Cytometry is defined as quantitative analysis of individual cells and cell constituents. Within the context of this definition, autoradiography and some other early methods of measuring cells such as microspectrophotometry, microfluorometry, and microinteferometry, represent branches of cytometry that preceded flow cytometry. Prior to the dawn of flow cytometry, during 1950s and 60s, autoradiography was the major methodology utilized to measure DNA replication and transcriptional activity in individual cells. Autoradiography relied on analysis of incorporation of the radioactive precursor of DNA (e.g., 3H- or 14C- labeled thymidine) or RNA (3H- or 14C-uridine). The application of nuclear emulsion to the cells on slides followed by photographic silver-grain development provided a means to detect the radiation associated with precursor incorporation. Despite of cumbersome nature of autoradiographic procedures, which required long exposure times and dark-room photographic emulsion processing followed by visual silver grain counts, important discoveries were made. Thus, it could be determined that DNA replication is discontinuous, occurring within a distinct time interval during interphase. This finding led to identification of four major phases of the cell cycle: the DNA pre-synthetic interphase or the “post-mitotic gap” (G1), DNA synthesis phase (S), post-synthetic interphase or “pre-mitotic gap” (G2), and mitosis (M) (1). These phases are still recognized as a foundation of the cell cycle subdivision.
The very elegant technique based on pulse-labeling cells with 3H-thymidine followed by analysis of the percentage of labeled cells moving through mitosis was developed to assess kinetics of cell progression through the cell cycle (2). In this technique, the kinetics of progression of the cohort of cells labeled during the short pulse in S phase through a narrow time-window in M phase provided the means to accurately estimate the duration of each phase of the cycle (TG1, TS, TG2, TM) as wells as the cell cycle duration (Tc). This “frequency of labeled mitosis” (FLM) method was subsequently widely used in cell proliferation studies in vitro and in vivo (3, 4).
Another approach to assess kinetics of cell cycle progression relied on cell arrest in mitosis by mitotic blockers such as colcemid or vinblastine and quantification of frequency of mitotic cells as a function of duration of treatment with the blocker (5). This “stathmokinesis”-based methodology was an alternative to autoradiography in cell kinetics studies and was often utilized in studies of mechanism of action of antitumor drugs on cell cycle progression. In vitro and in vivo applications of 3H- or 14C-thymidine autoradiography yielded significant information on cell cycle distribution and the kinetics of cell proliferation of several normal and cancer cell models (3, 6–8). Continuous in vivo labeling with 3H-thymidine made it possible to define the tumor growth fraction as a cohort of proliferating cells responsible for tumor expansion (9), as well as to distinguish the non-proliferating (quiescent, G0) cells in normal bone marrow (10).
The rate of DNA transcription was estimated by counting silver grain densities in the nuclear emulsion over individual cells that were previously exposed in vitro or in vivo to 3H- or 14C-labeled uridine. The induction of RNA synthesis during mitogenic stimulation of peripheral blood normal lymphocytes, revealed by the synthesis of “rapidly labeled” RNA, detected by autoradiography, made it possible to characterize the early transition of cells from the quiescent G0 to the transcriptionally active G1 state (11). Transcriptional activity and cellular RNA content have subsequently become characteristic and widely used biomarkers to discriminate G1 from G0 cells (12).
Initial Measurements of DNA Synthesis by Flow Cytometry: Fluorescence Quenching upon Incorporation of Halogenated Precursors
The capability of rapid and accurate measurements of cell constituents offered by flow cytometry made it the instrument of choice in analysis of cell proliferation and cell physiology. Instead of radioactive markers, the incorporation of halogenated nucleotides has become the foundation providing new vistas to analyze DNA replication and RNA synthesis. The early progress in this area was characterized by methodological approaches stemming from the observation that the intensity of Hoechst 33358 fluorescence was reduced (quenched) upon binding to DNA containing 5-bromo-2-deoxyyridine (BrdU) (13). The BrdU-incorporating cells therefore could be identified based by their property of quenching Hoechst 33358 fluorescence and the degree of quenching was proportional to the extent of BrdU incorporation. Further progress has been marked by the combined staining of DNA with Hoechst 33358 and with a dye whose fluorescence was unaffected by the incorporated BrdU, such as ethidium bromide (EB). This approach made it possible to identify BrdU-incorporating cells and at the same time to reveal the distribution of cells across all phases of the cell cycle, i.e., G1 vs. S vs. G2M (14). An extension of this methodology involved long-term pulse labeling with BrdU, through one or more cell generations, followed by staining with Hoechst 33342 and EB. This approach made it possible to estimate several kinetic parameters of cell cycle progression during sequential cell division cycles (15, 16).
An elegant procedure that made it possible to measure cellular DNA content (cell cycle phases) concurrently with DNA replication was developed by Crissman and Steinkamp (17). In their approach following incorporation of BrdU, cellular DNA was stained with both, Hoechst 33358 and mithramycin. While Hoechst 33358 fluorescence of DNA containing incorporated BrdU was quenched, fluorescence of mithramycin was not. The differential fluorescence was obtained by subtracting the fluorescence signal of Hoechst 33358 from that of mithramycin, on cell by cell basis, which represented the extent of BrdU incorporation per given cell. When this differential fluorescence was plotted versus intensity of mithramycin fluorescence (DNA content) the bivariate distributions revealed the classical pattern of BrdU incorporation across the phases of the cell cycle (17).
Another approach employed the metachromatic properties of acridine orange (AO) which differentially stains double stranded versus single stranded (denatured) DNA and allows one to identify mitotic cells by flow cytometry (18). Since fluorescence of AO bound to DNA containing incorporated BrdU, as with Hoechst 33358, is quenched, by combining BrdU incorporation and staining with AO it become possible to distinguish BrdU-labeled and -unlabeled mitotic cells (18). In analogy to the autoradiographic method of Quastler and Sherman (2), this approach made it possible to measure, by cytometry, a variety of kinetic parameters of cell cycle progression based on the percentage of labeled mitoses (19).
Immunocytochemical Detection of Incorporated Halogenated Precursors: Analysis of DNA Replication
Further advances in methodological development to detect DNA replication and RNA synthesis by flow cytometry stemmed from the development of antibodies (Abs) targeting the halogenated precursor and immunocytochemical detection of the precursor incorporated into DNA or RNA (20, 21). Analysis of DNA replication by this method required the induction of partial DNA denaturation to make the incorporated BrdU or IdU accessible to the Ab (22,43). DNA denaturation was usually accomplished by treatment of cells with strong acid (2 M HCl) or heat (85–95°C). The bivariate analysis of BrdU incorporation and cellular DNA content provided a significant advance in the methodology (22, 23). In this approach, an antibody (Ab) tagged with green-fluorescing fluorophore (e.g., FITC or AlexaFluor® 488) was used to detect the incorporated BrdU or IdU on the single-strand (denatured) DNA sections while the non-denatured, double-stranded DNA sections were concurrently counterstained with a red fluorescing fluorochrome such as propidium iodide (PI) or 7-aminoactinomycin D (7-AAD). The bivariate distributions of cells stained in this way revealed cellular DNA content (and therefore the cell distribution or DNA ploidy) as well as the presence of incorporated BrdU or IdU, indicating the fraction of cells synthesizing DNA (22). For the next two decades this approach became the predominant methodology used to assess DNA replication in conjunction with cell cycle analysis. It was particularly useful in studies of the kinetics of cell cycle progression in normal cells and in tumors. Several mathematical models have been developed to analyze the movement of the cohort of BrdU pulse-labeled cells through subsequent cell cycle phases (24, 25). Of particular importance in clinical oncology was the method developed by Begg et al. (24) to measure the duration of S phase (TS) and the potential doubling time (Tpot) of tumor cells, from a single sample collected at a particular time after in vivo administration of BrdU.
Time-separated short pulse-exposures of cells to two different halogenated DNA precursors such as IdU and CldU, subsequently detected immunocytochemically using Abs labeled with fluorochromes emitting at different wavelengths, was proposed as a new strategy to significantly expand the analysis of the cell cycle kinetics through consecutive cell divisions in a relatively simple flow cytometric assay (26–28). Thiselegant double-labeling technique for cell cycle kinetic analysis was reminiscent of the autoradiographic approach utilizing short consecutive pulses of cells exposed to 3H- and 14C-labeled thymidine. These DNA precursors were subsequently distinguished on the autoradiograms by differences in length of the tracks of the silver-grains in the nuclear emulsion (29, 30).
As mentioned previously, immunocytochemical detection of incorporated BrdU typically required the harsh step of DNA denaturation by acid or heat. This treatment often ruined cell morphology, and also damaged the epitope of many proteins, precluding their immunocytochemical detection with fluorescence-labeled Abs. Therefore attempts were made to omit the denaturation step in order to make the detection of BrdU incorporation compatible with the use of other immunocytochemical probes. One approach toward this goal was partial DNA degradation with nucleases such as DNase I to make some DNA sections accessible to the BrdU Ab (31). This method did not receive wide acceptance perhaps due to the difficulty in obtaining reproducible (constant) levels of DNA degradation with DNase I when analyzing different cell types or samples.
Another approach for BrdU detection that omits the DNA denaturation step was based on DNA strand breaks induction by photolysis (SBIP) (32–34). This method relied on illumination of cells with UV light to selectively photolyse DNA (induce DNA breaks) at the sections containing the incorporated BrdU. DNA strand breaks generated by photolysis are subsequently labeled with a fluorochrome-tagged deoxynucleotide by utilizing exogenous terminal deoxynucleotidyl transferase, in analogy to the TUNEL assay, developed to detect DNA breaks in apoptotic cells (35, 36). Several kits that rely on this methodology are now commercially available (e.g., ABSOLUTE-SBIP; Molecular Probes, Eugene, OR or Phoenix Flow Systems, San Diego, CA).
Applications of the SBIP methodology made it possible to correlate the expression of cyclin A and cyclin B1, two key proteins driving the cell through the cell cycle, with initiation and termination of DNA replication during S phase (34). Figure 1 illustrates this approach. In this experiment MOLT-4 cells were exposed for 60 min to BrdU and then illuminated with UV light to selectively photolyse DNA at sites that had incorporated BrdU. The cells were then fixed and DNA strand breaks resulting from the photolysis were labeled again with BrdU using terminal transferase and BrdUTP; BrdU was subsequently detected immunocytochemically with anti-BrdU Ab tagged with FITC, without a need for DNA denaturation. Cyclin A or cyclin B1 was detected with phycoerythrin (PE)-conjugated Ab and DNA was counterstained with 7-aminoactinomycin D (7AAD) (34). Gating analysis revealed that the elevated level of expression of cyclin A coincided with the time when cells entered S phase during the 60-min pulse exposure to BrdU (SE). Likewise, the rise in expression of cyclin B1 coincided with entrance to G2 phase (G2E) (34).
Another application of the SBIP technique is shown in Figure 2. In this application, the SBIB approach was used to detect DNA replicating cells concurrent with cells undergoing apoptosis (33). Sequential labeling of DNA strand breaks, first those induced by apoptosis with PE, then the breaks induced by DNA photolysis with FITC and counterstaining DNA with 7-AAD made it possible to simultaneously detect DNA replication and apoptosis and to correlate them with cell cycle phase (33).
Analysis of RNA Synthesis
As in the case of DNA replication, the flow cytometric analysis of RNA synthesis was based on immunocytochemical detection of the incorporated halogenated RNA precursor, BrU or BrUTP (38, 39). There are several variants of this methodology, differing primarily in the procedure by which the cells were permeabilized to make the incorporated BrU accessible to Ab (40). All variants utilized Abs marketed as anti BrdU which happen to cross-reacts with BrU. This is expected since nearly all commercially available BrdU Abs were generated using 5-bromouridine rather than 5-bromo-2-deoxyuridine to immunize rodents or rabbits. Either direct or indirect immunocytochemical procedures employing Ab tagged with fluorochromes of any desired emission and excitation properties could be used. Depending on the duration of the BrU pulse during in vitro incubation, one could preferentially incorporate BrU either into the rapidly turning over RNA found in nucleoli and associated with synthesis and processing of pre-ribosomal RNA species, or incorporate into messenger or ribosomal RNA(41, 42).
Cytometric assessment of BrU incorporation combined with analysis of cellular DNA content using the respective DNA and RNA probes of different emission color provided the basis for correlating the overall transcriptional activity with cell cycle phase or cell ploidy (40–42). Detection of the incorporated BrU concurrent with cell surface immunophenotyping offered the possibility to correlate cell phenotype with its transcriptional activity (39).