The Novel IKK2 Inhibitor LY2409881 Potently Synergizes with Histone

Published OnlineFirst October 29, 2014; DOI: 10.1158/1078-0432.CCR-14-0384

Cancer Therapy: Preclinical Clinical Cancer Research The Novel IKK2 Inhibitor LY2409881 Potently Synergizes with Histone Deacetylase Inhibitors in Preclinical Models of Lymphoma through the Downregulation of NF-kB Changchun Deng1,2, Mark Lipstein1, Richard Rodriguez1, Xavier O. Jirau Serrano1, Christine McIntosh1,3, Wei-Yann Tsai4, Andrew S. Wasmuth2, Susan Jaken5, and Owen A. O'Connor1,2

Abstract

Purpose: To evaluate the pharmacologic activity of a novel tion and apoptosis in lymphoma cells. In models of diffuse large inhibitor of IkB kinase b (IKK2), LY2409881, in preclinical B-cell lymphoma (DLBCL), the cytotoxicity of LY2409881 corre- models of B- and T-cell lymphoma, as a single agent and in lated with the overall activation status of NF-kB, but not simply in combination with histone deacetylase (HDAC) inhibitors. a pattern predicted by the cell-of-origin classification of these cell Experimental Design: The in vitro activity of LY2409881 was lines. LY2409881 was safe to mice at three dose levels, 50, 100, determined using an ATP-based growth inhibition assay and flow and 200 mg/kg, all of which caused significant inhibition of cytometric assay of apoptosis in lymphoma cell lines. The in vivo tumor growth. LY2409881 suppressed the activity of the NF-kB activity of LY2409881 was determined using SCID-beige xeno- subunit p65 in lymphoma cells treated by the HDAC inhibitor graft mouse model. The mechanism of action was determined romidepsin, underlying a potential mechanism of the marked using immunoblotting, immuofluorescence, and electrophoretic synergy observed of these two drugs. mobility shift assay. Synergy of LY2409881 with other drugs Conclusion: Collectively, these data strongly suggest that active in lymphoma was determined by calculating relative risk targeting the NF-kB pathway in combination with romidepsin ratio (RRR) and combination index (CI). could represent a novel and potent regimen for the treatment of Results: LY2409881 inhibited constitutively activated NF-kB, B- and T-cell lymphoma. Clin Cancer Res; 21(1); 134–45. 2014 and caused concentration- and time-dependent growth inhibi- AACR.

Introduction of NF-kB activation, stimulation by a variety of signals such as B- cell receptor engagement leads to activation of the IkB kinase NFKB is a family of inducible transcription factors that com- b (IKK2). IKK2 phosphorylates IkB, which renders it a substrate prises five different proteins, RelA (p65), cRel, RelB, NFKB1 (p50/ for polyubiquitination and subsequent degradation by the 26S p105), and NFKB2 (p52/p100; refs. 1, 2). In their inactive state, proteasome. Degradation of IkB releases the p50-RelA dimer, the proteins are sequestered in the cytoplasm through their which is then translocated into the nucleus, where it regulates binding to the inhibitory protein IkB. In the classical pathway expression of genes important in inflammation, proliferation, survival, and DNA-damage repair. Among the target genes regulated by NFKB are IkB, cyclin D1, Bcl-xL, survivin, BCL2, and ATM. 1Center for Lymphoid Malignancies, Columbia University Medical Cen- In the alternative pathway of NF-kB activation, upstream stimuli ter, New York, New York. 2Division of Experimental Therapeutics, – 3 such as TNF superfamily cytokine B cell activating factor (BAFF) Columbia University Medical Center, New York, New York. Depart- k – ment of Molecular Biology and Genetics, Cornell University, Ithaca, activate NF- B inducing kinase (NIK). NIK, in turn, phosphor- New York. 4Department of Statistics, Mailman School of Public Health, ylates IkB kinase a (IKK1), leading to phosphorylation of p100. 5 Columbia University Medical Center, New York, New York. Division of Subsequent polyubiquitination and degradation of p100 allows Cancer Research, Lilly Research Laboratories, Indianapolis, Indiana. translocation of the RelB–p52 dimer into the nucleus, where it Note: Supplementary data for this article are available at Clinical Cancer activates the expression of prosurvival genes. Research Online (http://clincancerres.aacrjournals.org/). Constitutive activation of NFKB is frequently observed in a Current address for Richard Rodriguez: School of Pharmacy, University of number of lymphoma subtypes, including cutaneous T-cell lym- Washington, Seattle, Washington. phoma (CTCL; refs. 3, 4), mantle cell lymphoma (MCL; ref. 5), Corresponding Author: Changchun Deng, Center for Lymphoid Malignancies, diffuse large B-cell lymphoma (DLBCL; refs. 6–8), and adult T cell Columbia University Medical Center, 51 West 51st Street, 2nd Floor, New York, leukemia lymphoma (ATLL; ref. 9). The pathologic significance of NY 10019. Phone: 212-326-5720; Fax: 212-326-5725; E-mail: NF-kB is perhaps best established in DLBCL, as the activated B cell [email protected] (ABC) subtype of DLBCL is characterized by constitutive activation doi: 10.1158/1078-0432.CCR-14-0384 of NF-kB and a worse prognosis compared with those DLBCL 2014 American Association for Cancer Research. derived from the germinal center B cell (GCB; refs. 10, 11).

134 Clin Cancer Res; 21(1) January 1, 2015

IKK2 Inhibition Potently Synergizes with HDAC Inhibitors

inhibitors are now approved for the treatment of T-cell lympho- Translational Relevance ma. On a whole, however, this class of drugs produces a response Targeting histone deacetylases (HDAC) has become a vali- rate of only 25% to 30% across a diverse group of mature, albeit dated therapeutic strategy for the treatment of select subtypes of chemotherapy-resistant T-cell lymphoma. The HDAC inhibitors lymphoma, though the rationale remains strong across the are recognized as highly pleiotropic drugs that obviously in diversity of subtypes. In the peripheral T-cell lymphomas balance, effectively kill T-cell lymphoma. In reality, the pleiotro- (PTCL), HDAC inhibitors on the whole produce a response pic effects of these drugs have been long recognized to potentially rate of only 25% to 30%, albeit in heavily treated patient include a host of "unfavorable effects" that might well explain the populations, and at present are only approved for the treatment restricted range of activity observed across the known HDAC of relapsed or refractory PTCL and cutaneous T-cell lymphoma. inhibitors in diverse subtypes of lymphoma. We hypothesize that The HDAC inhibitors are recognized as highly pleiotropic drugs. if mechanistic strategies can be found to mitigate these unfavor- Although in balance they have therapeutic properties, they also able effects of HDAC inhibitors, then the efficacy of HDAC can produce a host of potentially "unfavorable effects" on inhibitors could be potently enhanced. For example, recent stud- growth and survival pathways that might well explain their ies have reported that acetylation of p65/RelA is promoted by restricted range of activity, and somewhat modest response rate HDAC inhibitors, including suberoylanilide hydroxamic acid in the clinical setting. The hypothesis of this work is based on (SAHA) and MS-275, resulting in increased transcriptional activ- understanding some of the undesirable effects of this important ity of NF-kB (17). If the acetylation of RelA is an important class of drugs, and using that information to develop rational mechanism to limit the anticancer activity of HDAC inhibitors, drug–drug combinations to mitigate the unfavorable effects, in then combining HDAC inhibitors and NF-kB inhibitors will be order to enhance their therapeutic effect. Our results suggest synergistic and may become a promising therapeutic strategy. HDAC inhibitors activate some elements of the NF-kBpathway, Because of the central role of IKK2 in the activation of NF-kB, which likely contributes to some of the undesirable effects of IKK2 has taken the center stage as a drug target for cancer and this class of drugs. Simultaneous targeting of the NF-kBpath- inflammatory disorders over the past two decades. Two of the ways with IkB kinase b (IKK2) inhibitors was found to be highly well-published IKK2 inhibitors in preclinical models include synergistic in models of diffuse large B-cell lymphoma MLN120B (18–21) and AS602868 (4, 22), both of which have (DLBCL). The mechanistic basis of the synergy was attributed been shown to possess moderate activity as single agents in to the mitigation of the unfavorable effects of HDAC inhibitors models of multiple myeloma and CTCL, respectively. However, on the NF-kB pathway. We propose that by understanding the no IKK2 inhibitors have reached clinical development, primarily unfavorable therapeutic features of pleiotropic drugs, rational because of the concern over hepatic toxicity seen in IKK2 knockout and synergistic combinations can be identified that can marked- mice. Novel IKK2 inhibitors that are structurally different from ly enhance the efficacy of HDAC inhibitor–based therapy across previously developed inhibitors remain an important area of drug most forms of cancer. development, and may be particularly effective when combined with HDAC inhibitors and other chemotherapeutic drugs. Our results demonstrate that a novel IKK2 inhibitor, LY2409881, inhibited the growth and survival of lymphoma in Moreover, the ABC subtype of DLBCL appears to be addicted to the vitro and in vivo. LY2409881 induced higher levels of cytotoxicity constitutively activated NF-kB signals in a number of preclinical in lymphoma models known to have constitutively activated NF- studies. Indirectly, the importance of NF-kB is also supported by a kB signals. LY2409881 was well tolerated in vivo. Furthermore, recent clinical study of a Bruton tyrosine kinase (BTK) inhibitor, LY2409881 potently synergized with antilymphoma drugs. In ibrutinib (Imbruvica; refs. 12, 13). BTK is a kinase that mediates B- particular, LY2409881 and the HDAC inhibitor romidepsin were cell receptor signaling, which also leads to activation of NF-kB. As a highly synergistic in every cell line model of T- and B-cell lym- single agent, ibrutinib produced rare partial response in GCB phoma examined. The synergy of LY2409881 and romidepsin lymphoma, and an impressive response rate of 40% in the ABC may derive from the antagonistic effect of LY2409881 on romi- lymphoma, including 8% of complete response (12, 13). These depsin-mediated activation of NF-kB. These data provide new results demonstrate that inhibition of NF-kBisanimportant insights into this synergistic interaction and suggest a therapeutic strategy to treat lymphoma, particularly those lymphoma types strategy for the treatment of lymphoma that is associated with where constitutively activated NF-kB plays critical pathologic roles. constitutively activated NF-kB. Another common pathologic feature of DLBCL involves deregulation of the acetylation and deacetylation state of critical Materials and Methods proteins. For example, BCL6 is a transcription repressor involved Materials in lymphomagenesis (14), and acetylation of BCL6 at the KKYK LY2409881 was obtained from Lilly Research Laboratories. – (376 379) motif inhibits its function as transcription repressor Merck IKK2 inhibitor VIII was ordered from Merck. Bay 11-7082 and oncoprotein (15). The balance of acetylation and deacetyla- was ordered from Sigma. Romidepsin, belinostat, and vorinostat tion is carefully orchestrated through the activity of two opposing were ordered from Selleck Chemicals. All chemicals were dissolved classes of enzymes, the histone/protein lysine acetyltransferases in dimethyl sulfoxide (DMSO) for in vitro studies. DMSO in (HAT) and histone deacetylases (HDAC). Inactivating mutations untreated control samples matched its concentrations in the treated of two of the HAT family genes, CBP and p300, are found in 40% samples. DLBCL patient samples (16). These mutations prevent BCL6 acetylation, leading to aberrantly activated BCL6 (14). From the Cell lines therapeutic perspective, lymphomas represent a disease entity OCI-Ly1, OCI-Ly7, and Su-DHL4 are GCB DLBCL cell lines; where there is a strong rationale to target HDAC, as three HDAC OCI-Ly3, OCI-Ly10, HBL1, and Su-DHL2 are ABC DLBCL lines.

www.aacrjournals.org Clin Cancer Res; 21(1) January 1, 2015 135

Deng et al.

These cell lines were grown in Iscove Modified Dulbecco Medium to the manufacturer's manual. The NF-kB consensus DNA was with 10% FCS. HH and H9 are CTCL cell lines. MT1 and TL are purchased from Promega, and labeled using the Biotin 30 End ATLL cell lines that do not express Tax; MT2 and C5MJ are HTLV1- DNA Labeling Kit (Thermo Scientific). The nuclear extracts from transformed cell lines that express Tax. These cell lines were grown different treatment groups were incubated with the biotin-labeled in RPMI medium with 10% FCS. Fresh medium was added every 2 NF-kB DNA, and resolved on a native gel, rapidly (30 minutes) to 3 days, and the cells were kept at a cell concentration of 0.1 to transferred to a positive nylon membrane, UV cross-linked, and 1 106/mL. probed with streptavidin–HRP conjugate using the LightShift Chemiluminescent EMSA Kit (Thermo Scientific). Competition Cytotoxicity assays reactions containing a 200-fold molar excess of unlabeled duplex Cytotoxicity was evaluated using the CellTiter-Glo Reagent were performed to confirm the specificity of the protein–DNA (Promega) according to the manufacturer's manual, as reported interactions. previously (23). Experiments were carried out in 96-well plates, with each treatment in triplicate. Samples were taken at typically 24, In vivo tumor model 48, and 72 hours after treatment. Cytotoxicity was expressed by the Mouse experiments were carried out in accordance with the decreasing percentage of live cells in each treatment relative to the principles of laboratory animal care under an Institutional Animal untreated control from the same experiment, as a function of time. Care and Use Committee–approved protocol. Five- to 7-week-old IC50 (half the maximal inhibitory concentration) for each cell line SCID beige mice (Taconic) were injected with 107 Ly10 cells was calculated using the CalcuSyn Version 2.0 software (Biosoft). mixed in Matrigel in the posterior flank subcutaneously. When the tumors approached 150 mm3, the mice were divided into four Flow cytometry groups of 8 mice: (i) control group, which received 5% dextrose in To study apoptosis, Yo-Pro-1 and propidium iodide (Vybrant water; (ii) LY2409881 at 50 mg/kg in D5W; (iii) LY2409881 at apoptosis assay kit #4; Invitrogen) were used, as previously 100 mg/kg in D5W; and (iv) LY2409881 at 200 mg/kg in D5W. described (23). A minimum of 1 104 events were acquired The drug or D5W was administered intraperitoneally on day 1 and from each sample. The fluorescence signals acquired by a FACS- 4 of every week for 4 weeks. The data were expressed as average Calibur System were resolved by detection in the conventional tumor volume (mm3) per group as a function of time. Tumor FL1 and FL3 channels. Cells were considered early apoptotic if Yo- volume was calculated using the formula: (4/3)pr3, where r ¼ Pro-1–positive but PI-negative, late apoptotic if Yo-Pro-1- and PI- (length þ width)/4. Care of mice was done according to institu- positive, and necrotic if only PI-positive. Alternatively, dead cells tional guidelines, as previously described (23). were detected by flow cytometry using the Alexa Fluor 488 annexin V/Dead Cell Apoptosis Kit from Invitrogen. Statistical analysis Western blotting Relative risk ratio (RRR) was calculated to measure the degree of Western blotting was performed according to standard proto- drug–drug synergism, and was previously shown to correlate well cols, using the chemiluminescence detection system from Thermo with combination index (CI; refs. 23, 24). RRR is based on Scientific. The following primary antibodies from Cell Signaling calculating the ratio between the actual value and expected value Technology were used: anti-p105/p50, anti-p100/p52, antipho- (EV) of percentage of surviving cells after treatment. EV is calcu- sphorylated IkB, anti-IkB, anti-p65, c-Rel, anti-Ku80, anti-actin, lated by formula: EV ¼ NA NB/100, where NA represents the anti-HDAC1, anti-HDAC2, anti-HDAC3 anti-HDAC4, anti- percentage of viable cells in the sample treated with drug A and NB HDAC6, and anti-HDAC8 (Sigma). Antiacetylated p65 was represents the percentage of viable cells in the sample treated with obtained from Abcam. Goat anti-rabbit or anti-mouse secondary drug B. RRR values below 1 represent synergistic effect of the two antibodies (Santa Cruz Biotechnology) were used. drugs, values equal to 1 indicate additive effect of the drugs; and values above 1 represent antagonistic effect of the two drugs. fl Immuno uorescence Calculation of the CI is based on the Chou and Talalay method, fl Immuno uorescence was performed as previously described using the free software at www.combosyn.com. (24). After incubation with LY2409881for 4 hours, the samples For evaluation of tumor growth in vivo, the generalized esti- fi were placed on the slides using the Cytospin. After xation in 10% mating equation (GEE) was used to assess the difference in the formalin and 100% methanol, the slides were transferred to the tumor volume among the treatment groups. It was assumed that blocking buffer (10% nonfat dry milk) and incubated with primary the tumor volume follows the model Volumeij ¼ B0 þ B1 timej þ B2 antibody in the humidity chamber overnight. The slides were then treatmenti þ B3 timej x treatmenti þ eij for the ith mouse at time timej. incubated with fluorochrome-conjugated secondary antibody for 0 The model assumes that the tumor volume is linear in time and 45 minutes and mounted with 4 ,6-diamidino-2-phenylindole allows different intercepts and slopes for different treatment (DAPI) using the VECTASHIELD Mounting Media from Vector groups. The error term eij is uncorrelated between different mice Labs.Anti-p105/p50primaryantibody(CellSignalingTechnology) but correlated between same individual mouse at different times and Cy3-conjugated donkey anti-rabbit IgG (Jackson ImmunoR- to account for within-mouse correlation. esearch Laboratories) secondary antibody were used. The images fl were collected using Nikon Eclipse TE 2000-E inverted epi uores- Results cent microscope and a Nikon Photometrics Coolsnap HQ2 camera. The images were analyzed using NIS-Elements AR 3.2 software. LY2409881 is a novel IKK2 inhibitor that inhibits TNFa- induced activation of NF-kB Electrophoretic mobility shift assay LY2409881 is the trihydrochloride salt of a novel pyrimidinyl Nuclear extract was made using the NE-PER Nuclear and benzothiophene compound (Fig. 1A). Using kinase profiling in Cytoplasmic Extraction Reagents (Thermo Scientific) according more than 300 kinases, LY2409881 was highly selective for IKK2.

136 Clin Cancer Res; 21(1) January 1, 2015 Clinical Cancer Research

IKK2 Inhibition Potently Synergizes with HDAC Inhibitors

AB125

100

% of Survival 25 LY LY + TNF 0 0.01 0.1 1 10 Compound, µmol/L C E Cyto Nuc Cyto Nuc Cyto Nuc p105

p50

c-Rel

RelA

D Actin Anti-p50 DAPI Merge HDAC1 LY10

Neg p105 LY1

p50

c-Rel

LY RelA

Actin

HDAC1

Control DMSO LY: 2 mmol/L LY: 10 mmol/L

Figure 1. Novel IKK2 inhibitor, LY2409881, blocked the activation of NF-kB in response to TNFa. A, structure of LY2409881. B, LY2409881 (LY) and TNFa (TNF) in combination demonstrated marked cytotoxicity that was not present in either LY2409881 or TNFa treatment (not shown) in the ovarian cancer cell line SKOV3. The cells were treated by TNFa at 10 ng/mL, LY2409881 at concentrations ranging from 0.02 to 10 mmol/L, combination of TNFa and LY2409881, or negative control for 72 hours. Percentage of survival in the treated sample relative to the control was expressed as a function of the concentration of LY2409881 in mmol/L. C, the DLBCL cell line. LY10, was grown in log phase, then pretreated with DMSO or LY2409881 at 10 mmol/L for 60 minutes. TNFa at 10 ng/mL was then added, and samples were removed after 2, 20, and 60 minutes of incubation and prepared for Western blot analysis using anti-IkB and antiphosphorylated IkB. Equal amount of proteins was loaded. The protein level of IkB in the LY2409881-treated samples served as an internal control for the level of phosphorylated IkB. D, SUDHL2 cells were treated by LY2409881 10 mmol/L or DMSO (Neg) for 4 hours, then prepared for epiﬂuorescence microscopy using the primary antibody anti-p105/p50, which recognized both p105 and p50, and DAPI staining that recognized DNA. In the untreated control (Neg), p50 was detected in the cytoplasm, but also in the nucleus, because of constitutively activated NF-kB. Treatment by LY2409881 for 4 hours resulted in exclusion of p50 from the nucleus. E, LY10 (top) and LY1 (bottom) cells were cultured in log phase, treated with LY2409881 at 2 and 10 mmol/L for 8 hours, then harvested. Nuclear (Nuc) and cytoplasmic (Cyto) extracts were made and electrophoresed, and probed with the indicated antibodies. Actin was used as a loading control. HDAC1 served as a loading control for nuclear proteins.

By in vitro kinase assay, LY2409881 potently inhibited IKK2, with upstream stimulus of NF-kB. TNFa stimulates both antiapoptotic aIC50 of 30 nmol/L. In contrast, the IC50 for IKK1 and other signals, mediated by NF-kB, and proapoptotic signals, mediated common kinases was at least one log higher. The speciﬁcity of by TNF receptor–associated death domain (TRADD) and FAS- LY2409881 for NF-kB signaling was further studied in a cell-based associated death domain (FADD) cascade pathways (25). As a assay, by examining the effect of LY2409881 in the TNFa-depen- result, TNFa is not a potent anticancer drug. In the ovarian cancer dent antiapoptosis function. TNFa is a well-characterized cell line SKOV3, LY2409881 demonstrated moderate cytotoxicity

www.aacrjournals.org Clin Cancer Res; 21(1) January 1, 2015 137

Deng et al.

(Fig. 1B), whereas TNFa at 10 ng/mL did not cause any cytotox- (9, 26). Figure 2C demonstrated that LY2409881 caused concen- icity. In contrast, coadministration of LY2409881 and TNFa tration-dependent apoptosis and cell death in MT2 cells after 48 resulted in markedly higher cell killing compared with hours of exposure. These results confirm that LY2409881 inhibits LY2409881 (Fig. 1B). We believe this is because TNFa-dependent constitutively activated NF-kB, leading to cell death in B- and T- activation of antiapoptotic signals mediated by NF-kB was cell lymphoma cell lines. blocked by LY2409881, while the proapoptotic TRADD and FADD cascade pathways activated by TNFa were not affected by LY2409881 causes cytotoxicity to both the ABC and GCB the compound. subtypes of DLBCL Next, we investigated the biochemical evidence that Previous studies have reported that the ABC subtype cell lines k LY2409881 inhibits TNFa-dependent activation of NF-kBin are more sensitive to disruption of the NF- B signals, through a models of lymphoma. In the DLBCL cell line LY10, TNFa small-molecule inhibitor, MLX105 (27), or a nondegradable k increased the level of phosphorylated-IkB significantly by 60 super-repressor I B (6). As we have shown, LY2409881 induced minutes, in the meantime reduced the level of total IkB through cytotoxicity in LY10 in a time- and concentration-dependent dynamic degradation of IkB (Fig. 1C). Pretreatment of LY10 with manner (Fig. 2B); next, the activity of LY2409881 in both ABC 10 mmol/L of LY2409881 prevented the formation of phosphor- and GCB cell lines was determined. Consistent with previous ylated IkB in samples treated with TNFa for 2, 20, and 60 minutes reports that ABC cell lines are more sensitive than GCB cells to k in the presence of abundant IkB protein. These results demon- inhibition of NF- B, two ABC cell lines, LY10 and LY3, were found strate that the IKK2 inhibitor LY2409881 specifically inhibits to be the most sensitive cell lines (Fig. 3A and Supplementary TNFa-dependent phosphorylation of IkB, an important marker Table S1). Two other ABC cell lines, HBL1 and SUDHL2, were fi for the canonical pathway of NF-kB. Surprisingly, when LY10 cells signi cantly more resistant to LY2409881 than LY10 and LY3. In were treated with the combination of LY2409881 and TNFa, there fact, HBL1 and SUDHL2 were even more resistant to LY2409881 was no increase in the cytotoxicity of LY2409881, in contrast to than the three GCB cell lines (Fig. 3A and Supplementary Table the results in the SKOV3 cell line (Fig. 1B). This may be because S1). We reasoned that the two resistant ABC cell lines might have k lymphoma cells such as LY10 are not sensitive to the TNF- different pathways of activation of NF- B from the sensitive ABC dependent proapoptotic signals such as TRADD and FADD. cell lines, and interrogated four key components of the NB-kB pathway: (i) nuclear p50, which reflects activation of the classical LY2409881 inhibits constitutively activated NF-kBin pathway, (ii) p65/RelA, which is required for the classical path- lymphoma cell lines and causes apoptosis in models of B- and way, (iii) nuclear p52, which reflects activation of the alternative T-cell lymphoma pathway, and (iv) IkB, which negatively regulates both p50 and The DLBCL cell line, SUDHL2, is known to have constitutively p52. activated NF-kB due to inactivating mutation of A20 (7). It has Figure 3B demonstrates that all four ABC cell lines (bold and been shown that introduction of wild-type A20 suppresses the italicized) had a higher level of activation of NF-kB through the activity of NF-kB, resulting in death of SUDHL2 cells. We eval- classical pathway, represented by a substantially higher ratio of uated whether LY2409881 was able to inhibit NF-kB and the p50/p105 in the nucleus, than the three GCB cell lines (under- survival of SUDHL2 cells. Figure 1D, top demonstrated that lined). Similarly, the four ABC cell lines had higher level of nuclear SUDHL2 cells had an abundant NFKB1/p50 signal in the nucleus, RelA than the three GCB cell lines. Note that the lower band in the consistent with previous reports of constitutively activated NF-kB. immunoblot of the nuclear protein extract was RelA/p65, and the When treated by LY2409881 at 10 mmol/L, the nuclear signals of upper band was nonspecific. In the cytoplasmic protein extracts, p50 in SUDHL2 were markedly reduced (Fig. 1D, bottom), there was only one band detected by the anti-p65 antibody, reflecting the sequestration of p50 in the cytoplasm when NF-kB corresponding in size to the lower band in the nuclear protein was inactivated by LY2409881. We further evaluated the effects of extract. Nuclear p52 was abundantly present in all four ABC cell LY2409881 on other subunits of NF-kB in two more DLBCL cell lines, and only present in one of the three GCB cell lines, LY7. The lines: LY10, representing the ABC subtype of DLBCL, and LY1, the level of IkB was lower in the two most sensitive cell lines, LY10 and GCB subtype. As shown in Fig. 1E, LY2409881 inhibited the LY3, and one of the most resistant cell lines, HBL1, than the other nuclear protein level of RelA/p65 and c-Rel in the LY10 cell line. four cell lines. The above results demonstrate that the ABC cell Surprisingly, in this cell line, nuclear p50 level was not signifi- lines consistently activate NF-kB through both the classical and cantly reduced by the compound. In the LY1 cell line, LY2409881 alternative pathways. The GCB cell lines have relatively lower level caused reduction of nuclear levels of p50 and RelA (p65), but not of activation through the classical pathway, and are infrequently c-Rel. These results suggest complex roles of NF-kB in different activated through the alternative pathway. Different DLBCL cell lymphoma cell lines, and that p65 may be a shared target by lines may use complex and distinct mechanisms to activate NF-kB LY2409881 in DLBCL cell lines from both the ABC and GCB signaling, and nevertheless remain sensitive to the IKK2 inhibitor subtypes. LY2409881. The pharmacologic activity of LY2409881 in diverse types of lymphoma models is shown in Fig. 2. Figure 2A demonstrates that LY2409881 is well tolerated and inhibits tumor growth in vivo the IKK2 inhibitor LY2409881 caused a concentration-dependent We used a well-established xenograft model of DLBCL to increase of apoptosis in the DLBCL cell line SUDHL2 cells after 24 confirm the activity of LY2409881 in vivo. SCID-beige mice hours of exposure. Similarly, the survival and proliferation of the implanted with LY10 cell-derived tumors were given intraperito- DLBLC cell line LY10 was potently inhibited by LY2409881, in a neal injections of LY2409881 twice weekly at three different time-dependent manner (Fig. 2B). The T-cell lymphoma MT2 cell doses: 50, 100, and 200 mg/kg. The treatments were well toler- line is established through transformation by HTLV-1, resulting in ated, resulting in no death or severe morbidity of the mice. The constitutive activation of the NF-kB pathway by the Tax oncogene average tumor volume was graphed as a function of time for each

138 Clin Cancer Res; 21(1) January 1, 2015 Clinical Cancer Research

IKK2 Inhibition Potently Synergizes with HDAC Inhibitors

Figure 2. A B LY2409881 inhibited constitutively Tx Dead 120 activated NF-kB and caused LY10-24 h apoptosis in models of B- and T-cell 100 LY10-48 h lymphoma. A, SUDHL2 cells were 80 LY10-72 h treated by LY2409881 (LY) at 10 and FL3 20 mmol/L or DMSO (Neg) for 24 60 hours. Apoptosis was determined None 14% 40 using ﬂow cytometric measurement of cells stained by propidium iodide 20 (FL3, Y-axis) and Yo-pro (FL1, X-axis), Survival % of control Survival using the Vybrant Apoptosis Assay 0 Kit. The percentage of apoptotic cells 0102030 in each treatment was summarized LY2409881 (µmol/L) next to each graph. B, the DLBCL cell

line, OCI-LY10, was treated by FL3 LY2409881 at the indicated concentrations. Samples were LY-10 29% collected after 24, 48, and 72 hours of mmol/L C continuous exposure. Surviving 80 percentage in the treated cells relative Dead cells to the untreated control cells was 60 Apoptotic cells expressed as a function of the drug concentration. C, the ATLL cell line, 40 MT2, was treated with LY2409881 at the indicated concentrations for 48 FL3 20 ﬂ

hours, and then processed for ow cells % of Total LY-20 53% cytometry. The percentage of mmol/L 0 apoptotic and dead cells was graphed Pos against the concentration. Positive mol/L FL1 control (Pos) was treated by romidepsin at 50 nmol/L. LY-0 µmol/LLY-1 µmol/LLY-2 µ LY-4 µmol/LLY-8 µmol/L

treatment group as shown in Fig. 3C. The rates of tumor volume of each drug as single agent (Table 1). The values of CI were growth of the treatment groups were all significantly slower than calculated for results in Fig. 4 and presented in Supplementary the untreated control group (P 0.01). There was no significant Table S3, demonstrating highly concordant results with the RRR difference among the three treatment cohorts. One animal in the values. 200-mg/kg cohort achieved a complete response. These results Next, we determined whether other IKK2 inhibitors also syner- demonstrate that LY2409881 is an active drug in DLBCL in vivo. gize with romidepsin, and examined the interaction of romidepsin with three other IKK2 inhibitors, including the Merck IKK2 LY2409881 synergizes moderately with chemotherapeutic inhibitor IV, Merck IKK2 inhibitor VIII, and Bay 11-7082 in the drugs and potently with HDAC inhibitors LY1 cell line. As shown in Fig. 4G, the Merck IKK2 inhibitor VIII If NF-kB responsive genes are involved in the proliferation and and romidepsin were markedly synergistic in the LY1 cell line, survival of cancer cells, then inhibiting the classical pathway of similar to the synergy of LY2409881 and romidepsin demon- NF-kB using the IKK2 inhibitor LY2409881 should theoretically strated in Fig. 4F. In contrast, neither Bay 11-7082 (Fig. 4H) nor potentiate the effect of other chemotherapeutic drugs. We inves- IKK2 inhibitor IV (not shown) was highly synergistic with romi- tigated the drug–drug interaction of LY2409881 with doxorubicin depsin in the LY1 cell line. We then sought to determine whether and cyclophosphamide in DLBCL cell lines. LY2409881 was the synergy observed is a class effect of HDAC inhibitors and highly synergistic with doxorubicin (Fig. 4A) and cyclophospha- LY2409881. As summarized in Table 1, high-level synergy was mide (Fig. 4C) in the ABC cell line SUDHL2. In the GCB cell line, rarely observed for LY2409881 and another pan-HDAC inhibitor LY1, the drug LY2409881 demonstrated no synergy with either belinostat, suggesting that the synergistic effect of IKK2 inhibitors doxorubicin (Fig. 4B) or cyclophosphamide (Fig. 4D). Supple- and HDAC inhibitors may not be class-dependent, but rather mentary Table S2 summarizes the drug–drug interaction of unique to the specific drug romidepsin. LY2409881 with doxorubicin and cyclophosphamide in a panel of lymphoma cell lines representing different disease entities, LY2409881 and romidepsin act antagonistically on p65/RelA indicating that synergy was uncommon in DLBCL cell lines when We explored the mechanism of the synergy of LY2409881 and LY2409881 was combined with these two chemotherapeutic HDAC inhibitors by determining how they differentially affect the drugs. In contrast, LY2409881 and the pan-HDAC inhibitor, p65 protein binding to the DNA probe containing the NF-kB romidepsin, were potently synergistic in both SUDHL2 and LY1 consensus sequence. As shown in Fig. 5A, LY2409881 decreased (Fig. 4E and F). In fact, marked and consistent synergy was the binding of p65 to the NF-kB consensus sequence, and romi- observed for LY2409881 and romidepsin in all seven DLBCL depsin increased the binding to the NF-kB consensus sequence in a cells irrespective of the ABC and GCB classification, with RRR concentration-dependent manner in the DLBCL cell line LY10. often below 0.1 at mildly or moderately cytotoxic concentrations Similarly, two other HDAC inhibitors, belinostat and vorinostat

www.aacrjournals.org Clin Cancer Res; 21(1) January 1, 2015 139

Deng et al.

A 120 HBL1-ABC 100 SUDHL2-ABC LY1-GCB 80 SUDHL4-GCB 60 LY7-GCB LY10-ABC 40 LY3-ABC 20 0 Survival % of control 0 5 10 15 20 25 30 LY2409881 (mmol/L) C B Dosage of Rate of growth Nuclear Cytoplasmic 2,500 LY2409881 mm3/d p100

200 mg/kg 20.59 (P < 0.01) p52 2,000 100 mg/kg 36.12 (P < 0.01) )

3 Actin 50 mg/kg 27.82 (P < 0.01) p105 1,500 Neg control 64.41 p50

1,000 p65

Tumor volume volume (mm Tumor IkB 500 Bcl-2 HDAC1 0 Actin 0 5 10 15 20 25 30 Cell Days from enrollment line

IC50 ABC cell lines: Bold and italicized; (mmol/L) GCB cell lines: underlined.

Figure 3. LY2409881 inhibited DLBCL cells in vitro and in vivo. A, DLBCL cell lines representing both ABC (HBL1, SUDHL2, LY10, and LY3) and GCB (LY1, SUDHL4, and LY7) subtypes were treated by LY2409881 for 48 hours. Surviving percentage in the treated cells relative to the untreated control cells was expressed as a function of the drug concentration. B, the same cell lines from A were cultured into log phase, and then harvested. Nuclear (Nu) and cytoplasmic (Cy) extracts were made and electrophoresed, and probed with the indicated antibodies. Actin was used as a loading control, and HDAC1 served as a loading control for nuclear proteins. The sample layout was the same for nuclear and cytoplasmic fractions. The ABC cell lines HBL1, LY10, SUDHL2 (SU2), and LY3 were labeled bold and italicized, and the GCB cell lines LY1, LY7, and SUDHL4 (SU4) underlined. The IC50 values of LY2409881 in these cell lines were noted below each cell line. The arrow indicated the lower band in the nuclear extract recognized by the anti-p65 antibody, which corresponded in size to the single band in the cytoplasmic extract recognized by the same antibody. The upper band in the nuclear extract was due to nonspeciﬁc binding. C, SCID-beige mice with LY10 implant were treated with LY2409881 at the indicated doses. The average tumor volume in different treatment groups was expressed as a function of time. Rate of tumor volume growth was calculated as in Materials and Methods, using the GEE. P value refers to the comparison with the untreated control mice.

(SAHA), also increased the binding capacity of p65 to the NF-kB romidepsin increased the binding of DNA and p65 in the DNA (Fig. 5B). setting of unchanged or even reduced nuclear protein level of Figure 5C demonstrated that LY2409881 markedly reduced p65 strongly suggests that romidepsin increased the afﬁnity of binding of p65 to its substrate DNA. In contrast, romidepsin p65 for the NF-kB DNA. markedly increased the binding of DNA and NF-kBina As acetylation of p65 has been shown to promote DNA-binding concentration-dependent manner. Combining LY2409881 and activity of p65, we investigated the effect of romidepsin on the romidepsin, on the other hand, was able to suppress the acetylation of p65 in DLBCL cell lines. Treatment with romidepsin increased binding of DNA and p65 caused by romidepsin. The resulted in increased acetylation of p65 in the nucleus in all of four amount of loaded nuclear protein, as demonstrated by HDAC1, DLBCL cell lines, irrespective of the ABC or GCB subtype to which they was relatively constant across the samples, as shown by the belong (Fig. 5E). The induction of acetylated p65 by romidepsin quantiﬁcation using densitometry (Fig. 5D). We also compared was concentration dependent, and was least in the ABC cell line the protein levels of p65 in the samples used for the electro- HBL1. These results demonstrated that treatment with romidep- phoretic mobility shift assay (EMSA) assay. Compared with the sin was associated with increased acetylation of p65 and increased untreated control, LY2409881 reduced the level of p65 by binding of p65 to the NF-kB consensus DNA, all of which might about 40%, whereas romidepsin at the 5-nmol/L concentration limit the antitumor properties of romidepsin due to transcription did not change the level of nuclear p65; and decreased it by of NF-kB–dependent genes. Treatment with LY2409881 was able about 30% at the 10-nmol/L concentration. The fact that to inhibit nuclear localization of p65 and suppress romidepsin-

140 Clin Cancer Res; 21(1) January 1, 2015 Clinical Cancer Research

IKK2 Inhibition Potently Synergizes with HDAC Inhibitors

AB 120 120 IK SUDHL2 IK LY1 D RRR = 1.16 D 100 RRR = 0.98 100 IK+D IK+D 80 80 RRR = 0.45 RRR = 1.22 60 60 RRR = 2.78 40 RRR = 0.07 40 20

LiveCell (% of ctrl) 20 LiveCell (% of ctrl) 0 0

CD 120 IK 120 IK SUDHL2 LY1 100 C 100 C RRR = 0.92 IK+C RRR = 0.99 IK+C 80 80 RRR = 0.95 RRR = 1.58 RRR = 0.43 60 60 40 RRR = 0.28 40 20 20 LiveCell (% of ctrl) LiveCell (% of ctrl) 0 0

EF 120 IK 120 SUDHL2 LY1 IK 100 R 100 R IK+R 80 80 IK+R RRR = 0.81 60 60 RRR = 0.48

40 RRR = 0.03 40 RRR = 0.40 20 20 RRR = 0.03 RRR = 0.04

LiveCell (% of ctrl) 0 LiveCell (% of ctrl) 0

GH 120 Mk 120 Bay LY1 LY1 100 R 100 R

Mk+R 88 Bay+R 80 80 . 60 60 RRR = 1.03 0.13 RRR = 0.48 = 40 40 RRR = 0 RRR = 0.11 RRR = 1.26 20 RRR 20 LiveCell (% of ctrl) 0 LiveCell (% of ctrl) 0

Figure 4. LY2409881 was synergistic with doxorubicin, cyclophosphamide, and HDAC inhibitors. Synergy was determined by calculating the RRR values, labeled in each graph. The CI values were calculated in the Supplementary Table S3, demonstrating the same level of synergy as revealed here by the RRR values. A, SU-DHL2 cells were treated by the following drugs: LY2409881 at 4, 8, and 15 mmol/L (IK1, IK2, and IK3, respectively), doxorubicin at 40, 160, and 800 nmol/L (D1, D2, and D3, respectively), and the two drugs in combination. B, LY1 cells were treated with LY2409881 at 2, 5, and 10 mmol/L (IK1, IK2, and IK3, respectively); doxorubicin at 100, 400, and 750 nmol/L (D1, D2, and D3, respectively), and the two drugs in combination. C, SU-DHL2 cells were treated by the following drugs: LY2409881 at 4, 8, and 15 mmol/L (IK1, IK2, and IK3, respectively), 4-hydroxycyclophosphamide at 2, 10, and 16 mmol/L (C1, C2, and C3, respectively), and the two drugs in combination. D, LY1 cells were treated with LY2409881 at 2, 5, and 10 mmol/L (IK1, IK2, and IK3, respectively), 4-hydroxycyclophosphamide at 2.5, 3, and 4 mmol/L (C1, C2, and C3, respectively), and the two drugs in combination. E, SU-DHL2 was treated with LY2409881 at 4, 8, and 15 mmol/L (IK1, IK2, and IK3, respectively), romidepsin at 0.8, 1.5, and 2.8 nmol/L (R1, R2, and R3, respectively), and the two agents in combination (IK1þR1, IK2þR2, and IK3þR3, respectively) for 48 hours. F, OCI-LY1 cells were treated with LY2409881 at 2, 5, and 10 mmol/L (IK1, IK2, and IK3, respectively), romidepsin at 3, 4.5, and 6 nmol/L (R1, R2, and R3, respectively), and the two agents in combination (IK1þR1, IK2þR2, and IK3þR3, respectively) for 48 hours. Note that the concentrations of drugs used for A and B were different, as the two cell lines differed in their sensitivity to these drugs. G and H were conducted in the LY1 cell line, as in F. Romidepsin was given at the concentrations of 3, 4.5, and 6 nmol/L (R1, R2, and R3, respectively), same as in F. Note that there was a small difference in the cytoxicity of romidepsin in the experiments done on different days. In G, the Merck IKK2 inhibitor VIII was given at 8, 10, and 15 mmol/L (Mk1, Mk2, and Mk3, respectively), replacing LY2409881 in F. In H, the IKK2 inhibitor Bay11-7082, in place of LY2409881, was given at 0.25, 0.375, and 0.75 mmol/L (Bay1, Bay2, and Bay3, respectively). www.aacrjournals.org Clin Cancer Res; 21(1) January 1, 2015 141

Deng et al.

Table 1. Synergy of LY2409881 with HDAC inhibitors the classical pathway has been shown to be the predominant IC20 IC40 IC60 pathway involved in tumorigenesis, targeting NF-kB has histor- (A) Synergy of LY2409881 and romidepsin ically focused on finding potent and selective inhibitors of IKK2. SUDHL2 0.81 0.4 0.03 Although dozens of IKK2 inhibitors have been reported, none RIVA 0.75 0.13 0.31 HBL1 1.02 0.51 0.42 have yet made it to the clinic. The long-standing concern over LY10 0.92 0.84 0.19 potential hepatotoxicity of IKK2 inhibitors originated from the LY1 0.48 0.03 0.04 observation that the IKK2 knockout mice died during embryo- LY7 0.75 0.05 0.01 genesis from massive liver necrosis. However, recent data dem- LY8 0.67 0.22 0.11 onstrated that hepatocyte-specific ablation of IKK2 did not lead to (B) Synergy of LY2409881 and belinostat increased apoptosis after TNFa stimulation (28, 29), suggesting SUDHL2 0.52 0.63 0.39 RIVA 1.03 0.84 1.09 that adult liver may not be prone to necrosis when exposed to HBL1 1.13 3.23 5.81 IKK2 inhibitors. Novel IKK2 inhibitors that are structurally dif- LY10 0.88 0.84 0.98 ferent from the previously developed inhibitors remain an impor- LY1 1.18 1.67 2.95 tant goal for the treatment of many diseases that are pathogen- LY7 0.93 0.36 0.07 etically associated with highly activated NF-kB signaling, such as LY8 0.76 0.99 0.4 many types of cancer. RRR Synergy 1.0 NO Here, we demonstrate that the novel IKK2 inhibitor 0.5–0.99 Mild LY2409881 is active in both B- and T-cell lymphoma models 0.10–0.49 Strong in which NF-kB is established to play important pathogenetic <0.1 High roles. For example, the sensitivity of DLBCL cells to LY2409881 NOTE: (A) DLBCL cells were treated with LY2409881 or romidepsin at the correlate with the activation status of NF-kB, with two ABC cell concentrations that were expected to cause 20%, 40%, and 60% inhibition, i.e., lines, LY3 and LY10, being most sensitive to LY2409881 (Fig. IC20,IC40,andIC60, respectively. In addition, cells were treated by the combi- 3A and Supplementary Table S1). Paradoxically, two other ABC nation of the two drugs at IC /IC ,IC /IC ,andIC /IC concentrations. 20 20 40 40 60 60 cell lines, SUDHL2 and HBL1, were actually more resistant than RRR was calculated as described in Materials and Methods, with lower RRR values indicating higher levels of synergy. (B) same as in (A), except belinostat the GCB cell lines studied (Fig. 3A and Supplementary Table replaced romidepsin. Four cell lines belong to the ABC subtype of DLBCL, S1). We believe these results may underscore the complexity of including SUDHL2, RIVA, HBL1, and LY10. Three cell lines belong to the GCB the NF-kB pathway, and suggest that the cell of origin (COO) subtype of DLBCL, including LY1, LY7, and LY8. classification of DLBCL cell lines may not necessarily correlate so well with the status of NF-kBasinpatientswithDLBCL.For mediated enhancement of p65-DNA binding, thereby mitigating example, we demonstrated that the classical pathway was an undesired feature of the pleiotropic HDAC inhibitor activated in all four ABC cell lines in a higher degree and in romidepsin. all three GCB cell lines to a relatively lower degree, as demonstrated by the presence of p50 and p65 in the nucleus (Fig. The synergy of belinostat and LY2409881 varies in different cell 3B). The alternative pathway was activated in all four ABC cell lines with distinct expression level of HDAC isoforms lines and only one of three GCB cell lines, as demonstrated by Unlike romidepsin, belinostat synergized with LY2409881 the presence of p52 in the nucleus (Fig. 3B). Of note, both the only in select DLBCL cell lines, including LY10, LY7, and SUDHL2 classical and alternative pathways have previously been (Table 1). In contrast, belinostat and LY2409881 were antago- reported to be involved in DLBCL (30). We also observed nistic in HBL1 and LY1. We asked whether the difference in substantial difference in the protein level of nuclear IkB, which synergy in different cell lines may be secondary to different was least abundant in the most sensitive ABC cell lines LY3 and expression levels of HDAC isoforms that are targets of belinostat. LY10, and the more resistant ABC cell line HBL1. Collectively, Class I HDACs HDAC1, HDAC2, and HDAC3 have been shown to these results suggest that sensitivity to the IKK2 inhibitor localize primarily to the nucleus, and are potently inhibited by cannot be simply predicted by the COO classification, or any belinostat and romidepsin. Figure 5F demonstrated that the level of the subunits of NF-kB. Rather, the sensitivity of the different of nuclear HDAC3 in HBL1 was substantially lower than cell lines to IKK2 inhibition may also be explained by how NF- SUDHL4, SUDHL2, and LY10. Similarly, the protein level of kB is activated using available genetic data. For example, the HDAC1 was lower in HBL1 than the other three cell lines. In LY3 cell line is known to harbor a homozygous activating contrast, the cytosolic level of HDAC4 was highest in HBL1, and mutation of CARD11, leading to activation of IKK2 and the the nuclear level of HDAC6 is much higher in HBL1 than the other NF-kB pathway (8). Therefore, LY3 is expected to be highly cell lines. It is conceivable that the distinct expression pattern of dependent on IKK2-induced NF-kB–responsive prosurvival the HDAC isoforms in HBL1 may be responsible for the lower signals. Consistently, LY3 is the most sensitive DLBCL cell line. level of acetylated p65 induced by romidepsin (Fig. 5E), and In contrast, SUDHL2 has homozygous inactivating mutation of ultimately the weak synergy in the combination LY2409881 and TNFAIP3/A20 (7), which results in constitutive activation of romidepsin and lack of synergy between LY2409881 and belino- upstream regulatory proteins that stimulate the IKK1/IKK2/ stat (Table 1). IKK3 kinases (31). Although LY2409881 potently inhibits IKK2, it is not active against IKK1. The leaky inhibition of the NF-kB pathway by LY2409881 in SUDHL2 therefore may result Discussion in relative resistance of the cell line to LY2409881. Similarly, NF-kB has been a logical target for cancer treatment for many the activation NF-kBinHBL1isaresultoftheheterozygous years due to its frequent activation in many cancer models and its missense mutation Y196F in the immunoreceptor tyrosine- capability to promote cell proliferation and survival (2). Because based activation motif of CD79B (32). Mutation of Y196 in

142 Clin Cancer Res; 21(1) January 1, 2015 Clinical Cancer Research

IKK2 Inhibition Potently Synergizes with HDAC Inhibitors

A B E nuc cy nuc cy p65 bound Ac-p65 kB DNA Mcl-1 Free Actin kB DNA

mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L mol/L

C F p65 bound Nuclear Cytoplasmic kB DNA

mol/L mol/L mol/L mol/L mol/L HDAC 1 HDAC 2 HDAC 3 D 1.00 0.63 0.98 0.69 0.55 0.29 RelA HDAC 4 HDAC 6 1.00 0.90 0.99 0.95 0.94 1.05 HDAC1 HDAC 8 Actin

mol/L mol/L mol/L mol/L mol/L

Figure 5. The distinct effects of LY2409881 and HDAC inhibitors on p65/RelA, and the expression of HDACs in DLBCL. A–C, DNA-binding activity of p65 determined by EMSA. A, LY10 cells were treated by LY2409881 (LY), DMSO, or romidepsin (Romi) for 22 hours and prepared for nuclear extract and EMSA. B, same as in A, except that LY10 cells were treated with DMSO, belinostat (Bel), or vorinostat (Vor) for 22 hours. C, same as in A, except cells were treated with DMSO, LY2409881, and romidepsin at 5 or 10 nmol/L in combination with LY2409881 (R5þLY or R10þLY, respectively) for 8 hours to avoid excessive cell killing by the combination at 22 hours. D, nuclear protein samples, same as those used in C, were run for Western blot analysis probed by the indicated antibodies. The numbers represented normalized signal intensity with the HDAC1 level in the control DMSO sample as 1.00. RelA level in the control sample was also set to 1.00, then compared with HDAC1 level to derive a relative RelA level after normalized for the loading error, which was negligible. E, acetylation of p65 determined by Western blot analysis. Two ABC (LY10 and HBL1) and two GCB (LY7 and LY1) subtype DLBCL cells were treated by romidepsin at 2.3 and 10 nmol/L for 22 hours. Nuclear (nuc) and cytoplasmic (cy) extracts were electrophoresed and probed with the antiacetylated p65 antibody. F, the expression of HDAC isoforms varied in different cell lines. Four DLBCL cell lines were harvested in log phase. The nuclear and cytoplasmic extracts were separated by electrophoresis, and probed with the anti-HDAC antibodies as shown.

CD79B impairs association of the negative regulatory Lyn inhibit NF-kB may be a promising strategy to improve the kinase, leading to activationoftheBTK,whichisknownto outcome of patients with ATLL. positively regulate both the classical and alternative NF-kB The in vivo activity of LY2409881 was confirmed in a xenograft pathways (33). It has also been shown that the PI3K pathway model of DLBCL (Fig. 3C). All three treatment doses were well is constitutively activated in HBL1 (34). As a result, the HBL1 tolerated, with more weight loss in the cohort treated with 200 cells treated by the IKK2 inhibitor may survive using the mg/kg of the drug. The treatments resulted in significant reduction prosurvival signals from the alternative NF-kBandthePI3K– of the tumor growth rate compared with the vehicle-treated AKT pathways. It is possible that the cell culture system may not control. There was no significant difference in the tumor growth be an ideal tool to study the ABC/GCB distinction and their among the three treatment cohorts. We frequently observed representative phenotype. It remains unknown whether the fluctuation of tumor volume in the treated cohorts; in contrast, complexity of DLBCL sensitivity to IKK2 inhibition seen in and untreated control demonstrated persistent increase of tumor the cell lines also applies to patients. Although gene expression volume. It is highly possible that daily administration of profiling is able to define the ABC subtype of DLBCL as LY2409881 may lead to more sustained inhibition of NF-kB and characterized by activated NF-kB,thereisnomethodtopredict more consistent tumor shrinkage, and may demonstrate a dose- sensitivity to specificNF-kB inhibitors. As a number of drugs in dependent control of the tumor. Indeed, daily administration of clinical use or development are able to inhibit NF-kB, including kinase inhibitors has been more commonly used in cancer ther- BTK inhibitors, bortezomib, carfilzomib, revlimid, and NEDD8 apy, as in the case of the BTK inhibitors, PI3K inhibitors, and NAE activating enzyme (NAE) inhibitors, it will be important to use inhibitors. primary patient samples to study the relationship of NF-kB Although LY2409881 is not a highly potent drug as a single status and sensitivity to these drugs. The insight may help us agent, we have found that it markedly synergizes with other determine which patients with the ABC subtype of DLBCL will anticancer drugs. In particular, the synergy of LY2409881 and respond to these drugs that affect NF-kB. Finally, we demon- the HDAC inhibitor romidepsin was substantial and consistently strated that LY2409881 was active in models of T-cell lympho- observed in all the tested DLBCL cell lines. The mechanistic basis ma (Supplementary Table S1 and Fig. 2). As there is no highly of the synergy is at least partially explained by how LY2409881 effective treatment for ATLL, LY2409881 or similar drugs that and romidepsin exert opposite effects on the NF-kB pathway on

www.aacrjournals.org Clin Cancer Res; 21(1) January 1, 2015 143

Deng et al.

multiple aspects, particularly p65/RelA. LY2409881 inhibited romidepsin may synergistically increase p21, leading to cell-cycle phosphorylation of IkB (Fig. 1C), leading to accumulation of arrest and apoptosis. IkB in the cytoplasm. LY2409881 blocked the accumulation of In summary, we have demonstrated that a novel IKK2 inhibitor, p50 and p65 in the nucleus (Figs. 1D, 1E, and 5D), and decreased LY2409881, effectively inhibits models of B- and T-cell lympho- the capacity of p65 binding to its substrate DNA (Figs. 5A and C). ma, where NF-kB is a well-established pathogenetic factor. We Romidepsin, on the other hand, increased the binding capacity of validated the activity of LY2409881 using a xenograft mouse p65 to the NF-kB DNA (Figs. 5A and C). It is worth noting that model of DLBCL. LY2409881 was demonstrated to directly and acetylated p65 was accumulated in the nucleus of four DLBCL cell indirectly inhibit romidepsin-mediated activation of NF-kB, lines, irrespective of the ABC and GCB subtypes (Fig. 5E), but was underlying a strong and consistent synergism with romidepsin least abundant in the HBL1 cell line. It is well established that in models of DLBCL. We believe combining LY2409881 and acetylation of p65 facilitates its nuclear localization (35). Fur- romidepsin will be a highly effective strategy in treating thermore, acetylation of p65 is facilitated by its phosphorylation lymphoma. at Serine 536, which is mediated by IKK2 in models of multiple myeloma (36). These results demonstrate that romidepsin sti- Disclosure of Potential Conﬂicts of Interest mulates acetylation of p65 in DLBCL cells irrespective of the ABC No potential conﬂicts of interest were disclosed. and GCB subtypes, leading to increased p65 and DNA binding k and activation of the NF- B transcription program. LY2409881, Authors' Contributions on the other hand, inhibits nuclear localization and the DNA- Conception and design: C. Deng, R. Rodriguez, O.A. O'Connor binding activity of p65, likely due to the inhibitory effect of IKK2 Development of methodology: C. Deng, R. Rodriguez, X.O.J. Serrano, O.A. inhibition on the phosphorylation, acetylation, nuclear localiza- O'Connor tion, and activity of p65. Collectively, LY2409881 may prevent Acquisition of data (provided animals, acquired and managed patients, romidepsin-mediated NF-kB activation either indirectly, by provided facilities, etc.): C. Deng, R. Rodriguez, X.O.J. Serrano, C. McIntosh increasing the level of IkB, or directly, by inhibiting the phos- Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C. Deng, M. Lipstein, R. Rodriguez, W.-Y. Tsai, phorylation and acetylation of p65. We believe these effects A.S. Wasmuth, O.A. O'Connor underlie the high level and consistent synergism of LY2409881 Writing, review, and/or revision of the manuscript: C. Deng, M. Lipstein, and romidepsin in all studied DLBCL cell lines. The observation R. Rodriguez, C. McIntosh, S. Jaken, O.A. O'Connor that the synergism occurred in both the ABC and GCB cells Administrative, technical, or material support (i.e., reporting or organizing obviously raises the possibility that other uncharacterized data, constructing databases): C. Deng, M. Lipstein, O.A. O'Connor mechanisms may contribute to the ubiquitous synergy, possibly Study supervision: C. Deng, O.A. O'Connor Other (chemistry support): A.S. Wasmuth due to the pleiotropic effects of romidepsin. Other IKK2 inhibitors were not synergistic with romidepsin consistently (Fig. 4H), presumably because they may not cause the same biochemical Acknowledgments changes as LY2409881. LY2409881 also synergized with another The authors thank the Lymphoma Research Fund for the Center for Lym- phoid Malignancies, Columbia University, and Marcio Chedid and Victoria HDAC inhibitor, belinostat, in some DLBCL cell lines. The protein Robinson for help with providing LY2409881. level of HDAC isoforms may be a potential determinant of the The costs of publication of this article were defrayed in part by the payment of lack of synergy between LY2409881 and belinostat in some cell page charges. This article must therefore be hereby marked advertisement in lines such as HBL1 (Fig. 5F and Table 1). Finally, the IKK2 accordance with 18 U.S.C. Section 1734 solely to indicate this fact. inhibitors, CYL-19s and CYL-26z, were recently found to increase p53 stability and p21 expression (37). As romidepsin also sti- Received February 19, 2014; revised October 6, 2014; accepted October 6, mulates the expression of p21, the IKK2 inhibitor LY2409881 and 2014; published OnlineFirst October 29, 2014.

References 1. Karin M, Cao Y, Greten FR, Li ZW. NF-kappaB in cancer: from innocent 7. Compagno M, Lim WK, Grunn A, Nandula SV, Brahmachary M, Shen Q, bystander to major culprit. Nat Rev Cancer 2002;2:301–10. et al. Mutations of multiple genes cause deregulation of NF-kappaB in 2. Baud V, Karin M. Is NF-kappaB a good target for cancer therapy? Hopes and diffuse large B-cell lymphoma. Nature 2009;459:717–21. pitfalls. Nat Rev Drug Discov 2009;8:33–40. 8. Lenz G, Davis RE, Ngo VN, Lam L, George TC, Wright GW, et al. Oncogenic 3. Izban KF, Ergin M, Qin JZ, Martinez RL, Pooley RJ, Saeed S, et al. Consti- CARD11 mutations in human diffuse large B cell lymphoma. Science tutive expression of NF-kappa B is a characteristic feature of mycosis 2008;319:1676–9. fungoides: implications for apoptosis resistance and pathogenesis. Hum 9. Rauch DA, Ratner L. Targeting HTLV-1 activation of NFkappaB in mouse Pathol 2000;31:1482–90. models and ATLL patients. Viruses 2011;3:886–900. 4. Sors A, Jean-Louis F, Pellet C, Laroche L, Dubertret L, Courtois G, et al. 10. Alizadeh AA, Eisen MB, Davis RE, Ma C, Lossos IS, Rosenwald A, et al. Down-regulating constitutive activation of the NF-kappaB canonical path- Distinct types of diffuse large B-cell lymphoma identified by gene expres- way overcomes the resistance of cutaneous T-cell lymphoma to apoptosis. sion profiling. Nature 2000;403:503–11. Blood 2006;107:2354–63. 11. Rosenwald A, Wright G, Chan WC, Connors JM, Campo E, Fisher RI, et al. 5. Pham LV, Tamayo AT, Yoshimura LC, Lo P, Ford RJ. Inhibition of The use of molecular profiling to predict survival after chemotherapy for constitutive NF-kappa B activation in mantle cell lymphoma B cells diffuse large-B-cell lymphoma. N Engl J Med 2002;346:1937–47. leads to induction of cell cycle arrest and apoptosis. J Immunol 2003; 12. Staudt LM, Dunleavy K, Buggy JJ, Hedrick E, Lucas N, Pittaluga S, et al. The 171:88–95. Bruton's tyrosine kinase (BTK) inhibitor PCI-32765 modulates chronic 6. Davis RE, Brown KD, Siebenlist U, Staudt LM. Constitutive nuclear factor active BCR signaling and induces tumor regression in relapsed/refractory kappaB activity is required for survival of activated B cell–like diffuse large B ABC DLBCL [abstract]. In: Proceedings of the 53rd ASH Annual Meeting cell lymphoma cells. J Exp Med 2001;194:1861–74. and Exposition; 2011 Dec 10–13;San Diego, CA. Abstract nr 2716.

144 Clin Cancer Res; 21(1) January 1, 2015 Clinical Cancer Research

IKK2 Inhibition Potently Synergizes with HDAC Inhibitors

13. Wilson WH, Gerecitano JF, Goy A, Vos Sd, Kenkre VP, Barr PM, et al. The 25. Aggarwal BB, Gupta SC, Kim JH. Historical perspectives on tumor necrosis Bruton's tyrosine kinase (BTK) inhibitor, ibrutinib (PCI-32765), has factor and its superfamily: 25 years later, a golden journey. Blood 2012; preferential activity in the ABC subtype of relapsed/refractory de novo 119:651–65. diffuse large B-cell lymphoma (DLBCL): interim results of a multicenter, 26. Hironaka N, Mochida K, Mori N, Maeda M, Yamamoto N, Yamaoka S. Tax- open-label, phase 2 study [abstract]. In: 54th ASH Annual Meeting and independent constitutive IkappaB kinase activation in adult T-cell leuke- Exposition; 2012 Dec 8–11; Atlanta, GA. Abstract nr 686. mia cells. Neoplasia 2004;6:266–78. 14. Cattoretti G, Pasqualucci L, Ballon G, Tam W, Nandula SV, Shen Q, et al. 27. Lam LT, Davis RE, Pierce J, Hepperle M, Xu Y, Hottelet M, et al. Small Deregulated BCL6 expression recapitulates the pathogenesis of human molecule inhibitors of IkappaB kinase are selectively toxic for subgroups of diffuse large B cell lymphomas in mice. Cancer Cell 2005;7:445–55. diffuse large B-cell lymphoma defined by gene expression profiling. Clin 15. Bereshchenko OR, Gu W, Dalla-Favera R. Acetylation inactivates the Cancer Res 2005;11:28–40. transcriptional repressor BCL6. Nat Genet 2002;32:606–13. 28. Maeda S, Chang L, Li ZW, Luo JL, Leffert H, Karin M. IKKbeta is required for 16. Pasqualucci L, Dominguez-Sola D, Chiarenza A, Fabbri G, Grunn A, prevention of apoptosis mediated by cell-bound but not by circulating Trifonov V, et al. Inactivating mutations of acetyltransferase genes in B- TNFalpha. Immunity 2003;19:725–37. cell lymphoma. Nature 2011;471:189–95. 29. Luedde T, Assmus U, Wustefeld T, Meyer zu, Vilsendorf A, Roskams T, 17. Dai Y, Rahmani M, Dent P, Grant S. Blockade of histone deacetylase Schmidt-Supprian M, et al. Deletion of IKK2 in hepatocytes does not inhibitor-induced RelA/p65 acetylation and NF-kappaB activation potenti- sensitize these cells to TNF-induced apoptosis but protects from ischemia/ ates apoptosis in leukemia cells through a process mediated by oxidative reperfusion injury. J Clin Invest 2005;115:849–59. damage, XIAP downregulation, and c-Jun N-terminal kinase 1 activation. 30. Pham LV, Fu L, Tamayo AT, Bueso-Ramos C, Drakos E, Vega F, et al. Mol Cell Biol 2005;25:5429–44. Constitutive BR3 receptor signaling in diffuse, large B-cell lymphomas 18. Nagashima K, Sasseville VG, Wen D, Bielecki A, Yang H, Simpson C, et al. stabilizes nuclear factor-kappaB-inducing kinase while activating both Rapid TNFR1-dependent lymphocyte depletion in vivo with a selective canonical and alternative nuclear factor-kappaB pathways. Blood 2011; chemical inhibitor of IKKbeta. Blood 2006;107:4266–73. 117:200–10. 19. Hideshima T, Chauhan D, Kiziltepe T, Ikeda H, Okawa Y, Podar K, et al. 31. Shembade N, Harhaj EW. Regulation of NF-kappaB signaling by the A20 Biologic sequelae of I{kappa}B kinase (IKK) inhibition in multiple mye- deubiquitinase. Cell Mol Immunol 2012;9:123–30. loma: therapeutic implications. Blood 2009;113:5228–36. 32. Hailfinger S, Lenz G, Ngo V, Posvitz-Fejfar A, Rebeaud F, Guzzardi M, et al. 20. Al-Katib A, Arnold AA, Aboukameel A, Sosin A, Smith P, Mohamed AN, Essential role of MALT1 protease activity in activated B cell–like diffuse et al. I-kappa-kinase-2 (IKK-2) inhibition potentiates vincristine cytotox- large B-cell lymphoma. Proc Natl Acad Sci U S A 2009;106:19946–51. icity in non-Hodgkin's lymphoma. Mol Cancer 2010;9:228. 33. Shinners NP, Carlesso G, Castro I, Hoek KL, Corn RA, Woodland RT, et al. 21. Annunziata CM, Davis RE, Demchenko Y, Bellamy W, Gabrea A, Zhan F, Bruton's tyrosine kinase mediates NF-kappa B activation and B cell survival et al. Frequent engagement of the classical and alternative NF-kappaB by B cell-activating factor receptor of the TNF-R family. J Immunol pathways by diverse genetic abnormalities in multiple myeloma. Cancer 2007;179:3872–80. Cell 2007;12:115–30. 34. Kloo B, Nagel D, Pfeifer M, Grau M, Duwel M, Vincendeau M, et al. Critical 22. Sors A, Jean-Louis F, Begue E, Parmentier L, Dubertret L, Dreano M, et al. role of PI3K signaling for NF-kappaB-dependent survival in a subset of Inhibition of IkappaB kinase subunit 2 in cutaneous T-cell lymphoma activated B-cell–like diffuse large B-cell lymphoma cells. Proc Natl Acad Sci down-regulates nuclear factor-kappaB constitutive activation, induces cell U S A 2011;108:272–7. death, and potentiates the apoptotic response to antineoplastic chemo- 35. Chen L, Fischle W, Verdin E, Greene WC. Duration of nuclear NF- therapeutic agents. Clin Cancer Res 2008;14:901–11. kappaB action regulated by reversible acetylation. Science 2001;293: 23. Marchi E, Paoluzzi L, Scotto L, Seshan VE, Zain JM, Zinzani PL, et al. 1653–7. Pralatrexate is synergistic with the proteasome inhibitor bortezomib in in 36. Dai Y, Chen S, Wang L, Pei XY, Funk VL, Kramer LB, et al. Disruption of I vitro and in vivo models of T-cell lymphoid malignancies. Clin Cancer Res {kappa}B kinase (IKK)-mediated RelA serine 536 phosphorylation sensi- 2010;16:3648–58. tizes human multiple myeloma cells to histone deacetylase (HDAC) 24. Kalac M, Scotto L, Marchi E, Amengual J, Seshan VE, Bhagat G, et al. HDAC inhibitors. J Biol Chem 2011;286:34036–50. inhibitors and decitabine are highly synergistic and associated with unique 37. Yang PM, Huang WC, Lin YC, Huang WY, Wu HA, Chen WL, et al. Loss of gene-expression and epigenetic profiles in models of DLBCL. Blood IKKbeta activity increases p53 stability and p21 expression leading to cell 2011;118:5506–16. cycle arrest and apoptosis. J Cell Mol Med 2010;14:687–98.

www.aacrjournals.org Clin Cancer Res; 21(1) January 1, 2015 145

The Novel IKK2 Inhibitor LY2409881 Potently Synergizes with Histone Deacetylase Inhibitors in Preclinical Models of Lymphoma through the Downregulation of NF-κB

Changchun Deng, Mark Lipstein, Richard Rodriguez, et al.

Clin Cancer Res 2015;21:134-145. Published OnlineFirst October 29, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-14-0384

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2015/07/17/1078-0432.CCR-14-0384.DC1

Cited articles This article cites 35 articles, 18 of which you can access for free at: http://clincancerres.aacrjournals.org/content/21/1/134.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/21/1/134.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/21/1/134. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.