British Iournal of Harrnatolog!y, 1989. 72, 492-196

A flow cytometric assay for the determination of cell proliferation with a monoclonal antibody directed against DNA-methyltransferase

ANDREAS NEUBAUER, STEFAN SERKE, WOLFGANG SIEGERT, WERNER K R O L L , * REINHARD MUSCH A N D

DIETER HUHN Universitiitsklinikum Rudolf-Virchow, Abteilung lnnere Medizin. Spandauer Damm 130, D l 0 0 0 Berlin 19, and *Merck GmbH, Darmstadt, F.R.G.

Received 1 9 October 1988; accepted for publication 3 March 1989

Summary. The enzyme DNA-methyltransferase is respon- sible for the methylation of a newly synthesized DNA-strand. A monoclonal antibody directed against DNA-methyltrans- ferase was used to determine cell proliferation by means of flow cytometry. The reactivity of DNA-methyltransferase antibody was compared with the known proliferation markers transferrin-receptor and Ki67. All three methods showed comparable reactivity with the erythroblastic cell line K562 (86%. 81%. 76% respectively). In a second set of experiments peripheral blood mononuclear cells were stimu-

lated with phytohaemagglutinin; in three experiments, a mean of 63% ofthe cells reacted with DNA-methyltransferase antibody after 72 h of culture as compared to a mean of 6% in the case of unstimulated control cells. HL60 cells were incubated with DMSO and harvested on day 5 of culture. The results obtained show that in differentiated cells the fraction positive with DNA-methyltransferase antibody decreased to levels below 10%. It is concluded that the technique described is a fast and easy method for the flow cytometric determina- tion of cellular proliferation.

The enzyme DNA-methyltransferase is responsible for the methylation of cytosine residues and plays an important role in the regulation of cellular growth. Most of the methylated cytosine residues are found in cytosine-guanosine (CG) doublets (Lewin, 1987. p. 549). 70% of which-in animal cells-are methylated. The methylation of cytosine follows replication of DNA (Torsten. 1986). Replication of a fully methylated CG doublet produces two hemimethylated daughter duplexes (Lewin, 198 7. p. 5 50). There is evidence that DNA-methyltransferase acts only on hemimethylated DNA (Lewin. 1987, p. 550). Thus, in a cellular population with high proliferative activity (and high level of DNA- replication), such as a cell line in culture, the enzyme DNA-methyltransferase has a high activity. On the other hand, the methylation of DNA is important for the regulation of gene expression. IJndermethylated genes can be tran- scribed. whereas fully methylated genes are silent. The absence of methyl groups is therefore associated with the ability to be transcribed (Doeder. 1983: Lewin. 1987,

Correspondence: Dr A. Neubauer, Lineberger Cancer Research Center, The University of North Carolina at Chapel Hill. Chapel Hill. NC 27599-7295. U.S.A.

p. 550). There is evidence that this methylation of DNA is regulated by an entirely different enzyme activity (Lewin,

Measurement of cellular proliferation is of great impor- tance in cell biology. One standard method is based on the incorporation of radioactively labelled DNA-precursors. However, this method is quite complicated and necessitates the use of radioactive markers. The flow cytometric counter- part to thymidine uptake is the incorporation of bromodeoxy- uridine into newly synthesized DNA and subsequent staining with a monoclonal antibody against bromodeoxyuridine (Gratzner. 1982). However, this assay is not very simple to perform either. Recently, several alternative methods have been described, such as the demonstration of transferrin receptor (TRF) (Cotner et al, 1983; Serke et al, 1987) or the proliferation-associated antigen Ki67 (Gerdes et al, 1983; Palutkeet al. 1987; Schwartingetal. 1986) withmonoclonal antibodies. However, transferrin positive cells are not always proliferating (e.g. resting monocytes also stain positively) and staining with antibodies against Ki67 may be difficult to perform with stored specimens due to instability of the Ki67 antigen (Neubauer & Huhn. 1987). Here, a fast and simple method is described using a monoclonal antibody against the enzyme DNA-methyltransferase and flow cytometry.

1987. pp. 550-551).

492

Cell Proliferation and DNA-Methyl Transferase 493

MATERIALS AND METHODS

Monoclonal antibodies. The monoclonal antibody directed against DNA-methyltransferase (Drahovsky et al, 1985) was donated by Merck GmbH, Darmstadt. The dilution used was 1 : 200.

The monoclonal antibody directed against the nuclear proliferation antigen Ki67 was purchased from Dianova, Hamburg, F.R.G. (dilution 1 : 5). The monoclonal control antibody MPCll was a generous gift from Dr Holzmann, Institut fur Immunologie, Munchen, F.R.G. The fluorescein- conjugated monoclonal antibody directed against the transferrin receptor was purchased from Becton Dickinson, Heidelberg, F.R.G.

Cells. Various human cell lines were cultured under standard conditions in RPMI 1640 with 10% fetal calf serum (Gibco, Frankfurt, F.R.G.), penicillin and streptomycin. Ex- ponentially growing cells were harvested and washed twice in phosphate-buffered saline (PBS) (Seromed, Berlin, F.R.G.). The percentage of dead cells was evaluated using trypan-blue exclusion. Less than 5% dead cells were counted in all cell lines.

Peripheral blood mononuclear cells were obtained by means of Ficoll-Hypaque separation of anticoagulized whole blood from healthy volunteers. Cytospin preparations were performed under standard conditions.

Phytohaemagglutinin stimulation. Peripheral blood mono- nuclear cells were cultured under standard conditions in IMDM with 10% fetal calf serum containing phytohaem- agglutinin (PHA) (Seromed. Berlin, F.R.G.) (10 pg/ml) in a 24-well titre plate (Costar, Cambridge, Mass., U.S.A.). Control cells were incubated without PHA. Cells were harvested after 72 h of culture.

Cellular diflerentiation of H L 6 0 cells. HL60 cells (American type culture collection) were incubated at a concentration of 3 x 105/ml under standard conditions as mentioned above. 1% (v/v) DMSO (Serva, Heidelberg, F.R.G.) was used for the induction of granulocytic differentiation of the cells; control cells were incubated without DMSO. Cells were harvested after 5 d of culture.

Staining and fixation. DNA-methyltransferase and Ki67: The cells were fixed as follows: blood mononuclear cells or cells from culture were harvested, washed twice in PBS (Seromed. Berlin, F.R.G.) and the pellet fixed with ice-cold ethanol (Merck, Darmstadt, F.R.G.) (70% v/v) for 10 min at 4OC. Thereafter, cells were washed twice in PBS and lo6 cells were incubated with the appropriately diluted monoclonal antibody for 30 min at 4 O C . Cells were again washed twice in PBS and incubated with a fluorescein-conjugated goat anti- mouse antibody (Fab2) (Paesel, Frankfurt, F.R.G.) (dilution 1:20) for 30 min at 4OC. After two further washes in PBS, cells were analysed with a FACSCAN flow cytometer (Becton Dickinson). For double staining with the DNA dye propidium iodide, the pelleted cells were incubated in PBS containing 40 pg/ml propidium iodide and 100 pg/ml FWAse (Sigma, Munchen, F.R.G.) at room temperature for 30 min prior to flow cytometry.

Transferrin receptor antibody: lo6 cells were incubated for 30 min at 4OC (dilution according to the manufacturer),

Table 1. Reactivity of various human cell lines with the monoclonal DNA-methyltransferase antibody.

Cell line Origin Positive cells i%)

MOLT4 JURKAT CCRF.CEM REH

HRIK DAUDI NAMALWA RPMI8 2 2 6 u937

HL60 K5h2 HEL92.17

T-ALL T-ALL T-Cells ALL

Burkitt lymphoma Burkitt lymphoma B-cell lymphoma Myeloma (IgG kappa) Histiocytic lymphoma

AML CML, blast crisis Erythroleukaemia

95 85 93 95

86 97 97 92 97

97 93 94

PBL Control lymphocytes 2

AML acute myeloid leukaemia: CML: chronic rnye- loid leukaemia; ALL acute lymphocytic leukaemia; PBL peripheral blood lymphocytes (healthy volunteer).

washed twice in PBS and analysed by means of flow cytometry. A FITC-conjugated control antibody purchased from Becton Dickinson was used as a control.

Immunocytochemistry. The APAAP-technique was per- formed according to Cordell et a1 (1984). The following monoclonal antibodies were employed: mpcl 1, anti-DNA- methyltransferase and anti-transferrin-receptor (Becton Dickinson). Flow cytornetry. A FACSCAN flow cytometer equipped with

a 15 mW argon laser with excitation wavelength at 488 nm was used (Becton Dickinson). Fluorescence was analysed at 530 nm (FITC) and 585 nm (propidium iodide) using the FACSCAN research software program. Fluorescence I (FITC) was amplified logarithmically, fluorescence I1 (propidium iodide) linearly. The photomultiplier and the threshold for fluorescence I were adjusted such that the cells stained with the negative control antibody showed less than 2% of the cells to be positive. The software stores each of the four parameters (forward and side scatter fluorescence I and 11) in list mode (1024 channels each) on a 20 Mbyte hard disk of a microcomputer system (Hewlett Packard 3 10). Data were analysed in a second step using the software provided by the manufacturer. The DNA cycle analysis was conducted using the DNA-software from Becton Dickinson.

RESULTS Cell lines Various human cell lines were tested; results obtained are summarized in Table I.

The erythroblastic cell line K562 was used to compare: (1) transferrin receptor expression (81% positive); (2) stain- ing with the antibody against the nuclear proliferation antigen Ki-67 (76% positive); (3) staining with the antibody against DNA-methyltransferase (86% positive).

494 Andreas Neubauer et a1

- PHA + PHA

Individual TRF Methyl TRF Methyl

A 6 7 8 6 81 B 6 3 56 54 C 5 8 66 55

200

Table 11. Correlation between the per cent transferrin-receptor positive (TRF) cells and cells staining with DNA-methyltransferase antibody with or without PHA-stimulation. Data are presented as % positive cells.

LL 0

PI w m s 3 z

0 1

. -

. .

FLUORESCENCE INTENSITY

Fig 1. DNA-methyltransferase as determined by flow cytometry in human mononuclear cells incubated without (dotted line) or with (solid line) PHA after 72 h of culture. The fluorescence intensity is displayed in arbitrary units. In each sample 5000 cells were measured. Analysis gate was drawn around lymphocytes and blast cells. In the sample without PHA. 7% were positive compared to 8 1% in the sample incubated with PHA.

Immunocytochemistry of peripheral blood mononuclear cells No positive staining was observed in either lymphocytes or monocytes, whereas nearly 90% of MOLT4 cells showed dark cytoplasmic and nuclear staining using the anti-DNA- methyltransferase monoclonal antibody. The staining for transferrin-receptor was negative for lymphocytes and weakly positive for monocytes.

PHA stimulation The results of PHA-stimulation using peripheral blood from three different donors are displayed in Table 11. Fig 1 shows flow cytometric measurements of lymphocytes incubated with and without PHA after 72 h of culture.

DijJerentiation of H L 6 0 cells with DMSO The cells incubated with and without DMSO were harvested at day 5 of the culture. The cells harvested were stained with

Table 111. Induction of cellular differentiation of HL60 cells by 1% DMSO. Cells were incu- bated without or with DMSO for 5 d and stained with antibodies directed against trans- ferrin-receptor (TRF) and DNA-methyltrans- ferase. Furthermore the DNA cell cycle phase was evaluated from the list-mode data. The data are shown as a range of two experiments in per cent positive cells (TRF. methylase) and percentage of cells in S- or GZM-phase, res- pectively.

- DMSO + DMSO

TRF 81-89 11-15 Methyl 79-82 7-10 S 48-51 1 1 G2M 3-10 2-5

transferrin receptor antibody and DNA-methyltransferase antibody. Results of two experiments are listed in Table 111. Good correlation between transferrin-receptor expression and binding of the antibody against DNA-methyltransferase was obtained. A double staining of cells incubated with and without DMSO using DNA-methyltransferase antibody and propidium iodide is shown in Fig 2 . It is obvious that not all S-phase cells stain positive with anti-methyltransferase anti- body. This heterogeneity of s-phase cells was corroborated in a further set of experiments comparing Ki-67 staining and DNA-methyltransferase-staining. The heterogeneity was found for both Ki-67 and DNA-methyltransferase. Fig 3A shows two parameter staining for DNA-methyltransferase antibody vs DNA-content, Fig 3B the respective plot for Ki67 versus DNA-content (DMSO induced cells, day 5 ) .

DISCUSSION

A monoclonal antibody directed against the enzyme DNA- methyltransferase. an important enzyme in the regulation of cellular proliferation, was used as a flow cytometric marker to determine cellular proliferation. The assay correlated well with standard staining methods using the nuclear prolifer- ation marker Ki-67 (Gerdes et al, 1983: Schwarting et al. 1986: Palutke et al, 1987) and the transferrin-receptor expression, which is expressed mainly on proliferating cells (Cotner et al, 1983; Serke, 1988) (Table 11).

Compared to bromodeoxyuridine staining and flow cyto- metric staining with the Ki-67 antibody, the assay described here is simple and fast. Only one fixation process is needed (70% ethanol for 10 min) for single and double staining with the DNA-dye propidium iodide. When cells have to be processed in plastic titre plates, acetone fixation-used by Schwarting for the determination of growth fraction with Ki- 67 (Schwarting et d. 1986)-will dissolve in plastic. For this reason, Palutke isolated cell nuclei and fixed them for double staining with propidium iodide (Palutke et al, 1987). How- ever, this procedure is time-consuming compared to the simple fixation with 70% ethanol for 10 min used in the present study.

Cell Proliferation and DNA-Methyl Transferase 49 5

Fig 2. Induction of granulocytic differentiation of human cell line HL60. Cells were incubated without (A) or with (B) 1% DMSO for 5 d and stained with DNA-methyltransferase antibody. In (A) 82% were positively staining compared to 7% in (B). Shown is a two-parameter contour plot (FITC-fluorescence of DNA-methyltransferase-antibody versus DNA-content as determined with propidium-iodide). 5000 cells were measured in each sample. The nine contour lines are plotted so that the first contour line represents a single cell; the next contour line upwards represents two cells, the next upwards four cells, etc. (256 as the maximum channel content). The plots are divided into four quadrants. In (A) 43% of the cells were in quadrant 1 .53% in quadrant 2,2%in quadrant 3 and 0.5% in quadrant 4 compared to 5% 5%, 81% and 6% respectively in (B). Note that in (A) 51% of the cells were in cell-cycle phase S compared to 11% in (B) as determined by the DNA-software program provided by the manufacturer.

w m 4 J > I !- W II

I d z n

(A) (B) Fig 3. HL60 cells induced with DMSO, stained for: (A) anti-DNA-methyltransferase antibody versus DNA content, and (B) nuclear proliferation associated Ki-67 antibody versus DNA-content. The contour lines were drawn according to Fig 2. In (A) 5000 cells were measured compared to 3033 in (B). In (A) Z%ofthe cells were in quadrant 1.8% in quadrant 2.80% in quadrant 3, and 8% in quadrant 4 compared to 3% in quadrant 1 , 7% in quadrant 2, 81% in quadrant 3. and 8% in quadrant 4 in Fig 3B.

496 Andreas Neubauer et a1 A very sensitive fluorometric assay for the determination of

cellular proliferation has recently been described (Dotsika & Sanderson. 1987). This assay is based on the hydrolysis of fluorogenic substrates by cell esterases, and involves measur- ing the cellular population as a whole. The assay described here, however. is based on single cell measurement, making it possible to determine cellular heterogeneity of proliferation rates, which is of importance in tumour biology. As can be seen from Fig ZB. DMSO induced HL60 S-phase cells are heterogeneous with regard to staining for DNA-methyltrans- ferase. Within the S-phase population, about 50% of the cells stained negative with the antibody against DNA-methyl- transferase. This can be due to ( i ) failure of this antibody to detect all proliferating cells or (ii) due to the well-documented finding that, after DMSO induction (day 5). about 20% ofcells are in S-phase according to the DNA-content, whereas only about 7% show uptake for 3H-thyrnidine (Ross. 1985). Since we have shown that this heterogeneity can also be detected using the antibody Ki-67 (Fig 3 ) . we think that it is much more likely that the S-phase cells staining negative with DNA-methyltransferase antibody are arrested in the S-phase.

About 7% of the cells in Fig 28 stained positive with the DNA-methyltransferase antibody. This correlated well with the morphology of the cells. which revealed a blast cell content of about 5-10% (not shown). Summarizing, the method described here is a fast and reliable way to determine cellular proliferation.

ACKNOWLEDGMENTS

The excellent technical assistance of Jutta Laser. Antje van Lessen and Karin Langmach is gratefully acknowledged.

This study was supported by grant M40/85/Hul from the Deutsche Krebshilfe and from the Deutsche Krebsgesellschaft Berlin.

KEFEKENCES Cordell. J.L., Falini. B.. Erber. W.N.. Ghosh. A.K.. Abdulazix. Z.,

Macdonald. S.T.. Pulford. K.A.F.. Stein. H. & A4ason. D.Y. (1984) Immunoenzymatic labelling of monoclonal antibodies using immune complexes of alkaline phosphatase and monoclonal anti- alkaline phosphatase (APAAP complexes). Journal of Histo- chemistry and Cytochemistry. 32. 219-229.

Cotner. T.. Williams, J.M., Christenson. L., Shapiro, H., Strom, T.B. & Strominger, J. (1983) Simultaneous flow cytometric analysis of human T-cell activation antigen expression and DNA content. Journal of Experimental Medicine, 157, 4 6 1-472.

Doerfler. W. (1983) DNA methylation and gene activity. Annual Review of Biochemistrg. 52, 93-1 24.

Dotsika. E.N. & Sanderson, C.1. (1987) A fluorornetric assay for determining cell growth in lymphocyte proliferation and lympho- kine assays. Journal of Immunological Methods, 105, 55-62.

Drahovsky. G., Pfeifer, G.P., Grunwald. S.. Hith, H.P., Vogel, M., Brzsoka. M.J. & Palitti. S. (1985) Characterization of DNA- methylating enzyme by use of monoclonal antibodies. In: Progress in Clinical and Biochemical Research (ed. by G. N. Cantony and A. Razin). Vol. 198 (Biochemistry and Biology of DNA Methyla- tion). Alan R. Liss. New York.

Gerdes. J., Schwab. U., Lemke, H. & Stein. H. (1983) Production of a mouse monoclonal antibody reactive with a human nuclear antigen associated with cell proliferation. International Journal of Cancer. 31, 13-20.

Gratzner. H.G. (1982) Monoclonal antibody to i-bromo- and 5- iodode-oxyuridine: a new reagent for detection of DNA replication. Science. 218. 474-475.

Lewin. B. ( 1 98 7) Genes III. John Wiley and Sons, New York. Neubauer. A. & Huhn. D. (1987) Monoclonal antibody Ki67 as a

proliferation marker. (Letter). British Journal of Haematology, 67, 495 .

Palutke. M.. KuKuruga, D. & Tabaczka. P. (1987) A flow cytometric method for measuring lymphocyte proliferation directly from tissue culture plates using Ki-67 and propidium iodide. Journal of Immunological Methods. 105, 97-105.

Ross. D.W. (1985) Differences in cell cycle kinetics during induced granulocytic versus monocytic maturation of H L 6 0 leukemia cells. Cancer Research, 45, 1308-1 313.

Schwarting. R.. Gerdes. J.. Niehus. J., Jaeschke. L. &Stein, H. (1986) Determination of the growth fraction in cell suspensions by flow cytometry using the monoclonal antibody Ki-67. Journal of Immunological Methods, 90, 65-70.

Serke. S.. Serke. M. & Brudler. 0. (1987) Lymphocyte activation by phytohemagglutinin and pokeweed mitogen: identification of proliferating cells by monoclonal antibodies. Journal of Immuno- logical Methods. 99, 167-1 72.

Serke. S. (1988) In situ immunocytochemical staining method for haematopoietic cells grown in vitro. Journal of Immunological Methods. 1 1 5 , 187-193.

Torsten. U. ( 1 986) DNS Methylierung: Signal fur die Regulation von Genen. Deutsche Medizinische Wochenschrijt. 111, 1495-1496.