© 2000 Wiley-Liss, Inc.
`
`Cytometry 40:151–160 (2000)
`
`Flow Cytometry Detection of Caspase 3 Activation in
`Preapoptotic Leukemic Cells
`F. Belloc,1,2* M.A. Belaud-Rotureau,1,2 V. Lavignolle,1 E. Bascans,1 E. Braz-Pereira,1 F. Durrieu,1,2
`and F. Lacombe1,3
`1Laboratoire d’He´matologie, Hoˆpital Haut Le´veˆque, Pessac, France
`2Laboratoire Universitaire d’He´matologie, Universite´ Victor Segalen, Bordeaux, France
`3CNRS-UMR 5540, Universite´ Victor Segalen, Bordeaux, France
`
`Received 9 November 1999; Revision Received 10 February 2000; Accepted 20 February 2000
`
`Background: Procaspase 3 is a constitutive proenzyme
`that is activated by cleavage during apoptosis. The result-
`ing enzyme is able to cleave several target proteins after
`the second aspartate of a DEVD sequence common to all
`the substrates of caspases 3 and 7 (DEVDase). Because
`active caspase 3 is a common effector in several apoptotic
`pathways, it may be a good marker to detect (pre-)apop-
`totic cells by flow cytometry (FCM).
`Materials and Methods: Apoptosis was induced in U937
`or bone marrow mononuclear cells by daunorubicin
`(DNR), idarubicin (IDA), or camptothecin (CAM). Viable
`and apoptotic cells were sorted by FCM on the basis of
`either fluorescein isothiocyante (FITC)–annexin V binding
`or DiOC6(3) accumulation. DEVDase activity was mea-
`sured in sorted populations by spectrofluorometry.
`Cleaved caspase 3 was labeled in situ with phycoerythrin
`(PE)– conjugated anti–activated caspase 3 antibodies and
`analyzed by FCM.
`Results: DEVDase activity was detected in sorted viable
`CAM- and DNR-treated U937 cells, demonstrating that a
`partial caspase activation occurred earlier than phosphati-
`
`dyl-serine exposure and mitochondrial membrane poten-
`tial dissipation. The presence of a low amount of active
`caspase 3 in the treated viable cells was confirmed in situ
`with PE-conjugated anti–active caspase 3 antibodies. The
`same antibody was used in combination with FITC-an-
`nexin V and CD45–PC5 to study caspase 3 activation in
`acute leukemia blast cells after in vitro DNR and IDA
`treatment. Both anthracyclines induced a caspase 3– de-
`pendent apoptosis that was more efficient in blast cells
`than in contaminating lymphocytes.
`Conclusions: These results demonstrate that anti–active
`caspase 3 labeling can be an alternative to fluorogenic
`substrates to efficiently detect early apoptosis by FCM in
`heterogeneous samples. They also confirm that anthracy-
`clines induce blast cell apoptosis by a caspase 3– depen-
`dent pathway. Cytometry 40:151–160, 2000.
`© 2000 Wiley-Liss, Inc.
`
`Key terms: caspase; apoptosis; leukemic cells; flow cy-
`tometry; anthracyclines; camptothecin
`
`The efficiency of the cytotoxic agents used in the treat-
`ment of leukemia is attributed in part to their ability to
`induce apoptotosis (1– 4). Thus, it would be useful to
`quantify in vivo apoptosis for evaluation of the efficiency
`of a treatment (5). Because the tissues concerned in leu-
`kemia (blood and bone marrow) are composed of cell
`suspensions, it would be interesting to adapt this detec-
`tion to flow cytometric (FCM) analysis. Moreover, because
`they are heterogeneous tissues, a multivariate analysis
`must be used to focus apoptosis analysis on leukemic blast
`cells independently of contaminating normal cells. How-
`ever, most of the methods currently used to label apopto-
`tic cells fail to detect circulating apoptotic leukemic cells
`in patients under treatment. Some reports relating positive
`findings on this topic have used a very sensitive labeling of
`DNA breaks occurring during apoptosis (6 – 8), but this
`method is not compatible with immunotyping of blast
`cells and can generate false-positive results on FCM anal-
`
`ysis, which can only be recognized by laser scan cytom-
`etry (9). Two other methods have been described to
`recognize early apoptotic cells using FCM: alterations in
`mitochondrial membrane modify the uptake of the li-
`pophilic fluorochrome DiOC6(3) (10,11), and modifica-
`tions in cell membrane structure are detected with the
`binding of fluorescein isothiocyante (FITC)– conjugated
`annexin V to phosphatidylserines (12,13). In our hands,
`both methods failed to detect circulating apoptotic cells.
`Studies on neutrophil polymorphonuclear cells offer a
`possible explanation for this failure. Indeed, apoptosis is
`
`Grant sponsor: Conseil Regional d’Aquitaine; Grant sponsor: Ligue
`Nationale Franc¸aise Contre le Cancer; Grant sponsor: Association pour la
`Recherche sur le Cancer.
`*Correspondence to: Dr. Francis Belloc, Laboratoire d’He´matologie,
`Hoˆpital Haut Le´veˆque, 33604 Pessac, France.
`E-mail: francis.belloc@wanadoo.fr
`
`Page 1 of 10
`
`Fluidigm
`Exhibit 1007
`
`

`
`152
`
`BELLOC ET AL.
`
`responsible for the elimination of aging neutrophils and
`leads to recognition and uptake of intact senescent cells
`by macrophages (14 –16). Similarly, when leukemic cells
`induced to enter apoptosis were in contact with a phago-
`cyte model, none of the FCM methods was able to detect
`apoptotic cells before they were ingested by nonprofes-
`sional phagocytes (17).
`Caspase activation is now well known as a key event in
`the apoptotic process. In particular, caspase 3 activity is
`the common effector of most of the apoptotic pathways.
`This enzyme is constitutively expressed by most of the
`leukemic cells (18) as a proenzyme of 32 kDa. In response
`to several stimuli, caspase 3 is cleaved and generates two
`subunits of 17 and 12 kDa, which fit together in a double
`heterodimer exhibiting the activity (19). The active en-
`zyme is able to cleave several target proteins before the
`aspartate of a DEVD sequence common to all the sub-
`strates of caspases 3 and 7. We previously showed how
`DEVDase activity is activated in apoptotic cells before
`they are recognized by phagocytes (17). Attempts to de-
`velop cell-permeant substrates for caspase that would al-
`low FCM analysis of caspase 3 activity have not been
`successful, but polyclonal antibodies preferentially recog-
`nizing cleaved (activated) caspase 3 have been recently
`developed and are now commercially available.
`In the present work, we confirm that a DEVDase activ-
`ity attributable to a partial caspase 3 activation precedes
`the alterations in mitochondrial membrane demonstrated
`by DiOC6(3) incorporation and the loss of external mem-
`brane asymmetry revealed by annexin V binding during
`apoptosis induced by chemotoxic molecules. Further-
`more, we found that this caspase 3 activation can be
`detected by the anti– cleaved caspase 3 antibodies for the
`purpose of FCM analysis, thus providing an early marker
`of apoptosis. This labeling can be used in multivariate
`analysis of bone marrow samples to reveal in vitro apo-
`ptosis of leukemic blast cells. Moreover, this is the first
`report showing that anthracyclines efficiently and specif-
`ically induce the activation of the caspase pathway in
`these cells.
`
`MATERIALS AND METHODS
`Cells and Culture
`U937 cells (a human leukemic cell line from a histio-
`cytic lymphoma) or bone marrow mononuclear cells were
`cultured in RPMI 1640 medium (Gibco-BRL, Eragny,
`France) supplemented with 10% fetal calf serum (FCS) and
`1 mM L-glutamine, 10 mM Hepes (Gibco-BRL), 100 U/ml
`penicillin, and 50 mg/ml streptomycin in humidified 95%
`O2 and 5% CO2 atmosphere at 37°C. Exponentially grow-
`ing cells were used in all experiments concerning cell
`lines.
`Mononuclear cells were obtained from bone marrow
`aspirates that were taken from adult leukemic patients for
`routine immunophenotyping. They were prepared the
`same day from heparin or ethylene-diaminetetraacetic
`acid (EDTA) anticoagulated bone marrow samples by Fi-
`coll centrifugation (Lymphoprep, Nycomed Pharma, Oslo,
`
`Norway). The mononuclear cells were then resuspended
`at 106/ml in culture medium. The samples from 8 adult
`patients with either acute myeloid leukemia (6) or acute
`lymphoid leukemia (2) were processed. All preparation
`steps were performed in less than 3 h after withdrawal.
`
`Induction of Apoptosis
`Camptothecin (CAM; Sigma, St. Quentin, Fallavier
`France) was stored at 220°C as a 3 mMsolution in di-
`methylsulfoxide. Daunorubicin (DNR; RPR-Bellon, Neuilly,
`France) and idarubicin (IDA; Pharmacia, Saint Quentin en
`Yvelines, France) were aliquoted and stored at 220°C as
`1 mg/ml solutions in distilled water. Apoptosis was in-
`duced in U937 cells (3 3 105/ml) with 1 mM CAM for 3 h.
`In other experiments, U937 cells were treated with either
`1 mM DNR or 0.2 mM IDA for 1 h, washed, resuspended in
`drug-free medium, and further cultured for time course
`analysis. Mononuclear bone marrow cells at 106/ml were
`treated with either 1 mM DNR or 0.2 mM IDA for 1 h,
`washed, and further cultured for 18 h.
`
`Analysis and Sorting of Apoptotic Cells by FCM
`Variation in mitochondrial potential (DC
`m) was demon-
`strated with 100 nM DiOC6(3) (3,39 dihexyloxacarbocya-
`nine iodide) for 30 min at 37°C in culture medium, as
`previously described (9,10). Analysis was then performed
`at 525 nm with an ELITE cell sorter (Coulter-Immunotech,
`Margency, France). The highly fluorescent population was
`considered viable cells, and the dull population repre-
`sented apoptotic cells (Figs. 1B, 2A).
`FITC–annexin V (Coulter-Immunotech) was used as
`specified by the manufacturer to measure the exposure of
`the phosphatidylserines on the cytoplasmic membrane.
`Cells (2 3 105) were resuspended in the 13 binding
`buffer. FITC–annexin V solution (5 ml) was added to 500
`ml of the cell suspension. The samples were incubated on
`ice for 10 min before analysis at 525 nm with an ELITE cell
`sorter (Coulter-Immunotech). The FITC–annexin V bind-
`ing cells were considered as apoptotic and the negative
`cells as viable (Fig. 2B).
`In the cell sorting experiments, the cell flow was set to
`analyze 2,000 events/s, and cells in the windows of inter-
`est were sorted in sterile tubes containing 0.5 ml FCS.
`Cells (5 3 105) were sorted, immediately centrifuged, and
`processed for caspase activity or cleaved caspase 3 label-
`ing, as described below. Because the percentage of apo-
`ptotic cells was approximately 50% in all samples, the
`sorting step was performed in less than 15 min on ice. In
`all sorting experiments, the viable population of untreated
`cells was also sorted and used as a reference population.
`When DiOC6(3) was used as a probe, the whole popula-
`tion was sorted and reanalyzed to verify that no modifica-
`tion had occurred in the apoptotic frequency during sort-
`ing. This step was not feasible with annexin V because the
`lack of Ca21 ions in the sheath fluid induced the elution of
`the annexin V from the cell surface.
`
`Page 2 of 10
`
`Fluidigm
`Exhibit 1007
`
`

`
`CASPASE 3 ACTIVATION IN LEUKEMIC CELLS
`
`153
`
`FIG. 1. DEVDase activation precedes alterations of mitochondrial membrane potential during camptothecin (CAM)–induced apoptosis. U937 cells were
`incubated in the absence (A) or in the presence (B) of 1mM CAM for 3 h. The cells were then stained with 0.1 mg/ml of DiOC6(3) and analyzed for
`mitochondrial potential by flow cytometry (A,B). Cells with high DiOC6(3) content were viable, whereas cells with low DiOC6(3) content were apoptotic.
`C: Kinetic chart of mean DEVDase activity/105 cells of U937 cells incubated in the absence (black spots) or in the presence (white triangles) of 1 mM CAM
`for 3 h.D: Kinetic chart of DEVDase measured on sorted populations with low (white triangles) or high (white squares) DiOC6(3) content from
`CAM-treated samples (as shown in B). Untreated sorted U937 cells were used as a reference (black spots). Mean 6 S.D. of three independent culture and
`sorting experiments.
`
`Cleaved Caspase 3 Labeling
`Cells (5 3 105) were pelleted, resuspended in 1 ml
`Permeafix (Ortho Diagnostic Systems, Roissy, France) and
`incubated for 40 min at 20°C. The suspension was then
`centrifuged, and the pellet was washed twice with wash-
`ing buffer (0.2 mM EDTA, 5% FCS in phosphate buffered
`saline [PBS]). Labeling was performed by adding to the
`cells 100 ml of washing buffer containing 5 ml of poly-
`clonal phycoerythrin (PE)– conjugated anti–active caspase
`3 antibodies (PharMingen–Becton Dickinson, Le Pont de
`Claix, France). The samples were gently stirred for 1 h,
`washed in washing buffer, and analyzed by FCM at 575 nm
`with an XL cytometer (Coulter-Immunotech). A negative
`
`control sample incubated with PE-conjugated immuno-
`globulin G was run in parallel.
`For applications on bone marrow mononuclear cells,
`the cells were pelleted, incubated in 100 ml of culture
`medium containing 5 ml of PC5-conjugated anti-CD45 an-
`tibodies (Coulter-Immunotech) for 20 min. Binding buffer
`(500 ml) containing FITC–annexin V (Coulter-Immuno-
`tech) was added and further incubated for 20 min on ice.
`The cells were then fixed and labeled with anti– cleaved
`caspase 3 antibodies, as above, and analyzed using an XL
`cytometer at 525 nm for annexin V binding, at 575 nm for
`cleaved caspase 3 labeling, and at 675 nm for CD45 ex-
`pression.
`
`Page 3 of 10
`
`Fluidigm
`Exhibit 1007
`
`

`
`154
`
`BELLOC ET AL.
`
`FIG. 2. DEVDase activation precedes mitochondrial and membrane alterations during daunorubicin (DNR)–induced apoptosis. U937 cells were treated
`for 1 h with 1 mM DNR, washed, and incubated for 18 h at 37°C in drug-free medium. The cells were stained for apoptosis with either DiOC6(3) (A,C)
`or fluorescein-labeled annexin V (B,D), analyzed by flow cytometry (A,B) and sorted in intact and apoptotic (apop) populations. The DEVDase activity/105
`cells/20 min was measured in each sorted population (C,D) and divided by the activity of untreated intact sorted cells as a reference (Ref) to calculate the
`DEVDase activation ratio. Mean 6 S.D. of four independent culture and sorting experiments.
`
`The results shown in Figures 4 and 7 show the percent-
`age of positive (cleaved caspase 3 exhibiting or annexin V
`binding) cells. Results in Figure 5 are expressed as the
`specific labeling (the mean fluorescence of labeled sample
`minus the mean fluorescence of negative control) of
`treated sorted populations divided by the specific labeling
`of the reference sample constituted by sorted untreated
`viable cells. This ratio was called caspase 3 activation.
`
`Determination of DEVDase Activity
`
`DEVDase activity was evaluated as previously described
`with some modifications (17). Briefly, 3 3 105 cells were
`suspended in 50 ml of a permeabilizing buffer, pH 7.4,
`containing 10 mM HEPES, 5 mM dithiothrietol, 0.02%
`saponine, 1 mM phenyl methyl sulfonyl fluoride (PMSF),
`10 mg/ml pepstatin A, 10 mg/ml leupeptin, and 14 mM
`
`Page 4 of 10
`
`Fluidigm
`Exhibit 1007
`
`

`
`CASPASE 3 ACTIVATION IN LEUKEMIC CELLS
`
`155
`
`fluorogenic substrate Ac-DEVD aminomethylcoumarin
`(UBI, Euromedex, Souffelweyersheim, France). The cells
`were seeded in 96-well microplates, and fluorescence
`kinetics were recorded every 5 min at 37°C in a Wallac
`1420 Victor2 plate fluorometer (EG&G Instruments, Evry,
`France) for 30 min. The lamp power was set at 1,000,
`excitation filter at 355 nm, and emission filter at 460 nm.
`A blank was performed without the addition of cells in the
`substrate buffer, and its value was subtracted from all the
`measurements. The results were expressed as fluores-
`cence units generated by 105 cells. Figures 2C and 2D
`show the DEVDase activation calculated by the ratio of the
`specific fluorescence generated by the sorted populations
`over the specific fluorescence generated by the reference
`population in 20 min.
`
`Western Blot Analysis
`One million cells were washed in 0.9% NaCl and resus-
`pended in 100 ml of 0.9% NaCl containing the following
`protease inhibitors: 35 mg/ml PMSF, 0.3 mg/ml EDTA, 0.5
`mg/ml leupeptin, and 0.7 mg/ml pepstatin A. Cells were
`then lysed by the addition of 25 ml 53 sample buffer (60
`mM Tris-HCl, pH 6.8, 25% glycerol, 2% sodium dodecyl
`sulfate [SDS], 1/20 v/v 2-mercaptoethanol, 0.1% bromo-
`phenol blue), boiled for 2 min, and cleared by centrifuga-
`tion for 5 min at 14,000 rpm. Total protein content was
`quantified in the sample before SDS–polyacrylamide gel
`electrophoresis analysis of 40 mg using 12% acrylamide
`gel. After electrotransfer to nitrocellulose filters, blots
`were incubated with preblocking solution (0.5% low-fat
`dry milk in PBS) for 1 h at room temperature followed by
`a 1-h incubation with primary antibodies diluted to 1/500
`in PBS Tween (0.5% Tween 20). After extensive washing,
`the blots were incubated for 1 h with 1/10,000 diluted
`peroxidase-conjugated second antibody (Sigma) and then
`washed as described above. The antibodies were detected
`using a chemoluminescence kit (ECLt, Amersham, Little
`Chalfont, UK). Chemoluminescence was detected and
`quantified using a Kodak Image Station 440CF (NEN Life
`Science Products, Paris, France). The antibodies used
`were polyclonal rabbit anti– caspase 3, purified mouse
`anti– human caspase 7 monoclonal, and monoclonal
`mouse anti-PARP, all from PharMingen-Becton Dickinson.
`
`RESULTS
`DEVDase Activation Occurs in Preapoptotic
`Leukemic Cells
`CAM is a well-known inducer of apoptosis in leukemic
`cells. When U937 cells were treated with 1 mM CAM for
`3 h, approximately 50% of the cells exhibited a decreased
`mitochondrial membrane potential
`as detected by
`DiOC6(3) incorporation and FCM analysis (compare Fig.
`1B with 1A). These cells have previously been described
`as apoptotic cells (10,11). At the same time, the apoptosis-
`specific DEVDase activity was increased in the treated
`versus the untreated cells (Fig. 1C). To verify whether
`DEVDase activation and mitochondrial apoptosis were
`synchronous, cell sorting experiments were performed
`
`and DEVDase activity was measured in sorted high and
`low DiOC6(3) cell populations and compared with the
`activity in an untreated sorted high DiOC6(3) cell popu-
`lation (Fig. 1D). The bulk of the DEVDase activity was
`found in low DiOC6(3) apoptotic cells (55 6 17–fold the
`reference value in 20 min), but a slight significant DEV-
`Dase activation was also detected in sorted high DiOC6(3)
`viable cells (11 6 6 –fold), indicating that a partial activa-
`tion of DEVDase could precede the alterations of the
`mitochondrial membrane during CAM-induced apoptosis.
`Whether such an early activation was specific to the
`action of CAM or could be found with other chemo-
`inducers remained questionable. Thus, a similar experi-
`ment was performed using DNR to induce apoptosis.
`U937 cells were treated for 1 h with 1 mM DNR to mimic
`in vivo DNR administration to leukemic patients (4). The
`cells were then incubated for 18 h in drug-free medium
`and analyzed for variation in mitochondrial transmem-
`brane potential (DC
`m; Fig. 2A). Both apoptotic and viable
`populations were sorted and analyzed for their DEVDase
`activity. The results shown in Figure 2C confirm that DNR
`also induced DEVDase activation (5.3 6 3.6%–fold) in
`cells identified as viable on the basis of their DiOC6(3)
`content. The results appeared heterogeneous in view of
`the standard deviation, but an increase in DEVDase activ-
`ity of treated viable cells was found in all the four exper-
`iments performed.
`Phosphatidylserines exposure has also been described
`as an early event of apoptosis (12). Thus, it was interesting
`to check whether DEVDase activation preceded the loss
`of membrane asymmetry. Phosphatidylserines were
`probed with fluorescent annexin V on DNR-treated U937
`cells. After FCM analysis, annexin V–positive (apoptotic)
`and –negative (viable) populations were sorted and their
`DEVDase activation was measured in comparison with
`untreated annexin V–negative sorted cells (control). DEV-
`Dase was also found to be activated in DNR-treated viable
`cells when they were sorted on the basis of their annexin
`V binding, confirming that DEVDase activation is a very
`early event during chemo-induced apoptosis.
`
`DEVDase Results From Caspase 3 and 7 Activation
`in Chemo-Induced Apoptotic Leukemic Cells
`DEVDase activity is shared by caspases 3 and 7. Which
`of these caspases was responsible for the activity detected
`in viable DNR-treated U937 cells remained to be eluci-
`dated. Therefore, Western blot analysis of the cleavage of
`these caspases was performed early during DNR-induced
`apoptosis. The results shown in Figure 3A,B show that no
`cleavage product of caspase 3 or 7 was detected after 3 h
`of treatment, whereas cleavage of both caspases 3 and 7
`was clearly evident at 4 h. Quantification of the cleaved
`bands (performed by image analysis) indicated that 5% of
`caspase 3 was cleaved at 4 h and 11% at 5 h. For caspase
`7, these percentages were 9% at 4 h and 18% at 5 h. These
`results strongly suggest that activation of both caspases
`exhibiting DEVDase activity occurs between 3 and 4 h
`after the beginning of DNR treatment. This was further
`confirmed by the analysis of the endogenous DEVDase
`
`Page 5 of 10
`
`Fluidigm
`Exhibit 1007
`
`

`
`156
`
`BELLOC ET AL.
`
`FIG. 3. Caspases with DEVDase activity are activated early during daunorubicin (DNR)–induced apoptosis. U937 cells were incubated with 1 mM DNR
`for 1 h, washed, and further cultured in drug-free medium. A–C: Extracts were obtained at 1, 3, 4, and 5 h after the beginning of the experiment for Western
`blot analysis. After electrophoresis and transfer, the blots were labeled with anti– caspase 3 (A), anti– caspase 7 (B), or anti-PARP (C) antibodies. D: The
`DEVDase activity/105 cells/20 min was measured after 0, 1, 3, 4, and 5 h ofculturing. Mean 6 S.D. of three separate experiments.
`
`substrate PARP, whose cleavage followed cleavage of
`caspases 3 and 7 after 5 h oftreatment (Fig. 3C). In vitro
`measurement of DEVDase activity confirmed that signif-
`icant enzyme activation occurred between 3 and 4 h
`after the beginning of the experiment (Fig. 3D). The
`activities at 1 and 3 h did not differ from those of
`untreated samples.
`
`Cleaved Caspase 3 Is a Good Marker for
`Preapoptotic Cells
`These results were in agreement with the hypothesis
`that detection of cleaved caspase 3 could be a good
`marker for early chemo-induced apoptosis. This was con-
`firmed by the experiment shown in Figure 4, where the
`same samples as shown in Figure 3 were labeled with the
`anti– cleaved caspase 3 antibodies and analyzed by FCM.
`One and three hours after the beginning of DNR treat-
`ment, a low amount (about 8%) of active caspase 3–pos-
`itive cells was detected, which did not differ from the
`percentage in untreated U937 cells. At the fourth hour,
`this percentage of positive cells was increased twofold
`
`(14%) and this increase continued until the fifth hour
`(20%). These FCM results are in concordance with the
`batch analysis shown in Figure 3.
`The ability of anti– cleaved caspase 3 antibody to label
`preapoptotic cells was assayed after sorting viable and
`apoptotic cells on the basis of their DiOC6(3) incorpora-
`tion. The specific labeling of either CAM-treated (Fig. 5A)
`or DNR-treated (Fig. 5B) viable (preapoptotic) cells was
`found to be increased when compared with untreated
`viable cells. This increase was homogeneous and con-
`cerned all high DiOC6(3)-containing cells (Fig. 5C). This
`increase in preapoptotic cell labeling represented only a
`part of the final labeling observed in the CAM-induced
`apoptotic cells (Fig. 5A) but was almost as high as the
`labeling of DNR-induced apoptotic cells (Fig. 5B). The
`results obtained with anti– cleaved caspase 3 antibody
`paralleled that obtained by measuring DEVDase activation,
`but with a lower sensitivity than by the enzymatic
`method.
`The usefulness of this anti– cleaved caspase 3 antibody
`to recognize preapoptotic cells was also assayed on bone
`
`Page 6 of 10
`
`Fluidigm
`Exhibit 1007
`
`

`
`CASPASE 3 ACTIVATION IN LEUKEMIC CELLS
`
`157
`
`FIG. 4. Activation of caspase 3 by
`daunorubicin can be followed by
`flow cytometry (FCM). U937 samples,
`treated the same as those shown in
`Figure 3, were fixed and stained with
`anti– cleaved caspase 3 antibodies at 1
`(B), 3 (C), 4 (D), and 5 (E) h after the
`beginning of the experiment and an-
`alyzed by FCM. A negative control
`performed with an irrelevant anti-
`body is shown A. The percentage of
`cells exhibiting active caspase 3 is
`indicated on each histogram as the
`mean 6 SD. of four experiments.
`
`marrow mononuclear cells from leukemic patients. Blast
`cells were discriminated from contaminating normal lym-
`phocytes on the basis of their CD45 expression by taking
`advantage of the multivariate analysis capacity of FCM
`(Fig. 6A). These cells had also been labeled for their
`phosphatidylserine exposure by FITC–annexin V before
`their fixation, permeabilization, and labeling with PE/anti–
`cleaved caspase 3 antibody. When apoptosis was induced
`with either DNR (Fig. 5C) or IDA (Fig. 5D), the amount of
`annexin V–positive blast cells was increased as opposed
`to the untreated sample (Fig. 6B). The whole annexin
`V–positive population of blast cells was also positive for
`caspase 3 activation. In some samples, a labeling of an-
`nexin V–negative cells by anti– cleaved caspase 3 antibody
`could be observed (i.e.,
`in Fig. 6C), suggesting that
`caspase 3 could also be activated in leukemic blast cells
`before the loss of membrane asymmetry. However, this
`was not constantly observed, and from the 24 samples
`obtained from the 8 patients we analyzed, a very good
`correlation (but not identity) between the percentages of
`annexin V–positive and cleaved caspase 3–positive cells
`was found on both blast cells (Fig. 7A; r 5 0.96) and
`normal lymphocytes (Fig. 7B; r 5 0.94). In fact, a signifi-
`cantly (P , 0.01) higher frequency of annexin V–positive
`than of active caspase 3–positive cells was observed, and
`this could be due to a better discrimination between
`positive and negative cells in the annexin V–labeled sam-
`ples (Fig. 6). Figure 7 also demonstrates that spontaneous
`apoptosis and DNR- or IDA-induced apoptosis occurred
`more efficiently in blast cells than in lymphocytes, what-
`ever the labeling method.
`
`DISCUSSION
`In some apoptotic pathways, the activation of pro-
`caspase 3 is favored by the release into the cytoplasm of
`two mitochondrial
`factors, cytochrome c and Apaf-1,
`(20,21). For this reason, the decrease in mitochondrial
`(DC
`membrane potential
`m) as shown by DiOC6(3)
`(10,11) incorporation was expected to precede the acti-
`vation of caspase 3. We previously showed that DEVDase
`activation occurs earlier than the ingestion of apoptotic
`leukemic cells by phagocytes. This was not observed
`when DiOC6(3) was used to label apoptotic cells (17). In
`the present study, we have confirmed by sorting experi-
`ments of apoptotic and viable cells from CAM- or DNR-
`treated U937 cells that induced cells with a normal DC
`m
`are already primed for DEVDase activity. This activation of
`DEVDase in preapoptotic cells was higher in DNR- than in
`CAM-treated viable cells (compare Fig. 1D with 2C). Noth-
`ing is known of the relation, if any, or the chronology
`between caspase activation and phosphatidyl-serine expo-
`sure, which occurs during apoptosis. This loss of mem-
`brane asymmetry is thought to represent the signaling for
`apoptotic cell
`recognition by macrophages because
`phagocytosis is inhibited in part by annexin V binding
`(15,17). However, in our hands, annexin V binding was
`not found to precede ingestion by phagocytes (17). In the
`present work, DEVDase activation preceded phosphati-
`dyl-serine exposure on sorted annexin V–negative DNR-
`treated cells (Fig. 2B). Both the cleavage of caspase 3 and
`the generation of DEVDase activity were found to occur
`between the third and fourth hours after DNR treatment
`in U937. This phenomenon preceded the cleavage of the
`endogenous PARP substrate for caspase 3. However, there
`
`Page 7 of 10
`
`Fluidigm
`Exhibit 1007
`
`

`
`158
`
`BELLOC ET AL.
`
`FIG. 5. Caspase 3 is partially cleaved in viable
`cells during camptothecin- (CAM) and daunoru-
`bicin- (DNR) induced apoptosis. U937 cells were
`treated either with CAM, as in Figure 1 (A), or
`DNR, as in Figure 2 (B), to induce apoptosis.
`Viable or apoptotic cells were sorted by flow
`cytometry (FCM) on the basis of their DiOC6(3)
`content, as in Figure 1. Viable untreated cells
`were also sorted as a reference (Ref). Sorted
`populations were labeled with anti– cleaved
`caspase 3 antibodies and analyzed by FCM as
`indicated in Materials and Methods. The specific
`mean fluorescence of sorted populations was
`divided by the specific mean fluorescence of
`reference cells (Ref) to calculate the caspase 3
`activation ratio. Mean 6 SD. of three indepen-
`dent culture and sorting experiments. C: FCM
`analysis histograms of reference (Ref) CAM-
`treated viable (shaded in gray) and apoptotic
`(Apop)
`sorted populations after anti–active
`caspase 3 labeling.
`
`is a parallel with previous results obtained with DiOC6(3)
`incorporation or annexin V binding (17), suggesting that
`the moment of appearance of all these three markers
`occurs in a period shorter than 1 h.
`Because we were interested in measuring caspase 3
`activation by FCM, a kinetic study of cleaved caspase 3
`expression during DNR induction of apoptosis was per-
`formed on U937 cells. The percentage of activated
`caspase-positive cells was found to produce results similar
`to those of the batch analysis by Western blot. When the
`labeling of cleaved caspase 3 followed by FCM analysis
`was performed on cells sorted on the basis of their DC
`m,
`a global increase in labeling was found in the so-called
`viable preapoptotic cells as compared with untreated
`cells. The results on a one-cell basis (Fig. 5) reflected the
`batch analysis obtained by measuring DEVDase activity
`(Figs. 1D, 2C) in terms of relative activation between
`untreated, preapoptotic, and apoptotic cells during both
`CAM and DNR induction of apoptosis. However, the acti-
`vation ratios were lower by FCM analysis than by measur-
`ing enzymatic activity. DNR led to a level of caspase 3
`activation similar to that of CAM in preapoptotic cells.
`This early low-level increase was followed by a further
`increase at completion of apoptosis in CAM-treated cells
`
`but not in DNR-treated cells, suggesting that these two
`drugs act by different pathways converging in the caspase
`3 pathway.
`Although an increase in the rate of caspase 3 cleavage
`was found between 4 and 5 h of DNRtreatment by both
`FCM (Fig. 4) and Western blot analysis (Fig. 3A), it was not
`accompanied by an increase in DEVDase activity (Fig. 3D).
`We believe that the caspase 3 detected by both Western
`blot and FCM analysis loses its enzymatic activity after a
`certain time. This belief suggests that cleaved caspase 3
`has a short-lived enzymatic activity in DNR-treated cells. In
`fact, a progressive loss of DEVDase activity was observed
`in apoptotic samples when conserved at room tempera-
`ture, whereas the percentage of activated caspase 3–pos-
`itive cells was not modified (data not shown).
`When anti– cleaved caspase 3 antibodies were used to
`investigate in vitro anthracycline-induced apoptosis in
`bone marrow from leukemic patients, they were found to
`be compatible with the identification of blast cells on the
`basis of their CD45 expression (22). It was possible with
`this method to compare the respective apoptotic behavior
`of leukemic blast cells and normal lymphocytes in culture
`with or without anthracycline treatment. Several conclu-
`sions may be made from these experiments. First, a good
`
`Page 8 of 10
`
`Fluidigm
`Exhibit 1007
`
`

`
`CASPASE 3 ACTIVATION IN LEUKEMIC CELLS
`
`159
`
`FIG. 6. Multivariate analysis of apoptosis and caspase 3
`activation on blast cells in bone marrow of leukemic pa-
`tients. Bone marrow mononuclear cells were treated in
`vitro with either daunorubicin (DNR; 1 mM) or idarubicin
`(IDA; 0.2 mM), as described in Materials and Methods. After
`18 h of culture, the cells were sequentially labeled with
`PC5–anti-CD45 antibody, fluorescein isothiocyanate–an-
`nexin V and phycoerythrin/anti– cleaved caspase 3 antibod-
`ies and analyzed by flow cytometry. A: Bivariate analysis of
`mononuclear cells from a leukemic bone marrow with
`CD45 labeling on the x axis and side scatter on the y axis.
`Blast cells (BC) can easily be discriminated from contami-
`nating normal lymphocytes (L). B–D show bivariate analysis
`of annexin V binding (x axis) versus active caspase 3 con-
`tent (y axis) gated on the blast cell population in untreated
`control (B), DNR-treated (C), or IDA-treated (D) samples.
`
`FIG. 7. Caspase activation is related to phosphatidylserine exposure in apoptotic leukemic cells and lymphocytes. Mononuclear cells from 6 acute
`myeloid (white symbols) and 2 acute lymphoid (black symbols) leukemia patients were treated with either daunorubicin (circles) or idarubicin (triangles)
`or untreated (squares). The samples were processed as described in Figure 6. Annexin V and active caspase 3–positive cells were counted by flow
`cytometry in each sample after gating acquisition on either leukemic blast cells (A) or normal lymphocytes (B) on the basis of their CD45 expression (Fig.
`6A). The percentage of active caspase 3–positive cells was plotted as a function of the percentage of annexin V–positive cells. The correlation coefficient
`(r) is indicated on each plot.
`
`corr