffer; η = viscosity of the buffer; ω = angular velocity of the rotor; and r = radial position of the particle. Thus, SV is proportional to cell diameter and density, and since diameter is raised to the second power, it contributes more significantly than density which is generally constant through the cell cycle.
An important advantage of CCE is that cells are not subjected to harsh conditions of chemical treatment or nutritional deprivation and can be recovered in excellent yield essentially unperturbed. A requirement is that the cells in question undergo a significant increase in diameter, of at least 30%, as they traverse the cell cycle. The main disadvantage is the cost of the specialized centrifuge, rotor, and accessories. Nonetheless, the ability to prepare cells enriched in specific cell cycle phases by CCE has greatly facilitated cell cycle research in organisms from yeast to mammalian cells9,10. Moreover, recent advances are expanding uses to the separation of cells from healthy versus cancerous tissue, the separation of different cell types in heterogeneous mixtures, and to the production of specific cell types for immunotherapy9.
Our major interest has been in understanding the mechanism of action of microtubule targeting agents (MTAs) such as vinca alkaloids and taxanes. These drugs were thought to act exclusively in mitosis, blocking spindle microtubule function11,12, but recent evidence suggests they may also target interphase microtubules13,14. We recently reported that primary acute lymphoblastic leukemia (ALL) cells undergo death in either the G1 phase or in the M phase when treated with vincristine and other MTAs in a cell cycle-dependent manner15. The ability to separate ALL cells into different cell cycle phases by CCE was instrumental in reaching this conclusion. In this article, we describe the technical details and offer practical tips for the application of CCE to the preparation of primary ALL cells in different cell cycle phases.
NOTE: This protocol has been optimized for primary B-ALL cells which range in diameter from approximately 8 µm (G1 phase) to 13 µm (G2/M phases). Thus, the specific centrifuge speeds and pump flow rates may not be applicable to cells with different size ranges. Nevertheless, appropriate elutriation parameters can be estimated using the sedimentation rate equation. Cells are cultured in a defined serum-free medium as described16 at an optimal density of 1 – 3 x 106 cells/mL. With the exception of the collection of fractions which is conducted at 4 °C using ice-cooled tubes, all steps are conducted at room temperature and the centrifuge set at 20 °C with a maximum of 24 °C.
1. Preparation of Elutriation Buffer and Materials
2. Assembly of the Elutriation System
3. Priming of the Elutriation System
4. Preparation of the Cell Sample
5. Introduction of Cell Sample into the Elutriation System and Collection of Fractions
6. Flushing of the Elutriation System
7. Analysis of Fractions and Collection of the G1 or G2/M Pools.
Note: In order to demonstrate the effectiveness of elutriation, each fraction is analyzed for cell count, cell diameter, and DNA content, and pools are created of cells in the G1 or G2/M phases for immunoblot analyses, as described in Representative Results. This protocol section will describe how to create these pools and prepare them for experimental use.
Primary ALL cells (3 – 4 x 108) were subjected to centrifugal elutriation and collected into two wash fractions and twenty main fractions, as described in the protocol. Table 1 shows representative data where the total number of cells in each fraction as well as the corresponding rotor speed are presented. Overall yield was typically over 80%. Measurement of cell diameter in individual fractions confirmed that cell diameter increased with advancing elutriation (Figure 2). These results represent averages of two independent determinations and reflect the high degree of data reproducibility. Fractions were next subjected to propidium iodide staining and flow cytometry to determine DNA content (Figure 3). Early fractions (F1 – F5) contained almost exclusively cells with 2N DNA content representing G1 phase. A mixture of cells in G1 phase and S phase was observed in intermediate fractions (F6 – F9) and cells with 4N DNA content were observed starting in F10. Later fractions had mainly cells with 4N DNA content representing G2/M phases. These data together with the results of Figure 2 confirmed a relationship between cell size and cell cycle phase.
Based on these results, fractions F1 – F5 were combined to create a pool of cells in G1 phase, and fractions F17 – F20 were combined to create a pool of cells in G2/M phase. The proportion of cells with 2N DNA content in the G1 phase pool was typically 99%, and the proportion of cells with 4N DNA content in the G2/M pool was typically 85%, as determined by propidium iodide staining and flow cytometry. Under asynchronous conditions, 60 – 70% of primary ALL cells are in G1 phase and 15 – 20% are in G2/M phases15, thus the values obtained after elutriation reflect high levels of enrichment. To further authenticate the two pools with respect to cell cycle phase, extracts from each were subjected to immunoblotting with markers of either G1 or G2/M phases. First, we utilized a phospho-specific antibody which recognizes retinoblastoma (Rb) protein phosphorylated on either Ser-807 or Ser-811, since it is well established that Rb becomes increasingly phosphorylated as cells advance through the cell cycle17. As shown in Figure 4, levels of phospho-Rb were much higher in cells in the G2/M pool versus the G1 phase pool, consistent with expectations. We also probed for cyclin B1, which is expressed most highly in G2/M phase cells, and for cyclin D1, which is most highly expressed in G1 phase cells18 (Figure 4). The G2/M pool had much higher levels of cyclin B1 compared to the G1 pool, consistent with expectations. Cyclin D1 gave only a weak signal in these preparations, but nonetheless was detectable in the G1 pool and undetectable in the G2/M pool. GAPDH was used a control to confirm equal protein loading.
Modifications to the standard protocol were performed to assess their effects and to establish conditions for optimal performance. First, the number of fractions collected was decreased, from the standard twenty to only three, using speeds determined from Table 1 to represent the terminal points for collection of cells in the G1, S or G2/M phases, as shown in Table 2. Analysis of the DNA content of fractions F1 and F3 are shown in Figure 5A. Fraction 1 contained cells highly enriched in G1 phase (95%) and was thus of similar purity to that obtained under standard conditions (Figure 3). However, fraction 3, ostensibly cells highly enriched in G2/M phase, gave a mixed population, with cells in G1 and S phases also present. G2/M cells represented only 51% of the total, compared to 85% under standard conditions. These results indicate that attempts to simplify the procedure by reducing the number of fractions compromises sample quality. Next, we tested the effect of reducing fraction volume, as a means to reduce reagent consumption. A standard elutriation was performed except that 40 mL instead of 50 mL fractions were collected. Fractions F1 and F20 were analyzed for DNA content. As shown in Figure 5B, fraction F1 was highly enriched in G1 phase cells (98%), similar to that obtained under standard conditions. However, fraction F20 contained cells in all three phases, with cells in G2/M phase only representing 64% of the total, compared to 85% under standard conditions. These results indicate that reducing fraction volume also compromises sample quality.
Figure 1. Principle of counterflow centrifugal elutriation.
Note that elutriation of cells can be achieved either by increasing the rate of counterflow as indicated here or by decreasing rotor speed. Adapted by permission from Macmillan Publishers Ltd: [Nature Protocols] (Ref 8.), copyright (2008)8. Please click here to view a larger version of this figure.
Figure 2. Diameter of primary ALL cells increases with elutriation fraction.
The average diameter was determined for each fraction using a cell analyzer which records data from 500 – 4,500 individual cells per sample. The data shown are mean ± S.D. from two independent experiments. Note that error bars for many points are smaller than the symbol and were omitted for clarity. Please click here to view a larger version of this figure.
Figure 3. Analysis of DNA content confirms successful separation of primary ALL cells in different cell cycle phases.
10 mL of each fraction was fixed in 70% EtOH, stained in propidium iodide and analyzed by flow cytometry, as described15. Propidium iodide integrated intensity (x-axis) was plotted versus cell count (y-axis) to determine the portion of cells in each phase, where 2N DNA content indicates G1 phase and 4N DNA content indicates cells in either G2 or M phases. The data shown are from a representative experiment and essential identical results have been obtained in four repeats. Please click here to view a larger version of this figure.
Figure 4. Immunoblot analysis of cell cycle-related proteins in pooled fractions.
Following a standard elutriation of primary ALL cells, fractions F1 – F5 were combined to create a G1 phase pool and fractions F17 – F20 were combined to create a G2/M phase pool. Extracts were prepared and 40 µg/lane subjected to immunoblot analysis, as described previously15, with antibodies to phospho-Rb (Ser807/811), cyclin B1, or cyclin D1, all used at 1:2,000 dilution. GAPDH (1:10,000) was used as a loading control. Please click here to view a larger version of this figure.
Figure 5. Illustrative data from suboptimal elutriations.
A) Effect of decreasing the number of fractions collected. Only three main fractions (F1 – F3), instead of the standard twenty, each of 100 mL, were collected, at speeds as indicated in Table 2. Fractions F1 and F3 were analyzed for DNA content as in Figure 3. B) Effect of reducing fraction volume. Twenty fractions were collected, under the same conditions as in Table 1, but only 40 mL of elutriation buffer instead of the standard 50 mL was used for each. Fractions F1 and F20 were analyzed for DNA content as in Figure 3. Please click here to view a larger version of this figure.
Fraction | Speed (rpm) | G-Force | Total Cells (x 10^6) |
W1 | 3000 | 867 x g | 17.75 |
W2 | 2920 | 821 x g | 7.00 |
F1 | 2860 | 788 x g | 4.60 |
F2 | 2800 | 755 x g | 10.75 |
F3 | 2740 | 723 x g | 15.25 |
F4 | 2680 | 692 x g | 22.50 |
F5 | 2620 | 661 x g | 26.75 |
F6 | 2560 | 631 x g | 25.00 |
F7 | 2500 | 602 x g | 22.75 |
F8 | 2440 | 573 x g | 18.00 |
F9 | 2380 | 546 x g | 14.00 |
F10 | 2320 | 518 x g | 10.75 |
F11 | 2260 | 492 x g | 7.43 |
F12 | 2200 | 466 x g | 8.63 |
F13 | 2140 | 441 x g | 6.63 |
F14 | 2100 | 425 x g | 4.98 |
F15 | 2060 | 409 x g | 3.46 |
F16 | 2020 | 393 x g | 3.43 |
F17 | 1980 | 378 x g | 2.63 |
F18 | 1940 | 363 x g | 3.17 |
F19 | 1900 | 248 x g | 2.78 |
F20 | 1860 | 333 x g | 1.66 |
STOP | 0 | 0 x g | 5.19 |
Total Yield | 245.06 | ||
Percent Yield | 81.69 |
Table 1. Cell yield and corresponding rotor speed for each fraction.
Cell count was determined using a automated cell counter. Total yield was determined by the sum of individual fractions versus input (3 x 108 cells). Data are an average of two representative experiments. Centrifugal force and corresponding rotor speed are given.
Fraction | Speed (rpm) | G Force |
W1 | 3000 | 867 x g |
W2 | 2920 | 821 x g |
F1 | 2620 | 661 x g |
F2 | 2020 | 393 x g |
F3 | 1860 | 333 x g |
Stop | 0 | 0 x g |
Table 2. Speed parameters for suboptimal elutriation.
Shown are centrifugal forces and corresponding rotor speeds for the washes and three fractions collected in a compressed suboptimal elutriation. See Figure 5A for corresponding DNA content and text for details.
We have described a method for obtaining primary ALL cells in different phases of the cell cycle using CCE. Under optimal conditions an essentially pure population of cells in G1 phase and a highly enriched population of cells in G2/M phases could readily be obtained in excellent yield, and cells highly enriched in S phase can also be obtained if desired. The results presented here were obtained using one primary ALL culture (specifically, ALL-5), and essentially identical results have been obtained with an independent culture, ALL-215. In addition, all other primary ALL cultures examined to date have similar density ranges during asynchronous growth, and will likely perform similarly when subjected to CCE. Our prior studies have also shown that other leukemia cell lines, including HL60 and K562, can be separated into different cell cycle phases using CCE19,20. Critical steps in the procedure include ensuring that the cells are mono-dispersed and not clumped at the start of the run; optimization of the conditions of counterflow rate and rotor speed; ensuring the components of the system are clean and free of debris; and avoiding the introduction of air bubbles into the flow lines. CCE works best with cells that significantly increase in cell size as they advance through the cell cycle. For a given population of cells in suspension, this can be established initially and easily by examining the relationship between DNA content and side-scatter, the latter dependent on cell size. Both parameters can be measured simultaneously by flow cytometry after propidium iodide staining, as we have described15. A linear relationship between side-scatter and DNA content provides confidence that cell size increases with cell cycle advance and thus that CCE has the potential to separate cells based on cell cycle phase.
As we have shown (Figure 5), deviations from the optimized procedure result in compromised sample quality, especially for cells in G2/M phases. However, reduction in the number of fractions or fraction volume did not appreciably affect the purity of cells in G1 phase (Figure 5). Thus, shortcuts to the procedure can be tolerated if only purified G1 phase cells are needed. Although CCE is readily amenable to suspension cell cultures, it can also be applied to subcellular fractions. For example, nuclei from murine pre-B cells have been successfully size- separated using CCE8. Adherent cell types pose potential problems due to their propensity for clumping and possible adherence to the surfaces of the tubing and elutriation chamber and thus should be investigated with caution.
The main limitation of CCE is the cost of the equipment; a possible solution is to collaborate with a lab which has the necessary apparatus. Note however that while the rotor is a dedicated instrument, the centrifuge can be used for general centrifugation purposes, thus reducing the cost invested solely in CCE. Another limitation is that the scheduling of experiments is more problematic compared to the flexibility of other methods such as serum starvation or chemical synchronization. Elutriation runs typically take 5 – 6 hours, and if the cells obtained are subsequently needed for short-term or multi-time point experiments, this can result in scheduling inconveniences. The biggest advantage is that cells obtained by this method are unperturbed and can be cultured for many days without any signs of distress, as we have shown15. This contrasts with cells synchronized by chemical means which may harbor drug-induced damage. A recent study, using NB4 cells derived from acute myeloid leukemia blasts, directly compared CCE and other methods, including arrest with serum starvation, hydroxyurea, or treatment with a cyclin-dependent kinase 1 inhibitor, RO-330621. Elutriated cells and cells arrested at comparable phases appeared similar with respect to DNA content and cyclin expression. However, when the proteome was examined, major changes in protein abundance, particularly stress-related and arrest-specific proteins, were observed in the arrested cells and not in the elutriated cells at equivalent cell size positions. These results highlight the utility of CCE for the production of unperturbed cells and emphasize that caution should be used when interpreting data from cells synchronized by other means.
The authors have nothing to disclose.
This work was supported by NIH CA109821 (to TCC). We thank Beckman-Coulter for generously providing funds to cover publication costs and for technical assistance in the set-up and operation of the elutriation system. We thank Dr. Fred Falkenburg for providing initial primary ALL cultures.
Hanks' balanced salt solution | Lonza | 10-508Q | 1 L bottle |
Fetal Plus bovine serum | Atlas Biologicals | FP-0500-A | 500 mL bottle |
2-napthol-6,8-disulfonic acid dipotassium salt | Acros Organics | 212-672-4 | 100 g powder |
50 mL concial tubes | Denville Scientifics | C1062-P | 500/case |
PES Membrane 0.22 µm filter unit | Millipore | SLGP033RB | 250/pack |
Avanti J-26S XPI w/ Elut. Non-IVD – 50/60 Hz, 200/208/240V | Beckman Coulter | B14544 | centrifuge compatible with elutriation system |
JE 5.0 elutriator rotor kit | Beckman Coulter | 356900 | rotor kit which includes the rotor, T-handle hex wrench, the quick release assembly (w/o the elutriation chamber), anchoring cable, and the strobe assembly |
Chamber, standard, 4-mL, "A" | Beckman Coulter | 356943 | standard elutriation chamber |
Masterflex L/S Easy-Load pump head | Cole-Parmer | EW-07518-10 | |
Masterflex L/S Variable-Speed Drive | Cole-Parmer | EW-07528-10 | |
Masterflex BioPharm platinum-cured silicone pump tubing, L/S 16 | Cole-Parmer | EW-96420-16 | |
25 G x 1.5 Precision Glide needle | BD | 305127 | sterile, single-use needles |
10 mL Luer-Lok syringe | BD | 309604 | sterile, single-use syringes |
Vi-Cell XR | Beckman Coulter | 383556 | |
PI/RNase staining buffer | BD Pharmingen | 550825 | propidium iodide/RNase staining buffer |
cyclin B1 (GNS1) mouse monoclonal IgG antibody | Santa Cruz Biotechnology | sc-245 | |
cyclin D1 (DCS-6) mouse monoclonal IgG antibody | Santa Cruz Biotechnology | sc-20044 | |
P-Rb (S807/811) (D20B12) XPR rabbit monoclonal antibody | Cell Signaling Technology | 8516s | |
GAPDH (14C10) rabbit monoclonal antibody | Cell Signaling Technology | 2118s | |
Goat anti-mouse IgG (H+L)-HRP conjugate | BioRad | 170-6516 | |
Goat anti-rabbit IgG (H+L)-HRP conjugate | BioRad | 170-6515 |