Effect of the histone deacetylase inhibitor LBH589 against epidermal growth factor receptor–dependent human lung cancer cells

The epidermal growth factor receptor (EGFR) is an important therapeutic target in non–small cell lung cancer (NSCLC) given its role in regulating diverse oncogenic pathways, such as cell proliferation, survival, angiogenesis, and invasion (1). Gefitinib and erlotinib are small-molecule EGFR tyrosine kinase inhibitors (TKI) that act by inhibiting EGFR kinase activity and can in some instances inhibit tumor cell growth and/or induce tumor cell apoptosis. Erlotinib monotherapy has been shown to extend survival for previously treated patients with NSCLC (2). Importantly, somatic mutations in the kinase domain of EGFR are important components of oncogenic survival of lung cancer cells by regulating important antiapoptotic pathways, including Akt and signal transducers and activators of transcription (STAT) pathways (3–7). Despite initial success with EGFR TKI therapy, patients develop resistance to these agents, yet in some cases, the tumor cells seem to maintain dependence on EGFR for growth and/or survival (8, 9). Therefore, identifying alternative therapeutic approaches to disrupting EGFR-dependent tumor cell growth and/or survival is an important and clinically relevant problem.

One alternative therapeutic approach to lung cancer cells harboring EGFR mutations is to target EGFR for degradation. Many kinases that contribute to deregulated signaling and cell growth in human cancers rely on the heat shock protein 90 (Hsp90) chaperone function for conformational maturation (10, 11). The molecular chaperone complex containing Hsp90 is essential for the stability and function of client proteins necessary for cellular homeostasis, including EGFR, Akt, and B-Raf (12–15). Hsp90 inhibitors interact specifically with a single molecular target causing instability and degradation of client proteins. Hsp90 inhibitors have shown promising antitumor activity in preclinical models. Geldanamycin and derivatives 17-(allylamino)-17-demethoxygeldanamycin and 17-dimethylaminoethylamino-17-demethoxygeldanamycin hydrochloride (17DMAG) can inhibit the activity of oncogenic proteins that are Hsp90 clients, and this can in some instances reverse the malignant phenotype of cancer cells (11, 16, 17). Among the functions of Hsp90 in cellular chaperoning, it also maintains a high level of expression in EGFR-expressing NSCLC particularly for EGFR-expressing cells with mutations in the kinase domain (15). Hsp90 was also recently shown to play a key role in maintaining the active confirmation of EGFR mutants and preventing ligand-induced receptor down-regulation (18). Importantly, both of these studies also showed activity of Hsp90 inhibitors against lung cancer cells with acquired resistance to EGFR TKI (15, 18).

Hsp90 activity can be regulated by reversible acetylation, a posttranslational modification often associated with histones and chromatin and an important mechanism by which protein activities are regulated (19–21). Histone deacetylase inhibitors (HDACi) can induce acetylation of Hsp90, resulting in the inhibition of both ATP binding and chaperone function (22). In human leukemia cells, this promotes the polyubiquitylation and degradation of the progrowth and prosurvival client proteins Bcr-Abl, c-Raf, and Akt (22). In addition, HDACi inhibits the chaperone function of Hsp90 promoting polyubiquitylation and degradation of the growth and survival proteins in human leukemia cells dependent on mutant FLT-3 (23). Importantly, combination of the HDACi or a direct Hsp90 inhibitor with kinase inhibitors can result in synergistic tumor death in tumors dependent on mutant kinases for survival (24). These results suggest that inhibition of Hsp90 function may provide alternative antitumor mechanisms to HDACi beyond their more traditional transcriptional mechanisms.

Based on these results, we hypothesized that the HDACi may selectively induce apoptosis in lung cancer cells dependent on EGFR for survival through Hsp90 acetylation and disruption of Hsp90 chaperone function with EGFR and other oncogenic proteins. The ability of HDACi to deplete Akt, Src, and Raf in cell types of other models also suggested to us possible efficacy against cells with acquired resistance to EGFR TKI. Finally, combined depletion of EGFR and downstream proteins by HDACi in conjunction with inhibition of EGFR kinase function by EGFR TKI may have a more profound antitumor effect than EGFR TKI alone.

All cell lines were obtained from the American Type Culture Collection and maintained in RPMI 1640 plus 5% bovine calf serum (Life Technologies) with the following exceptions. HCC827 cells were provided by Dr. Jon Kurie (M. D. Anderson, Houston, TX) and grown in 10% fetal bovine serum; PC9 cells were provided by Dr. Matt Lazzara (Massachusetts Institute of Technology, Boston, MA); H820, H226, and H157 cells were provided by Dr. John Minna (UT Southwestern, Dallas, TX); and H322 cells were provided by Dr. Paul Bunn (University of Colorado Health Sciences Center, Denver, CO). H1975 and H1650 cells were grown in RPMI 1640 supplemented with 0.15% sodium bicarbonate, 0.45% glucose, 10 mmol/L HEPES, 1 mmol/L pyruvate, and 10% bovine calf serum.

Cytotoxicity assays [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] were done according to the manufacturer's recommendations (Cell Proliferation kit, Roche). Cells (1 × 10³ to 5 × 10³) in 5% bovine calf serum complete medium were placed into single wells in a 96-well plate from FALCON and exposed to indicated agents, and viability was assessed following 120-h incubation. The IC₅₀ was defined as the drug concentration that induced a 50% reduction in cellular viability in comparison with DMSO controls normalized to 1 and was calculated by nonlinear regression analyses using MATLAB scripts that pair data points with sigmoidal curves that predict a signal response based on a four-variable fit. Data presented represent three separate experiments with at least 10 data points separating each dose per condition. Data are expressed as mean ± SD.

For combination assays using both LBH589 and erlotinib, cells (1 × 10³) in 5% bovine calf serum complete medium were placed into 96-well plates and exposed to indicated agents. Cells were treated with LBH589 and erlotinib at a fixed ratio with concentrations for 120 h, after which viable cells were assayed using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide. Analysis of the dose-effect relationship for LBH589- and erlotinib-induced apoptosis of NSCLC cells was done according to the median effect method of Chou and Talalay using CalcuSyn software (Biosoft; ref. 25). The combination index (CI) values were calculated for three independent experiments. CI < 1, CI = 1, and CI > 1 represent synergism, additivity, and antagonism of the two agents, respectively.

Stock solutions of 17DMAG were prepared in DMSO at a concentration of 10 mmol/L and maintained at −20°C. Drugs were diluted to 1 mmol/L in DMSO for a working solution and used at concentrations ranging from 3.9 to 1,000 nmol/L. Stock solutions of erlotinib and 17DMAG in 100% DMSO and LBH589 in sterile PBS were diluted directly into the medium to indicated concentrations. Erlotinib was provided by Genentech. PS341 was obtained from Millennium Pharmaceuticals. 17DMAG was a gift from Kosan Biosciences. LBH589 was provided by Novartis Pharmaceuticals.

Briefly, 5 × 10⁵ cells were resuspended in 70% ethanol for at least 12 h at 4°C. After washing with PBS, the cells were suspended in high-molarity phosphate-citrate buffer at room temperature for 5 min for fractional (sub-G₁) DNA content assay, treated with 50 μg/mL of RNase A (Sigma) at 37°C for 30 min, and then exposed to 50 μg/mL of propidium iodide (Roche) at 4°C for 20 min in the dark. Cells were analyzed in a flow cytometer (FACS 420, Becton Dickinson) based on the manufacturer's instructions. Apoptosis was assayed using a cleaved poly(ADP-ribose) polymerase (PARP) antibody from Cell Signaling. The cleaved PARP (Asp²¹⁴) antibody detects endogenous levels of the large fragment (89 kDa) of human PARP1 produced by caspase cleavage. The antibody does not recognize full-length PARP1 or other PARP isoforms.

Cell lysates were prepared using radioimmunoprecipitation assay buffer [10 mmol/L Tris (pH 7.4), 100 mmol/L NaCl, 1 mmol/L EDTA, 1 mmol/L EGTA, 1 mmol/L NaF, 20 mmol/L Na₄P₂0₇, 2 mmol/L Na₃VO₄, 0.1% SDS, 0.5% sodium deoxycholate, 1% Triton X-100, 10% glycerol, 1 mmol/L phenylmethylsulfonyl fluoride, 60 μg/mL aprotinin, 10 μg/mL leupeptin, 1 μg/mL pepstatin], normalized for total protein content (30 μg), and subjected to SDS-PAGE. Detection of proteins was accomplished using horseradish peroxidase–conjugated secondary antibodies and enhanced chemiluminescence purchased through Amersham Biosciences.

Primary antibodies used in these studies consisted of total EGFR, phosphorylated Ser⁴⁷³ Akt, total Akt, phosphorylated Tyr⁷⁰⁵ STAT3, total STAT3, c-Src, PIM-1, antiacetyl lysine antibody, antiacetyl tubulin antibody, and cleaved PARP, Bcl-2, Bcl-xL, and Mcl-1 (Santa Cruz Biotechnology), Hsp90 (StressGen Biotechnologies Corp.), and β-actin (Sigma).

Following the designated treatments, cells were lysed in the lysis buffer [20 mmol/L Tris (pH 8), 150 nmol/L sodium chloride, 1% NP40, 0.1 mol/L sodium fluoride, 1 mmol/L phenylmethylsulfonyl fluoride, 1 mmol/L sodium orthovanadate, 2.5 μg/mL leupeptin, 5 μg/mL aprotinin] for 30 min on ice, and the nuclear and cellular debris were cleared by centrifugation. Total cellular proteins were then quantified using the Bradford protein assay. Cell lysates (200 μg) were incubated with the Hsp90-specific monoclonal antibody for 1 h at 4°C. Washed protein G-agarose beads were added and incubated overnight at 4°C. The immunoprecipitates were washed thrice in the lysis buffer, and proteins were eluted with the SDS sample loading buffer before the immunoblot analyses with specific antibodies against monoclonal Hsp90 or other signaling proteins.

Significant differences between values obtained in a population of EGFR-expressing cells treated with different experimental conditions were determined using the Student's t test. P values of <0.05 were assigned significance.

We first determined the effects of LBH589 on Hsp90 acetylation in HCC827 lung cancer cells. LBH589 is a novel cinnamic hydroxamic acid analogue HDACi (26). HCC827 cells containing an activating deletion mutant of EGFR were exposed to 100 nmol/L LBH589 for 0, 8, 16, and 24 h, after which immunoprecipitation of Hsp90 followed by immunoblotting with an antibody that recognizes an acetylated lysine residues was done (antiacetyl lysine antibody; Fig. 1A). We observed acetylation of Hsp90 in a time-dependent manner without significant change on the levels of Hsp90 immunoprecipitated. Because acetylation is associated with functional inactivation of Hsp90, we next evaluated the effect of these modifications on the function of Hsp90 as a chaperone protein (19). Following exposure of HCC827 cells to LBH589 for 0, 8, 16, and 24 h, Hsp90 was immunoprecipitated from the cell lysates, and the binding of Hsp90 to EGFR, Akt, c-Src, and STAT3 was determined (Fig. 1B). At 8 h and beyond, EGFR as well as other codependent proteins were no longer associated with Hsp90 following treatment with LBH589, whereas Hsp90 levels remained unchanged. Although these results strongly suggest that LBH589 can inhibit Hsp90 function and deplete Hsp90 client proteins, HDACi could also be acting by modifying gene expression through acetylation of histones. To evaluate this possibility, we inhibited proteasome degradation with PS341 and again evaluated for changes in protein expression with LBH589 (Fig. 1C). Blockage of proteasomal degradation by PS341 prevents degradation of EGFR, c-Src, Akt, and STAT3 by LBH589, showing that the mechanism of LBH589-induced changes in protein expression requires protein degradation.

We next determined the effects of LBH589 on cell viability using a large panel of NSCLC with defined EGFR status and sensitivity to EGFR TKI. Cells were exposed to increasing concentrations of LBH589 or the EGFR TKI erlotinib, and cell viability was assessed (Fig. 2A). All of the tested cells are sensitive to LBH589 with approximate IC₅₀ ranging between 9 and 54 nmol/L. There was no obvious correlation between the EGFR mutation status and erlotinib IC₅₀ with the IC₅₀ for LBH589. Importantly, LBH589 inhibits cell viability in cells containing EGFR mutation that are nonetheless resistant to EGFR TKI, such as H1975, H1650, and H820, with IC₅₀ ranging between 10 and 20 nmol/L (27).

To assess how changes in downstream signaling relate to sensitivity of the lung cancer cells to LBH589, we evaluated changes in key signaling proteins as well as determined the effect on cell cycle progression and apoptosis. Cells were exposed to 100 nmol/L LBH589 for 24 h, and total proteins were evaluated for total EGFR, c-Src, Akt, and STAT3 as well as cleaved PARP indicative of apoptosis (Fig. 3A and B). The choice of these molecules was based on their role in mutant EGFR-dependent survival signaling (6, 7, 28–31). Cells were also exposed to 250 nmol/L 17DMAG to show the effects of a direct Hsp90 inhibitor and allow comparison with LBH589. These results show that LBH589 depletes EGFR, STAT3, c-Src, and Akt proteins in cells with EGFR mutation (Fig. 3A). In addition, all cells with EGFR mutation show apoptosis indicated by appearance of cleaved PARP. Conversely, wild-type (WT) EGFR cells showed minimal changes in protein levels and no appearance of cleaved PARP was evident, arguing against significant apoptosis at least at this time point (Fig. 3B). Instead, we observed reduced percentages of cells undergoing DNA synthesis with accumulation of cells in both G₁ and G₂-M phases of the cell cycle (Fig. 3C).

Interestingly, EGFR and Akt, both well-characterized Hsp90 clients, were not depleted in cells lacking activating EGFR mutations with the tested concentrations of LBH589 or 17DMAG (32). To explain this occurrence, we treated both A549 and H460 cells with either 100 or 1,000 nmol/L of LBH589 and observed changes on a 4-h incremented time course in protein expression of EGFR and Akt. To show changes in global acetylation within the cell and provide an indicator for drug activity within the cells, we used an antiacetyl tubulin antibody (Fig. 3D). Our results show that cells treated with 100 nmol/L LBH589 have transient depletion of both EGFR and Akt evident at 4 h but soon after the levels have recovered back to baseline. These effects mimic the global acetylation of the cell where acetyl tubulin peaks at 4 h and returns to baseline at 24 h. However, cells exposed to 1,000 nmol/L LBH589 have durable down-regulation of both EGFR and Akt expression that parallels the durable increase in acetyl tubulin expression. These results could suggest that metabolism of either LBH589 or 17DMAG could be different between the cells. Because prior studies indicated a positive relationship between DT-diaphorase gene expression and growth inhibition by Hsp90 inhibitors, we investigated the relative gene expression using previously reported microarray data from lung cancer cells with defined EGFR status and sensitivity to EGFR TKI (33, 34). We find an ∼3-fold higher DT-diaphorase (NQO1) expression in EGFR TKI-resistant cell lines as opposed to EGFR TKI-sensitive cell lines (Fig. 3E). Thus, in our lung cancer cells, we do not find a positive relationship between DT-diaphorase and either indirect (LBH589) or direct (17DMAG) Hsp90 inhibitors.

We next evaluated the effect of LBH589 on STAT3 survival protein status because STAT3 has been previously shown to regulate Bcl-2 family gene expression, including Bcl-xL, Mcl-1, Bcl-2, Pim-1, and others (35). NSCLC cells were exposed to either 100 nmol/L LBH589 or 250 nmol/L 17DMAG to show the effects of a direct Hsp90 inhibitor for 24 h, and total proteins were evaluated for total levels of STAT3 survival proteins (Bcl-xL, Bcl-2, and Mcl-1). In cells harboring EGFR mutations, levels of these proteins are depleted following treatment with LBH589 and 17DMAG (Fig. 4A). However, in cells lacking EGFR mutations, the levels of STAT3 survival proteins were minimally affected by LBH589 or 17DMAG treatment at the concentrations studied (Fig. 4B).

We next determined if similar events occur in other cell lines that are also dependent on STAT3. We chose to study fibroblast cells transformed by v-Src that require STAT3 for survival (36, 37). As shown in Fig. 4C, exposure of v-Src–transformed NIH-3T3 cells to LBH589 results in a concentration-dependent inhibition of cell viability. As seen previously in the EGFR-dependent lung cancer cells, Western blots run after cells were treated with 100 nmol/L LBH589 for 24 h and harvested for analysis suggests a reduction in STAT3 and STAT3-dependent survival protein expression was observed after exposure of the cells to 100 nmol/L LBH589 (Fig. 4D). Similar results are observed in cells exposed to the direct Hsp90 inhibitor 17DMAG. These experiments also show apoptosis in these cells evident by the appearance of cleaved PARP. These findings suggest that LBH589 can affect STAT3 levels and function in cell lines dependent on upstream tyrosine kinases and loss of STAT3 contributes to apoptosis (29).

Previous studies in kinase-dependent leukemia cells showed synergistic apoptosis with Hsp90 inhibitors combined with kinase inhibitors (24). Similarly, other studies have shown synergy with combined EGFR TKI and HDACi (38). To determine if similar events occur in EGFR-dependent lung cancer cells, we determined the effects of cotreatment with LBH589 and erlotinib, an EGFR TKI. We selected four cell lines sensitive to both LBH589 and erlotinib (HCC827, H292, H358, and H322) and exposed them to increasing concentrations of erlotinib, LBH589, or the combination in a fixed ratio. Cell viability was assessed using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, and the effect was analyzed using median effect analysis (25). Importantly, exposure to the combination of erlotinib and LBH589 exerted additive or synergistic effect in all cell lines as determined by the median dose-effect analysis, which revealed CI values of generally <1.0 ± 0.02. Synergy between LBH589 and erlotinib was most evident at a dose range from 7.8 to 125 nmol/L, with most CI values below 0.5 ± 0.02 (Fig. 5B).

We next evaluated the effects of LBH589, erlotinib, or the combination on downstream signaling proteins. HCC827 cells were exposed to concentrations of erlotinib, LBH589, or the combination of both agents for 24 h and protein lysates were evaluated for phosphorylated as well as total levels of Akt and STAT3 as well as cleaved PARP indicative of apoptosis (Fig. 5C). Erlotinib inhibits phosphorylated Akt while having no effect on total Akt, has a modest effect on phosphorylated STAT3 without any effect on total STAT3, and induces apoptosis evident by PARP cleavage. LBH589 results in reduced levels of phosphorylated Akt and total Akt and phosphorylated STAT3 and total STAT3 and induces apoptosis evident by PARP cleavage. The combination of both erlotinib and LBH589 results in complete inhibition of phosphorylated Akt and phosphorylated STAT3 along with depletion of total Akt and STAT3; this results in increased amounts of cleaved PARP corresponding to the reduced levels of cell viability seen with combined treatment. This result showed that combined EGFR TKI, along with depletion of EGFR and other key downstream proteins, such as Akt and STAT3, results in synergistic apoptosis. These results along with the result showing an effect of LBH589 on STAT3 and STAT3-dependent survival proteins may suggest a mechanism in which LBH589 and an EGFR TKI synergistically induce apoptosis.

Our results show that the HDACi LBH589 can induce acetylation of Hsp90, resulting in reduced association of Hsp90 with key chaperone proteins, including EGFR, c-Src, STAT3, and Akt. Analyses of apoptosis, cell cycle, and key signaling proteins necessary for survival of NSCLC cells showed a potential mechanism of action through the Hsp90 chaperone function. In NSCLC harboring EGFR mutations, LBH589 treatment results in depletion of Hsp90-dependent chaperone clients important in antiapoptotic signaling and STAT3-dependent proteins along with the induction of cell death. In NSCLC WT cell lines, LBH589 treatment results in no evident apoptosis after 24 h, minimal changes in key Hsp90 clients or STAT3-dependent proteins, but rather decreased cell numbers due to cell cycle arrest. The depletion of Hsp90 clients was observed to have a transient effect depending on the timing, drug concentration, and levels of global acetylation. The effect of LBH589 on NSCLC cell lines was observed at an IC₅₀ in the nanomolar range and combined LBH589 and erlotinib results in synergistic effects. Therefore, LBH589 represents a novel hydroxamic acid-derived HDACi with strong activity against multiple NSCLC with EGFR-mutant cells undergoing depletion of EGFR and other client proteins resulting in apoptosis.

Although acetylation is commonly known for modifications that occur on the histone proteins important for the regulation of chromatin structure and chromatin-directed activities such as transcription, it has become more evident that proteins other than histones are regulated by protein acetylation. Posttranslational modifications, such as acetylation, can affect protein functions, including cellular distribution as well as the ability to interact with other proteins, DNA, and RNA (21). HDAC inhibition results in acetylation not only of histones but also of transcription factors such as p53 that can modify their function (39–41). Our results in EGFR-mutant cells show that inhibition of HDAC function by LBH589 results in increased acetylation of Hsp90 and disrupts its association with EGFR and other important oncogenic proteins resulting in cell death. Maintaining the equilibrium between binding and unbinding of the Hsp90 molecular chaperone complex to critical oncogenic proteins is likely to be important for proper progression of oncogenic cellular reactions. Here, we have shown that LBH589 mediated acetylation and inhibition of Hsp90 as a chaperone protein as a mechanism for depletion of EGFR in NSCLC cell lines. The findings presented here show that EGFR is coimmunoprecipitated with Hsp90 and that this association is inhibited following treatment with LBH589. Data presented here also show that increased acetylation of Hsp90 has other functional consequences, including the depletion of specific progrowth- and prosurvival-associated Hsp90 client proteins, such as Akt, c-Src, and STAT3. We also considered that RAS mutation status could be relevant to the effects seen with LBH589 in these lung cancer cells. All the cells tested with EGFR mutation lack a RAS mutation with the exception of the H820 cell that contains an activating EGFR mutation as well as K-RAS mutant (codon 12; refs. 42–44). However, the H322 and H292 cells have been reported to have WT EGFR and RAS and we observe no apoptosis in these cells. Therefore, EGFR mutation status rather than RAS mutation status seems to be the more relevant biomarker indicative of apoptosis in these cells.

It is well known that HDACis can have profound effects on signaling proteins through effects on transcription in addition to effects mediated through Hsp90 client proteins (22, 23, 45). The effects of HDACis on gene expression are thought to be highly selective, leading to transcriptional activation of genes, including the cyclin-dependent kinase inhibitor p21, but also repression of other genes important in cell growth control. These effects likely explain the universal sensitivity of our cells to LBH589 and explain changes seen in cell cycle progression. However, our results clearly show that inhibition of proteasomal degradation prevents depletion of EGFR and other Hsp90-dependent client proteins, further strengthening the posttranscriptional mechanism of action in EGFR-mutant cells (Fig. 1C).

Our studies on the mechanism of action of LBH589 thus indicate that this compound affects several routes involved in the control of signaling pathways, cellular growth, and apoptosis. Nonetheless, other mechanisms beyond Hsp90 inhibition likely exist and Hsp90 acetylation may not be the only mechanism by which the compounds exert antitumor effects.

Our results also suggest that Hsp90 inhibition from LBH589 suppressed the levels of STAT3 as well as STAT3-dependent survival protein expression in both lung cancer cells dependent on EGFR and fibroblast cells dependent on v-Src. Inhibitory effects of Hsp90 on STAT3 activation were confirmed using 17DMAG, a Hsp90-specific inhibitor. Our studies are supported by previous studies showing that Hsp90 directly interacted with STAT3 via its NH₂-terminal region (14). Given the important role of STAT3 in the pathophysiology of disease states, such as cancer, or autoimmune diseases, indirect inhibition of STAT3 via Hsp90 inhibition may significantly affect the progression of these diseases (35, 46). More detailed understanding of the cross-talk between Hsp90 and STAT3 as well as other STAT proteins is required as this new information may provide new therapeutic approaches for these and other pathologic conditions.

There are some possible clinical implications of our work. First, our data show that LBH589 has potent activity especially against lung cancer cells dependent on EGFR for survival. Notably, this includes lung cancer cells with acquired resistance to EGFR TKI (H1975) and, along with other reports, suggests that targeting Hsp90 may be an effective salvage regimen for patients relapsing following successful treatment with EGFR TKI (15, 18). Similar results have been shown in myeloma cells where LBH589 can overcome drug resistance and potentiate the effect of other drugs (47). Our results suggest that dosing and schedule of LBH589 may be relevant to clinical trial outcomes. Low concentrations of LBH589 (100 nmol/L) can affect Hsp90 function, depletion of EGFR, and apoptosis only in EGFR-mutant cells. This effect may be related to “oncogenic shock” where tumor cells dependent on EGFR undergo apoptosis through an imbalance between prosurvival and prodeath pathways (48, 49). An initially transient depletion of EGFR and other prosurvival signaling proteins by LBH589 is sufficient to shift the balance between survival and death signals and triggers rapid apoptosis. Thus, rates of overt tumor regressions with continuous dosing that obtains these levels in patients would be correlated best with rates of EGFR mutation (∼10–20%) based on these preclinical data. Patients whose tumors do not harbor activating EGFR mutations would be predicted to have stable disease under similar dosing regimens but possibly tumor regressions via apoptosis mediated by changes in expression in signaling proteins only with higher concentrations of LBH589. Similar to previous findings, the effects of combined erlotinib and LBH589 may also suggest a benefit of this combination for treatment of patients with EGFR kinase mutations (38). Early-phase trials of LBH589 in patients with hematologic malignancies have identified both biological effects and effects on tumor cell burden, and studies in patients with solid tumors are ongoing (26). The action of LBH589 on resistant NSCLC cell lines and on cell lines harboring EGFR TKI-sensitive mutations, together with its capacity to potentiate other EGFR inhibitor agents, would support the clinical evaluation of LBH589 in NSCLC.

We thank Drs. Jon Kurie, Matthew Lazzara, John Minna, and Paul Bunn for providing cell lines; members of the Bhalla laboratory with assistance with the Hsp90 immunoprecipitation studies; Dr. Ed Seto (Moffitt Cancer Center, Tampa, FL) for providing acetylation reagents; Dr. Penni Black (University of Kentucky, Lexington, KY) for providing microarray analysis; and Tiffany Dyn for administrative assistance.