Receptor tyrosine kinase inhibitors have recently become important therapeutics for a variety of cancers. However, due to the heterogeneous and dynamic nature of tumors, the effectiveness of these agents is often hindered by poor response rates and acquired drug resistance. To overcome these limitations, we created a novel small molecule, CUDC-101, which simultaneously inhibits histone deacetylase and the receptor kinases epidermal growth factor receptor (EGFR) and human epidermal growth factor receptor 2 (HER2) in cancer cells. Because of its integrated histone deacetylase inhibition, CUDC-101 synergistically blocked key regulators of EGFR/HER2 signaling pathways, also attenuating multiple compensatory pathways, such as AKT, HER3, and MET, which enable cancer cells to escape the effects of conventional EGFR/HER2 inhibitors. CUDC-101 displayed potent antiproliferative and proapoptotic activities against cultured and implanted tumor cells that are sensitive or resistant to several approved single-targeted drugs. Our results show that CUDC-101 has the potential to dramatically improve the treatment of heterogeneous and drug-resistant tumors that cannot be controlled with single-target agents. Further, they provide a framework to create individual small molecules that simultaneously antagonize multiple biochemically distinct oncogenic targets, suggesting a general paradigm to surpass conventional, single-target cancer therapeutics. Cancer Res; 70(9); 3647–56. ©2010 AACR.
Single-target agents are often limited in their clinical utility because tumors harbor multiple misregulated growth and survival pathways, which can evolve during the course of treatment. In that light, combinatory approaches to targeted therapy offer great promise, endowed with a greater likelihood of blocking tumor survival and metastasis through the rectification of multiple oncogenic pathways. In some instances, multitargeting can be achieved by simply combining several drugs against just one specific target. The risk with such combination therapies is that multiple drugs can introduce adverse effects related to pharmacokinetics, toxicity, and patient compliance. Alternatively, single drugs with multiple inhibitory activities offer the advantages of pharmacokinetic simplicity and reduced cost. Current drugs of this type derive their various inhibitory activities by targeting related members of the same gene family due to the cross-reactivity of a single small molecule. To further explore the versatility of this idea, we proposed a novel strategy with a single small molecule designed to concurrently act on two or more biochemically distinct targets.
In this first proof-of-principle study, we chose to integrate into one compound the crucial structural elements required to inhibit histone deacetylase (HDAC) and human epidermal growth factor receptor (HER) kinases. Many selective inhibitors of HER family receptor tyrosine kinases (RTK), including erlotinib, gefitinib, and lapatinib, have become important therapeutics against multiple solid tumor cancers (1, 2). However, due to molecular heterogeneity among and within tumors, their efficacy is restricted to only a small subset of patients (2). The efficacy of RTK inhibitors is also limited by the drug resistance that frequently emerges following treatment (3, 4). Several strategies have been proposed to overcome the low response rate and acquired resistance to RTK inhibitors. One particularly promising approach is the modulation of RTK pathways by the inhibition of HDACs. By modulating the acetylation of both histone and nonhistone substrates (5–8), HDAC inhibitors can affect a variety cell functions through subsequent regulation of indirect downstream targets. Importantly, many of these targets are key regulators of RTK signaling pathways (6, 7, 9). Several reports also suggest a synergy between RTK and HDAC inhibition in cancer cells (10–13).
Here, we report CUDC-101, a small molecule rationally designed to directly inhibit epidermal growth factor receptor (EGFR) and HER2 as well as class I and class II HDACs. We show that CUDC-101 effectively suppresses the progression of a broad range of tumor types in both in vitro and in vivo xenograft models, including lapatinib- and erlotinib-resistant cancer cell lines. Mechanistic studies reveal that CUDC-101 not only directly inhibits both EGFR and HER2 signaling but also indirectly attenuates the survival signaling pathways Akt, HER3, and MET. This work also shows a practical approach to create multitarget anticancer agents, based on a single small molecule, which could significantly enhance cancer therapy.
Materials and Methods
CUDC-101, vorinostat (SAHA), erlotinib, and lapatinib were synthesized as described (14). The PCR primer pairs used for detection of Her2 (Genbank accession number: NM_004448) and glyceraldehyde-3-phosphate dehydrogenase (Genbank accession number: M33197) were obtained from R&D Systems.
HDAC activity and EGFR and HER2 kinase assay
The activities of class I and II HDACs were assessed using the Biomol Color de Lys system. EGFR and HER2 kinase activity was measured using the HTScan EGF receptor and HER2 kinase assay kits (Cell Signaling Technology).
Cell growth, viability, and apoptosis assay
Cancer cell lines were obtained from the American Type Culture Collection and were maintained according to the supplier's instructions. Cancer cell lines were plated at 5,000 to 10,000 cells per well in 96-well flat-bottomed plates with varying concentrations of compounds. The cells were incubated with compounds for 72 hours in the presence of 0.5% of fetal bovine serum. Growth inhibition was assessed by an ATP content assay using the Perkin-Elmer ATPlite kit. Apoptosis was routinely assessed by measuring the activities of caspase-3 and caspase-7 using the Promega Apo-ONE Homogeneous Assay kit.
Immunoblotting, immunocytochemistry, and immunohistochemistry
Immunoblotting, immunocytochemistry, and immunohistochemistry were done using standard procedures with a blocking solution (Li-Cor Bioscience) containing the indicated primary and IRDye 680–, 800CW-, or peroxidase-conjugated secondary antibodies.
Efficacy study in human cancer xenograft model
Four- to 6-week-old female athymic mice (nude nu/nu CD-1, Charles River) were inoculated s.c. into the right hind flank region with 1 to 5 × 106 cells in a medium suspension of 100 to 200 μL. For orthotopic implantation of breast cancer cells, a cell suspension in 100 μL of medium was injected directly into the mammary fat pads through a 27G needle. Different doses of CUDC-101, standard anticancer agents, and vehicle were administered orally, i.p., or through tail vein injection as indicated.
CUDC-101 inhibits EGFR, HER2, and HDAC
We designed and synthesized a multitarget inhibitor, CUDC-101, which integrates HDAC, EGFR, and HER2 inhibitory functional groups into one single small molecule (molecular weight, 434; Fig. 1A). We introduced hydroxamic acid with a flexible side chain onto the methoxyethoxy group of the phenylaminoquinazoline backbone of the RTK inhibitors (Fig. 1A). We anticipated that these engineered molecules would fit well into the substrate-binding pocket of HDACs and would also interact with zinc at the active site to disrupt enzyme activity, yet still retain their ability to compete with ATP binding and function as RTK inhibitors (14, 15). Confirming the effectiveness of this approach, we showed that one of the best compounds, CUDC-101, inhibits EGFR and HER2 kinase as well as HDAC enzyme activities with high potency (2.4, 16.4, and 4.2 nmol/L, respectively; Table 1). It inhibits both class I– and class II–purified HDACs (see Supplementary Table S1). Interestingly, we note that CUDC-101 inhibits the erlotinib-resistant EGFR mutant T790M (2, 3), although its effects are incomplete with an Amax of ∼60% of peak enzyme activity after inhibition (Supplementary Fig. S1).
We next determined the effects of CUDC-101 on histone acetylation in cancer cell lines. Exposing cells to varying concentrations of CUDC-101 for 5 to 24 hours increased the acetylation of histone H3 and H4 in a dose-dependant manner in various cancer cell lines (Supplementary Fig. S2; Fig. 1B). Similarly, we observed that CUDC-101 increased the acetylation of p53 and α-tubulin, nonhistone substrates of HDAC, in treated cancer cells (Supplementary Fig. S3; Fig. 2B). We also noted that CUDC-101 inhibits EGFR autophosphorylation in a dose-dependent fashion (Supplementary Fig. S4; Fig. 1C) and inhibits HER2 phosphorylation in cultured cancer cells (Fig. 1D). Taken together, these results show that CUDC-101 displays activities against multiple targets, inhibiting both RTK and HDAC pathways in a variety of cancer cell types.
To assess the selectivity of its effects, CUDC-101 was tested against a panel of 69 other kinases. It exhibited only weak inhibition of KDR (VEGFR2), Lyn, Lck, Abl-1, FGFR-2, Flt-3, and Ret with an IC50 of 0.85, 0.84, 5.9, 2.89, 3.43, 1.5, and 3.2 μmol/L, respectively (Supplementary Table S2). These results confirm that CUDC-101 is a selective EGFR and HER2 kinase inhibitor.
HDAC and RTK inhibition are synergistic
We next asked whether there is any additive or synergistic effect between the HDAC and RTK inhibition mediated by CUDC-101. In this study, two reference compounds, vorinostat (SAHA, a HDAC inhibitor) and erlotinib (an EGFR kinase inhibitor), were used to achieve HDAC and EGFR pathway inhibition, respectively. A well-established mathematical model for studying multidrug interactions was applied to the analysis (16). Our results show that the combined inhibition of HDAC and EGFR pathways achieves considerable synergy, as the calculated combination index (16) is significantly smaller than 1 in an experiment using the MDA-MB-468 breast cancer cell line as a model (Fig. 3A). This result suggests that CUDC-101, with its combined RTK and HDAC inhibitory activities, could act synergistically in treated cancer cells. In this study, we also observed a similar combination index in experiments using varying ratios of vorinostat and erlotinib. These results suggest that a specific dosing ratio is not required to achieve a synergistic effect on these two pathways and further validate the idea of using one compound directed against multiple targets for improved therapeutic outcomes.
Growth-inhibition of CUDC-101 in vitro
To further confirm its apparent synergistic effects, we tested the growth-inhibitory activity of CUDC-101 in a total of 54 human cancer cell lines (see Table 2). Our results show that the compound is effective against a broad range of cancer cell types, including lung, pancreas, liver, colon, breast, prostate, and head and neck. Of note, CUDC-101 exhibited equal or greater potency in these assays than vorinostat, erlotinib, lapatinib, or combinations (1:1 ratio) of vorinostat and erlotinib, or vorinostat and lapatinib. CUDC-101 also suppressed the growth of the lapatinib-insensitive, triple-negative (estrogen receptor negative, progesterone receptor negative, and HER2 negative) breast cancer cell line MDA-MB-231 as effectively as it did the lapatinib-sensitive, HER2-overexpressed lines BT-474 and SkBr-3. Similarly, CUDC-101 inhibits the growth of lung cancer cell lines that are not sensitive to erlotinib treatment, including H1975, which harbor an EGFR-T790M mutation. These results suggest that CUDC-101, with its synergistic inhibitions of multiple pathways, has the potential to improve the response rates to traditional kinase inhibitors.
To further dissect the growth-inhibitory effects of CUDC-101, we monitored caspase activity in treated cells. We found that CUDC-101 effectively induced the activities of caspases 3 and 7 in HCT-116 colon cancer cells, which harbor mutations downstream of RTK in k-Ras and phosphoinositide 3-kinase (PI3K), in a dose-dependent manner (Fig. 2A). In addition, CUDC-101 reduced the levels of the antiapoptotic protein survivin and Bcl-xL in MDA-MB-231 breast cancer cells that did not respond well to lapatinib (Fig. 2B). We also observed that CUDC-101 induced the expression of p21, a cyclin-dependent kinase inhibitor and key regulator of cell proliferation (Fig. 2B), providing further evidence of the antiproliferative and proapoptotic roles of the compound in nonresponder cell lines.
CUDC-101 modulates RTK activity and expression
To reveal the potential mechanisms behind this synergy, we examined the effects of CUDC-101 on RTK activity, transcription, and protein levels. Using breast cancer cell lines as a model, we showed CUDC-101 simultaneously inhibited kinase activity and suppressed RNA transcription and protein levels (Figs. 1D and 3B and C). Both lapatinib (an EGFR/HER2 kinase inhibitor) and erlotinib inhibited HER2 kinase activity yet neither showed effects on Her2 transcription (Fig. 3B). In contrast, vorinostat had no effect on kinase activity but inhibited Her2 transcription (Fig. 3B). Both CUDC-101 and vorinostat also reduced HER2 receptor protein levels in BT-474 and SK-Br-3 cells (Fig. 3C). Taken together, these results suggest that CUDC-101 could achieve its synergistic effects through the modulation of multiple variables affecting RTK signaling, including activity and protein level of RTKs.
CUDC-101 exhibits immediate and stable inhibition of RTK and downstream Akt signaling
Many tumors invariably develop drug resistance against RTK inhibitors. One of the proposed underlying mechanisms is a compensatory shift in the HER3 phosphorylation-dephosphorylation equilibrium that leads to reactivation of the downstream PI3K/Akt survival pathway (17–19). Consistent with this hypothesis, we observed Akt reactivation in cultured cancer cells within 8 hours of erlotinib treatment (Fig. 2C). Similarly, we detected increased levels of phosphorylated HER3 despite an initial downregulation following exposure to erlotinib. In contrast, CUDC-101 showed a rapid and prolonged inhibition of Akt signaling in cultured cancer cells, with no subsequent reactivation during 24 hours of continuous exposure. The levels of phosphorylated HER3 also remained low throughout the CUDC-101 treatment period (Fig. 2C). Vorinostat also showed similar late-stage inhibition of HER3 and Akt signaling (Fig. 2C), but the latter effect was observed only after 8 hours of treatment (Supplementary Fig. S5). These data suggest that CUDC-101, with its combined inhibitory effects on RTK and HDAC, can rapidly and sustainably suppress RTK and Akt signaling. Furthermore, these results suggest that the multiple targeted activities of CUDC-101 may hinder the ability of heterogeneous tumors to acquire resistance to RTK inhibitor therapy.
CUDC-101 reduces levels of phosphorylated and total MET
MET amplification has been detected in a significant percentage (22%) of lung tumor samples that exhibit resistance to RTK inhibitors (20–22). Therefore, we measured the erlotinib and CUDC-101 sensitivity of a non–small cell lung carcinoma (NSCLC) cell line (H1993) that harbors an amplification of MET. As expected, the cells failed to respond to erlotinib treatment, showing an IC50 of ∼20 μmol/L in a proliferation assay (Table 2). However, the H1993 cells were sensitive to CUDC-101, with an IC50 of ∼300 nmol/L. Moreover, CUDC-101, but not erlotinib, consistently inhibited Akt signaling in these cells. We subsequently confirmed that CUDC-101 can downregulate the levels of phosphorylated and total MET, presumably due to its HDAC-inhibitory activity (Fig. 2D). This result provides further support for CUDC-101 as a novel agent to mitigate RTK inhibitor drug resistance.
CUDC-101 reduces estrogen receptor α protein levels
Acquired resistance to HER2 RTK inhibitors also limits the clinical efficacy of this important class of cancer therapeutics. Recent evidence indicates that increased estrogen receptor signaling may underlie lapatinib resistance in HER2-positive breast cancers (23). Based on these data, we decided to investigate the effects of CUDC-101 on estrogen receptor α expression. Our studies revealed that CUDC-101, but not lapatinib, can effectively suppress estrogen receptor α protein levels with its HDAC inhibitory activity (Supplementary Fig. S6), a finding that underscores the potential of CUDC-101 to overcome drug resistance in HER2-positive breast cancers.
CUDC-101 displays broad activity in in vivo xenograft models
Next, we evaluated whether CUDC-101 is efficacious in human cancer cell xenograft models. Because preclinical studies of EGFR inhibitors in hepatocellular carcinoma cell lines and phase II studies in human hepatocellular carcinoma have been encouraging, we decided to compare the effects of CUDC-101 to those of erlotinib and vorinostat in the Hep-G2 liver cancer model (Fig. 4A). At a daily dose of 120 mg/kg, CUDC-101 induced 30% tumor regression (P < 0.0001) and was more efficacious than erlotinib at its maximum tolerated dose and vorinostat at an equimolar concentration. Notably, one of the seven treated mice displayed complete tumor regression at the end of the dosing cycle, an effect that lasted for at least 6 months after treatment (Fig. 4A, bottom). CUDC-101 also exhibited potent inhibition of tumor growth in the erlotinib-resistant, A549 NSCLC xenograft model (P < 0.001 compared with control; Fig. 4B). In the lapatinib-resistant, HER2-negative, EGFR-overexpressing MDA-MB-468 breast cancer model, CUDC-101 caused significant tumor regression compared with control or lapatinib-treated animals (P < 0.0001; Fig. 4B). Similar effects were observed in the EGFR-overexpressing, CAL-27 head and neck squamous cell carcinoma xenograft model (Fig. 4B). In all of the above xenograft experiments, no significant weight loss or other side effect was observed in the CUDC-101–treated animals (Supplementary Fig. S7). Moreover, we also observed that the combination of CUDC-101 with the chemotherapeutic agent paclitaxel could achieve further enhanced effects and induce tumor regression in the MDA-MB-468 breast cancer xenograft model (Fig. 4B). Overall, our in vivo experiments provide direct support for the rationale for improving RTK inhibitor efficacy through the synergy of multitargeted inhibition. To determine if the observed activity in xenograft models is target specific, we asked whether CUDC-101 can directly inhibit EGFR and HER2 in vivo. In this experiment, we observed an increase in acetylated histone H3 levels (Fig. 4C) and a decrease in phosphorylated EGFR and HER2 levels (Fig. 4C and D) in tumor tissue isolated from CUDC-101–treated animals. These tumors also exhibited reduced staining of Ki67 and increased staining of activated caspase-3 (Fig. 4D). Our results provide evidence that CUDC-101 inhibits specific molecular targets in vivo to block proliferation and induce apoptosis in drug-resistant cancer cells.
CUDC-101, a potent HDAC, EGFR, and HER2 inhibitor, is rationally designed to synergistically suppress multiple oncogenic signaling pathways. The compound displays potent antiproliferative and proapoptotic activities in in vitro as well as in vivo drug-resistant tumor models. Our mechanistic studies consistently showed that it can rapidly and stably inactivate EGFR, HER2, and other survival signaling pathways, such as Akt, HER3, and MET. Moreover, CUDC-101 has the potential to achieve further synergy in tumor suppression when combined with chemotherapeutics. Collectively, our results indicate that CUDC-101, with its unique and novel multitarget activities, is an attractive candidate for future clinical development. Because of its potential to improve conventional EGFR and HER2 kinase inhibitor therapy in various cancer indications, CUDC-101 has now entered phase I clinical trials.
A specific multitargeted inhibitor with a favorable safety profile
CUDC-101 robustly and selectively inhibits EGFR/HER2 kinases and HDAC. To further support this, in vitro pharmacology profiling studies revealed that CUDC-101 shows no significant activity or binding affinity to key enzymes, receptors, or transporters (Supplementary Table S3). CUDC-101 consistently displays a favorable safety profile both in efficacy and toxicity studies in animals (Supplementary Table S4 and Fig. S7). The specificity and tolerability of CUDC-101 were further supported by the preliminary data from its ongoing phase I clinical trial. Currently, dose levels up to 275 mg/m2 have been well tolerated with the most frequent adverse events being dry skin/rash, nausea, fatigue, constipation, dyspnea, and pyrexia. The type and frequency of adverse events seem comparable with those observed with single agent erlotinib or vorinostat therapy. No new CUDC-101–specific toxicities have been reported. Plasma exposure over the dose levels tested seems dose proportionate. Taken together, these results suggest that the observed effects of CUDC-101 are mediated specifically through the inhibition of HDAC, EGFR, and HER2.
Synergy between RTK and HDAC pathway inhibition
The simultaneous inhibition of RTK and HDAC pathways through CUDC-101 can achieve synergistic effects on tumor suppression. Indeed, a synergy between RTK and HDAC inhibition has been previously showed by others and also confirmed in this report at several levels. First, cotreatment with HDAC and RTK inhibitors in cultured cancer cells has a calculated combination index lower than 1 (10–13). Second, CUDC-101 reduces RTK transcription and/or protein levels as well as those of downstream regulators, suggesting that direct kinase inhibition is enhanced by indirect effects on expression. Finally, CUDC-101 as well as other HDAC inhibitors reduce HIF-1α protein levels (Supplementary Fig. S8), suggesting that antitumor activity could be accomplished by a combination of antiproliferative and antiangiogenic effects (9).
A viable approach to improve response rate
CUDC-101 displays high potency in both in vitro and in vivo models. The finding that both erlotinib-sensitive and erlotinib-resistant NSCLC cell lines are highly responsive to CUDC-101 treatment suggests that future clinical applications may improve patient response. Similarly, CUDC-101 is effective in both lapatinib-sensitive (HER2 positive) and lapatinib-resistant (HER2 negative) breast cancer models. These results provide further support for the therapeutic rationale to simultaneously inhibit HDAC and EGFR/HER2 to boost antiproliferative effects in cancer cells. CUDC-101 also achieves greater potency than combinations of HDAC and EGFR or HER2 inhibitors. The most parsimonious explanation is that HDAC and EGFR activities are each more effectively inhibited by CUDC-101 than by vorinostat and erlotinib, respectively.
Inroads to overcoming drug resistance
Acquired resistance to tyrosine kinase inhibitor therapy likely results from further mutations of the RTK gene and/or the compensatory effects of survival pathways through additional or reactivated downstream signaling. CUDC-101 could overcome such drug resistance through its ability to both inhibit RTK activity and reduce protein levels as well as suppress survival signaling pathways. Patients with NSCLC typically relapse within a year of treatment. In many cases, relapse correlates with a secondary mutation, T790M, in the EGFR kinase domain. Interestingly, we found that CUDC-101 can inhibit the EGFR-T790M mutant to an Amax of ∼60% of peak enzyme activity (Supplementary Fig. S1). In addition, CUDC-101 downregulates EGFR protein levels in H1975 cells, which harbor an EGFR-T790M mutation (Supplementary Fig. S9). Consistent with this finding, we observed that H1975 is sensitive to CUDC-101, but not erlotinib treatment (Table 2). By virtue of its effects on both the enzymatic activity and protein levels of the mutated EGFR, our results support that CUDC-101 may overcome drug resistance in NSCLC tumors resulting from secondary EGFR kinase domain mutations. CUDC-101 could also overcome drug resistance by inhibiting other survival signaling pathways. For instance, in breast cancer and NSCLC cells, CUDC-101 suppressed Akt, HER3, and MET signaling. CUDC-101 also reduced the levels of estrogen receptor protein in breast cancer cells. These results further support the potential of CUDC-101 to escape drug resistance in its future clinical application.
In summary, our results indicate that CUDC-101, with its unique and novel multitarget activities, is suited for future clinical development as a potential means to improve marketed EGFR/HER2 kinase inhibitors in various cancer indications. Importantly, this report provides a proof-of-principle that one can create a single small molecule with multiple targeted specificities to improve the effectiveness of current anticancer therapeutics.
Disclosure of Potential Conflicts of Interest
All authors are employees of Curis, Inc.
We thank Mark Noel, Joseph Davie, Stuart Aaronson, Kenneth Pienta, and George Vande Woude for their review of this manuscript and Nicole Davis for her assistance with the manuscript preparation.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
- Received September 9, 2009.
- Revision received January 26, 2010.
- Accepted February 4, 2010.
- ©2010 American Association for Cancer Research.