Doublet Discrimination in DNA Cell-Cycle Analysis
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Cytometry (Communications in Clinical Cytometry) 46:296 –306 (2001) Doublet Discrimination in DNA Cell-Cycle Analysis Robert P. Wersto,1* Francis J. Chrest,1 James F. Leary,2 Christa Morris,1 MaryAlice Stetler-Stevenson,3 and Edward Gabrielson4 1 Flow Cytometry Unit, National Institute on Aging, National Institutes of Health, Bethesda, Maryland 2 Molecular Cytometry Unit, University of Texas Medical Branch at Galveston, Galveston, Texas 3 Laboratory of Pathology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 4 Department of Pathology, Johns Hopkins School of Medicine, Baltimore, Maryland Differences in doublet analysis have the potential to alter DNA cell-cycle measurements. The techniques for doublet determination are often used interchangeably without regard for the complexity in cell shapes and sizes of biological specimens. G0/1 doublets were identified and quantitated using fluorescence height versus area and fluorescence width versus area pulse measurements, by enumerating the proportion of G2 ⴙ M cells that lack cyclin B1 immunoreactivity, and modeled in the DNA histograms by software algorithms. These techniques were tested on propidium iodide-stained whole epithelial cells or nuclei from asynchronous cultures, or after exposure to chemotherapeutic agents that induced cell-cycle arrest and were extended to human breast tumor specimens having DNA diploid patterns. G0/1 doublets were easily discernible from G2 ⴙ M singlets in cells or nuclei that are generally homogenous and spherical in shape. Doublet discrimination based on pulse processing or cyclin B1 measurements was nonconcordant in some nonspherical cell types and in cells following cell cycle arrest. Significant differences in G0/1 doublet estimates were observed in breast tumor specimens (n ⴝ 50), with estimates based on pulse width twice those of pulse height and nearly five times greater than computer estimates. Differences between techniques are attributed to difficulties in the separation of the boundaries between G0/1 doublets and G2 ⴙ M singlet populations in biologically heterogeneous specimens. To improve reproducibility and enhance standardization among laboratories performing cell cycle anal- ysis in experimental cell systems and in human breast tumors, doublet discrimination analysis should best be accomplished by computer modeling. Shape and size heterogeneity of tumor and arrested cells using pulse-processing can lead to errors and make interlaboratory comparison difficult. Cytometry (Comm. Clin. Cytometry) 46:296 –306, 2001. Published 2001 Wiley-Liss, Inc.† Key terms: G0/1 doublets; DNA cell cycle; flow cytometry; cylin B1; pulse width; pulse height; modeling Debris and aggregates can be prominent components of removed by nylon mesh filtration (typically 35–53 m), DNA histograms, affecting the accuracy and reproducibil- sheared apart by passage through small gauge needles, or ity of cell-cycle estimates (1– 4). Debris originate from the identified on the basis of forward-angle light scattering (2). damage and disintegration of cells following apoptosis or Strategies to separate overlapping G0/1 doublets from the the fragmentation associated with the slicing of cells or G2 ⫹ M1 population have utilized the gating of correlated nuclei during mechanical disaggregation. Clumping may measurements of integral DNA fluorescence pulse (⬇ be due to the incomplete disruption of tissues by mechan- area) with either peak pulse height (brightness) (5) or ical or enzymatic means into single-cell or nuclear suspen- pulse duration (width) (6 –9), gating on the G2 ⫹ M cells sions, by the use of alcohol-based fixatives that yield DNA that lack cyclin B1 protein expression (10), and computer histograms with low coefficients of variation of the G0/1 peak but induce clumping, or by centrifugation. In addi- tion, clumping may be an inherent attribute of some cell types, e.g., keratinocytes. 1 Single–parameter DNA histograms cannot resolve G2 from M phase Aggregates can be composed of large clusters of cells or cells, and in this report are considered to be a single entity. *Correspondence to: Robert Wersto, Ph.D., Flow Cytometry Unit, Geron- nuclei or two or more G0/1 (2N) events adhered together tology Research Center, National Institute on Aging, 5600 Nathan Shock (G0/1 doublets) that are indistinguishable from particles Drive, Baltimore, MD 21224. E-mail: werstor@grc.nia.nih.gov with 4N, 6N, or 8N DNA content. Large clumps can be Received 5 February 2001; Accepted 24 August 2001 † Published 2001 WILEY-LISS, INC. This article is a US government work and, as such, is in the public domain in the United States of America.
DOUBLET ANALYSIS OF DNA HISTOGRAMS 297 FIG. 1. G0/1 doublet discrimination in A549 cells, using fluorescence pulse processing and cyclin B1 protein expression. A: Pulse height versus area (left) and pulse width versus area (right) dot plots. G0/1 doublets are depicted by rectangular region. B: Cyclin B1 immunofluorescence versus DNA content (area). Immunofluorescence of cells stained with FITC-conjugated mouse IgG1 is shown by the trapezoidal area and reflects the staining of cells lacking cyclin B1 protein expression. G0/1 doublets are identified in the ellipsoidal region. Insets: Positions of cyclin B1-identified G0/1 doublets in either the height versus area or width versus area dot plots. C: Comparison of percentage of G0/1 doublets identified by three techniques using different flow cytometers. N ⫽ 15 for each determination. Dot plots represent 25,000 cells. No differences are observed (P ⬎ 0.8). algorithms to model aggregate probability distributions in the laser beam. Although a G2 ⫹ M cell has twice the DNA histograms (4,11–14). volume of a G0/1 cell, diameter only increases by ⬇26% G0/1 doublet discrimination based on signal processing (based on the formula relating the volume V of a sphere to (pulse, width, area) is dependent on the geometry of the its diameter D, i.e., V ⫽ (3.14/6)D3, since if VG2 ⫽ 2VG1, illumination beam and particle size (review in Shapiro DG2 ⫽ (2)1/3 DG1 ⫽ 1.26DG1), by contrast, the combined (15)). Under conditions where the cell size approximates diameter of a G0/1 doublet is at least twice that of a single laser beam width, G0/1 doublets are not in the beam at the G0/1 event, providing that it aligns in the direction of flow. same time, resulting in a broadened height signal (that In the case of uninucleate cells, the diameter of a G2 may be bimodal, depending on nuclear size and beam nucleus compared to a G0/1 nucleus correlates very well width) comparable to the height pulse of a G0/1 singlet, with that predicted by this theory (16). but with the integrated fluorescence of a G2 ⫹ M event. The width signal is the sum of the width of the laser The width signal increases with increasing particle diam- beam and cell or nuclear diameter. When the vertical eter, since it takes longer for a larger particle to traverse beam size is large as compared to particle width (used to
298 WERSTO ET AL. improve measurement sensitivity by increasing the dura- cells were centrifuged at 400g for 10 min, resuspended in tion of particle illumination and collection), the ability to either 0.5 ml FCS, and stored at ⫺70°C, or 70% ethanol or discriminate differences in pulse widths between cells or acetone/methanol mixture, and stored at ⫺20°C until between G0/1 singlets and doublets may be marginal. DNA staining. While this problem can be solved by real-time subtraction of the beam width using high-precision biased amplifiers, DNA Staining this is rarely done with the needed degree of precision on Cells frozen in FCS and frozen solid tumor specimens most commercial instruments. Similarly with a large ver- were stained for DNA cell-cycle analysis, using the rapid tical beam, the resolving power of peak height versus area nuclear isolation medium (NIM) procedure (3). Nuclear measurements is minimized, since the height signals are preparations yield higher quality DNA histograms in terms optically integrated, which decreases its difference from of the coefficient of variation (CV) of the G0/1 peak than the pulse area signal that is electronically integrated. those based on whole cells, and are the method of choice G0/1 doublet discrimination using pulse height versus for DNA cell-cycle analysis of tumor specimens. The NIM area and pulse width versus area measurements are some- technique was used due to its stability (2 days after stain- times interchangeably used among laboratories. In sam- ing versus 3 h using detergent-trypsin prepared nuclei ples where there is substantial variability in cell/nuclear technique (26)). Tumor samples were thawed in approx- size or shape, doublet determination might be variable. imately 0.5 ml of cold NIM buffer and finely minced using This study critically evaluates G0/1 doublet discrimination a pair of iridectomy scissors. The nuclear suspension was signal processing techniques in homogenous and hetero- filtered through a piece of 47-m nylon mesh, and geneous biological specimens, using commercial instru- counted with a hemocytometer, and the concentration ments that have standard wide-beam geometry. Results was adjusted to 2 ⫻ 106 nuclei/ml. For DNA staining, 0.5 are compared to those obtained from the modeling of ml of the nuclear suspension were mixed with an equal doublets, triplets, and quadruplets in software doublet volume of NIM buffer containing 50 g/ml propidium correction modeling. Human breast tumor specimens, iodide PI (Calbiochem-Behring, LaJolla, CA) and 15 l of having diploid DNA profiles, were examined because of freshly prepared RNAse (dissolved in water; 500 U/ml; the strong association between increased proliferative ac- Type IV, Sigma Chemical Co., St. Louis, MO). Stained tivity (S-phase fraction) and DNA ploidy and poor clinical samples were incubated at room temperature in the dark outcome (17–24). for a minimum of 1.5 h or then stored overnight at 4°C before analysis. FCS-frozen tissue culture cells were MATERIALS AND METHODS thawed at room temperature, after which NIM-PI buffer Sample Acquisition and RNAse were added as described above. Human breast carcinoma specimens were collected Ethanol or acetone-methanol fixed samples were cen- from samples submitted to surgical pathology in accor- trifuged (300g, 10 min), washed in PBS (lacking Ca2⫹, dance with Institutional Review Board (IRB) regulations. Mg2⫹ salts, Gibco/BRL, Rockville, MD) , and counted, and Tissue samples were analyzed for DNA ploidy and the 1 ⫻ 106 cells were stained in 1 ml of PBS containing 50 histograms classified as aneuploid when two separate G0/1 g/ml PI and 1 mg/ml RNAse for 1 h at room temperature. peaks were present (25). Tumor specimens with diploid For evaluation of cyclin B1 protein expression across DNA patterns were used in this study (n ⫽ 50). Histolog- the cell cycle and enumeration of G0/1 doublets, the cells ical examination of tissue sections verified the predomi- were fixed for at least 24 h in a cold (⫺20°C) 1:1 acetone- nance of tumor cells (⬎85%) in these samples from pa- methanol mixture (HPLC grade solvents), followed by tients who presented with primary untreated node- negative breast adenocarcinoma. All unique patient identifiers were removed prior to analysis, making addi- ™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™3 tional clinical follow-up parameters, such as survival, un- FIG. 2. G0/1 doublet discrimination in complex samples, using (left to right) cyclin B1 immunofluorescence, pulse height versus area, and available for further study. Specimens were rapidly frozen pulse width versus area. Insets: Positions of G0/1 doublets in the two other at ⫺70°C and stored at this temperature until analysis. corresponding dot plots after gating of the primary (larger) dot plot (e.g., width versus area and cyclin B1 versus DNA content, gated on height A549 and LoVo cells, derived from a human lung carci- versus area G0/1 doublets). Shaded areas in height versus area and width noma and metastatic colorectal adenocarcinoma, respec- versus area dot plots depict positions of singlet events (G0/1, S, and G2 ⫹ tively, were obtained from the American Type Culture M). Immunofluorescence of cells stained with FITC-conjugated mouse IgG1 is shown in trapezoidal areas, and reflects staining of cells lacking Collection (Manassas, VA) and grown in Iscove’s modified cyclin B1 protein expression. G0/1 doublets are identified in the ellipsoi- Eagle’s medium supplemented with 10% fetal calf serum dal region in the cyclin B1 or height versus area dot plots, or in the (FCS), 25 mM Hepes, 100 U/ml penicillin G, and 100 rectangular region in width versus area dot plots. Dot plots represent 25,000 cells and are representative of at least three experiments. A: g/ml streptomycin. Both cell lines have epithelial mor- LoVo cells. B: Human epidermal keratinocytes. When gated on the basis phology. Human epidermal keratinocytes were purchased of height versus area, G0/1 doublets contained both cyclin B1-positive (arrow) and -negative populations. C: A549 exposed to SKF-96365, an from Clonetics (Walkersville, MD) and grown in KGM-2 M-phase blocker. D: Comparison of percentage of G0/1 doublets in SKF- media. To ensure proper sampling of all cell cycle phases, 96365-treated cells by instrument. N ⫽ 9 separate determinations. P ⬍ media containing nonadherent floating cells were com- 0.05, comparing G0/1 doublets estimated on the basis of cyclin B1 staining and pulse-processing techniques. E: Comparison of percentage bined with the cells harvested by trypsinization (0.05% of G0/1 doublets in taxol-treated A549 cells, using FACScan and FACS- trypsin, 0.02% EDTA). For DNA cell-cycle analysis, 1 ⫻ 106 calibur. N ⫽ 5 separate determinations.
DOUBLET ANALYSIS OF DNA HISTOGRAMS 299 FIG. 2.
300 WERSTO ET AL. washing in 70% (v/v) ethanol, and PBS, and permeabilized scale at a flow rate of 12 l/min (low), corresponding to for 5 min on ice with 0.25% Triton X-100 in PBS contain- 50 –150 events/s. Some data were acquired on an older ing 1% bovine serum albumin (BSA). Cells (1 ⫻ 106) were FACScan™ that lacked the ability to acquire width mea- incubated with a fluorescein-isothiocyante (FITC)-conju- surements in the FL3 channel. In this case, the high volt- gated antibody to cyclin B1 (Pharmingen, San Diego, CA) age and signal cables from FL2 and FL3 were exchanged for 1 h at room temperature, then washed once with PBS prior to alignment and standardization with CENs. PI flu- containing 1% BSA and resuspended in PBS containing PI orescence was measured thru a 660 ⫾ 20 nm bandpass (10 g/ml) and RNAse (1 mg/ml) for 30 min prior to filter on the FACSstar™ on linear scale at similar sample analysis (27). FITC-conjugated mouse IgG (Pharmingen) rates, and excited at 488 nm with 200 mW of laser power. was used as negative control for antibody staining. Com- For dual-parameter flow cytometric measurements of cy- parable results could be obtained with acetone-methanol clin B1 expression and DNA content, green FITC fluores- fixed cells stored at ⫺20°C for up to 2 weeks or cells fixed cence was measured with a 530 ⫾ 30 nm bandpass filter in 70% ethanol and stored at ⫺20°C. Attempts were made (FL1) in addition to the FL3 signal. to perform cyclin B1 staining on isolated nuclei and solid In the course of this study, it was noted that chick red tumor specimens previously frozen at ⫺70°C. In the blood cells, frozen in FCS at ⫺70°C and stained using the former case, cyclin B1 labeling was absent in NIM-pre- NIM procedure could be used in place of the detergent- pared nuclei from A549 cells, and the discrimination be- treated, alcohol-fixed CEN supplied by Becton-Dickinson. tween G1 doublets and G2 ⫹ M was minimal for A549 Peripheral blood lymphocytes were stained in a similar nuclei isolated using a sucrose-buffer-based technique manner as the tumor samples, and cell lines were used as (www.dcs.nci.nih.gov/branches/medicine/flowcore/ diploid DNA standards. The coefficient of variation (CV) nuclei.htm) and stained with cyclin B1 antibody. Cyclin of the G0/1 peak was calculated by the full-width method B1 staining was present in nuclei prepared using the and averaged ⬍1.5% for diploid human lymphocytes. washless KI-67 technique (28) (data not shown); how- ever, the number of G1 doublets was low (⬍0.5%) as Histogram Analysis measured by cyclin B1 staining, fluorescence peak/width List mode files, consisting of forward and side scatter, signals, and microscopic evaluation, making statistical integrated (area), peak (height) and pulse (width) red comparisons difficult. Thawed, previously frozen solid fluorescence, were analyzed using CellQuest™ (Becton- tumor specimens that were mechanically disaggregated Dickinson). Raw data were minimally gated by excluding and fixed in acetone/methanol showed no reactivity with only channels 1–5 and 1020 –1024, respectively, during the cyclin B1 antibody (data not shown). sample acquisition. Cell-cycle analysis of the DNA histo- grams of integrated red fluorescence was performed with Flow Cytometry Modfit (Verity Software, Inc., Topsham, ME), using the DNA content was measured using three different flow F_DIP_T1 model (fresh, diploid, aggregrates) with the cytometers: FACScan™, FACScalibur, and FACSStar Plus™ S-phase fraction modeled as multiple broadened trape- (Becton-Dickinson, Mountain View, CA). FACScan™ and zoids, and Multicycle (Phoneix Flow Systems, San Diego, FACScalibur™ are the most commonly used nonsorting CA), using a zero-order polynomial to model the S-phase flow cytometers in the research and clinical environ- fraction. Both software packages include algorithms for ments. Both instruments utilize a quartz flow cell for estimating multicut debris, and in the case of doublets, sensing (430 ⫻ 180 m, the only difference being a slight they model doublets, triplets, and quadruplets as probabi- change in the internal geometry of the flow cell on the listic events (29). Debris subtraction was performed in all FACScalibur威). The FACSStar Plus™ was equipped with samples prior to cell-cycle analysis. List mode data were jet-in air sensing, using a 70-m nozzle. Beam geometries displayed as two-parameter correlated histograms of peak for all three instruments are similar: 20 ⫻ 64 m (FACS- width, peak height, and peak area for reprocessing, using can™), 22 ⫻ 66 m (FACScalibur™), and 20 ⫻ 60 m different gating strategies. (FACSStar Plus™). Sheath fluids were either isotonic sa- Statistical comparison of data was performed using the line (Fisher Diagnostics, Swedesboro, NJ) or Biosure Statgraphics statistical software package (STSC, Rockville, sheath fluid (Riese Enterprises, Grass Valley, CA) on the MD). Variance between doublet estimates and cell-cycle FACSstar™, FACSflow on the FACScalibur™, or isotonic fractions was tested using the chi-square test at the 95% saline or Haema-Line 2 (Serono-Baker Instruments, Allen- confidence limit. Data are reported as the mean ⫾ stan- town, PA) on the FACScan™. Instruments were aligned dard deviation. with unstained and stained fluorescent polystyrene micro- spheres (CaliBRITE beads, Becton-Dickinson), or, in the RESULTS case of FACSStar Plus™, 6.4-m (6 peaks) Rainbow Cali- Doublet Discrimination in Biologically Homogenous bration Particles (Spherotech, Inc., Libertyville, IL). Samples Chicken erythrocyte nuclei (CENs) were used to check Identification of G0/1 doublets in asynchronously grow- the linearity of the fluorescence pulse detectors. On FAC- ing A549 cells cultured in vitro (Fig. 1) or human periph- Scan™ and FACScalibur™, red fluorescence of PI-stained eral blood lymphocytes (data not shown) stained for cell- nuclei was collected through a 650-nm longpass filter into cycle analysis was straightforward, using fluorescence the fluorescence 3 (FL3) photomultiplier tube on linear pulse processing signals. In dot plots of height versus area
DOUBLET ANALYSIS OF DNA HISTOGRAMS 301 or width versus area, doublets were clearly discriminated ination). In A549 cells, exposure to 10 nM taxol induced from other stained cells (Fig. 1A). Independent of the apoptosis (30); again, the number of G0/1 doublets esti- peak-processing methods, doublets can be detected using mated from cyclin B1 protein expression varied (Fig. 2E). cyclin B1 staining; this technique is based on the rationale that cyclin B1 protein expression is restricted to the G2 ⫹ Doublet Analysis in Biologically Heterogeneous M phases of the cell cycle, hence, G2 ⫹ M cells lacking Specimens- Human Breast Tumor Specimens cyclin B1 immunofluorescence are G0/1 doublets (Fig. 1B). The poor separation between G0/1 doublets and G2 ⫹ M Doublets identified by cyclin B1 staining fell within the cells in the height versus area and width versus area dot “doublet” regions of height versus area or width versus plots from cells treated with cell cycle inhibitors might area dot plots (Fig. 1B, insert). Overall, all three strategies have been due to heterogeneity in nuclear size and/or yielded comparable results (Fig. 1C), irrespective of the shape. To test this hypothesis, PI-stained nuclei were type of flow-chamber sensing (jet-in-air or closed quartz). prepared from normal human breast tissue and tumors, and the percentage of G0/1 doublets was evaluated using Doublet Analysis in Biologically Complex Specimens the pulse height and width techniques. Breast tumors are In some tumor cell lines, primary cultures, and cells heterogeneous (31) and represent a human cancer where exposed to cell-cycle inhibitors, pulse processing and cy- DNA cell-cycle and ploidy measurements have clinical clin B1 immunofluorescence G0/1 doublet detection tech- relevance (reviewed in Wenger and Clark (24)). As shown niques were not mutually inclusive. G0/1 doublets could in Figure 3, G0/1 doublets present in nuclear preparations be clearly identified in PI-stained unperturbed LoVo cells, of normal breast tissue are clearly separated from the G2 ⫹ using either height versus area or width versus area mea- M fraction in either the height versus area and width surements (6.5 ⫾ 1.8% height versus area, as compared to versus area dotplots. By contrast, the width signals of G0/1 6.9 ⫾ 1.3% width versus area; n ⫽ 5, P ⬎ 0.8) (Fig. 2A). doublets in tumor nuclei samples are broad, making the However, cyclin B1 protein expression was diminished in clear identification of the boundary between the G0/1 these cells, resulting in an overestimation of number of doublets and G2 ⫹ M population difficult. To exclude the G0/1 doublets by this technique (24.2 ⫾ 2.8% versus an possibility that these observations were due to the use of average of 6.2 ⫾ 1.6% for the height versus area or width nuclei for cell-cycle analysis, height and width pulse-pro- versus area measurements combined; n ⫽ 7, P ⬍ 0.05). cessing doublet-detection techniques were tested on PI- Reanalysis of the position of doublets identified by cyclin stained nuclei from A549 cells prepared using the NIM B1 staining in width versus area dot plots showed that method. No differences were observed in the number of both contained G2 ⫹ M and G0/1 populations (Fig. 2A, G0/1 doublets estimated from height versus area and width inserts). versus area dot plots analyzed either on the FACScalibur™ Conversely, while width versus area and cyclin B1 dou- (4.7 ⫾ 2.2% height versus area; 5.1 ⫾ 2.2% width versus blet detection methods yielded nearly equivalent numbers area; n ⫽ 20, P ⬎ 0.8) or FACScan™ (3.5 ⫾ 2.3% height of G0/1 doublets in samples of PI-stained primary human versus area; 3.3 ⫾ 2.1% width versus area; n ⫽ 20, P ⬎ keratinocytes (Fig. 2B; 4.5 ⫾ 0.8% width versus area, as 0.8), nor do the results significantly differ from those compared to 4.9 ⫾ 0.3% cyclin B1; n ⫽ 5, P ⬎ 0.8), or obtained with whole cells (Fig. 1C; P ⬎ 0.5). Hs578T cells (a human breast cancer cell line that exhibits Microscopically, nuclei prepared from breast tumor tis- a high degree of chromosomal instability; data not sue using the NIM technique appear as a continuum of shown), doublet discrimination using peak versus area various shapes, ranging from the spherical to the markedly detection was poor (23.4 ⫾ 4.4%.). Not surprisingly, pu- elongated. Consequently, this might account for the indis- tative height versus area-identified G0/1 doublets con- tinct separation of G0/1 doublets from the G2 ⫹ M fraction tained mixtures of both G0/1 doublets and G2 ⫹ M cells in by pulse-width signals. To further explore this possibility, the width versus area or cyclin B1 versus DNA content dot the percentage of G0/1 doublets in 50 breast tumor spec- plots (Fig. 2B, inserts). imens having diploid DNA histogram patterns was evalu- The correlation between the doublet detection tech- ated, using the pulse-processing techniques, and com- niques was poorest in cells exposed to agents that induce pared to the results from computer modeling of doublets. cell cycle arrest and may alter cyclin B1 protein expres- Figure 4 illustrates the height versus area and width versus sion. Compared to untreated cells (above), in both the gating strategies used for doublet analysis in a representa- height versus area or width versus area dot plots, G0/1 tive tumor specimen. In height versus area dot plots (shad- doublets and G2 ⫹ M cells populations overlapped after ed area, Fig. 4A), region R1 contained DNA singlet events treatment with SKF-96365, an agent that induces M-phase (G0/1, S, G2 ⫹ M; upper red fluorescence area histogram), arrest (27) (Fig. 2C; similar patterns were observed in while region R2 contained G0/1 doublets (lower red fluo- colcemid-, vinblastine-, and nocodazole-treated cells, data rescence area histogram). Analysis of the position of G0/1, not shown), and showed large variation with the number S, G2 ⫹ M singlets and G0/1 doublets in the width versus of G0/1 doublets estimated from cyclin B1 protein expres- area dot plot, based on the R1 and R2 height versus area sion (Fig. 2D). Notably, this uncertainty in positioning of gates (upper right dot plot, Fig. 4A), indicated that R1 the G0/1 doublet gate in height versus area or width versus contained only G0/1, S, and G2 ⫹ M singlets. Importantly, area dot plots underestimated the number of doublets R2 contained a mixture of both G2 ⫹ M singlets and G0/1 (average ⬇1%, as compared to 4% by microscopic exam- doublets (lower right dot plot, Fig. 4A). By contrast, the
302 WERSTO ET AL. FIG. 3. G0/1 doublet discrimina- tion in human breast normal and tumor tissue using fluorescence pulse processing. Left: Height ver- sus area and width versus area dot plots from normal tissue. In both dot plots, G2 ⫹ M singlets and G0/1 doublets are clearly separated (ar- row in width versus area). Dot plots represent 25,000 cells and are representative of at least three ex- periments. Right: Tumor sample with near-diploid aneuploid stem- line. Doublets appear better sepa- rated from singlet events in the height versus area dot plot than in width versus area (large arrow). This sample also illustrates that both the diploid and aneuploid G0/1 (grey arrow) and G2 ⫹ M sin- glet width signals are a broad con- tinuum. width versus area dot plot of the same sample (shaded area, Fig. 4B), was divided into three regions: R1 con- tained DNA singlet events (G0/1, S, and G2 ⫹ M; upper red ™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™3 FIG. 4. Comparison of fluorescence pulse height versus area and width fluorescence area histogram); R2 contained putative G0/1 versus area doublet-discrimination techniques in DNA diploid human doublets (middle red fluorescence area histogram); and R3 breast tumor specimens. A: Height versus area. B: Width versus area. The primary dot plot is shown in the shaded area and depicts gating regions. was composed of a mixture of singlet, doublets, and Dot plots at right show the corresponding gated dot plot (i.e., width aggregates (lower red fluorescence area histogram). In the versus area for gating on height versus area regions). The integrated red fluorescence (area) histograms from the gating regions are shown below width versus area dot plots, no discrete boundary was each shaded dot plot. The primary dot plot represents data from 25,000 present clearly separating G0/1 doublets from G2 ⫹ M nuclei. B: Region R3 corresponds to right side of dot plot and includes nuclei. Corresponding analysis of the position of G0/1, S, R2. The rationale for positioning R2 and R3 with respect to singlet events was based on the uppermost width signal of G0/1 singlets and was and G2 ⫹ M singlets and G0/1 doublets in the height versus approximately channel 560. Data are representative of results from 50 area dot plots, based on the R1, R2, and R3 width versus tumor samples.
FIG. 4.
304 WERSTO ET AL. when the cells are spherical in shape, e.g., peripheral blood lymphocytes or trypsinized A549 cells, and express normal cyclin B1 protein levels. By contrast, the methods are not mutually inclusive and produce variable G0/1 dou- blet estimates in growth-arrested cells or in human breast tumor tissue, presumably due to heterogeneity in cell/ nuclear shape and size, or when cyclin B1 protein expres- sion levels are abnormal, as in perturbed cells. From the data presented here, G0/1 doublet discrimination in bio- logically complex breast tumor specimens, using width versus area pulse analysis, is confounded by uncertainty in the placement of the doublet gates due to the lack of a discrete boundary between G2 ⫹ M singlets and G0/1 doublets. The net result is overestimation of the number of G0/1 doublets and underestimation of the G2 ⫹ M fraction. G0/1 doublet estimates, using height versus area dot plot analysis, were nearly twice those obtained from computer modeling of DNA histograms. Rabinovitch (11) showed good correlation between computer modeling of G0/1 doublets and visual observation. For the breast tumor specimens studied here, the percentage of doublets varied FIG. 5. Comparison of fluorescence pulse height versus area and width from 1–3%, well within the range estimated from com- versus area doublet discrimination techniques and computer modeling in puter modeling. These results further support the objec- DNA diploid human breast tumor specimens. Data obtained from 50 tumors. *Observations that are statistically significant (P ⬍ 0.001). tiveness and reliability of computer models for G0/1 dou- Regions in Figure 4 were used for doublet estimates, based on pulse- blet detection. processing gates (R2 for height and width doublets). Narrow and wide Likewise, instrumental problems can affect G0/1 doublet designations refer to changing R2 in the width dot plot from channels 560 –575 to channel 700, respectively. estimates. Doublets may be correctly identified only if they are aligned sequentially and perpendicular to the laser beam. Consequently, errors in doublet detection may result when G0/1 doublets are oriented in parallel (side-by- area gates (right dot plots, Fig. 4B) indicated that R1 side) or off axis to the excitation beam, as occurs in contained both G0/1, S, and G2 ⫹ M singlets and G0/1 momentary hydrodynamic instability caused by clogs. doublets (upper right dot plot, Fig. 4B), R2 both G0/1 However, under normal conditions of stable laminar flow, doublets and G2 ⫹ M singlets (middle right dot plot, Fig. this does not happen, and doublets align well in the 4B), and R3 G0/1, S, and G2 ⫹ M singlets, and G0/1 and direction of flow. Likewise, the results here were obtained G2 ⫹ M doublets (lower right dot plot, Fig. 4B). on instruments having a fixed, standard wide-beam geom- The results from these analyses are depicted in Figure 5 etry (FACScan, FACSCalibur, and FACStar). Instruments and show significant differences between the pulse-pro- having non-comparable beam and cell sizes might yield cessing and computer-modeling techniques. On average, dissimilar results. the number of G0/1 doublets estimated from the width With respect to DNA cell-cycle analysis, does the inclu- versus area dot plots was nearly three times those from sion or removal of G0/1 doublets have any real effect on height versus area measurements. While no differences cell cycle estimates? Without G0/1 doublet subtraction, the were observed using different modeling software pro- G2 ⫹ M fraction of the DNA diploid breast tumor speci- grams (Modfit and Multicycle), these values were gener- mens studied here (n ⫽ 50) averaged 12.1 ⫾ 1.7%. Height ally one-half and one-sixth those of the height versus area versus area or width versus area doublet subtraction de- and width versus area measurements, respectively. For creased the percentage of G2 ⫹ M cells to 5.6 ⫾ 2.3% and width versus area dot plots, decreasing the size of R2 3.8 ⫾ 2.6%, respectively (P ⬍ 0.05, comparing with and (from ⬇channel 560 to ⬇channel 700) decreased the without subtraction), which was similar to the values from percentage of G0/1 doublets (to 7.3 ⫾ 6.0% from 11.6 ⫾ computer modeling (3.6 ⫾ 2.4%, Multicycle; 4.6 ⫾ 2.2%, 9.3%; Fig. 5), nearly equivalent the doublet estimates (R2) Modfit). The lower number of G2 ⫹ M cells in width based on the height versus area dot plot. versus area-corrected DNA histograms, compared to height versus area doublet subtraction, is due to the in- DISCUSSION clusion of G2 ⫹ M singlets in the “doublet” region. Be- In cell preparations stained with DNA-specific fluoro- cause cell-cycle percentages are usually calculated relative chromes, G0/1 doublets can be identified and removed to each other (i.e., the sum of the G0/1, S, and G2 ⫹ M prior to cell-cycle analysis using fluorescence pulse height phases equals 100%), decreasing the number of G2 ⫹ M or width signals, or by the absence of cyclin B1 immuno- events (by overestimating G0/1 doublets) will affect the reactivity, or modeled in the DNA histogram by computer estimates of other cell-cycle compartments. The sum of algorithms. These techniques yield comparable results the S ⫹ G2 ⫹ M phases is frequently used as a measure of
DOUBLET ANALYSIS OF DNA HISTOGRAMS 305 the proliferative/growth fraction (30), and here ranged Infrequently, light scatter has reportedly been used to from 14.5 ⫾ 4.0% (no doublet discrimination) to 8.8 ⫾ detect G0/1 doublets (38,39). Backgating of pulse-identi- 4.0% (height versus area) and 6.9 ⫾ 3.9% (width versus fied or cyclin B1-identified G0/1 doublets indicated that area; P ⬍ 0.05 compared to height versus area). S-phase neither forward nor 90° light scatter combined with a estimates after height versus area or width versus area fluorescence pulse (area/height/width) measurement doublet subtraction were similar (4.9 ⫾ 3.5% and 5.6 ⫾ could be used to identify or exclude G0/1 doublets from 3.6%, respectively) and comparable to nonsubtracted val- DNA histogram analysis (data not shown). Consequently, ues (4.9 ⫾ 2.9%). Interestingly, S-phase estimates were this process most likely excludes nondoublet cells from also comparable, even though different models were used DNA histograms, and may alter cell-cycle analysis. (2.8 ⫾ 1.6%, Multicycle; 3.7 ⫾ 0.8%, Modfit). This result Guidelines for DNA cell-cycle analysis of clinical sam- may be dependent on the low CVs and percentage of ples (25) suggest that specimens be excluded from anal- S-phase cells in these specimens, and could be a source of ysis when the combined levels of sample debris and ag- variation in the cell-cycle analysis of tumor samples that gregates exceed 20% of the total number of histogram do not share these characteristics. S-phase estimates based events. Surprisingly, doublet discrimination is often ab- on computer modeling using Multicycle were significantly sent or used sporadically in multicenter studies comparing lower (P ⬍0.05) than those obtained from pulse-process- interlaboratory variations in DNA measurements and S- ing techniques. As similarly noted by Bergers et al. (32), phase fraction analysis (40 – 42). The discrepancies be- the decrease in S-phase estimates following computerized tween doublet-discrimination techniques presented here doublet correction has the clear potential to introduce and the differences in SPF estimates obtained by different interlaboratory discrepancies in the placement of clinical models and debris corrections (3,4,32,40,43) add yet an- specimens into prognostic subgroupings. Alternatively, other level of complexity to the transposition of prognos- variations in doublet estimates could alter the ploidy clas- tic variables obtained from DNA cell-cycle analysis of sification of diploid/tetraploid DNA histograms (33) or breast tumors between laboratories (24). affect the prognostic value of the percentage of G2 ⫹ M cells in prostate cancer (34). Because of the subjectivity ACKNOWLEDGMENTS involved in positioning the boundary between G2 ⫹ M The authors thank Leopold G. Koss for his critical eval- events and G0/1 doublets in complex heterogeneous sam- uation of the manuscript and his role in stimulating nu- ples, particularly with fluorescence pulse width versus merous thought-provoking issues related to improving the area measurements in fixed-beam commercial instru- potential of DNA flow cytometry as a reproducible, clini- ments, we suggest that G0/1 doublets be estimated by cally relevant prognostic tool. computer modeling. This appears to be the most conser- vative approach for improving reproducibility and en- LITERATURE CITED hancing standardization of doublet discrimination analy- 1. Deitch AD. The effect of cytotoxicity on DNA histograms. Cytometry sis. Perhaps electronic doublet-correction techniques, 1987;8:339. 2. Bauer KD. Quality control issues in DNA content flow cytometry. Ann except for the simplest cases, should be reserved for NY Acad Sci 1993;677:59 –77. situations such as cell sorting requiring real-time, rather 3. Wersto RP, Stetler-Stevenson MA. Debris compensation of DNA his- tograms and its effect on S-phase analysis. Cytometry 1995;20:43–52. than off-line, corrections for doublets. More complex 4. Heiden T, Castro J, Graf B-M, Tribukait B. Comparison of routine flow types of cells, such as epithelial cells, make electronic cytometric DNA analysis of fresh tissues in two laboratories: effects of doublet-correction difficult to reliably reproduce between differences in preparation methods and background models of cell cycle calculation. Cytometry 1998;34:187–197. laboratories that have different electronics of differing 5. Dressler LG, Bartow SA. DNA flow cytometry in solid tumors: prac- sensitivity and precision of calibration. tical aspects and clinical applications. Semin Diagn Pathol 1989;6:55– 82. Human breast tumors are composed of mixtures of 6. Bach BA, Knape WA, Edinger MG, Tubbs RR. Improved sensitivity epithelial tumor cells, stromal tissue, and to a lesser de- and resolution in the flow cytometric DNA analysis of human solid gree, mononuclear cells. The use of cytokeratin antibodies tumor specimens. Am J Clin Pathol 1991;96:615– 627. 7. Sharpless TK, Melamed MR. Estimation of cell size from pulse shape to identify and remove diploid nontumor tissue or infil- in flow cytofluorimetry. J Histochem Cytochem 1976;24:257–264. trating lymphocytes from DNA histograms has been advo- 8. Sharpless T, Traganos F, Darzynkiewicz Z, Melamed MR. Flow cyt- ofluorimetry: discrimination between single cells and cell aggregrates cated as a method to improve histopathologic correlates by direct size measurements. Acta Cytol 1975;19:577–581. in clinical DNA cell-cycle analysis (35–37). To examine 9. Steinkamp JA. Flow cytometry. Rev Sci Instr 1984;55:1375–1400. whether this technique might improve the correlation 10. Gong J, Traganos F, Darzynkiewicz Z. Simultaneous analysis of cell cycle kinetics at two different DNA ploidy levels based on DNA between pulse-height and width-doublet discrimination, a content and cyclin B measurements. Cancer Res 1993;53:5096 –5099. limited number (n ⫽ 5) of mechanically disaggregated, 11. Rabinovitch PS. Practical considerations for DNA content and cell cycle analysis. In: Bauer KD, Duque RE, Shankey TV, editors. Clinical ethanol-fixed intact (37) fresh breast tumor specimens flow cytometry. Principles and application. Baltimore: Williams and were studied. Even after removal of the cytokeratin-nega- Wilkins; 1993. p 117–142. tive elements from the height versus area or width versus 12. Funai S, Kurooka K, Fujii Y, Inufusa H, Yamada H, Imanishi Y, Sato S, Shinko K, Yasutomi M. Multicycle software for cell cycle analyses. area dotplots, G0/1 doublets still did not constitute a dis- Gan To Kagaku Ryoho 1994;21:102–107. crete population clearly separated from G2 ⫹ M cells. 13. Torres FX, Mackowiak PG, Brown RD, Linden MD, Zarbo RJ. Com- parison of two methods of mechanical disaggregation of scirrhous (data not shown). Again, software modeling appears to be breast adenocarcinomas for DNA flow cytometric analysis of whole the most objective rationale for doublet discrimination. cells. Am J Clin Pathol 1995;103:8 –13.
306 WERSTO ET AL. 14. Bergers E, Montironi R, van Diest PJ, Prete E, Baak JP. Interlaboratory 30. Torres K, Horwitz SB. Mechanisms of taxol-induced cell death are reproducibility of semiautomated cell cycle analysis DNA histograms concentration dependent. Cancer Res 1998;58:3620 –3626. obtained from fresh material of 1,295 breast cancer cases. Hum 31. Fuhr JE, Frye A, Kattine AA, Vanmeter S. Flow cytometric determina- Pathol 1996;27:553–560. tion of breast-tumor heterogeneity. Cancer 1991;67:1401–1405. 15. Shapiro HM. Practical flow cytometry. 3rd ed. New York: Wiley-Liss; 32. Bergers E, van Diest PJ, Baal JPA. Comparison of five cell cycle 1995. analysis models applied to 1,414 flow cytometric DNA histograms of 16. Rudolph NS, Ohlsson-Wilhelm BM, Leary JF, Rowley PT. Single-cell fresh frozen breast cancer. Cytometry 1997;30:4 – 60. analysis of the relationships between transferrin receptors, prolifera- 33. Wersto RP, Liblit RL, Koss LG. Flow cytometric DNA analysis of tion, and cell cycle phase in K562 cells. Cytometry 1985;6:151–158. human solid tumors: a review of the interpretation of DNA histo- 17. McDivitt RW, Stone KR, Craig RB, Palmer JO, Meyer JS, Bauer WC. A grams. Hum Pathol 1991;22:1085–1098. proposed classification of breast cancer based on kinetic information. 34. Shankey TV, Kallioniemi O-P, Koslowski JM, Lieber ML, Mayall BH, Cancer 1996;75:269 –276. Miller G, Smith GJ. Consensus review of the clinical utility of DNA 18. Dressler LG, Seamer LC, Owebs MA, Clark GM, McGuire WL. DNA content cytometry in prostate cancer. Cytometry 1993;14:487–500. flow cytometry and prognostic factors in 1331 frozen breast cancer 35. Kimmig R, Spelsberg H, Kapsner T, Untch M, Hepp H. Flow cytomet- specimens. Cancer 1988;61:420 – 427. ric DNA analysis of breast cancer by two colour method using cyto- 19. Clark GM, Dressler LG, Owens MA, Pounds G, Oldaker T, McGuire keratin labeling for the identification of cells. Anal Cell Pathol 1994; WL. Predication of relapse or survival in patients with node-negative 7:205–215. breast cancer by DNA flow cytometry. N Engl J Med 1989;320:627– 36. Wingren S, Stal O, Nordenskjold B. Flow cytometric analysis of 633. S-phase fraction in breast carcinomas using gating on cells containing 20. Sigurdsson H, Baldetorp B, Borg A, Dalherg M, Ferno M, Killander D, cytokeratin. Br J Cancer 1994;69:546 –549. Olsson H. Indicators of prognosis in node-negative breast cancer. 37. Zarbo RJ, Visscher DW, Crissman JD. Two-color multiparametric 1990. N Engl J Med 1990;322:1045–1053. method for flow cytometric DNA analysis of carcinomas using stain- 21. O’Reilly SM, Richards MA. Is DNA flow cytometry a useful investiga- ing for cytokeratin and leukocyte-common antigen. Anal Quant Cytol Histol 1989;11:391– 402. tion in breast cancer? Eur J Cancer 1992;28:504 –507. 38. Graham ML, Bunn P, Jewett PB, Gonzalez-Aller C, Horwitz KB. Simul- 22. Balslev I, Christensen IJ, Brunn Rasmussen B, Larsen JK, Lykkesfedlt taneous measurement of progesterone receptors and DNA indices by AE, Thorpe SM, Rose C, Briand P, Mouridsen HT. Flow cytometric flow cytometry: characterization of an assay in breast cancer cell DNA ploidy defines patients with poor prognosis in node-negative lines. Cancer Res 1989;49:3934 –3942. breast cancer. Int J Cancer 1994;56:16 –25. 39. Lee T-K, Wiley AL, Esinhart JD, Riley RS, Blackburn LD. Variations 23. Bryant J, Fisher B, Gunduz N, Costantino JP, Emir B. S-phase fraction associated with disaggregation methods in DNA flow cytometry. Anal combined with other patient characteristics for the prognosis of Quant Cytol Histol 1993;15:195–200. node-negative, estrogen-receptor-positive breast cancer. Breast Can- 40. Baldetrorp B, Bendahl P-O, Ferno M, Alanen K, Delle U, Falkmer U, cer Res Treat 1998;51:239 –253. Hansson-Aggesjo B, Hockenstrom T, Lindgren A, Mossberg L, Nor- 24. Wenger CR, Clark GM. S-phase fraction and breast cancer—a decade dling S, Sigurdsson H, Stal O, Visakorpi T. Reproducibility in DNA of experience. Breast Cancer Res Treat 1998;51:255–265. flow cytometric analysis of breast cancer: comparison of 12 labora- 25. Shankey TV, Rabinovitch PS, Bagwell B, Bauer KD, Duque RE, Hedley tories’ results for 67 sample homogenates. Cytometry 1995;22:115– DW, Mayall BH, Wheeless L. Guidelines or implementation of clinical 127. DNA cytometry. Cytometry 1993;14:472– 477. 41. Danesi DT, Spano M, Altavista P, Pantarotto M, Filippini C, Mazzini G, 26. Vindelöv LL, Christensen IJ, Nissen NI. A detergent trypsin method Erba E, Ubezio P. Quality control study of the Italian group of for the preparation of nuclei for flow cytometric DNA analysis. cytometry on flow cytometry DNA content measurements: II. Factors Cytometry 1983;3:323–327. affecting inter- and intralaboratory variability. Cytometry 1997;30:85– 27. Wersto RP, Rosenthal ER, Seth PK, Eissa NT, Donahue RE. Recombi- 97. nant, replication-deficient adenoviral gene transfer vectors induce 42. Wheeless LL, Coon JS, Cox C, Deitch AD, de Vere White RW, Fradet cell cycle dysregulation and inappropriate expression of cyclin pro- Y, Koss LG, Melamed MR, O’Connell MJ, Reeder JE, Weinstein RS, teins. J Virol 1998;72:9491–9502. Wersto RP. Precision of DNA flow cytometry in interinstitutional 28. Larsen JK. Washless double staining of a nuclear antigen (Ki-67 or analysis. Cytometry 1991;12:405– 412. bromodeoxyuridine) and DNA in unfixed nuclei. Methods Cell Biol 43. Baldetorp B, Stal O, Ahrens O, Cornelisse C, Corver W, Falkmer U, 1990;33:227–234. Ferno M. Different calculation methods for flow cytometric S-phase 29. Rabinovitch PS. Multicycle: a program for DNA content and cell cycle fraction: prognostic implications in breast cancer? The Swedish So- analysis. Product manual. San Diego: Phoenix Flow Systems; 1991. ciety of Cancer Study Group. Cytometry 1998;33:385–393.
You can also read