Synergistic Activity of Bortezomib and Hdaci in Preclinical Models of B-Cell Precursor Acute Lymphoblastic Leukemia Via Modulation of P53, PI3K/AKT, and NF-Kb

Published OnlineFirst January 28, 2013; DOI: 10.1158/1078-0432.CCR-12-1511

Clinical Cancer Cancer Therapy: Preclinical Research

Synergistic Activity of Bortezomib and HDACi in Preclinical Models of B-cell Precursor Acute Lymphoblastic Leukemia via Modulation of p53, PI3K/AKT, and NF-kB

Lorenz Bastian1, Jana Hof1,2, Madlen Pfau1, Iduna Fichtner3, Cornelia Eckert1,Gunter€ Henze1, Javier Prada1, Arend von Stackelberg1, Karl Seeger1, and Shabnam Shalapour1

Abstract Purpose: Relapse of disease and subsequent resistance to established therapies remains a major challenge in the treatment of childhood B-cell precursor acute lymphoblastic leukemia (BCP-ALL). New therapeutic options, such as proteasome and histone deacetylase inhibitors (HDACi) with a toxicity profile differing from that of conventional cytotoxic agents, are needed for these extensively pretreated patients. Experimental Design: Antiproliferative and proapoptotic effects of combined HDACi/proteasome inhibitor treatments were analyzed using BCP-ALL monocultures, cocultures with primary mesenchymal stroma cells from patients with ALL, and xenograft mouse models. The underlying molecular mechanisms associated with combined treatment were determined by gene expression profiling and protein validation. Results: We identified the proteasome inhibitor bortezomib as a promising combination partner for HDACi due to the substantial synergistic antileukemic activity in BCP-ALL cells after concomitant application. This effect was maintained or even increased in the presence of chemotherapeutic agents. The synergistic effect of combined HDACi/BTZ treatment was associated with the regulation of genes involved in cell cycle, JUN/MAPK, PI3K/AKT, p53, ubiquitin/proteasome, and NF-kB pathways. We observed an activation of NF-kB after bortezomib treatment and the induction of apoptosis-related NF- kB target genes such as TNFaRs after concomitant treatment, indicating a possible involvement of NF-kBas proapoptotic mediator. In this context, significantly lower NF-kB subunits gene expression was detected in leukemia cells from patients who developed a relapse during frontline chemotherapy, compared with those who relapsed after cessation of frontline therapy. Conclusion: These results provide a rationale for the integration of HDACi/BTZ combinations into current childhood BCP-ALL treatment protocols. Clin Cancer Res; 19(6); 1445–57. 2013 AACR.

Introduction (1). New treatment approaches with a different toxicity Despite continuous improvement of therapeutic options proﬁle in comparison to conventional cytotoxic agents for children with acute lymphoblastic leukemia (ALL), the are required for these extensively pretreated patients. As a prognosis for patients who suffer a relapse remains poor promising alternative, histone deacetylase inhibitors with a probability of even-free survival of 30% at 10 years (HDACi) have been suggested (2, 3). We have previously shown that HDACi are able to limit the expansion of B-cell precursor (BCP)-ALL cells in vivo (4). The inclusion of

Authors' Affiliations: 1Department of Pediatric Oncology/Hematology, HDACi into current ALL chemotherapy concepts requires 2Department of Pediatrics, Division of General Pediatrics, Charite-Univer- the identification of suitable combination partners, as first sitatsmedizin€ Berlin; and 3Max-Delbruck-Center€ for Molecular Medicine, clinical trials achieved only limited therapeutic responses Experimental Pharmacology & Oncology, Berlin, Germany (5, 6) while establishing a good tolerance of HDACi in Note: Supplementary data for this article are available at Clinical Cancer pediatric patients with malignant diseases. So far, only Research Online (http://clincancerres.aacrjournals.org/). additive or even antagonistic interactions have been K. Seeger and S. Shalapour are co-last authors. observed after combined treatment with HDACi and dif- Corresponding Author: Shabnam Shalapour, Charite-Universit atsmedi-€ ferent conventional cytostatic drugs in myeloid leukemia zin Berlin, Berlin 13353, Germany. Phone: 49-30-450-559568; Fax: 49-30- and T-ALL cells (7). In BCP-ALL leukemia cells, we have 450-566946; E-mail: [email protected] recently observed that the sequence of drug application doi: 10.1158/1078-0432.CCR-12-1511 determined synergistic or antagonistic responses to com- 2013 American Association for Cancer Research. bined treatment with HDACi and methotrexate (8).

www.aacrjournals.org 1445

Bastian et al.

of Microorganisms and Cell cultures (DSMZ). The cell Translational Relevance lines were routinely tested for the expression of the Cure rates for children with B-cell precursor acute corresponding surface markers and fusion genes by fluo- lymphoblastic leukemia (BCP-ALL) have continuously rescence-activated cell sorting (FACS), quantitative real- improved, but treatment is still associated with the risk of time PCR (QRT-PCR), and FISH. serious acute and late effects and the outcome of patients All patients were enrolled in the relapse trial ALL-REZ with relapses is poor. Therapeutic alternatives with sub- BFM 2002, approved by the Institutional Review Board of stantial antileukemic potential and toxicity profiles dif- the Charite-Universit€atsmedizin Berlin, Berlin, Germany fering from that of conventional chemotherapeutics are (ClinicalTrials.gov identifier: NCT00114348). Written needed, as suggested for histone deacetylase inhibitors informed consent was obtained from patients or guardians. (HDACi). Proteasome inhibitors were found to increase the activity of HDACi in mature B-cell malignancies. In Substances BCP-ALL, we identified bortezomib as a promising com- Bortezomib (Janssen-Cilag International) and valproic bination partner for HDACi due to the synergistic anti- acid (VPA, in-vitro: Sigma-Aldrich, in-vivo: Desitin Arznei- leukemic activity. Our analyses indicate that active pro- mittel) were dissolved in isotonic saline or aqua and sub- teasomal processing and histone deacetylation function eroylanilide hydroxamic acid (SAHA, Merck Pharmaceuti- are indispensable for the survival of immature B-cell cals) in dimethyl sulfoxide (final concentration < 0.02%). malignancies, although the underlying mechanisms dif- Unless not otherwise stated, the final drug concentrations fered in part from mature B-cell malignancies. The were 1 mmol/L for VPA (6), 1.2 mmol/L for SAHA (23), synergism of BTZ/HDACi was maintained or even and 10 nmol/L for bortezomib (in Nalm6: 7 nmol/L BTZ; increased in the presence of chemotherapeutic agents, ref. 14), as chosen according to clinically achievable plasma particularly anthracyclines, establishing a rationale for concentrations. the inclusion of HDACi/BTZ combinations into current chemotherapy regimens for childhood BCP-ALL. Gene expression analysis Reh cells where treated either with bortezomib, VPA, or the combination of both for 12 hours. Reh cells incubated with medium served as control. After total RNA extraction Proteasome inhibitors such as bortezomib (BTZ) were (Qiagen), samples were hybridized on Agilent Whole found to increase substantially the activity of HDACi in Human Genome Oligo Microarrays (Miltenyi Biotec). different hematologic and solid malignancies (9–13). To Microarray normalization was conducted by Miltenyi Bio- date, no systematic studies have been conducted to analyze tec GmbH, whereas for further data analysis, the Partek the response of B-cell precursor leukemias to HDACi/pro- Genomics Suits software, version 6.5 beta 2009 (Partek teasome inhibitor combinations. A well-tolerated bortezo- Inc.) was used. Gene expression data are available in NCBI’s mib dose level in patients with childhood leukemia (14) Gene Expression Omnibus data base (accession number and described additive interactions of bortezomib with GSE41951). conventional cytostatics in vitro indicate bortezomib as a promising agent for treatment of BCP-ALL (15). Mouse models The synergistic induction of apoptosis after HDACi/ Male nonobese diabetic/severe combined immunode- BTZ treatments has been partially associated with an ficient (NOD/SCID) mice were injected either intravenous- inhibition of NF-kB signaling in mature B-cell malignan- ly with leukemia cells derived from a pediatric patient with cies, myeloid leukemia, and T-ALL (9, 10, 13, 16). These BCP-ALL at first relapse (ALL-SCID2) or subcutaneously observations are consistent with the finding that NF-kB with Nalm6 cells (1þ1 mixture with Matrigel) by transfer- signaling is indispensable for the homeostasis of mature B ring 1 107 leukemia cells per mouse (24). Animal experi- cells and contributes to the survival of mature B-cell ments were approved by the local responsible authorities malignancies (17, 18). The significance of NF-kB signal- (G0221/03) and carried out according to UICC guidelines ing for the survival and differentiation of precursor B cells (25). is still less clear (17, 19), and its role in the leukemo- genesis and treatment response of BCP-ALL remains con- Combination effects and statistical analyses troversial (20–22). Student t test, Mann–Whitney U test, and Kruskal–Wallis Therefore, we studied the effect of combined HDACi/ test were used for the analysis of statistical significance. To proteasome inhibitor treatments in BCP-ALL in vivo analyze the effects of combined treatment with regard to and in vitro and analyzed the underlying molecular synergism, additivity, or antagonism, either the combina- mechanisms. tion index (CI) according to Chou and Talalay was determined (CalcuSyn software; Biosoft) or the fractional prod- Material and Methods uct was calculated using the method of Webb (26). Cell lines and patient samples Further detailed information on the applied material and The human BCP-ALL cell lines Reh, Nalm6, SD-1, 697, methods is provided in the Supplementary Material and and SEM were purchased from the German Collection Methods.

1446 Clin Cancer Res; 19(6) March 15, 2013 Clinical Cancer Research

Bortezomib and HDACi in Childhood BCP-ALL

Results ness to bortezomib and HDACi. Simultaneous applica- Concomitant treatment with bortezomib and HDACi tion of both drugs was required for the synergistic induc- has synergistic effects in BCP-ALL cells tion of cell death (Fig. 1C; Supplementary Table S4). In To assess the antileukemic effect of combined treat- contrast, sequential applications of bortezomib and ments with HDACi and either established ALL therapy HDACi induced in most cases apoptosis in a near-addi- elements (idarubcin, vincristine, cytarabine) or the pro- tive manner, with the exception of VPA followed by teasome inhibitor bortezomib, BCP-ALL cell lines were bortezomib, which even resulted in antagonistic interac- exposed to the respective compounds either alone or in tions. Antagonism was not observed in any of the tested combination. Leukemia cells proliferation and the induc- concomitant combinations, thus confirming a broad syn- tion of apoptosis were measured by MTS assay and ergistic activity of bortezomib and HDACi in BCP-ALL FACS analysis after Annexin-V/propidium iodide (PI) cells, when applied concomitantly. staining, respectively (Reh, Fig. 1A and B). After concomitant HDACi/BTZ treatment inhibition of proliferation The synergistic induction of apoptosis involves the and induction of apoptosis was markedly increased, as regulation of cell cycle, death receptor, and p53 compared with bortezomib alone (Fig. 1A and B, left; signaling pathways Supplementary Fig. S1A). Combination index analysis To determine the molecular mechanisms involved and/ identified the antileukemic activity of HDACi/BTZ com- or regulated by the synergisticactivityofVPA/BTZtreat- binations as synergistic (Fig. 1A and B, right). This syn- ment, we sought to identify genes whose expression ergistic effect was also observed, when serial dilutions of changed after combined treatments. Gene expression HDACi were combined with a constant concentration of profiles were analyzed from the cell line Reh after single bortezomib (Supplementary Fig. S1B–S1D) and when treatment with either VPA or bortezomib as well as after apoptosis was assessed 24 and 48 hours after drug appli- concomitant treatment, compared with untreated control cation (Supplementary Fig. S1E). samples. The obtained results were further validated in The combined treatment of HDACi with idarubicin four cell lines. We identified 919 genes differentially induced mostly synergistic interactions, which were less expressed exclusively after concomitant treatment (Sup- pronounced when compared with HDACi/BTZ combina- plementary Tables S7 and S8). Pathway analyses con- tions (Fig. 1A and B). The combinations of HDACi with firmed, as was observed by aforementioned in vitro tests, vincristine or cytarabine were additive to antagonistic, the involvement of genes related to cell cycle, apoptosis, depending on tested concentrations and cell lines (Fig. and death receptor signaling. Furthermore, changes in the 1A and B). Similar results were obtained in Nalm6 cells signaling pathways involving retinoic acid, NF-kB, p53, (Supplementary Fig. S2A and Supplementary Table S1 and TLR/TREM1, PI3K/AKT/mTOR, mitogen-activated pro- S2), revealing bortezomib as the most promising combi- tein kinase (MAPK), endoplasmic reticulum (ER) stress nation partner for HDACi. response, and protein ubiquitination were identified. For further validation the amount of cell death in 697, Heatmap analyses for 7 of these pathways and validations SEM and SD-1 cells were analyzed after application of 2 of the corresponding genes are shown in Figs. 2–4 and different concentrations of bortezomib with and without Supplementary Figs. S6 and S7. VPA at 3 time points. Fractional product analysis indicated Upregulation of proapoptotic genes, such as caspases and synergism for the majority of tested conditions (Supple- TNFaR (TNF-related apoptosis-inducing ligand receptors), mentary Table S3). In addition, primary BCP-ALL cells and downregulation of antiapoptotic genes, such BCL2, obtained from 3 patients at relapse diagnosis were exposed elucidate the synergistic proapoptotic effect of concomitant to VPA and bortezomib treatments. Fractional product and treatment (Fig. 2A). The increase of DR-5 (TNFRSF-10B; Fig. combination index analysis showed that concomitant treat- 2B) and cleaved caspase-3 (Supplementary Fig. S6A) was ment induced apoptosis and reduced the amount of vital confirmed by flow cytometry. Analysis of cell-cycle distribu- þ CD19 leukemia cells in either additive or synergistic tions indicated a G0–G1 phase arrest after treatment with manner (Supplementary Fig. S2B–S2D). VPA, whereas after treatment with bortezomib or the com- To evaluate the effects of VPA and bortezomib treat- bination of VPA/BTZ no relevant changes in the cell-cycle ments on normal B and T cells, PBMCs of 3 healthy phases were observed, despite the synergistic induction of adults, murine splenocytes and murine BM-MNC were apoptosis after combined treatment (Supplementary Fig. analyzed in vitro (Supplementary Figs. S3–S5). Analyses S6B). The gene expression signature of concomitant treat- of subpopulations revealed a significant reduction of ment revealed the modulation of the p53 pathway, which þ CD19 B cells after single and concomitant treatment. we have identified to be associated with the response to þ The proportion of human CD3 T cells remained therapy of relapsed ALL (27). Key genes of the p53 pathway unchanged in vitro. such as MDM2 and CDKN1A were differentially expressed Changes in the order of application can substantially in leukemia cells after concomitant treatment (Fig. 2C). alter the effect of drug combinations, as we have These genes were further validated by immunocytochem- previously shown for the combination of HDACi and istry (Fig. 2D) and QRT-PCR after simultaneous and methotrexate (8). Therefore, we analyzed whether the sequential exposure to both drugs (Supplementary Fig. sequence of application has an effect on drug responsive- S6C and S6D). CDKN1A mRNA level was increased after

www.aacrjournals.org Clin Cancer Res; 19(6) March 15, 2013 1447

Bastian et al.

ABProliferation Apoptosis Bortezomib Bortezomib 2.0 100 2.0 BTZ + VPA 100 BTZ + VPA BTZ + SAHA BTZ + SAHA 80 1.5 80 1.5 antagonistic antagonistic

60 20 40 60 80 100 60 20 40 60 80 100 1.0 1.0 CI CI 40 40 synergistic synergistic BTZ 0.5 BTZ 0.5 20 20 BTZ + VPA BTZ + VPA BTZ + SAHA 0.0 BTZ + SAHA

Apoptotic cells (% of total) 0 0.0

(% of untreated control) 0 Inhibition of proliferation 0 5 10 15 20 25 Inhibition of proliferation (%) 0 5 10 15 20 25 Apoptotic cells (% of total) BTZ (nmol/L) BTZ (nmol/L)

Idarubicin Idarubicin 2.0 2.0 100 IDA + VPA 100 IDA + VPA IDA + SAHA IDA + SAHA 80 1.5 80 1.5 antagonistic antagonistic 60 20 40 60 80 100 60 20 40 60 80 100 1.0 1.0 CI 40 CI 40 IDA synergistic IDA synergistic 0.5 0.5 20 IDA + VPA 20 IDA + VPA IDA + SAHA IDA + SAHA

0.0 Apoptotic cells (% of total) 0.0 (% of untreated control) 0 0 Inhibition of proliferation 0 20 40 60 80 Inhibition of proliferation (%) 0 20 40 60 80 Apoptotic cells (% of total) IDA (nmol/L) IDA (nmol/L)

Cytarabine Cytarabine 2.0 100 2.5 100

80 2.0 80 1.5 antagonistic 60 1.5 60 20 40 60 80 100 antagonistic 1.0 20 40 60 80 100 CI CI 40 1.0 40 ARAC ARAC synergistic synergistic 0.5 20 ARAC + VPA 0.5 20 ARAC + VPA ARAC + VPA ARAC + VPA ARAC + SAHA ARAC + SAHA ARAC + SAHA 0.0 ARAC + SAHA 0.0 Apoptotic cells (% of total)

(% of untreated control) 0 0 Inhibition of proliferation 0.0 0.2 0.4 0.6 0.0 0.2 0.4 0.6 Apoptotic cells (% of total) AraC (µmol/L) Inhibition of proliferation (%) AraC (µmol/L)

Vincristine Vincristine 2.0 100 100 2.0

80 80 1.5 1.5 antagonistic antagonistic

60 20 40 60 80 100 60 20 40 60 80 100 1.0 1.0 CI 40 40 CI synergistic VCR 0.5 VCR 0.5 synergistic 20 20 VCR + VPA VCR + VPA VCR + VPA VCR + VPA VCR + SAHA VCR + SAHA VCR + SAHA VCR + SAHA 0.0 Apoptotic cells (% of total) 0 (% of untreated control) 0 0.0 Inhibition of proliferation 0 2 4 6 Inhibition of proliferation (%) 0 2 4 6 Apoptotic cells (% of total) VCR (nmol/L) VCR (nmol/L) C 100 *** ***

total) 80 f

(%o 60 s l el

cc 40 i tot 20 pop A 0

A A Z trol PA TZ H V VP B A BT on S C TZ + T P B B V H T BT VP A BTZ + SAHA S B SAH

Figure 1. Concomitant combinations of HDACi and BTZ synergistically inhibit proliferation and induce apoptosis in BCP-ALL cell lines. A and B, Reh cells were treated for 72 hours with BTZ or chemotherapeutic agents alone or in combination with HDACi (VPA, SAHA). A, the inhibition of proliferation (MTS assay). B, the induction of apoptosis (Annexin-V/PI FACS), together with CI analyses (right). Duplicates SD of 2 to 3 experiments. C, induction of apoptosis in Reh cells after treatment with single drugs for 36 hours followed by 36 hours in fresh media or sequential combinations of 36 hours with the ﬁrst and 36 hours with the second drug or concomitant combinations of both drugs for 72 hours. Triplicates SD of one experiment representative for Reh and Nalm6. , P < 0.001 (Student t test).

1448 Clin Cancer Res; 19(6) March 15, 2013 Clinical Cancer Research

Bortezomib and HDACi in Childhood BCP-ALL

A Cell-cycle regulation and B C death receptor pathway p53 signaling

BTZ VPA BTZ + VPA BTZ VPA BTZ + VPA

SUV39H1 contr (MFI: 8.42) E2F1 VPA (MFI: 7.65) (MFI: 9.39) CCNA2 BTZ MDM2 VPA+BTZ (MFI: 13.45) PRKDC PRKDC CCNE2 JMY Cell count MCM7 BBC3 MCM5 C12orf5 PLK1 CDKN1A BCL2 DR-5 (TNFRSF-10B) fluorescence intensity JUN HUS1 GADD45G CFLAR CCND2 TNFRSF10B

CASP9 -3 17

CASP10 D

TNFRSF25 2.0 × 10 7 ** 100

TNFRSF10A TCF) 80 1.5 × 10 7 PPM1J (C n 60 ssio 1.0 × 107 EBF3 40

cells (% of total) pre ex × 6

5.0 10 +

-3 17 2 20 DM p53

M 0 0 Control VPA + BTZ Control VPA +BTZ

MDM2 p53

MDM2 p53 DAPI DAPI

Control BTZ + VPA Control BTZ + VPA

Figure 2. Induction of cell-cycle, death receptor, and p53 pathways after concomitant treatment. Reh cells were treated with either BTZ or VPA alone or in concomitant combination. After 12 hours, mRNA was isolated for gene expression analyses. A, heatmap depicts differentially expressed genes in Reh cells after concomitant VPA/BTZ treatment, which regulate cell-cycle and death receptor signaling. Expression levels of those genes after single treatments. B, FACS analyses of DR-5 (TNFRSF10B) expression on Reh cells after treatments. Representative for 2 experiments. C, heatmap depicts differentially expressed genes involved in the p53 pathway. D, Reh cells after treatments were stained with mAbs against MDM2 (green) or p53 (red) and 40,6-diamidino-2- phenylindole (DAPI). Magniﬁcation, 400. Shown are the corrected total cell ﬂuorescence measurements of MDM2 expression in 30 nuclei SD and the presence of nuclear p53 after counting of 40 to 70 nuclei. , P < 0.01 (Mann–Whitney U test).

VPA single, concomitant, and sequential treatments. The Concomitant treatment involves the modulation of MDM2 mRNA and protein level decreased signiﬁcantly only PI3K/AKT/mTOR, MAPK, and protein ubiquitination after concomitant treatment, as conﬁrmed by QRT-PCR and signaling pathways immunocytochemistry. In line with this, only after con- The gene expression signature of concomitant treatment comitant treatment the upregulation and nuclear localiza- revealed a differential regulation of genes involved in PI3K/ tion of p53 protein was detectable (Fig. 2D), revealing the AKT/mTOR (e.g., ITGA4, BCL-2) and MAPK signaling path- involvement of the p53 pathway in the synergistic induc- ways (e.g., MAP2K3, CREB5; Fig. 3A). The effects of 2 PI3K/ tion of apoptosis. AKT inhibitors (LY294002, MK-2206) on leukemia cell

www.aacrjournals.org Clin Cancer Res; 19(6) March 15, 2013 1449

Bastian et al.

PI3K/AKT/mTOR and AB3 * CProtein ubiquitination pathway MAPK pathway *

n * BTZ VPA BTZ + VPA 12 h BTZ VPA BTZ + VPA 24 h ITGA4 2 UBE2C essio r CDC20 GPLD1 PSMB8 MLST8 UCHL1 1 tive exp SOCS1 EIF4G2 a of VLA-4 mRNA el USP6

YWHAE R HSPA8 GAB1 0 PSMB2 RPS6KA4 Control BTZ VPA BTZ + VPA PSMC6 HSP90AA1 PSMA1 D PSMC1 PLD1 Control BTZ + SAHA PSMD12 RHOB Ubi PSMA3 PRKAG2 PSMD7 RHOQ PSMD4 PSMB4 PIK3C2B UBC PRKCB UBE4B MKNK1 PSMD2 PSMD3 HSP90B1 Ubi DAPI USP45 MAP2K3 HSPA5 4.0 ) *** DNAJB6 MEF2B 7 3.0 RHOF DNAJC3 2.0 UBE2H IL1RAP 1.0 USP30 ubiquitin (CTCFubiquitin x 10 MAPK10 0 IFNGR1 control VPA+BTZ RPS6KA2 DNAJB9 DNAJC18 IL18RAP 30 *** HSPA6 CREB5 ** HSPA1A NA ** -3 17 -3 R 17 20 1A m A P

S 10 H ative expression l of Re

0 Control BTZ VPA BTZ + VPA

Figure 3. Differential regulation of genes involved in PI3K/AKT/mTOR/MAPK and protein ubiquitination pathway after concomitant treatment. A, heatmap depicts differentially expressed genes involved in the PI3K/AKT/mTOR/MAPK pathway. B, relative ITGA4 (VLA-4) mRNA expression in Reh cells after treatments. C, heatmap depicts differentially expressed genes involved in the protein ubiquitination pathway. D, above, SD-1 cells were stained with mAb against ubiquitin (green) and 40,6-diamidino-2-phenylindole (DAPI) after treatment for 12 hours. Magniﬁcation, 400. Graph, corrected total cell ﬂuorescence measurements of ubiquitin expression in 30 nuclei SD; below, relative HSPA1A mRNA expression in Reh cells after treatment for 12 hours. Triplicates SD of 1 to 2 experiments. , P < 0.05; , P < 0.01; , P < 0.001 (Student t test). ABL1 was used as housekeeping gene for both QRT-PCR experiments.

response to VPA/BTZ treatment were further determined. have previously found to be regulated via VLA-4 signaling Although the treatment with PI3K/AKT inhibitors did not in leukemia cells (29). induce apoptosis in leukemia cells, the triple combination Gene expression proﬁling further revealed the modu- with VPA/BTZ showed clearly a decrease of apoptosis com- lation of the protein ubiquitination pathway after appli- pared to VPA/BTZ treatment (Supplementary Fig. S7A– cation of bortezomib either alone or in combination with S7E), revealing the involvement of the PI3K/AKT pathway VPA (Fig. 3C). Bortezomib treatment markedly increased in the synergistic induction of apoptosis. the expression of regulatory proteasome subunits (19S) as Moreover, ITGA-4 (VLA-4)-mediated signaling during well as structural and catalytic subunits (20S). After the interaction with stroma cells has been shown to concomitant treatment the upregulation of 19S and support the survival of leukemia cells mostly through the 20S subunits was less pronounced and the immunopro- activation of the PI3K pathway (28, 29). VLA-4 expression teasome subunit b5i was downregulated. Thus, concom- was validated by QRT-PCR at 2 time points, showing that itant treatment with VPA shifted the regulation of protea- the VLA-4 mRNA expression decreased signiﬁcantly after some subunits, which was induced by bortezomib single concomitant treatments (Fig. 3B; Supplementary Fig. treatment. Ubiquitin mRNA and protein level increased S6E). Furthermore, concomitant treatment downregu- after bortezomib treatment, which was even more lated the expression of NLRXL, an inhibitor of NF-kB pronounced after concomitant treatment (Fig. 3C and activation, and upregulated CRY1, a circadian clock gene D; Supplementary Fig. S6G). This is consistent with (Supplementary Tables S7 and S8), both of which we previously published results reporting the accumulation

1450 Clin Cancer Res; 19(6) March 15, 2013 Clinical Cancer Research

Bortezomib and HDACi in Childhood BCP-ALL

of polyubiquitinated proteins as a consequence of analyzed by the gene expression profiling data available for reduced proteasome activity (11). Furthermore, gene a total cohort of 52 patients (39). We compared the expres- expression analysis showed an upregulation of different sion levels of NF-kB subunits in relevant clinical and bio- heat shock protein mRNAs (Fig. 3C; e.g., HSPA1A and logic subgroups of ALL relapse (Supplementary Table S5). HSPA5). The highest induction of HSPA1A was observed NF-kB subunits expression (RelA, RelB, NFKB2, NFKB1)in after concomitant treatment, as validated by QRT-PCR leukemia cells correlated significantly with the time point of (Fig. 3D; Supplementary Fig. S6F), indicating the activa- relapse. Furthermore, patients who suffered a relapse of ALL tion of the ER stress/unfolded protein response pathway. during the frontline chemotherapy (relapse on treatment) had significantly lower NF-kB subunits expression than Concomitant treatment activates the NF-kB signaling those who suffered a relapse after cessation of frontline pathway therapy (relapse off-treatment; Fig. 4D; Supplementary The proapoptotic effect of bortezomib in different malig- Table S5). nancies has been attributed, in part, to the inhibition of the constitutive and induced activation of NF-kB signaling Concomitant treatment has synergistic effects in pathway (9, 13, 30, 31). In contrast, an activation of the leukemia cell cocultures with mesenchymal stroma NF-kB pathway after bortezomib treatment has also been cells described in multiple myeloma (32) and endometrial car- To ascertain, whether the stroma might exert a cytopro- cinoma cells (33). tective effect on leukemia cells, as it was observed when In the tested BCP-ALL model, gene expression profiles treated with some chemotherapeutics (29, 40), the effects of clearly showed a modulation of NF-kB signaling pathway combined VPA/BTZ treatments were analyzed in cocultures genes after bortezomib single and concomitant treat- of leukemia cells with primary stroma cells. Mesenchymal ments (Fig. 4A). Importantly, the gene expression signa- stroma cells (MSC) were isolated from bone marrow (MSC; ture of NF-kB pathway varied between single and n ¼ 4) or testis (T-MSC; n ¼ 4) of patients with BCP-ALL. concomitant treatments, for example, regarding the Both organs represent major sites of leukemia cells infiltra- expression of interleukin (IL)-6, IL-8,andTLR7. However, tion in relapsed ALLs. Reh and 697 cells were cocultured from the 5 NF-kB subunits, only RelB mRNA expression with MSCs and treated with VPA and bortezomib as significantly increased after concomitant treatment. We described above. The amount of cell death was analyzed assessed the NF-kB activation by FACS and Western blot by flow cytometry. Combined treatment induced synergistic analysis of p65 (RelA) phosphorylation in Reh and 697 cell death in leukemia cells both in monocultures and in cells at multiple time points after VPA/BTZ treatment, cocultures (Fig. 5A; Supplementary Fig. S8A and S8B). revealing a marked p-p65 increase 9 hours after drug No significant induction of cell death in primary MSCs and application (Fig. 4B, above; Supplementary Fig. S7E and T-MSCs was observed both in monocultures and cocultures S7F). Accordingly, the activation of 4 NF-kB subunits (Supplementary Fig. S8C and S8D, S8G and S8H). How- (p65, RelB, p50, p52) was determined in 3 BCP-ALL cell ever, antileukemic activity of combined treatment in cocul- lines 9 hours after single and combined treatments using tures was reduced about 15% to 20% when compared with DNA-binding ELISA-based analysis (Fig. 4B, below). A monocultures (Fig. 5A). Moreover, culturing primary leu- significant activation of p65 was observable after borte- kemia cells with MSCs confirmed the prosurvival effect of zomib and concomitant treatments, which increased even MSCs, whereas combined treatment was able to overcome more after concomitant treatment in 2 of 3 cell lines (Fig. the supportive effect of MSCs (Supplementary Fig. S8E and 4B). Immunocytochemical analysis confirmed the nucle- S8F). In accordance with the aforementioned results, we ar translocation of p65 after concomitant treatment (Fig. observed an increase of RelA mRNA expression, transloca- 4C; Supplementary Fig. S9B). These results indicated an tion of p65 in nuclei and decrease of MDM2 in leukemia activation of the canonical NF-kB pathway. Furthermore, cells after combined treatment in cocultures (Supplemen- DNA-binding activity of RelB subunits increased slightly tary Fig. S9A–S9C). after concomitant treatment in 697 and Reh cells, sug- gesting also a possible involvementofthenoncanonical Combined treatment reduces leukemia cell burden in a pathway. BCP-ALL i.v. xenograft mouse model It has been assumed that tumor cells activate NF-kB To assess the effect of combined VPA/BTZ treatment on signaling in response to chemotherapy as a prosurvival leukemia cells survival and engraftment in vivo, an estab- factor (34). In contrast, new findings (35–37) and our lished i.v. xenograft mouse model of chemotherapy-resis- results indicate that NF-kB has also a tumor-suppressive tant childhood BCP-ALL was used (24). The protocol function. In ALLs, a constitutive activation of NF-kBasa included four groups, which were treated with isotonic survival factor but also an involvement of NF-kB signaling saline (control), bortezomib, VPA or VPA/BTZ, respectively in chemotherapy-induced apoptosis was reported (Supplementary Fig. S10A). Dosages and treatment sche- (20, 37, 38). A clear correlation between NF-kB expression dules for xenograft experiments were determined within a and BCP-ALL pathophysiology remains to be established. prior dose finding study (Supplementary Fig. S11A–S11C). Therefore, the contribution of NF-kB subunit expression in No significant weight loss was observed in any of the leukemia cells of patients with BCP-ALL at first relapse was treatment groups (Supplementary Fig. S10B). Also, spleen

www.aacrjournals.org Clin Cancer Res; 19(6) March 15, 2013 1451

Bastian et al.

A NF-κB signaling B Reh (VPA + BTZ) 2 h 4 h 9 h 24 h 100 BTZ VPA BTZ + VPA 100 100 100 100 80 IL1B 80 80 80 80 MYD88 60 60 60 60 60 NFKB1 40 40 40 40 40 20 20 20 20 IL1F9 20 0 0 0 0 p-p65 (p-RelA) mean (p-RelA) p-p65

% of Max 0 1 2 3 4 0 1 2 3 440 1 2 3 0 1 2 3 4 0 IRAK1 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 fluorescence intensity (MFI) cont. 2 h 4 h 9 h 24 h TLR7 p-p65 (p-RelA) MAP2K6 9 h TLR2 p65 (RelA) RelB NGFR ** 2.0 IL1F8 6 ** ** 1.5 TRAF1 TIRAP 4 1.0

RELB B activation κ B activation B RELA κ

NF 2

(x - fold change) fold - (x 0.5 NF NFKB2 change) - fold (x

MAP3K14 0 0.0 TLR3 Control VPA BTZ VPA + BTZ Control VPA BTZ VPA + BTZ CSNK2A2 p50 p52 TNF 2.0 2.0 IL6 TBK1 1.5 1.5 IL1R1 1.0 1.0 TLR6 B activation B activation B κ

IL8 κ

0.5 NF 0.5 (x - fold change) - fold (x NF IL15 change) - fold (x

-3 17 0.0 0.0 Control VPA BTZ VPA + BTZ Control VPA BTZ VPA + BTZ

Reh Nalm6 697 C Control VPA + BTZ p65

D RelA (p65) RelB

9.5 ** 10 * n n o o i 9.0 si 9 res 8.5 p p65

Aex 8

DAPI NA express 8.0 R

mRN 7 Am B l 7.5 l e 60 Re R 40 7.0 6 20 RRelapse on elapse off + cells (% of total) RRelapse on elapse off

p65 0 treatment treatment control VPA + BTZ treatment treatment

Figure 4. Analysis of NF-kB signaling pathway in response to concomitant treatment and in primary BCP-ALL cells from patients at first relapse. A, heatmap depicts differentially expressed genes involved in the NF-kB pathway. B, analysis of NF-kB activation by FACS in Reh and 697 cells with mAbs against phospho-p65 (blue line, control; red line, VPAþBTZ; right: bars, mean fluorescence intensity) at indicated time points. Results from Reh cells are shown. DNA-binding activity of NF-kB subunits after treatments for 9 hours. Measurements from 1 to 2 experiments in each cell line with group medians. Median fold change >1.5 and P < 0.05 was considered statistically significant (Mann–Whitney U test). C, Nalm6 cells were stained with a mAb against p65 and 40,6-diamidino-2-phenylindole (DAPI) after treatments for 12 hours. Arrow indicates nuclear translocation of p65 (magnification, 400). Insert, mean þ proportions of nuclear p65 cells after counting of 40 to 80 nuclei. D, RelA and RelB mRNA expression levels in BCP-ALL cells from patients at first relapse (n ¼ 52) as obtained by microarray analysis. The patient cohort was subdivided according to the time point of relapse, during the frontline chemotherapy (on treatment) or off treatment. Signal intensity (SI) of gene expression in individual patients with group medians. , P < 0.05; , P < 0.01 (Mann–Whitney U test).

weights did not differ significantly between treatment when compared to controls or to the single drug treatments groups (Supplementary Fig. S10C). To study the ALL cell (Fig. 5B; Supplementary Fig. S10D). burden in the whole body, the amount of leukemia cells was Treatment with bortezomib alone had no significant analyzed in different lymphatic tissues (spleen and bone impact on testis and brain involvement (Supplementary marrow) by flow cytometry and in extramedullary nonlym- Fig. S10E and S10F). Leukemia cell infiltration increased phatic tissues (brain and testis) by QRT-PCR. Quantifica- significantly after VPA treatment in testis and slightly tion of the leukemia cells burden in bone marrow and in brain (Supplementary Fig. S10E and S10F). This is in spleen showed a significant reduction of the proportion of accordance with recently published results, showing that þ þ hCD19 hCD10 leukemia cells after combined treatment HDACi treatments augment the migration and metastasis in

1452 Clin Cancer Res; 19(6) March 15, 2013 Clinical Cancer Research

Bortezomib and HDACi in Childhood BCP-ALL

A B i.v. - mouse model C s.c. - mouse model * * 100 MSC-1 600 ** *** *** MSC-2 * 6 * ) * MSC-3 3 80

T-MSC-3 cells

+ 400 60 4

40 hCD19 + 200 2

20 Tumor volume (cm hCD10 dead LC (% of total LC)

0 per 2,000 cells of spl and BM 0 0 Control BTZ VPA BTZ + VPA Reh + + + + Control BTZ VPA BTZ + VPA stroma - - + + BTZ + VPA - + - +

D Control BTZ + VPA ** 30 *

10 (% surface area)

cl. Casp-3 Cleaved caspase-3 CD10 0 DAPI Control BTZ VPA BTZ +VPA

300 *

) * el d l ie ess v

0xf 200 0 ro c /4 ( mi y +

sit 100 n D31 e d C

CD31 0 DAPI Control BTZ VPA BTZ + VPA

Figure 5. Antileukemic activity of combined treatment in leukemia cells (LC)/MSC cocultures and BCP-ALL xenograft mouse models. A, cell death in Reh cells was analyzed by FACS in mono- and cocultures after treatments. MSC were isolated either from BM (n ¼ 3) or from testis (T-MSC, n ¼ 1) of patients with þ þ BCP-ALL (details: Supplementary Fig. S8). Results from independent experiments with group means (Student t test). B, FACS analysis of hCD19 hCD10 LC in whole spleen and BM of the i.v. xenograft mouse model (details: Supplementary Fig. S10). C, tumor volumes in the s.c. xenograft mouse model on the final day of the experiment (details: Supplementary Figs. S11 and S12). D, tumor sections were stained with mAbs either against cleaved caspase-3 and CD10 and DAPI or CD31 and DAPI (magnification, 400), apoptosis was measured as detailed in Supplementary Fig. S12 and the microvessel density (MVD) was quantified by counting CD31þ microvessels within representative 400 magnification fields. Measurements of individual mice with group medians. (Mann–Whitney U test). , P < 0.05; , P < 0.01; , P < 0.01. different tumor models (41). The combined treatment cell burden observed in the i.v. xenograft mouse model and significantly reduced leukemia cell infiltration in testis, to analyze the impact of combined treatment on angiogen- only when compared with VPA (Supplementary Fig. esis and tumor-associated fibroblasts, we used an s.c. xeno- S10E). Most importantly, the combined treatment signifi- graft mouse model bearing Nalm6 cells. The treatment cantly decreased the leukemia cells infiltration in brain, protocol is provided in Supplementary Fig. S11D. Tumor when compared with controls and single VPA treatment volumes were significantly reduced after treatment with (Supplementary Fig. S10F). VPA (P ¼ 0.015) or after combined treatment with VPA/BTZ (P ¼ 0.008) in comparison to untreated controls Combined treatment reduces tumor volumes in a BCP- (Fig. 5C). Furthermore, tumor growth curves showed a ALL s.c. xenograft mouse model through induction of delay of tumor growth after single treatments with borte- apoptosis and changes the stroma microenvironment zomib or VPA, which was more pronounced after combi- The interaction of tumor cells with their microenviron- nation treatment (Supplementary Fig. S12A). Body weight ment plays an important role in mediating the effect of development indicated a significant weight loss in the antineoplastic drugs (29, 40, 42). To assess to which extent bortezomib- and the VPA/BTZ-treated groups compared induction of apoptosis might contribute to the reduced ALL with control, which resolved after completion of treatment

www.aacrjournals.org Clin Cancer Res; 19(6) March 15, 2013 1453

Bastian et al.

(Supplementary Fig. S11E). Analysis of complete blood increase of cytostatic treatment efficacy (Supplementary Fig. counts after 3 days of treatment revealed a thrombocyto- S13C). penia after bortezomib and VPA/BTZ treatment, whereas leukocyte counts and hemoglobin levels remained unaf- Discussion fected (Supplementary Fig. S11F). This is in accordance This study shows that HDACi and the proteasome inhib- with the known clinical risk profile of bortezomib. No itor bortezomib interact in a synergistic manner to induce significant differences in body weight and hematotoxicity apoptosis in BCP-ALL cells and to inhibit the expansion data were observed between bortezomib and combined of leukemia cells in monocultures and cocultures with treatment groups. Overall statistical analysis of events stroma cells. In addition, leukemia disease was significantly (death) indicated no significant differences between treat- reduced upon combined treatment in vivo, which was ment and control groups (Supplementary Table S6), show- associated with a decrease of leukemia cell infiltration in ing an enhancement of the therapeutic efficacy for the extramedullary tissues, reduction of angiogenesis and combined treatment without an increase in toxic effects changes in the supportive stroma microenvironment. Syn- in vivo. ergistic induction of cell death required simultaneous appli- Furthermore, tumor sections were stained for the cation of both drugs, which is well applicable in the clinical expression of cleaved caspase-3 and hCD10 (leukemia setting. cells; Fig. 5D; Supplementary Fig. S12B). After combined Despite the relatively uniform synergistic induction of VPA/BTZ treatment, a significant increase in cleaved cas- apoptosis by HDACi/proteasome inhibitors, the molecular pase-3 amount was observed, whereas application of pathways involved in the response to combined treatment single drugs showed no significant changes (Fig. 5D). To are still not fully understood and vary between different examine a possible effect on angiogenesis, the density of malignancies (9, 10, 13). The synergistic induction of þ CD31 microvessels (MVD) was quantified in tumor apoptosis in our models of BCP-ALL was associated with sections (Fig. 5D) (40). A significant decrease of MVD modulation of cell cycle, death receptor, JUN/MAPK, and was observed after single treatment with VPA, which was PI3K/AKT pathways. These pathways have also been impli- more pronounced after combined treatment (Fig. 5D). cated either in combined or single agent activity of HDACi The analysis of leukemia cell–associated fibroblasts indi- and proteasome inhibitors in other malignancies þ cated a decreased infiltration of ER-TR7 fibroblastoid (9, 13, 16, 43, 44). cells in tumor tissue, as well as changes in reticular The differential regulation of p53 pathway genes such as structures and extracellular matrix after combined treat- BBC3 (PUMA), MDM2, CDKN1A (p21), and nuclear p53 ment when compared with control samples (Supplemen- accumulation indicated that concomitant treatment tary Fig. S12C). induced p53-dependent apoptosis. However, treatment with VPA alone induced CDKN1A gene expression and a The synergistic effect of concomitant HDACi/BTZ G0–G1 phase cell-cycle arrest, whereas levels of p53 and its treatment is maintained or even enhanced in the inhibitor MDM2 remained unchanged. This indicates a presence of chemotherapeutic agents p53-independent CDKN1A regulation after VPA single To establish a rationale for the integration of HDACi/BTZ treatment, which has also been reported (45). These find- combinations into current BCP-ALL treatment protocols, ings are supported by previous reports showing that HDACi and bortezomib were combined with established CDKN1A-induced cell-cycle arrest can result either in senes- ALL chemotherapeutic elements at a moderately effective cence or in apoptosis and thereby influences the therapeutic dose level (Fig. 6; Supplementary Fig. S13A and S13B). efficacy of antineoplastic treatments (45, 46). A possible These combinations were analyzed in 3 BCP-ALL cell lines, explanation for the regulation of p53 and MDM2 only after and the induction of cell death was measured by FACS. concomitant treatment might be found in the ubiquitin/ Combination of VPA/BTZ with the anthracyclines (mitox- proteasome system, which plays a central role in the protein antrone, idarubicin; Fig. 6B), cytarabine (Fig. 6C), or dexa- homeostasis of MDM2, p53, and various other-cell cycle methasone (Fig. 6D) significantly increased the antileuke- regulators (47). Accordingly, after concomitant treatment, mic efficacy of the VPA/BTZ combination (Fig. 6A and D) as we observed complex changes in the gene expression of well as of single chemotherapeutic treatments. The induc- proteasome subunits, which might contribute to the stabi- tion of cell death was more pronounced in combination lization of p53, MDM2, and regulation of NF-kB pathway with anthracyclines, possibly due to the synergistic poten- (30). Moreover, the observed induction of ER stress signal- tial of the corresponding double combinations (Figs. 1B ing is in accordance with previous results, showing its and 6A and B). Combination of VPA/BTZ with vincristine involvement in p53 activation (48) and the synergistic or methotrexate significantly increased the antileukemic induction of apoptosis after HDACi/proteasome inhibitor efficacy of single chemotherapeutic treatments whereas treatment (11, 12). the effect of VPA/BTZ remained unchanged (Fig. 6C). To NF-kB activation has frequently been found to promote further exclude possible adverse effects of VPA/BTZ on the cancer cell survival and resistance to apoptosis (18, 30). efficacy of chemotherapeutic treatments, submaximal effec- Thus, the clinical activity of proteasome inhibitors such as tive concentrations of cytostatic agents were combined bortezomib has, at least in part, been attributed to the with VPA/BTZ, showing no reduction and mostly a slight inhibition of NF-kB activity (30). The differential regulation

1454 Clin Cancer Res; 19(6) March 15, 2013 Clinical Cancer Research

Bortezomib and HDACi in Childhood BCP-ALL

A Reh Nalm6 697 B Mitoxantrone Idarubicin 100 *** 100 * 100 * *** *** * * 80 80 80 otal) t ftotal) f total) o 60 o

%of 60 60 % ls ( ls (

el 40 40 el 40 cells (% dc a 20 ad 20 20 De Dead c De

0 0 0 A Z Z TZ Z P TZ PA B T B V BT ntrol VPA BTZ mol/L) + o +V +BTZ +BTZ + n + C A+B 0 DA PA + BT PA + 10 nmol/L)ITO PA (1 IDA I V V ( VPA VP O M MITO +V + IDA A MIT ITO ID M C Cytarabine Vincristine Methotrexate n.s. n.s. 100 * 100 100 * *** * 80 80 80 total)

60 of 60 60 % ( (%of total) s 40 lls 40 40

20 20 20 Dead ce Dead cells (%of total) Dead cell

0 0 0 Z Z Z Z TZ T T T TZ T PA TZ B B B B V B BTZ mol/L) mol/L) + + + + + n nmol/L) + R+VPA R (2 0 TX PA (50 n PA C 4 VPA + B V V VC VPA + B MTX M V ARAC + VPAARAC + +VPA X( AC VCR T CR M AR RAC+VPA+BTZ V MTX + A

D Dexamethasone 100 * *** 80

20 Dead cells (% of total) Dead cells (%

0 Z PA mol/L) + V

VPA + BTZ DEX DEX + BT

EX + VPA + BTZ DEX (100 n D

Figure 6. The synergistic effect of VPA/BTZ treatment is maintained or increased in combination with established ALL therapy elements. A, proportions of dead cells 48 hours after treatments as determined by FACS analysis of PI-stained cells. B to D, proportions of dead cells 48 hours after treatment either with 1 of 6 established ALL therapeutics at moderately effective concentrations, or double or triple combinations with VPA and BTZ. Results obtained after VPA/BTZ treatment, as shown in A, are presented in B and C as gray data points. Combined results from cell lines with similar sensitivity to chemotherapeutic agents are shown. Analyses from cell lines with different sensitivities and Annexin-V/PI FACS results after 24-hour single and triple combination treatments are provided in Supplementary Fig. S13A and B. Measurements from 1 to 2 independent experiments in each cell line with group medians. , P < 0.05; , P < 0.01; , P < 0.01 (Mann–Whitney U test). of NF-kB target genes in our model indicated that this gistic induction of apoptosis was only observed after con- pathway is also involved in the treatment response of comitant treatment. Our ﬁndings are consistent with recent- BCP-ALL cells. However, we observed an activation of ly published studies showing the tumor-suppressive role of particular NF-kB subunits (e.g., RelA) both after bortezo- the NF-kB signaling pathway, especially during chemother- mib single and concomitant treatment, although the dif- apy-induced senescence that contributes to the outcome of ferential regulation of apoptosis-related genes and syner- cancer therapy (35–37). In our study, patients with BCP-

www.aacrjournals.org Clin Cancer Res; 19(6) March 15, 2013 1455

Bastian et al.

ALL who suffered a relapse during frontline treatment, Disclosure of Potential Conflicts of Interest which is generally associated with a poor outcome, showed No potential conflicts of interest were disclosed. lower expression of NF-kB subunits in leukemia cells. Nevertheless, we do not want to exclude a tumor-promoting Authors' Contributions Conception and design: L. Bastian, G. Henze, K. Seeger, S. Shalapour role of NF-kB in BCP-ALL, depending on the context of its Development of methodology: L. Bastian, I. Fichtner, S. Shalapour activation (49). Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L. Bastian, M. Pfau, I. Fichtner, C. Eckert, G. Only minor responses have previously been observed Henze, A. von Stackelberg, K. Seeger, S. Shalapour after single treatments with either bortezomib (14) or Analysis and interpretation of data (e.g., statistical analysis, biosta- HDACi (5, 6) in phase I studies, revealing thereby a good tistics, computational analysis): L. Bastian, J. Hof, M. Pfau, I. Fichtner, C. Eckert, J. Prada, S. Shalapour tolerability. Bortezomib has recently shown favorable Writing, review, and/or revision of the manuscript: L. Bastian, J. Hof, I. results in the treatment of children with relapsed ALL Fichtner, G. Henze, J. Prada, A. von Stackelberg, K. Seeger, S. Shalapour when combined with conventional chemotherapy (50). Administrative, technical, or material support (i.e., reporting or orga- nizing data, constructing databases): C. Eckert, G. Henze, S. Shalapour HDACi are promising candidates for combination with Study supervision: J. Prada, A. von Stackelberg, K. Seeger, S. Shalapour conventional chemotherapy, as they have a moderate and completely different toxicity profile. Our analyses Acknowledgments revealed that the synergistic antileukemic effect of BTZ/VPA The authors thank the valuable discussions and support from Hagen Graf Einsiedel, Renate Kirschner-Schwabe, Thomas Kammertons,€ Kathrin Haupt- combination indeed was maintained or even increased in mann, Nina Weichert, Maya Schreiber, and Marzena Sosnowski. the presence of particular chemotherapeutic agents. How- ever, these need to be carefully selected to avoid antagonistic Grant Support effects of single agents such as methotrexate or PI3K/AKT This work was supported, in part, by a grant of the Berliner Krebsge- sellschaft e.v. (BAFF200812) to L. Bastian, K. Seeger, and S. Shalapour, the inhibitors. The presented findings establish a biologic basis Federal Ministry for Education and Research in the National Genome for the clinical evaluation of concomitant HDACi/BTZ Research Network (NGFN2, 01GS0870) to J. Hof, and a grant of the € treatment not only due to the synergistic potential of this Deutsche Jose Carreras Leukamie-Stiftung e.V. (SP10/01). The costs of publication of this article were defrayed in part by the combination but also because this enhanced therapeutic payment of page charges. This article must therefore be hereby marked efficacy was achieved without an increase of toxic effects. In advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate summary, our results provide new perspectives for consid- this fact. ering combinations of bortezomib and HDACi as thera- Received May 12, 2012; revised December 22, 2012; accepted January 21, peutic alternative in childhood BCP-ALL. 2013; published OnlineFirst January 28, 2013.

References 1. Tallen G, Ratei R, Mann G, Kaspers G, Niggli F, Karachunsky A, et al. in childhood acute lymphoblastic leukemia cells. Leukemia 2011;25: Long-term outcome in children with relapsed acute lymphoblastic 359–61. leukemia after time-point and site-of-relapse stratification and inten- 9. Bhalla S, Balasubramanian S, David K, Sirisawad M, Buggy J, Mauro L, sified short-course multidrug chemotherapy: results of trial ALL-REZ et al. PCI-24781 induces caspase and reactive oxygen species- BFM 90. J Clin Oncol 2010;28:2339–47. dependent apoptosis through NF-kappaB mechanisms and is syner- 2. Bhatla T, Wang J, Morrison DJ, Raetz EA, Burke MJ, Brown P, et al. gistic with bortezomib in lymphoma cells. Clin Cancer Res 2009;15: Epigenetic reprogramming reverses the relapse-specific gene expres- 3354–65. sion signature and restores chemosensitivity in childhood B-lympho- 10. DaiY,ChenS,WangL,PeiXY,KramerLB,DentP,etal.Bortezomib blastic leukemia. Blood 2012;119:5201–10. interacts synergistically with belinostat in human acute myeloid 3. Pui CH, Jeha S. New therapeutic strategies for the treatment of leukaemia and acute lymphoblastic leukaemia cells in association acute lymphoblastic leukaemia. Nat Rev Drug Discov 2007;6: with perturbation in NF-kappaB and Bim. Br J Haematol 2011;153: 149–65. 222–35. 4. Einsiedel HG, Kawan L, Eckert C, Witt O, Fichtner I, Henze G, et al. 11. Hideshima T, Bradner JE, Wong J, Chauhan D, Richardson P, Schrei- Histone deacetylase inhibitors have antitumor activity in two NOD/ ber SL, et al. Small-molecule inhibition of proteasome and aggresome SCID mouse models of B-cell precursor childhood acute lymphoblas- function induces synergistic antitumor activity in multiple myeloma. tic leukemia. Leukemia 2006;20:1435–6. Proc Natl Acad Sci U S A 2005;102:8567–72. 5. Fouladi M, Park JR, Stewart CF, Gilbertson RJ, Schaiquevich P, Sun J, 12. Nawrocki ST, Carew JS, Pino MS, Highshaw RA, Andtbacka RH, et al. Pediatric phase I trial and pharmacokinetic study of vorinostat: a Dunner K Jr, et al. Aggresome disruption: a novel strategy to enhance Children's Oncology Group phase I consortium report. J Clin Oncol bortezomib-induced apoptosis in pancreatic cancer cells. Cancer Res 2010;28:3623–9. 2006;66:3773–81. 6. Su JM, Li XN, Thompson P, Ou CN, Ingle AM, Russell H, et al. Phase 1 13. Zhang QL, Wang L, Zhang YW, Jiang XX, Yang F, Wu WL, et al. The study of valproic acid in pediatric patients with refractory solid or CNS proteasome inhibitor bortezomib interacts synergistically with the tumors: a children's oncology group report. Clin Cancer Res 2011;17: histone deacetylase inhibitor suberoylanilide hydroxamic acid to 589–97. induce T-leukemia/lymphoma cells apoptosis. Leukemia 2009;23: 7. Kano Y, Akutsu M, Tsunoda S, Izumi T, Kobayashi H, Mano H, et al. 1507–14. Cytotoxic effects of histone deacetylase inhibitor FK228 (depsipep- 14. Horton TM, Pati D, Plon SE, Thompson PA, Bomgaars LR, Adamson tide, formally named FR901228) in combination with conventional anti- PC, et al. A phase 1 study of the proteasome inhibitor bortezomib in leukemia/lymphoma agents against human leukemia/lymphoma cell pediatric patients with refractory leukemia: a Children's Oncology lines. Invest New Drugs 2007;25:31–40. Group study. Clin Cancer Res 2007;13:1516–22. 8. Bastian L, Einsiedel HG, Henze G, Seeger K, Shalapour S. The 15. Horton TM, Gannavarapu A, Blaney SM, D'Argenio DZ, Plon SE, Berg sequence of application of methotrexate and histone deacetylase SL. Bortezomib interactions with chemotherapy agents in acute leu- inhibitors determines either a synergistic or an antagonistic response kemia in vitro. Cancer Chemother Pharmacol 2006;58:13–23.

1456 Clin Cancer Res; 19(6) March 15, 2013 Clinical Cancer Research

Bortezomib and HDACi in Childhood BCP-ALL

16. Pei XY, Dai Y, Grant S. Synergistic induction of oxidative injury and 33. Dolcet X, Llobet D, Encinas M, Pallares J, Cabero A, Schoenenberger apoptosis in human multiple myeloma cells by the proteasome inhib- JA, et al. Proteasome inhibitors induce death but activate NF-kappaB itor bortezomib and histone deacetylase inhibitors. Clin Cancer Res on endometrial carcinoma cell lines and primary culture explants. J Biol 2004;10:3839–52. Chem 2006;281:22118–30. 17. Gerondakis S, Siebenlist U. Roles of the NF-kappaB pathway in 34. Wang CY, Cusack JC Jr, Liu R, Baldwin AS Jr. Control of inducible lymphocyte development and function. Cold Spring Harb Perspect chemoresistance: enhanced anti-tumor therapy through increased Biol 2010;2:a000182. apoptosis by inhibition of NF-kappaB. Nat Med 1999;5:412–7. 18. Staudt LM. Oncogenic activation of NF-kappaB. Cold Spring Harb 35. Chien Y, Scuoppo C, Wang X, Fang X, Balgley B, Bolden JE, et al. Perspect Biol 2010;2:a000109. Control of the senescence-associated secretory phenotype by NF- 19. Derudder E, Cadera EJ, Vahl JC, Wang J, Fox CJ, Zha S, et al. kappaB promotes senescence and enhances chemosensitivity. Development of immunoglobulin lambda-chain-positive B cells, but Genes Dev 2011;25:2125–36. not editing of immunoglobulin kappa-chain, depends on NF-kappaB 36. Jing H, Kase J, Dorr JR, Milanovic M, Lenze D, Grau M, et al. Opposing signals. Nat Immunol 2009;10:647–54. roles of NF-kappaB in anti-cancer treatment outcome unveiled by 20. Min DJ, Moskowitz NP, Brownstein C, Lee H, Horton TM, Carroll WL. cross-species investigations. Genes Dev 2011;25:2137–46. Diverse pathways mediate chemotherapy-induced cell death in acute 37. Riz I, Zweier-Renn LA, Toma I, Hawley TS, Hawley RG. Apoptotic role lymphoblastic leukemia cell lines. Apoptosis 2006;11:1977–86. of IKK in T-ALL therapeutic response. Mol Cancer Res 2011;9:979–84. 21. Wuchter C, Krappmann D, Cai Z, Ruppert V, Scheidereit C, Dorken B, 38. Kordes U, Krappmann D, Heissmeyer V, Ludwig WD, Scheidereit C. et al. In vitro susceptibility to TRAIL-induced apoptosis of acute Transcription factor NF-kappaB is constitutively activated in acute leukemia cells in the context of TRAIL receptor gene expression and lymphoblastic leukemia cells. Leukemia 2000;14:399–402. constitutive NF-kappa B activity. Leukemia 2001;15:921–8. 39. Kirschner-Schwabe R, Lottaz C, Todling J, Rhein P, Karawajew L, 22. Weston VJ, Austen B, Wei W, Marston E, Alvi A, Lawson S, et al. Eckert C, et al. Expression of late cell cycle genes and an increased Apoptotic resistance to ionizing radiation in pediatric B-precursor proliferative capacity characterize very early relapse of childhood acute lymphoblastic leukemia frequently involves increased NF-kap- acute lymphoblastic leukemia. Clin Cancer Res 2006;12:4553–61. paB survival pathway signaling. Blood 2004;104:1465–73. 40. Shalapour S, Zelmer A, Pfau M, Moderegger E, Costa-Blechschmidt C, 23. Rubin EH, Agrawal NG, Friedman EJ, Scott P, Mazina KE, Sun L, et al. A van Landeghem FK, et al. The thalidomide analogue, CC-4047, study to determine the effects of food and multiple dosing on the induces apoptosis signaling and growth arrest in childhood acute pharmacokinetics of vorinostat given orally to patients with advanced lymphoblastic leukemia cells in vitro and in vivo. Clin Cancer Res cancer. Clin Cancer Res 2006;12:7039–45. 2006;12:5526–32. 24. Fichtner I, Paal K, Borgmann A, Badiali L, Wurm R, Henze G. Chemo- 41. Lin KT, Wang YW, Chen CT, Ho CM, Su WH, Jou YS. HDAC inhibitors and radiation sensitivity of xenografted acute lymphoblastic leuke- augmented cell migration and metastasis through induction of PKCs mias–correlation to the expression of multidrug resistance proteins. leading to identification of low toxicity modalities for combination Anticancer Res 2003;23:2657–64. cancer therapy. Clin Cancer Res 2012;18:4691–701. 25. Workman P, Aboagye EO, Balkwill F, Balmain A, Bruder G, Chaplin DJ, 42. Iwamoto S, Mihara K, Downing JR, Pui CH, Campana D. Mesenchymal et al. Guidelines for the welfare and use of animals in cancer research. cells regulate the response of acute lymphoblastic leukemia cells to Br J Cancer 2010;102:1555–77. asparaginase. J Clin Invest 2007;117:1049–57. 26. Webb JL. Effect of more than one inhibitor. Enzymes and metabolic 43. Befani CD, Vlachostergios PJ, Hatzidaki E, Patrikidou A, Bonanou S, inhibitors. New York: Academic Press; 1963. p. 66–79 & 487–512. Simos G, et al. Bortezomib represses HIF-1alpha protein expression 27. Hof J, Krentz S, van Schewick C, Korner G, Shalapour S, Rhein P, et al. and nuclear accumulation by inhibiting both PI3K/Akt/TOR and MAPK Mutations and deletions of the TP53 gene predict nonresponse to pathways in prostate cancer cells. J Mol Med (Berl) 2011;90:45–54. treatment and poor outcome in first relapse of childhood acute lym- 44. Kodani M, Igishi T, Matsumoto S, Chikumi H, Shigeoka Y, Nakanishi H, phoblastic leukemia. J Clin Oncol 2011;29:3185–93. et al. Suppression of phosphatidylinositol 3-kinase/Akt signaling path- 28. Matsunaga T, Takemoto N, Sato T, Takimoto R, Tanaka I, Fujimi A, way is a determinant of the sensitivity to a novel histone deacetylase et al. Interaction between leukemic-cell VLA-4 and stromal fibronectin inhibitor, FK228, in lung adenocarcinoma cells. Oncol Rep 2005;13: is a decisive factor for minimal residual disease of acute myelogenous 477–83. leukemia. Nat Med 2003;9:1158–65. 45. Ocker M, Schneider-Stock R. Histone deacetylase inhibitors: signal- 29. Shalapour S, Hof J, Kirschner-Schwabe R, Bastian L, Eckert C, Prada ling towards p21cip1/waf1. Int J Biochem Cell Biol 2007;39:1367–74. J, et al. High VLA-4 expression is associated with adverse outcome 46. Abbas T, Dutta A. p21 in cancer: intricate networks and multiple and distinct gene expression changes in childhood B-cell precursor activities. Nat Rev Cancer 2009;9:400–14. acute lymphoblastic leukemia at first relapse. Haematologica 2011; 47. Wickliffe K, Williamson A, Jin L, Rape M. The multiple layers of 96:1627–35. ubiquitin-dependent cell cycle control. Chem Rev 2009;109:1537–48. 30. Baud V, Karin M. Is NF-kappaB a good target for cancer therapy? 48. Li J, Lee B, Lee AS. Endoplasmic reticulum stress-induced apoptosis: Hopes and pitfalls. Nat Rev Drug Discov 2009;8:33–40. multiple pathways and activation of p53-up-regulated modulator of 31. Dai Y, Rahmani M, Dent P, Grant S. Blockade of histone deacetylase apoptosis (PUMA) and NOXA by p53. J Biol Chem 2006;281:7260–70. inhibitor-induced RelA/p65 acetylation and NF-kappaB activation 49. Perkins ND. The diverse and complex roles of NF-kappaB subunits in potentiates apoptosis in leukemia cells through a process mediated cancer. Nat Rev Cancer 2012;12:121–32. by oxidative damage, XIAP downregulation, and c-Jun N-terminal 50. Messinger YH, Gaynon PS, Sposto R, van der Giessen J, Eckroth E, kinase 1 activation. Mol Cell Biol 2005;25:5429–44. Malvar J, et al. Bortezomib with chemotherapy is highly active in 32. Hideshima T, Ikeda H, Chauhan D, Okawa Y, Raje N, Podar K, et al. advanced B-precursor acute lymphoblastic leukemia: Therapeutic Bortezomib induces canonical nuclear factor-kappaB activation in Advances in Childhood Leukemia & Lymphoma (TACL) Study. Blood multiple myeloma cells. Blood 2009;114:1046–52. 2012;120:285–90.

www.aacrjournals.org Clin Cancer Res; 19(6) March 15, 2013 1457

Synergistic Activity of Bortezomib and HDACi in Preclinical Models of B-cell Precursor Acute Lymphoblastic Leukemia via Modulation of p53, PI3K/AKT, and NF-κB

Lorenz Bastian, Jana Hof, Madlen Pfau, et al.

Clin Cancer Res 2013;19:1445-1457. Published OnlineFirst January 28, 2013.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-12-1511

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2013/01/28/1078-0432.CCR-12-1511.DC1

Cited articles This article cites 49 articles, 26 of which you can access for free at: http://clincancerres.aacrjournals.org/content/19/6/1445.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/19/6/1445.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/19/6/1445. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.