Leukemia (2001) 15, 1564–1571  2001 Nature Publishing Group All rights reserved 0887-6924/01 $15.00 www.nature.com/leu Phosphodiesterase type 4 inhibitor suppresses expression of anti-apoptotic members of the Bcl-2 family in B-CLL cells and induces caspase-dependent apoptosis B Siegmund1, J Welsch1, F Loher1, G Meinhardt2, B Emmerich2, S Endres1 and A Eigler1

1Division of Clinical Pharmacology and 2Division of Haematology and Oncology, Medizinische Klinik Innenstadt, Klinikum of the Ludwig- Maximilians-University Munich, Munich, Germany

B cell chronic lymphocytic leukemia (B-CLL) is an incurable synergizes with chlorambucil in vitro and a phase II clinical clonal disease which shows initial responsiveness to a number trial suggested that this synergism may be clinically rel- of chemotherapeutic drugs. However, in most patients the dis- 13,14 ease becomes resistant to treatment. Rolipram, a specific evant. However, therapy with broad-spectrum PDE inhibi- inhibitor of phosphodiesterase (PDE) type 4, the PDE predomi- tors such as theophylline or pentoxifylline in patients with B- nantly expressed in B-CLL cells, has been shown to induce CLL is limited due to side-effects partially mediated by their cAMP-dependent apoptosis in these cells. In the present study, activity as adenosine receptor antagonists.15 The expression of we demonstrate that the extent of rolipram-induced apoptosis specific phosphodiesterases in B-CLL cells favors a more tar- is similar to fludarabine-induced apoptosis in vitro. The combi- geted strategy by using specific PDE inhibitors. Kim and nation of rolipram and fludarabine results in an enhancement 16 in the number of apoptotic cells compared to apoptosis Lerner described recently that CLL cells contain transcripts induced by either agent alone. Second, rolipram suppresses for PDE 1B1, PDE 4A and PDE 4B. They further demonstrated the expression of anti-apoptotic members of the Bcl-2 family the cAMP-dependent induction of apoptosis by the PDE 4 and induces the pro-apoptotic protein Bax, thereby shifting the inhibitor rolipram in B-CLL cells.16 New PDE 4 inhibitors have balance between pro- and anti-apoptotic members of the Bcl- a favorable ratio of desired action and side-effects, and are 2 family towards a pro-apoptotic direction. Finally rolipram- under clinical investigation in phase III trials for chronic induced apoptosis is caspase-dependent. PDE 4 inhibitors are currently under investigation for chronic obstructive pulmon- obstructive pulmonary disease and asthma, with encouraging 17 ary disease and asthma in phase III clinical trials showing preliminary results. promising results with tolerable side-effects. In conclusion, by To evaluate the function of PDE in the pathogenesis of B- inducing apoptosis, by enhancing apoptosis induced by fluda- CLL and the therapeutic potency of PDE 4 inhibitors such as rabine, by suppressing Bcl-2, Bcl-X and by inducing Bax rolipram further in vitro investigations are needed. A clinically expression, PDE 4 inhibitors may add a new therapeutic option relevant question is, whether rolipram synergizes with cur- for patients with B-CLL. Leukemia (2001) 15, 1564–1571. Keywords: rolipram; Bcl-X; Bax; apoptosis; B-CLL; PDE 4 rently used chemotherapy in induction of apoptosis. Combi- nation with PDE 4 inhibitors could allow a dose reduction of toxic cytostatic drugs. The resistance to chemotherapy in B- CLL can partly be mediated by the high expression of anti- Introduction apoptotic proteins such as Bcl-2 in B-CLL. If rolipram shifts the balance of pro- and anti-apoptotic members of the Bcl-2 B cell chronic lymphocytic leukemia (B-CLL) is the most com- family towards apoptosis cytotoxic agents could become mon adult leukemia in the western hemisphere and accounts more effective. for 25% of all leukemias. The indolent natural course of both In the present study we investigated the apoptosis-inducing early stage and smoldering CLL has left many physicians with potency of rolipram alone and in combination with either the perception that because of this ‘favorable course’ the dis- fludarabine or mitoxantrone as determined by annexin- ease can be ignored until the advanced stage at which time V/propidium iodide flow cytometric analysis and by DNA palliative therapy is indicated.1,2 A diverse interpretation of fragmentation. We studied the expression of Bcl-2 and Bcl-X the clinical data on CLL regards this disease as incurable, with expression as anti-apoptotic members and Bax expression advanced stage patients having a median survival of 18 as a pro-apoptotic member of the Bcl-2 family either by months to 3 years.3–5 The majority of circulating cells are non- flow cytometry or by Western blot analysis. We further investi- dividing, and it has been shown that the clonal excess of B gated the role of caspase activation in rolipram-induced cells results from decreased cell death rather than prolifer- apoptosis. ation.6,7 In lymphoid cells, cytolysis induced by phosphodiesterase (PDE) inhibition results from an increase in protein kinase A- mediated phosphorylation of unknown lymphoid target pro- Materials and methods teins which eventually induces apoptosis.8,9 Cyclic AMP is Ј Ј Ј catabolized within cells to 5 -AMP by 3 :5 cAMP-PDE. The Patients family of PDE includes 10 classes of enzymes which are differ- entially expressed in diverse cell types.10,11 The unspecific Nineteen patients (13 men and six women) with B-CLL who phosphodiesterase inhibitor theophylline has been reported to had not received treatment for the previous 6 months with a induce apoptosis in B-CLL cells.12 In addition, theophylline median age of 69 years (range, 53 to 79 years) were studied. B-CLL had been diagnosed according to standard clinical and laboratory criteria. Blood sampling was approved by the local Correspondence: S Endres, Division of Clinical Pharmacology, Mediz- ethics committee. The median peripheral blood leukocyte inische Klinik Innenstadt, Klinikum of the Ludwig-Maximillians-Uni- × 9 × 9 versity of Munich, Ziemssenstraβe 1, 80336 Mu¨nchen, Germany; Fax: count was 76 10 leukocytes/l (range, 12 to 159 10 089–5160–4406 leukocytes/l). The median ␤2-microglobulin concentration Received 20 November 2001; accepted 21 May 2001 was 4 mg/l (range, 2 to 14 mg/l) and the median thymidine Rolipram-induced apoptosis in CLL B Siegmund et al 1565 kinase concentration was 18 U/l (range 5 to 80 U/l) both inhibitor zVAD-fmk from R&D Systems (Wiesbaden, determined as negative predictors for disease-free Germany). survival.18–20 Leukemic cells were positive in all cases for CD5 and CD19 by flow cytometry. According to Binet’s classi- fication,2 at the time of inclusion, seven patients were in stage A, three patients were in stage B, and nine patients were at Analysis of apoptosis by annexin binding stage C (Table 1). Exposure of phosphatidylserine at the outer plasma cell mem- brane of apoptotic cells was quantified by surface annexin V staining as described.23 Briefly, one million cells were incu- Preparation of B-CLL cells bated for 48 h with the substances to be studied. Cells were then washed in phosphate-buffered saline (PBS), were resus- Mononuclear cells were isolated from peripheral blood by pended in 200 ␮l binding buffer (10 mmol/l Hepes, pH 7.4, gradient centrifugation over Ficoll–Hypaque (Biochrom, 2.5 mmol/l CaCl2, 140 mmol/l NaCl) and were incubated with Berlin, Germany), as described previously.21,22 As a modifi- 0.5 ␮g/ml of annexin V-fluorescein isothiocyanate (FITC; cation of the protocol, the isolation was performed in tubes Bender MedSystem, Vienna, Austria) for 15 min in the dark. containing a horizontal porous filter disc over the Ficoll layer Cells were washed again, were resuspended in binding buffer (Leucosep tubes; Greiner, Frickenhausen, Germany) in order and propidium iodide (PI, 5 ␮g/ml) was added. Samples were to facilitate layering of blood. RPMI 1640 culture medium was analyzed by a FACSCalibur (Becton Dickinson, Heidelberg, supplemented with 2 mML-glutamine, 10 mM Hepes buffer, Germany). For the determination of CD19 and CD5 100 U/ml penicillin and 100 ␮g/ml streptomycin (all from expression cells were labeled with either a FITC- or PE-labeled Sigma, Munich, Germany). The cells were suspended and antibody (Becton Dickinson). Data were analyzed using were cultured immediately after isolation at a final concen- FlowJo software (Version 2.7.8). tration of 2.5 × 106 cell/ml in RPMI 1640 culture medium further supplemented with 2% heat-inactivated sterile human serum. B-CLL cells were incubated at 37°C in a humidified Bcl-X analysis atmosphere containing 5% carbon dioxide. One million cells were fixed using a commercially available Fix and Perm kit (Caltag, Burlingame, CA, USA), washed in PBS and centrifuged at 300 g for 5 min. The pelleted cells Preparation of compounds were resuspended in permeabilization solution and incubated with 10 ␮l of anti-Bcl-XL (clone 7B2.5; epitope Bcl-XS; South- Rolipram (racemate of 4-[3Ј-cyclopentyloxy-4Ј-methoxy- ern Biotechnology Associates, Birmingham, AL, USA) or iso- phenyl]-2-pyrrolidone, from Schering, Berlin, Germany), sup- type-negative control (clone B 10; Southern Biotechnology plied in powder form, was dissolved in RPMI 1640 medium Associates). The cells were washed, centrifuged at 300 g for by vigorous vortexing. Fludarabine was obtained from Scher- 5 min and resuspended in 0.5 ml of 1% paraformaldehyde. ing, chlorambucil and mitoxantrone were obtained from Led- All samples were studied using a FACS Calibur (Becton erle Laboratories (Gosport, UK). Propidium iodide (PI) was Dickinson). Data were analyzed using FlowJo software obtained from Sigma (Munich, Germany) and the pan-caspase (Version 2.7.8).

Table 1 Characteristics of patients with B-CLL

Patient No. Age/Sex Stagea Leukocytes ␤2-microglobulin Thymidine Previous (109/l) (␮g/ml) kinase (U/l) treatment

1 58/F B 66.8 2.2 8.6 None 2 62/M C 51.7 1.9 6.7 Chlorambucil 3 79/M C 56.4 14.3 18.0 None 4 67/M C 84.3 1.7 5.6 Not known 5 70/M C 89.8 3.5 42.0 None 6 63/M A 36.5 3.3 6.3 None 7 79/F A 12.4 6.3 6.8 None 8 64/M C 101.9 9.9 69.8 None 9 76/F B 159.0 6.9 29.2 None 10 56/M A 26.3 2.1 11.2 None 11 53/M A 75.7 3.7 80.0 None 12 77/F C 110 4.8 28.9 Chlorambucil 13 64/F A 19.7 2.0 9.1 None 14 47/M C 76.0 4.1 5.3 None 15 60/M A 61.7 2.4 6.7 None 16 76/M C 134.6 7.3 18.0 Not known 17 69/M A 34.8 2.0 16.0 None 18 71/M C 93.6 4.2 45.6 Not known 19 72/F B 22.0 5.9 27.2 None aAccording to Binet’s classification.

Leukemia Rolipram-induced apoptosis in CLL B Siegmund et al 1566 Western blot Results

Rolipram enhances fludarabine-induced apoptosis in Cells were lysed in 10 mmol/l TRIS-HCl, pH 7.4, 1 mM EDTA, B-CLL cells 0.1% sodium dodecyl sulfate, 0.1% Triton-x 100, 1 ␮M aproti- nine and 0.1 mM phenylmethylsulfonyl fluoride. Protein con- centrations were adjusted using a colorimetric assay. Proteins To investigate the potency of rolipram alone or in combi- were separated by gel electrophoresis and transferred to a nation with fludarabine or mitoxantrone, freshly isolated B- nitrocellulose membrane (Millipore, Eschborn, Germany). The CLL cells were incubated in the presence or absence of the transfer buffer contained 14.4 g glycine, 3.03 g TRIS, and 20% different agents for 48 h. We first confirmed the previously described dose-dependent induction of apoptosis by rolipram methanol per 1000 ml. The membrane was blocked with PBS ␮ + containing 3.5% dried milk and 0.1% Tween 20 (PBS/Tween). (Figure 1). Ten M rolipram led to an increase of annexin V PI− cells (19 ± 2%) compared to the control cells (5.2 ± 1.2%; After washing three times with PBS/Tween, the membranes = = were incubated with anti-Bcl-2 at a dilution of 1:1000 (clone n 19, P 0.001). Second, we tested the apoptosis-inducing ⌬ potency of the combination of rolipram (1 ␮M) with fludarab- C 21; Santa Cruz Biotechnology, Heidelberg, Germany) or + ine (1.7 ␮M). Both rolipram alone (14.9 ± 3.1% annexin V anti-Bax at a dilution of 1:1000 (clone N-20; Santa Cruz − = ± + Biotechnology) and were visualized by autoradiography using PI cells; n 19) and fludarabine (20.5 1.2% annexin V PI− cells; n = 19) increased the proportion of apoptotic cells the ECL detection kit (Amersham Pharmacia Biotech, ± + − Freiburg, Germany). compared to control cells (5.2 1.2% annexin V PI cells; n = 19; P Ͻ 0.05). The combination of 1.7 ␮M fludarabine plus 1 ␮M rolipram further increased apoptosis (25.0 ± 1.4% annexin V+ PI− cells) compared to rolipram (n = 19, *P = 0.030) or fludarabine alone (n = 19, *P Ͻ 0.05). Third, mitox- antrone (0.5 ␮g/ml) alone induced more apoptosis (19.7 ± Caspase 3 colorimetric assay + − 5.0% annexin V PI cells; n = 19) than rolipram (1 ␮M). The combination of 1 ␮M rolipram with mitoxantrone (23.5 ± Caspase 3 activity was determined following the instructions 8.2% annexin V/PI-positive cells; n = 19) did not significantly of the commercial caspase 3 colorimetric assay (R&D Sys- increase apoptosis compared to mitoxantrone alone. tems, Minneapolis, MN, USA). As suggested by the manufac- One representative sample of the flow cytometric analysis is turer, caspase 3 activity is expressed as fold increase in shown in Figure 2. The right-shift in the presence of rolipram, caspase activity over that of control cells. fludarabine or the combination of rolipram and fludarabine indicates an increased mean fluorescence intensity (MFI) for annexin V. Additionally, in the presence of fludarabine, roli- pram or the combination there is an increased number of cells DNA fragmentation gel in the right upper quadrant (uptake of PI), indicating necrotic cells. Apoptosis was confirmed by DNA gel electrophoresis showing DNA fragmentation (Figure 3) for the different incu- × 6 Twenty 10 cells were washed with PBS and lysed in 0.1 M bation conditions. Even in the presence of 10 ␮M rolipram EDTA (pH 8.0), 0.01 M TRIS-HCl (pH 7.6), 0.02 M NaCl. Fifty alone DNA fragmentation occurred. mM sodium acetate and 0.1 mg/ml proteinase K (Roche-Diag- nostics, Mannheim, Germany) was added and incubated at 50°C for 5 h. One-fold volume phenol/chloroform/ Rolipram induces pro-apoptotic and suppresses anti- isoamylalcohol (25:24:1; Roth, Karlsruhe, Germany) was apoptotic members of the Bcl-2 family added, mixed and centrifuged for 5 min at 12000 r.p.m. The upper aqueous layer was transferred to a new microcentrifuge Bcl-2 is highly expressed in B-CLL cells contributing to the tube and 1 ml of cold isopropanol was added and incubated inhibition of apoptosis. However, not the expression of only − ° for 1 h at 70 C. After centrifugation the supernatant was one anti-apoptotic protein seems to determine apoptosis but decanted and the pellet was rinsed with 70% cold ethanol. the balance or disbalance of several pro- and anti-apoptotic After centrifugation and decanting of the supernatant the pel- members of the Bcl-2 family. Flow cytometry or Western blot ␮ let was dissolved in H2O. Ten to 15 g DNA were separated analysis was performed to determine Bcl-2 and Bcl-X in a 1.8% agarose gel. The DNA was visualized by UV expression as members of the anti-apoptotic family and Bax illumination after ethidium bromide staining. expression as a member of the pro-apoptotic family. The expression of the anti-apoptotic protein Bcl-X was dose dependently suppressed in the presence of rolipram. Figure 4a shows one representative flow cytometric analysis out of four performed. Figure 4b demonstrates the MFI ± s.e.m. of Statistical analysis these four independent experiments studying B-CLL cells of four different patients. The maximal MFI of 11.4 ± 2.2 could All data are expressed as means ± s.e.m. Statistical signifi- be reduced to 4.6 ± 0.3 in the presence of 10 ␮M rolipram (n cance of differences between treatment and control groups = 4, P = 0.027). Figure 5 presents the expression of Bcl-2 and was determined by a factorial ANOVA and a Bonferroni– Bax by Western blot analysis. In the upper panel one represen- Dunn procedure as post-hoc test. Differences were considered tative Western blot analysis out of seven different patients statistically significant for P Ͻ 0.05. Statistical analysis was shows a suppression of Bcl-2 and an induction of Bax performed using Stat-View 4.51 software (Abacus Concepts, expression. For quantitative analysis the densitometric data of Calabasas, CA, USA). Bcl-2 and Bax expression were determined and a ratio was calculated. The ratio of Bcl-2/Bax expression could signifi-

Leukemia Rolipram-induced apoptosis in CLL B Siegmund et al 1567

Figure 1 Rolipram enhances fludarabine-induced apoptosis. Mononuclear cells from patients with B-CLL were cultured for 48 h in medium with rolipram, fludarabine, mitoxantrone or a combination of rolipram with either fludarabine or mitoxantrone. Apoptosis was determined by annexin V positivity and PI negativity in flow cytometry. Rolipram dose dependently increased apoptosis (n = 19; left panel). The combination of fludarabine with rolipram significantly increased apoptosis compared with each condition alone (n = 19; middle panel). One representative flow cytometric analysis of this combination is shown in Figure 2. Incubation with the combination of mitoxantrone and rolipram did not increase apoptosis compared to mitoxantrone alone (n = 19, right panel). Bars represent means ± s.e.m.

Figure 2 Flow cytometric analysis of rolipram and fludarabine-induced apoptosis. Mononuclear cells from patients with B-CLL were cultured for 48 h in medium (control), rolipram, fludarabine or a combination of both. The abscissa reflects annexin V and the ordinate PI fluorescence. Apoptotic cells are characterized in the early stage by an increase in annexin V fluorescence and at the later stage by an additional increase in PI fluorescence. Control cells show only a small amount of annexin V positivity and PI uptake in flow cytometry (left panel; n = 19). In the presence of either rolipram (second panel; n = 19) or fludarabine (third panel; n = 19) there was a significant increase in annexin V-positive cells and PI uptake. Incubation with the combination of both significantly increased apoptosis compared with each condition alone (right panel; n = 19). One representative flow cytometric analysis out of 19 performed is depicted.

cantly be reduced from 1.0 ± 0.0 to 0.6 ± 0.1 (n = 7; P Ͻ almost completely be attenuated, comparing 12.5 ± 2.2% + 0.001) in the presence of 10 ␮M rolipram and to 0.5 ± 0.1 (n annexin V in the absence of rolipram, vs 14.8 ± 3.0% + = 7; P Ͻ 0.001) in the presence of 5 ␮M fludarabine. The annexin V in the presence of 1 ␮M rolipram and 14.6 ± 1.9% + combination of rolipram and fludarabine led to a further annexin V in the presence of 10 ␮M rolipram. When the con- decrease of the Bcl-2/Bax ratio to 0.4 ± 0.1. centration of zVAD-fmk was further increased to 200 ␮M, apoptosis induction was further suppressed below spon- taneous occurring apoptosis in untreated cells. In agreement with theses findings was the caspase 3 activity determined. In Rolipram-induced apoptosis is caspase-dependent the absence of zVAD-fmk rolipram dose dependently increased the caspase 3 activity from 1.0 ± 0.2 relative cas- Apoptosis of B-CLL cells by rolipram has been shown to corre- pase activity without rolipram vs 1.3 ± 0.2 relative caspase late with intracellular cAMP concentrations.16 However, it activity in the presence of 1 ␮M rolipram and 1.9 ± 0.3 relative remains elusive as to which apoptosis pathway is activated by caspase activity (P Ͻ 0.05 vs control) in the presence of 10 rolipram. To evaluate the role of caspase-mediated apoptosis, ␮M rolipram. In the presence of 50 ␮M zVAD-fmk increasing isolated B-CLL cells were coincubated with increasing con- concentrations of rolipram still resulted in a small increase of centrations of rolipram in the presence or absence of the pan- caspase 3 activity while in the presence of 200 ␮M zVAD-fmk caspase inhibitor zVAD-fmk. Apoptosis was determined by the caspase 3 activity determined was below the caspase 3 annexin-V positivity by flow cytometry after a 24 h incubation activity measured in the control cells. Interestingly, when (Figure 6A). In order to evaluate whether the caspase inhibitor apoptosis induction was analyzed in the same samples after is sufficiently inhibiting caspase activity, the caspase 3 activity 48 h incubation the presence of increasing concentrations of was determined as described in the Methods section (Figure rolipram resulted in increased apoptosis and increased cas- 6b). As shown in Figure 6a, in the absence of zVAD-fmk, pase 3 activity suggesting a loss of activity of zVAD-fmk after apoptosis was induced dose dependently. However, in the 24 h incubation in our experimental setting (data not shown presence of 50 ␮M zVAD-fmk apoptosis induction could as figure).

Leukemia Rolipram-induced apoptosis in CLL B Siegmund et al 1568

Figure 3 DNA fragmentation as apoptosis control marker. Mono- nuclear cells from patients with B-CLL were cultured for 72 h with medium alone (control) or rolipram, fludarabine, mitoxantrone or a combination of rolipram with either fludarabine or mitoxantrone. DNA was isolated and 12 ␮g per lane were separated by gel electro- phoresis and visualized by ethidiumbromide staining. Cells cultured in medium for 72 h did not show DNA fragmentation. However, all agents under investigation did induce DNA fragmentation.

Figure 4 Rolipram dose dependently inhibits Bcl-X expression in Discussion B-CLL cells. Mononuclear cells from patients with B-CLL were cul- tured in the presence of increasing concentrations of rolipram for 48 In the present study we were able to demonstrate for the first h. In the top panel the abscissa reflects Bcl-X expression and the ordi- time an enhancing effect of the PDE type 4 inhibitor rolipram nate cell counts. Bcl-X positivity was determined by isotype staining. Cells incubated with medium alone show a higher Bcl-X expression on fludarabine-induced apoptosis in freshly isolated human B- than cells incubated with rolipram. The flow cytometric analysis in CLL cells. Apoptosis was determined by annexin V positivity the top panel is one representative experiment out of four different and PI uptake in flow cytometry and by DNA fragmentation. patient samples. A more detailed analysis is shown in the bottom As a possible mechanism we could show a dose-dependent panel, the asterisk (*) indicating a significant reduction in Bcl-X suppression of the anti-apoptotic proteins Bcl-X and Bcl-2 by expression by 10 ␮m rolipram (n = 4; P = 0.027). rolipram and fludarabine by flow cytometry or by Western blot analysis. There was a dose-dependent increase of the pro- apoptotic protein Bax as evaluated by Western blot analysis. Both effects shift the balance of the members of the Bcl-2 fam- However, there are currently several PDE 4 inhibitors under ily towards an increased susceptibility to apoptosis. Rolipram- investigation in phase III clinical trials for the indications induced apoptosis was in parallel with increased caspase 3 asthma and chronic obstructive pulmonary disease showing activity and was completely suppressed in the presence of the encouraging results and tolerable side-effects.17,27,28 pan-caspase inhibitor zVAD-fmk. Even more remarkable is that the extent of rolipram-induced The broad-spectrum phosphodiesterase inhibitor theophyl- apoptosis is comparable to the apoptosis induced by fludara- line has been shown to induce apoptosis in B-CLL cells in bine. We found that rolipram significantly enhanced fludara- vitro and to exert some beneficial effect in vivo as currently bine-induced apoptosis. For the combination of mitoxantrone evaluated in clinical trials.12–14 A more specific therapeutic and rolipram we could not detect an increase in apoptosis. It strategy would be to take advantage of the preferential is not known in which way rolipram initiates the apoptotic expression of the PDE subtypes PDE 1B1, PDE 4A and PDE 4B pathway and whether there is a synergism with chemothera- in B-CLL cells.16 Kim and Lerner16 observed a dose-dependent peutic agents used in the treatment of B-CLL. induction of apoptosis by the specific PDE 4 inhibitor rolip- We subsequently investigated the mechanism by which rol- ram. Because of its high specificity rolipram leads to a 300- ipram induces apoptosis. We examined members of the Bcl- fold stronger induction of apoptosis than theophylline (Ref. 12 2 family as B-CLL cells have been consistently shown to con- and this paper, also Refs 14, 16). This corresponds to earlier tain high levels of the anti-apoptotic Bcl-2 protein. Several investigations by our group comparing rolipram and theophyl- studies have correlated high Bcl-2/Bax ratios with diminished line in suppression of tumor necrosis factor-␣ synthesis.24,25 apoptotic responses to chemotherapy.29–33 Consistent with its Rolipram is widely used as a reference agent for PDE 4 inhibi- therapeutic effect fludarabine decreases Bcl-2 and Mcl-1 tors in in vitro experiments. It was evaluated in clinical trials expression and increases Bax and p53 expression.33,34 In the in the 1980s as an antidepressant, when its clinical use was present study, rolipram suppressed the expression of the anti- found to be limited by side-effects like nausea and emesis.26 apoptotic members of the Bcl-2 family Bcl-X and Bcl-2 and

Leukemia Rolipram-induced apoptosis in CLL B Siegmund et al 1569

Figure 5 Bcl-2 and Bax expression determined by Western blot analysis. Mononuclear cells from patients with B-CLL were cultured for 48 h with medium alone or rolipram, fludarabine or a combination of both. In the top panel one representative experiment out of seven different patient samples is shown. For quantitative analysis the densi- tometric ratio between Bcl-2 and Bax was calculated and are presented in the buttom panel (n = 7). Asterisk (*) indicates statistically significant difference between cells incubated with medium and cells Figure 6 Rolipram-induced apoptosis is caspase-dependent. incubated with either rolipram (10 ␮M), fludarabine (5 ␮M) or a combi- nation of both. Bars respresent means ± s.e.m. Mononuclear cells from patients with B-CLL were cultured in the pres- ence of increasing concentrations of rolipram in the presence or absence of the pan-caspase-inhibitor zVAD-fmk for 24 h. Panel a rep- resents annexin V positivity in the presence or absence of rolipram and the zVAD-fmk as indicated (n = 7). Panel b shows relative caspase 3 activity in the conditions indicated (n = 4). Bars represent means ± s.e.m. increased the expression of the pro-apoptotic member Bax, thereby shifting the balance of the Bcl-2 family towards the more pro-apoptotic side. At this point, we cannot exclude that the downregulation of Bcl-2 and Bcl-x is a consequence rather The state of lymphocyte differentiation or activation may be a than the cause of apoptosis. The concomittant upregulation of major determinant whether cAMP promotes survival or death. Bax argues for the latter. In order to resolve this question To answer the question whether rolipram-induced further experiments are needed. We could not demonstrate apoptosis is caspase-mediated we incubated B-CLL cells with an association between the expression of apoptosis-regulating rolipram in the presence of the pan-caspase inhibitor zVAD- proteins and Binet’s stage. This is consistent with the findings fmk, which has been utilized for this approach recently.40 In described previously by Kitada and colleagues.35 the presence of zVAD-fmk the dose-dependent induction of Our findings, that rolipram, which dose dependently apoptosis and caspase 3 activity by rolipram could be com- increases intracellular cAMP levels induces apoptosis in pletely abrogated, indicating a caspase-dependent mechanism resting B-CLL cells is in agreement with the literature.9,36,37 of rolipram-induced apoptosis in B-CLL cells. This differs from reports concerning non-resting B cells, T cell In conclusion, since defective cell death mechanisms rather hybridomas and T lymphoblasts, where the increase of cyto- than dysregulation of cell cycle predominate in B-CLL, PDE 4 solic cAMP was associated with prevention of apoptosis.38,39 inhibitors as pro-apoptotic agents may provide a therapeutic

Leukemia Rolipram-induced apoptosis in CLL B Siegmund et al 1570 principle either as a single therapy in early-stage patients phodiesterase as a therapeutic target in chronic lymphocytic leu- or as combination therapy in advanced-stage patients with kemia. Blood 1998; 922: 484–494. B-CLL. 17 Barnes PJ. Therapeutic strategies for allergic diseases. Nature 1999; 402: B31–38. 18 Hallek M, Wanders L, Ostwald M, Busch R, Senekowitsch R, Stern S, Schick HD, Kuhn-Hallek I, Emmerich B. Serum beta(2)- Acknowledgements microglobulin and serum thymidine kinase are independent pre- dictors of progression-free survival in chronic lymphocytic leuke- We thank Rosemarie Kiefl for excellent technical assistance, mia and immunocytoma. Leuk Lymphoma 1996; 22: 439–447. 19 Hallek M, Kuhn-Hallek I, Emmerich B. Prognostic factors in Dr Wachtel (Schering AG, Berlin) for making available roli- chronic lymphocytic leukemia. Leukemia 1997; 11 (Suppl. 2): pram and Dr Uta Emmerich for completed investigations that 4–13. laid the ground for the present study. The experimental data 20 Hallek M, Langenmayer I, Nerl C, Knauf W, Dietzfelbinger H, of this study are part of the dissertation of cand. med. Julia Adorf D, Osterwald M, Busch R, Kuhn-Hallek I, Thiel E, Emmerich Welsch (Medizinische Klinik Innenstadt, Klinikum of the Lud- B. Elevated serum thymidine kinase levels identify a subgroup at wig-Maximilians-University, Munich, in preparation). These high risk of disease progression in early, nonsmoldering chronic lymphocytic leukemia. Blood 1999; 93: 1732–1737. studies were supported by a Novartis Stiftung fu¨r Therapeutis- 21 Bo¨yum A. Isolation of mononuclear cells and granulocytes from che Forschung and by a grant from the Deutsche Forschung- human blood (paper IV). Scand J Clin Lab Invest 1968; 21: 77–89. sgemeinschaft DFG SI 749/2–1 (to BS). 22 Endres S, Fu¨lle H-J, Sinha B, Stoll D, Dinarello CA, Gerzer R, Weber PC. Cyclic nucleotides differentially regulate the synthesis of tumor necrosis factor-␣ and interleukin-␤ by human mono- References nuclear cells. Immunology 1991; 72: 56–60. 23 Bellosillo B, Colomer D, Pons G, Gil J. Mitoxantrone, a topoiso- merase II inhibitor, induces apoptosis of B-chronic lymphocytic 1 Rai KR, Sawitsky A, Cronkite EP, Chanana AD, Levy RN, Pas- leukaemia cells. Br J Haematol 1998; 100; 142–146. ternack BS. Clinical staging of chronic lymphocytic leukemia. 24 Semmler J, Wachtel H, Endres S. The specific type IV phosphodie- Blood 1975; 46: 219–234. sterase inhibitor rolipram suppresses tumor necrosis factor-␣ pro- 2 Binet JL, Auquier A, Dighiero G, Chastang C, Piguet H, Goasguen duction by human mononuclear cells. Int J Immunopharmacol J, Vaugier G, Potron G, Colona P, Oberling F, Thomas M, Tchernia 1993; 15: 409–413. G, Jacquillat C, Boivin P, Lesty C, Duault MT, Monconduit M, 25 Semmler J, Gebert U, Eisenhut T, Moeller J, Scho¨nharting MM, Belabbes S, Gremy F. A new prognostic classification of chronic Alle´ra A, Endres S. Xanthine derivates: comparison between sup- lymphocytic leukemia derived from a multivariate survival analy- pression of tumor necrosis factor-␣ production and inhibition of sis. Cancer 1981; 48: 198–206. cAMP phosphodiesterase activity. Immunol 1993; 78: 520–525. 3 Lee JS, Dixon DO, Kantarjian HM, Keating MJ, Talpaz M. Prog- 26 Wachtel H. Potential antidepressant activity of rolipram and other nosis of chronic lymphocytic leukemia: a multivariate regression selective cyclic adenosine 3Ј,5Ј-monophosphate phosphodiester- analysis of 325 untreated patients. Blood 1987; 69: 929–936. 4 Montserrat E, Rozman C. Chronic lymphocytic leukaemia treat- ase inhibitors. Neuropharmacology 1983; 222: 67–72. ment. Blood Rev 1993; 7: 164–175. 27 Barnette MS. Phosphodiesterase 4 (PDE4) inhibitors in asthma and 5 Byrd JC, Rai KR, Sausville EA, Grever MR. Old and new therapies chronic obstructive pulmonary disease (COPD). Prog Drug Res in chronic lymphocytic leukemia: now is the time for a reassess- 1999; 53: 193–229. ment of therapeutic goals. Semin Oncol 1998; 25: 65–74. 28 Schudt C, Gantner F, Tenors H, Hatzelmann A. Therapeutic poten- 6 Gale RP, Caligaris-Cappio F, Dighiero G, Keating M, Montserrat tial of selective PDE inhibitors in asthma. Pulm Pharmacol Ther E, Rai K. Recent progress in chronic lymphocytic leukemia. Inter- 1999; 12: 123–129. national workshop on chronic lymphocytic leukemia. Leukemia 29 Schena M, Larsson LG, Gottardi D, Gaidano G, Carlsson M, Nils- 1994; 8: 1610–1614. son K, Caligaris-Cappio F. Growth- and differentiation-associated 7 Meinhardt G, Wendtner CM, Hallek M. Molecular pathogenesis expression of bcl-2 in B-chronic lymphocytic leukemia cells. of chronic lymphocytic leukemia: factors and signaling pathways Blood 1992; 79: 2981–2989. regulating cell growth and survival. J Mol Med 1999; 77: 282–293. 30 Hanada M, Delia D, Aiello A, Stadtmauer E, Reed JC. Bcl-2 gene 8 Coffino P, Bourne HR, Tomkins GM. Mechanism of lymphoma hypomethylation and high-level expression in B-cell chronic lym- cell death induced by cyclic AMP. Am J Pathol 1975; 81: 199– phocytic leukemia. Blood 1993; 82: 1820–1828. 204. 31 Pepper C, Bentley P, Hoy T. Regulation of clinical chemoresist- 9 McConkey DJ, Orrenius S, Jondal M. Agents that elevate cAMP ance by bcl-2 and bax oncoproteins in B-cell chronic lymphocytic stimulate DNA fragmentation in thymocytes. J Immunol 1990; leukaemia. Br J Haematol 1996; 95: 513–517. 145: 1227–1230. 32 Pepper C, Hoy T, Bentley P. Elevated bcl-2/Bax are a consistent 10 Beavo JA. Cyclic nucleotide phosphodiesterases: functional impli- feature of apoptosis resistance in B-cell chronic lymphocytic leu- cations of multiple isoforms. Physiol Rev 1995; 75: 725–748. kaemia and are correlated with in vivo chemoresistance. Leuk 11 Soderling SH, Bayuga SJ, Beavo JA. Isolation and characterization Lymphoma 1998; 28: 355–361. of a dual-substrate phosphodiesterase gene family: PDE10A. Proc 33 Pepper C, Thomas A, Hidalgo de Quintana J, Davies S, Hoy T, Natl Acad Sci USA 1999; 96: 7071–7076. Bentley P. Pleiotropic drug resistance in B-cell chronic lympho- 12 Mentz F, Merle-Beral H, Dalloul AH. Theophylline-induced B-CLL cytic leukaemia – the role of Bcl-2 family dysregulation. Leukemia apoptosis is partly dependent on cyclic AMP production but inde- Res 1999; 23: 1007–1014. pendent of CD38 expression and endogenous IL-10 production. 34 Bellosillo B, Villamor N, Colomer D, Pons G, Montserrat E, Gil J. Leukemia 1999;13: 78–84. In vitro evaluation of fludarabine in combination with cyclophos- 13 Binet JL, Mentz F, Leblond V, Merle-Beral H. Synergistic action phamide and/or mitoxantrone in B-cell chronic lymphocytic leu- of alkylating agents and methylxanthine derivatives in the treat- kemia. Blood 1999; 94: 2836–2843. ment of chronic lymphocytic leukemia. Leukemia 1995; 9: 35 Kitada S, Andersen J, Akar S, Zapata JM, Takayama S, Krajewski 2159–2161. S, Wang HG, Zhang X, Bullrich F, Croce CM, Rai K, Hines J, Reed 14 Mentz F, Mossalayi MD, Ouaaz F, Baudet S, Issaly F, Ktorza S, JC. Expression of apoptosis-regulating proteins in chronic lympho- Semichon M, Binet JL, Merle-Beral H. Theophylline synergizes cytic leukemia: correlations with in vitro and in vivo chemores- with chlorambucil in inducing apoptosis of B-chronic lymphocytic ponses. Blood 1998; 91: 3379–3389. leukemia cells. Blood 1996; 88: 2172–2182. 36 Alessi DR, Cuenda A, Cohen P, Dudley DT, Saltiel AR. PD98050 15 Schwabe U, Ukena D, Lohse MJ. Xanthine derivatives as antagon- is a specific inhibitor of the activation of mitogen-activated protein ists at A1 and A2 adenosine receptors. Naunyn Schmiedebergs kinase kinase in vitro and in vivo. J Biol Chem 1995; 270: Arch Pharmacol 1985; 330: 212–221. 7489–7494. 16 Kim DH, Lerner A. Type 4 cyclic adenosine monophosphate phos- 37 Kizaki H, Suzuki K, Tadakuma T, Ishimura Y. Adenosine receptor

Leukemia Rolipram-induced apoptosis in CLL B Siegmund et al 1571 mediated accumulation of cyclic AMP-induced death through phosphate rescue B-cell antigen receptor-induced apoptosis internucleosomal DNA cleavage. J Biol Chem 1990; 265: 280– through independent pathways and converge to prevent caspase 284. activation. J Allergy Clin Immunol 2000; 105: 522–531. 38 Goetzel EJ, An Z, Zeng L. Specific suppression of prostaglandin 40 Mateo V, Lagneaux L, Bron D, Biron G, Armant M, Delespesse E2 of activation-induced apoptosis of human CD4+CD8+T lym- G, Sarfati M. CD47 ligation induces caspase-independent cell phoblasts. J Immunol 1995; 154: 1041–1047. death in chronic lymphocytic leukemia. Nat Med 1999; 5: 39 Sakata N, Kawasome H, Terada N, Johnson GL, Gelfand EW. 1277–1284. CD40 and adenosine A2 receptor agonist-cyclic adenosine mono-

Leukemia