An improved mathematical method to estimate DNA synthesis time of bromodeoxyuridine-labelled cells, using FCM-derived data

Abstract. Chinese hamster ovary cells in vitro were pulse-labelled with bromodeoxyuridine (BrdUrd and were then allowed to progress through the cell cycle. Every half hour after labelling, cells were harvested and prepared for simultaneous flow cytometric determination of DNA content and incorporated BrdUrd, with the intercalating dye propidium iodide and with a monoclonal antibody against incorporated BrdUrd, respectively. The relative movement (RM), i.e. the relative mean DNA content of the moving cohort of BrdUrd-labelled cells in relation to that of G₁ and G₂ cells, was calculated. RM was then used to calculate DNA synthesis time (T_S), at all post-labelling times (t). Since labelled cells in G₂ and mitosis (M) in addition to S phase cells, are included in the cohort of moving labelled cells, and since the time of G₂ and M (T_g2+M) phases is finite, a non-linear relationship exists between RM and post-labelling time. Because of this, the use of a linear formula in the calculation of T_S yields results that are affected by t. We found that RM data can be corrected with regard to T_G2+M resulting in the derivation of a non-linear T_S formula. This non-linear T_S formula gave results that were nearly independent of t. Moreover, windows were set in the mid DNA distributions for G₁, S and G₂+ M cells in the bivariate DNA v. BrdUrd cytograms, to estimate the fraction of BrdUrd-labelled cells in each window at every post-labelling time. Plots of the fraction of BrdUrd-labelled cells v. post-labelling time were then made for each window. T_S obtained in this way was in agreement with T_S obtained with the corrected RM method. In conclusion, we present a method to calculate T_s which theoretically first makes the determination of RM independent of T_G2+M, and secondly compensates for the non-linear function of RM with post-labelling time caused by accumulation of BrdUrd-labelled cells in G₂+ M.     

Flow cytometric estimation of cell cycle parameters using a monoclonal antibody to bromodeoxyuridine

An estimation of cell kinetic parameters was made by simultaneous flow cytometric measurements of DNA and bromodeoxyuridine (BrdUrd) contents of cells. The procedure described in this paper involves the incorporation of BrdUrd by S phase cells, labeling the BrdUrd with an indirect immunofluorescent technique using a monoclonal anti-BrdUrd antibody, and staining DNA with propidium iodide (PI). The amount of incorporated BrdUrd in HeLa cells was proportional to that of synthesized DNA through S phase. For all cell lines examined, the pattern of BrdUrd incorporation was essentially the same and the rate of DNA synthesis during S phase was not constant. The bivariate BrdUrd/DNA distributions showed a horse-shoe pattern, maximum in the mid S phase and minimum in the early and late S phases. Furthermore, the durations of cell cycle (Tc) and S phase (Ts) were estimated from a FLSm (fraction of labeled cells in mid S phase) curve that was generated by plotting the percentage of BrdUrd pulse-labeled cells in a narrow window defined in the mid S phase of the DNA histogram. The values of these parameters in NIH 3T3, HeLa S3, and HL-60 cells were in good accordance with the reported data. This FCM method using the monoclonal anti-BrdUrd antibody allows rapid determination of both cell cycle compartments and also Ts and Tc without the use of radioactive DNA precursors.     

Flow cytometric analysis of the kinetic effects of cisplatin on lung cancer cells

The effects of cisplatin on the cell cycle and DNA synthesis of human lung adenocarcinoma cell line PC-9 were examined by flow cytometry. The cellular DNA content and the bromodeoxyuridine (BrdUrd) incorporation rate were measured simultaneously using a monoclonal anti-BrdUrd antibody. Following exposure to cisplatin (1.0 micrograms/ml) for 1 and 24 hr, the bivariate DNA/BrdUrd distributions revealed a delayed S-phase transit and an accumulation of cells in the G2M phase. The BrdUrd-linked green fluorescence intensity continued to decrease with the lapse of time. However, early- and mid-S-phase cells soon recovered DNA synthesis activity, and the former showed higher activity than the control cells. These findings suggested the vigorous DNA synthesis of cells in early S phase. However, for quantitative analysis of chemotherapeutic effects, some problems remained to be resolved regarding the condition for DNA denaturation and its alteration by the agents.     

Evaluation of four methods of DNA distribution data analysis based on bromodeoxyuridine/DNA bivariate data

Four published methods of DNA-content histogram analysis (those of Fried, Dean and Jett, simplified Dean, and Fox) were compared using a double labeling of different cell populations. Partially synchronized and asynchronous cell populations were incubated with bromodeoxyuridine (BrdUrd) and then stained with an anti-BrdUrd monoclonal antibody and propidium iodide (PI). The fractions of cells in the G1, S, and G2 + M phases were calculated by each method and compared with those derived from G1, S, and G2 + M areas plotted on BrdUrd/DNA bivariate histograms, taken as the "true" values. This procedure enabled an optimal choice of method for a given cell population.     

Cell cycle perturbation by sodium butyrate in tumorigenic and non-tumorigenic human urothelial cell lines assessed by flow cytometric bromodeoxyuridine/DNA analysis

The effect of sodium butyrate on cell proliferation was studied in eight human urothelial cell lines differing in transformation grade (TGr): Hu 1752 (mortal, TGr I); HCV29 (immortal and tumorigenic, TGr II); HCV29T, T24, T24A, T24B, Hu 961A and Hu 1703He (tumorigenic, TGr III). In all cell lines, except Hu 1752, addition of 4 mm sodium butyrate at 18 h after replating resulted in a significantly decreased population of adherent cells after a further 24–48 h. This might partially be explained by detachment of cells, probably mainly S phase cells, from the substrate in the lines HCV 29, HCV29T, Hu 961A and Hu 1703He. Flow cytometric DNA analysis of the adherent cell population showed that all TGr II and III urothelial cell lines were DNA aneuploid, and that butyrate perturbed the cell cycle distribution in these cell lines, mainly by a decrease of the S phase fraction. Flow cytometric bromodeoxyuridine (BrdUrd)/DNA analysis of continuously BrdUrd labelled cultures, using a ‘washless’ BrdUrd/DNA staining technique, showed that butyrate inhibited the G_0/1-S phase transition, indicated by a delayed depletion of BrdUrd negative G_0/1 cells in the cell lines HCV29, HCV29T, T24B, Hu 961A and Hu 1703He. BrdUrd/DNA analysis further showed that butyrate inhibited the G₂M-G_0/1 phase transition, indicated by a pronounced accumulation of BrdUrd positive G₂M cells in the cell lines HCV29T, T24B, Hu 961A and Hu 1703He. Microscopy of HCV29T and Hu 961A cells indicated that this block did not occur in mitosis. The parental cell line T 24 and the cell line T 24 A did not show an accumulation of BrdUrd negative G_0/1 cells or BrdUrd positive G₂M cells like that occurring in the derived cell line T 24B.     

Macrophage-induced cytostasis: kinetic analysis of bromodeoxyuridine-pulsed cells

The effect of tumoricidal macrophages on the cell cycle progression of six different cell lines was studied using an anti-bromodeoxyuridine (BrdUrd) monoclonal antibody to follow the traverse of BrdUrd-labeled cells. Exponentially growing cultured mammalian cells, from six different cell lines, were prepulsed with BrdUrd before exposure to tumoricidal macrophages. The cultured cells were then analyzed as a function of time for DNA content (by propidium iodide staining) and for BrdUrd incorporation (using a fluoresceinisothiocyanate [FITC]-conjugated anti-BrdUrd monoclonal antibody). The position of the cells in cycle and the progression of the BrdUrd-labeled cohort was followed using flow cytometry. The cell lines examined were: Colon 26, BALB/c-3T3, ST3T3 (a spontaneously transformed, tumorigenic clone of 3T3), WCHE5 (a clone of whole Chinese hamster embryo cells), RIF (a radiation-induced fibrosarcoma), and A101D (a human melanoma). The bivariate distributions showed that for all six cell lines the BrdUrd-labeled cohort in the control cultures progressed around the cell cycle during the first 12 h of culture, as the cells exponentially increased. In contrast, when each cell line was incubated with tumoricidal macrophages, the BrdUrd-labeled cohort did not progress through cell cycle but remained in S phase throughout the 12-h culture period. There was also no evidence for progression of cells out of G1. The data show that cells were arrested in every phase of cell cycle. This study suggests that cytostasis, as manifested by the termination of progression in all phases of the cell cycle, is a universal phenomenon induced by tumoricidal macrophages.     

Moderate and severe malnutrition alters proliferation of spleen cells in rats

Objectives

Previous studies have shown alterations in bone marrow cell proliferation in malnourished rats, during lactation. The objective of this study was to determine in vivo effects of moderate and severe malnutrition on spleen cell proliferation in 21‐day‐old rat pups.

Materials and methods

Spleen cell proliferation was determined following administration of bromodeoxyuridine (BrdUrd) over a time course of 2, 4, 6 and 8 h. Incorporation of BrdUrd was detected using FITC‐conjugated anti‐BrdUrd monoclonal antibodies and total DNA content was detected and evaluated using propidium iodide using flow cytometry.

Results

Proportions of cells in S and G₂/M were reduced in the rats with moderate (MN2^nd) and severe (MN3^rd) malnutrition. BrdUrd incorporation was lower in both groups of malnourished rat. In cells of MN2nd individuals, length of G₁ became shorter, while length of S‐phase increased. In contrast, fraction of cells in proliferation was significantly lower in both groups of malnourished rat, with MN3rd group having lowest percentage of cell population growth. In this study, severe malnutrition did not significantly affect duration of phases of the cell cycle, although fractions of proliferating cells were dramatically reduced.

Conclusion

Moderate malnutrition increased time of cells in DNA synthesis and time of total cell cycle and severe malnutrition reduced growth fraction of spleen cells in malnourished rats during lactation.

     

Unit gravity sedimentation analysis of basic nuclear protein synthesis during the HeLa cell cycle

Summary Exponentially growing HeLa cells have been separated according to their cell cycle age by sedimenting at unit gravity for 3 hr on a phosphate-buffered sucrose density gradient. Measurements of cell size, cell number, DNA content, and tritiated thymidine incorporation in consecutive portions of the gradient showed that cells in upper fractions were in G₁, cells in middle fractions were in S, and cells in lower fractions were in G₂. Basic amino acids were rapidly incorporated into nuclear protein during late G₁ and S; some incorporation also took place during G₂. This work is supported by grant A-3458 from the National Research Council of Canada.     

Progressive Increase In Polyamine Levels In 9l Cells in vitro During the Cell Cycle: Comparison Between Cells Isolated By Centrifugal Elutriation and Cells Grown In Synchrony

Centrifugal elutriation was used to separate 9L rat brain tumour cells into fractions enriched in the G₁, S, or G₂/M phases of the cell cycle. Cells enriched in early G₁, phase were recultured, grown in synchrony, and harvested periodically for analysis of their DNA distribution and polyamine content. Mathematical analysis of the DNA distributions indicated that excellent synchrony was obtained with low dissersion throughout the cell cycle. Polyamine accumulation began at the time of seeding, and intracellular levels of putrescine, spermidine, and spermine increased continuously during the cell cycle. In cells in the G₂/M phase of the cell cycle, putrescine and spermidine levels were twice as high as in cells in the G₁, phase. DNA distribution and polyamine levels were also analysed in cells taken directly from the various elutriation fractions enriched in G₁, S, or G₂/M. Because we did not obtain pure S or G₂/M populations by elutriation or by harvesting synchronized cells, a mathematical procedure—which assumed that the measured polyamine levels for any population were linearly related to the fraction of cells in the G₁, S, and G₂/M phases times the polyamine levels in these phases and that polyamine levels did not vary within these phases—was used to estimate ‘true’ phase-specific polyamine levels (levels to be expected if perfect synchrony were achieved). Estimated ‘true’ phase-specific polyamine levels calculated from the data obtained from cells either sorted by elutriation or obtained from synchronously growing cultures were very similar.     

Synchronization of 9L rat brain tumor cells by centrifugal elutriation

Asynchronous 9L cells were separated into relatively homogeneously-sized populations using centrifugal elutriation with both a conventional collection method and a long collection method. A substantial increase in the homogeneity of the volume distributions and in the degree of synchrony of the separated fractions was obtained using the long collection method. Autoradiographic data indicated that fractions containing ≥97% G₁ cells, ≥80% S cells, and 70–75% G₂ cells could be routinely recovered with this procedure. Recovery in these fractions varied from 5 to 8% of the total number of cells elutriated. The colony forming efficiency (CFE) of cells from fractions representing each phase of the cell cycle was a constant 60–70%, which was comparable to the 60–80% usually found for asynchronous 9L cells. The percentage of cells in the G₁, S, and G₂ phases in the elutriated fractions was more accurately determined from the volume distribution than from computer fits of the DNA histogram obtained from flow cytometry. In general, the degree of synchrony was related to the coefficient of variation (CV) of the volume distributions of the elutriated fractions. The CV was about 14% for all elutriated fractions. When the ≥97% G₁ population was allowed to progress to S and G₂, the CVs were about 17 and 20.2%, respectively. Thus, the best nonperturbing method for obtaining synchronous 9L cells in the S or G₂ phases was direct elutriation with the long collection method.     

Flow cytometric BrdUrd-pulse-chase study of X-ray-induced alterations in cell cycle progression

To better understand how the flow cytometric bromodeoxyuridine (BrdUrd)-pulse-chase method detects perturbed cell kinetics we applied it to measure cell cycle progression delays following exposure to ionizing radiation. Since this method will allow both the use of asynchronous cell populations and the determination of the alterations in cell cycle progression specific to cells irradiated in given cell cycle phases, it has a significant advantage over laborious synchronization methods. Exponentially growing Chinese hamster ovary (CHO) K1 cells were irradiated with graded doses of X-rays and pulse-labelled with BrdUrd immediately thereafter. Cells were subcultured in a BrdUrd-free medium for various time intervals and prepared for flow cytometric analysis. Of five flow cytometric parameters examined, only those that involved cell transit through G₂, i.e. the fraction of BrdUrd-negative G₂ cells and the fraction of BrdUrd-positive cells that had not divided, showed radiation dose-dependent delays. The magnitude of the effects indicates that the cells irradiated in G₂ and in S are equally delayed. S phase transit of cells irradiated in S or in G₁ did not appear to be affected. There were apparent changes in flow of cells out of G₁, which could be explained by the delayed entry of G₂ cells into the compartment because of G₂ arrest. Thus, in asynchronous cells the method was able to detect G₂ delay in those cells irradiated in S and G₂ phases and demonstrate the absence of cell-cycle delays in other phases.     

Cell cycle analysis using numerical simulation of bivariate DNA/bromodeoxyuridine distributions

This report describes a mathematical model of cell proliferation for simulation of bivariate DNA/bromodeoxyuridine (BrdUrd) distributions. The model formulates the change with time in the frequency of cells with any DNA content and in the amount of incorporated BrdUrd, according to given cytokinetic parameters, i.e., durations and dispersions of cell cycle phases and DNA synthesis rate during S-phase. We have applied this model to sequential DNA/BrdUrd distributions measured for Chinese hamster ovary cells asynchronously grown in vitro, 1) for 30 min in 10 microM BrdUrd followed by growth in BrdUrd-free medium for 0 to 24 h, or 2) during continuous incubation in 3 microM BrdUrd plus 30 microM thymidine for 2 to 24 h. The matches between the experimental and simulated distributions give the G1, S, G2M, and total cell cycle durations (and coefficients of variation) of 5.6 h (0.08), 7.0 h (0.07), 1.4 h (0.16), and 14.0 h (0.05), respectively. The model is shown to be useful for quantitative interpretation of the bivariate distributions.     

Mathematical model analysis of mouse epidermal cell kinetics measured by bivariate DNA/anti-bromodeoxyuridine flow cytometry and continuous [3H]-thymidine labelling

Abstract. In a previous study the epidermal cell kinetics of hairless mice were investigated with bivariate DNA/anti-bromodeoxyuridine (BrdU) flow cytometry of isolated basal cells after BrdU pulse labelling. The results confirmed our previous observations of two kinetically distinct sub-populations in the G₂ phase. However, the results also showed that almost all BrdU-positive cells had left S phase 6–12 h after pulse labelling, contradicting our previous assumption of a distinct, slowly cycling, major sub-population in S phase. The latter study was based on an experiment combining continuous tritiated thymidine ([³H]TdR) labelling and cell sorting. The purpose of the present study was to use a mathematical model to analyse epidermal cell kinetics by simulating bivariate DNA/BrdU data in order to get more details about the kinetic organization and cell cycle parameter values. We also wanted to re-evaluate our assumption of slowly cycling cells in S phase. The mathematical model shows a good fit to the experimental BrdU data initiated either at 08.00 hours or 20.00 hours. Simultaneously, it was also possible to obtain a good fit to our previous continuous labelling data without including a sub-population of slowly cycling cells in S phase. This was achieved by improving the way in which the continuous [³H]TdR labelling was simulated. The presence of two distinct sub-populations in G₂ phase was confirmed and a similar kinetic organization with rapidly and slowly cycling cells in G₁ phase is suggested. The sizes of the slowly cycling fractions in G₁ and G₂ showed the same distinct circadian dependency. The model analysis indicates that a small fraction of BrdU labelled cells (3–5%) was arrested in G₂ phase due to BrdU toxicity. This is insignificant compared with the total number of labelled cells and has a negligible effect on the average cell cycle data. However, it comprises 1/3 to 1/2 of the BrdU positive G₂ cells after the pulse labelled cells have been distributed among the cell cycle compartments.     

Flow cytometric BrdUrd-pulse-chase study of heat-induced cell-cycle progression delays

The flow cytometric, bromodeoxyuridine (BrdUrd)-pulse-chase method was extended by analysing five kinetic parameters to study perturbed cell progression through the cell cycle. The method was used to analyse the cell-cycle perturbations induced by heat shock. Exponentially growing, asynchronous Chinese hamster ovary (CHO) cells were pulse labelled with BrdUrd and simultaneously heated at 43°C for 5,10 or 15 min. The cells were then incubated in a BrdUrd-free medium and, at various times thereafter, were prepared for flow cytometry. Five compartments (BrdUrd-labelled divided and undivided, and unlabelled G₁, G₁S, and G₂) were defined in the resulting dual-parameter histograms. The fraction of cells and the mean DNA content, when appropriate, were calculated for each compartment. The rates of cell-cycle progression were assessed as time-dependent changes in the fraction of cells in a given compartment and/or the relative DNA content of cells within a given compartment. Linear regression analysis of the data revealed two distinct modes of alteration in cell progression: 1 a delay in cell transit (either out of or into a given compartment), and 2 a decrease in the rate of cell transit. Hyperthermia produced a delay in the exit of cells from the G₁ compartment of ≈ 16 min per minute of heat at 43°C with no threshold. In contrast, the delay in the exit of cells from all other compartments showed a threshold of from 3 to 5 min at 43°C. Above this threshold the delay in exit of cells from the BrdUrd-labelled, undivided compartment was 25 min per minute of heat at 43°C. The more complex dose-response function of this latter compartment may reflect the fact that it includes two cell-cycle phases, S and G₂+ M. The decrease in the rate of transit out of G₂ for cells heated in G₂ was significantly larger than that for any other compartment, consistent with previous studies, which showed a G₂ accumulation following hyperthermia. These results indicate that heat exposure induces very complex alterations in cell-cycle progression and that this flow cytometric method offers a straightforward approach for observing such alterations.     

Bromodeoxyuridine: a diagnostic tool in biology and medicine, Part I: Historical perspectives, histochemical methods and cell kinetics

Summary Bromodeoxyuridine (BrdUrd), a thymidine analogue incorporated into DNA, can be quantified by fluorescent or chromophoric quenching of dyes bound to DNA or with antibodies to BrdUrd. These technologies have been used since the 1970s as tools for measuring DNA synthesis in isolated chromosomes and in cells and tissues. This paper is Part I of a three-part comprehensive review of the literature over the last 20 years (to the end of 1993) describing the histochemical methods for measuring BrdUrd in cells and tissues. Fixation, denaturation and staining procedures are compared for quantifying BrdUrd for microscopy and flow cytometry. Non-immunochemical methods related to the quenching of fluorescent DNA stains by BrdUrd are also described. Methods are described for the comparative assay of cell kinetic parameters by tritiated thymidine and bromodeoxyuridine. The multivariate BrdUrd/DNA assay of T_s, and T_c, and a comparison of recent methods based on the single biopsy bivariate analysis of T_pot, is presented. Recent developments in the use of double halopyrimidine label to determine kinetic parameters are also reviewed.     

Proliferation of germ cells and somatic cells in first trimester human embryonic gonads as indicated by S and S+G2+M phase fractions

Objectives: The number of germ cells and somatic cells in human embryonic and foetal gonads has previously been estimated by stereological methods, which are time‐ and labour‐consuming with little information concerning cell proliferation. Here, we studied whether flow cytometry could be applied as an easier method, also enabling estimation of the fraction of cells in S or S+G₂+M (SG₂M) cell‐cycle phases as indicators of cell proliferation. Methods: Cell suspensions from 35 human embryonic gonads at days 37 to 68 post‐conception (pc) were immunomagnetically sorted into C‐KIT positive (germ) cells and negative (somatic) cells. They were stained for DNA content and analysed by flow cytometry. S and SG₂M fractions could be measured for 13 of the female and 20 of the male gonads. The number of cells was estimated using fluorescent reference beads. Results: During the period from 37 to 68 days pc, female germ and somatic cells had a stable S and SG₂M fractions indicating steady growth of both subpopulations, whereas they decreased in both male germ and somatic cells. The number of germ and somatic cells estimated by flow cytometry was significantly lower than in stereological estimates, suggesting loss of cells during preparation. Conclusions: Cell proliferation as indicated by S and SG₂M fractions could be estimated specifically for primordial germ and somatic cells. Estimation of total number of germ and somatic cells was not feasible.     

FLOW CYTOMETRIC CELL CYCLE ANALYSIS USING THE QUENCHING OF 33258 HOECHST FLUORESCENCE BY BROMODEOXYURIDINE INCORPORATION

Cultures of L cells were grown in medium containing 2.0 mg/l bromodeoxyuridine (BUdR) and stained with the fluorescent dye 33258 Hoechst for flow cytometric analysis. During exposure to BUdR, the cells replace thymidine by BUdR in the newly synthesized DNA. The new DNA is not stainable with 33258 Hoechst, which is highly specific for thymidine. The temporal development of the fluorescence distributions after addition of BUdR to the growth medium has been investigated in the flow cytometer, and the data were used to calculate the mean durations of the phases G₁, S and G₂+ M in exponentially growing cultures as well as the cycle transit times in synchronized cultures. The percentage of non-cycling cells was determined in each experiment.     

The Circadian Variations In the Eithelial Growth of the Hamster Cheek Pouch: Quantitative Analysis of Dna Distributions*

The pronounced diurnal rhythm in DNA distribution of the hamster check pouch epithelium both in the S fraction and in the (G₂+ M) fraction was compared with previous studies of the changes in tritiated thymidine labelling index and mitotic activity. the DNA distributions were obtained by flow cytometry after ultrasonic disaggregation of the isolated epithelium into a suspension of single nuclei. the DNA distributions were analysed with the computer program of J. Fried (1976) and by planimetry. the S fraction was higher than the autoradiographic labelling index during the whole 24 hr period. Only the computer fitted S fraction and the labelling index had the same difference between maximal and minimal values, and maxima at the same time of day. the DNA distributions showed a diurnal release of G₁ cells into S phase proceeding through (G₂+ M) phase and returning to G₁ phase within a 24 hr period.     

Kinetic heterogeneity of an experimental tumour revealed by BrdUrd incorporation and mathematical modelling

In the present paper we propose a method of analysis of the cell kinetic characteristics of in vivo experimental tumours, that uses DNA-BrdUrd flow cytometry data at various times after the bromodeoxyuridine (BrdUrd) injection and mathematical modelling. The model of the cell population takes into account the cell-cell heterogeneity of the progression rate across cell cycle phases within the tumour, and assumes a strict correlation between the durations of S and G2M phases. The model also allows for a nonconstant DNA synthesis rate across S phase. In addition, the measurement process is modelled, considering the possibility of nonimpulsive labelling and providing a representation of the time course of the bivariate DNA-BrdUrd fluorescence distribution. Sequential DNA-BrdUrd distributions were obtained in vivo from a human ovarian carcinoma transplanted in mice and, for comparison, in vitro from a cell line of the same origin. From these data, that included the fractional density and the mean BrdUrd-fluorescence of BrdUrd-positive cells as a function of the DNA-fluorescence, kinetic parameters such as the potential doubling time (T _pot) and the mean and variance of the transit times in S and G2M phases, were estimated. This study revealed the presence of a substantial heterogeneity in S and G2M phases within the in vivo cell population and of a lower heterogeneity in the in vitro population. Moreover, our analysis suggests a nonnegligible effect of the BrdUrd pharmacokinetics in the in vivo cell labelling.