Antitumor activity and mechanism of resistance of the novel HDAC and PI3K dual inhibitor CUDC‑907 in pancreatic cancer
Shuang Liu1 · Shoujing Zhao1 · Yang Dong1 · Tingting Wang1 · Xiaojia Niu1 · Lijing Zhao2 · Guan Wang1

Received: 24 July 2020 / Accepted: 24 November 2020 / Published online: 3 January 2021
© Springer-Verlag GmbH Germany, part of Springer Nature 2021

Abstract
Purpose Pancreatic cancer is a highly malignant disease with an extremely poor prognosis. The benefit of chemotherapy treatment for pancreatic cancer is very limited. Therefore, new therapeutic targets and approaches are urgently needed for this deadly disease. Multi-target therapy is a potential and feasible treatment strategy. Given the important roles that histone deacetylases (HDACs) and phosphoinositide-3-kinase (PI3K) play in pancreatic cancer, we investigated the antitumor activ- ity and mechanism of novel HDAC and PI3K dual inhibitor CUDC-907 in pancreatic cancer.
Methods and results MTT assay and flow cytometric analysis were used to examine the in vitro antitumor activity of
CUDC-907. A BxPC-3-derived xenograft mouse model was used to determine CUDC-907 in vivo efficacy. The TUNEL assay as used to determine apoptosis in tumors in vivo post CUDC-907 treatment. Western blots were used to determine the effect of CUDC-907 on protein levels. Our results show that CUDC-907 decreased viable cells and induced cell death in a concentration-dependent manner. Furthermore, CUDC-907 showed promising in vivo antitumor activity in the BxPC- 3-derived xenograft mouse model while exhibiting tolerable toxicity. Furthermore, long-term treatment with CUDC-907 induced phosphorylation of AKT, S6 (ribosomal protein S6), and ERK (extracellular regulated protein kinase), and inhibition of PI3K (phosphatidylinositol 3-kinase), mTOR (mammalian target of rapamycin), or ERK significantly enhanced CUDC- 907-induced cell death in pancreatic cell lines.
Conclusion Taken together, these findings support the clinical development of CUDC-907 for the treatment of pancreatic
cancer and identify compensatory activation of mTOR and MEK/ERK as a possible mechanism of resistance to CUDC-907.
Keywords CUDC-907 · Antitumor activity · Combination treatment · VS-5584 · Pancreatic cancer

Introduction
Because of the lack of early diagnosis and effective treat- ment, pancreatic cancer remains a highly lethal disease with 5 year survival rates < 10% [1, 2]. Gemcitabine-based chem- otherapy is the current standard of care, but with little effect [3]. Therefore, there is an urgent need for new therapeutic regimens to improve the survival rate of patients with pan- creatic cancer. Histone deacetylases (HDACs) are commonly

 Guan Wang [email protected]
1 Key Laboratory for Molecular Enzymology and Engineering, National Engineering Laboratory for AIDS Vaccine, The Ministry of Education, School of Life Sciences, Jilin University, 2699 Qianjin Street, Changchun, Jilin, China
2 Department of Rehabilitation, School of Nursing, Jilin University, Changchun, China

overexpressed in pancreatic cancer cells and are considered potential targets for the treatment of pancreatic cancer. The use of HDAC inhibitors (HDACIs) has been widely studied in preclinical and clinical models of pancreatic cancer, but these inhibitors have shown limited antitumor efficacy when administered as monotherapy [4]. Hyperactivation of phos- phatidylinositol 3-kinase (PI3K) is one of the most frequent events in human cancers, making PI3Ks important targets for cancer therapy [5, 6]. PI3Ks are also important targets in pancreatic cancer and are critical for pancreatic cancer development [7–9]. On the basis of the synergistic antitumor activity of HDACIs and PI3K inhibitors, CUDC-907 was designed to simultaneously inhibit HDAC (class I, II, and IV) and PI3K (class α, β, and δ). CUDC-907 inhibits the PI3K-AKT-mTOR (mammalian target of rapamycin) path- way, inactivates RAF-MEK (mitogen-activated extracellular signal-regulated kinase)-ERK (extracellular regulated pro- tein kinase), and has shown potent antitumor activity against

a variety of tumors [10]. Clinical trials on relapsed/refrac- tory diffuse large B-cell lymphoma, lymphoma, and multiple myeloma have revealed that CUDC-907 has a satisfactory toxicity profile as well as clinical activity [11, 12]. There is only one published report on CUDC-907 in pancreatic cancer [13]. The knowledge of the antitumor mechanism of CUDC-907 is very limited and the mechanism of resistance is unknown in pancreatic cancer.
In this study, we confirmed the antitumor activity of CUDC-907 both in vitro and in vivo in preclinical pancreatic cancer models. Furthermore, we identified the mechanism of resistance to CUDC-907, which we believe involves the mTOR and MEK/ERK pathways. Our results support the clinical evaluation of CUDC-907 for treating pancreatic cancer.

Materials and methods
Drugs

CUDC-907, rapamycin (an mTOR inhibitor), KU-0063794 (an mTORC1 and mTORC2 inhibitor), VS-5584 (a PI3K and mTOR dual inhibitor), and SCH772984 (an ERK- selective inhibitor) were purchased from Selleck Chemicals (Houston, TX, USA).
Cell lines and treatments

AsPC-1, BxPC-3, CFPAC-1, HPAC, MIAPaCa-2, and
PANC-1 human pancreatic cancer cell lines were purchased from the American Type Culture Collection (ATCC; Manas- sas, VA, USA) and were authenticated by the University of Arizona Genetics Core Facility (Tucson, AZ, USA). Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen, Carlsbad, CA, USA; HPAC and MIA- PaCa-2), RPMI-1640 medium (Invitrogen; AsPC-1, BxPC- 3, and PANC-1), or Iscove’s modified Dulbecco’s medium (IMDM, Invitrogen; CFPAC-1) with 10% heat-inactivated fetal bovine serum (FBS, HyClone Laboratories, Logan, UT, USA), 100 U/mL penicillin, and 100 µg/mL streptomycin in a 37 °C humidified atmosphere containing 5% CO2/95% air. Cell lines were tested for the presence of mycoplasma every 2 weeks.
MTT assays

The in vitro effect of CUDC-907 on the number of viable cells was determined using 3-(4,5-dimethylthiazol-2-yl)- 2,5-diphenyltetrazolium bromide (MTT) (Sigma-Aldrich, St. Louis, MO, USA) reagent, as previously described [14]. Briefly, pancreatic cancer cell lines were cultured in 96-well plates with 0–320 nM CUDC-907 for 72 h, and viable cells
were determined using MTT reagent by measuring absorb- ance at OD 590 nm. IC50 values were calculated as drug concentrations necessary to inhibit cell growth by 50% com- pared to that in vehicle control-treated cells using GraphPad Prism 5.0 software (GraphPad Software, San Diego, CA, USA).
Detection of cell death

Pancreatic cancer cell lines were treated with the indicated drugs for 48 h. The cells were fixed with ice-cold 80% (v/v) ethanol for 24 h at 4 °C. The cells were then pelleted, washed with phosphate-buffered saline (PBS, pH 7.4), and resus- pended in PBS containing propidium iodide (PI; 50 µg/mL), Triton X-100 (0.1%, v/v), and DNase-free RNase (1 µg/mL). DNA content was determined by flow cytometric analysis using a FACSCalibur flow cytometer (Becton–Dickinson, San Jose, CA, USA), as previously described [14]. Cell death events were expressed as the mean percentage of cells with sub-G1 DNA content ± standard error of triplicates. The experiment was repeated three times, and data from one representative experiment are shown.
Trypan blue exclusion test

Pancreatic cancer cell lines were treated with CUDC-907 alone or in combination with VS-5584 for 48 h. The cells were harvested and trypan blue solution (Sigma-Aldrich) was added to the cell suspensions in a ratio of 1:1. Total cells and dead cells (stained in blue) were counted using haema- cytometer. The percentage of dead cells was calculated. The experiments were repeated three times and the results are present as mean ± standard error.
Western blot analysis

Soluble proteins were extracted in the presence of protease and phosphatase inhibitors (Roche Applied Sciences China Inc., Shanghai, China) and then subjected to SDS–poly- acrylamide gel electrophoresis. Separated proteins were electrophoretically transferred onto polyvinylidene dif- luoride (PVDF) membranes (Thermo Fisher Inc., Rock- ford, IL, USA) and immunoblotted with anti-H4 (ab10158),
-ac-α-tubulin (ab24610), -p-ERK (T202/Y204) (ab214362) (Abcam, Cambridge, MA, USA), -ac-H4 (06-598) (Milli- pore, Billerica, MA, USA), -p-AKT (S473) (#9271), -p-AKT (T308) (#9275), -AKT (#9272), -p-S6 (S240/244) (#5364),
or -GAPDH (#97166) (Cell Signaling Technology, Danvers, MA, USA) antibodies, as previously described [14]. Pri- mary antibodies were diluted 1:1000 in an Odyssey blocking buffer (Li-Cor, Lincoln, NE, USA). Immunoreactive proteins were visualized using the Odyssey Infrared Imaging Sys- tem (Li-Cor). Fold changes of densitometry measurements,

normalized to GAPDH and then compared to the no-drug control (set as 1), are indicated below the corresponding blot.
Establishment of a pancreatic cancer cell line‑derived xenograft mouse model

Female BALB/c nude mice (18–22 g) were purchased from Vital River Laboratories (Beijing, China). BxPC-3 cells were adjusted to a density of 2 × 107 cells/mL with Matrigel (BD Biosciences, San Jose, CA, USA) and then subcutaneously inoculated into the right axilla of the mice (0.1 mL/mouse). Once the tumor diameter reached approximately 0.5 cm, the tumor was isolated and cut into small pieces (1 mm in diam- eter). The right axilla skin of the mice was then punctured to form a 5-mm-long sinus tract, where the tumor fragment was subcutaneously inserted. When the xenografts reached a vol- ume of 169.8 ± 7.1 mm3, the mice were randomized into two groups (7 animals in the vehicle control group and 6 animals in the CUDC-907 group). The mean tumor volumes were
171.9 ± 10.6 and 167.3 ± 10.1 mm3 for the vehicle control
and CUDC-907 group, respectively. The mice in the vehicle control and CUDC-907 groups were treated with (i) vehicle control (0.5% methyl cellulose (W/V), 1% DMSO (V/V) and sterile water) and (ii) 10 mg/kg CUDC-907, adminis- tered by intraperitoneal injection, for 20 days, respectively. Both control and experimental mice were treated once daily on Monday and Wednesday. The tumor diameters and mice body weights were measured every 1–3 days. Once a single tumor from all 13 mice reached 1500 mm3, the mice were sacrificed and the tumors were removed and fixed for immu- nohistochemical analysis.
Hematoxylin and eosin (H&E) staining, PCNA immunohistochemical analysis, and TUNEL staining

Fixed tumor tissues were subjected to H&E staining, pro- liferating cell nuclear antigen (PCNA) immunohistochemi- cal analysis, and TUNEL staining, as previously described [14]. Slides were then analyzed using a microscope, and brown staining was scored using Image-Pro Plus 6.0 (Media Cybernetics, Inc., Bethesda, MD, USA). The TUNEL assay was performed using the DeadEnd™ Fluorometric TUNEL System kit (Promega, Madison, WI, USA) according to the manufacturer’s protocol. The proliferation index was cal- culated as PCNA positive cells/observed cells × 100%. The apoptosis index was calculated as TUNEL-positive cells/ observed cells × 100%.
Statistical analysis

Differences were compared using the two-sample t test. Sta- tistical analyses were performed using GraphPad Prism 5.0.
Error bars represent the mean ± SEM. The level of signifi- cance was set at p < 0.05.

Results
Antitumor activity of CUDC‑907 in pancreatic cancer cell lines

We first aimed to investigate the effect of CUDC-907 on the acetylation of histone H4 and alpha-tubulin and on the PI3K- AKT-mTOR signaling pathway in three pancreatic cancer cell lines. Treatment of the cell lines with 1 μM CUDC-907 for 1 h resulted in a decrease in phosphorylated ribosomal protein S6 (S6) at Ser240/244, a downstream target of the PI3K-AKT-mTOR pathway. Hyperacetylation of histone H4 and α-tubulin, substrates of HDACs (Fig. 1a) was also observed. This confirms the dual inhibitor properties of CUDC-907 on HDAC and PI3K.
To determine the antitumor activity of CUDC-907, the effect of CUDC-907 on cell viability was determined in a panel of pancreatic cancer cell lines using MTT assays after 72 h of drug treatment. CUDC-907 treatment led to a concentration-dependent decrease in viable cells with IC50 values ranging from 36 to 318 nM (Fig. 1b, c). To determine the effect of CUDC-907 on cell death, pancreatic cancer cells were treated with CUDC-907 for 72 h, stained with PI and then subjected to flow cytometry analyses to detect the percentage of cells with sub-G1 DNA content. As shown in Fig. 1d, CUDC-907 treatment significantly induced cell death in all 3 cell lines, although to a lesser extent in MIA- PaCa-2 and BxPC-3 cell lines than in CFPAC-1 cells. The same trends were observed using trypan blue exclusion test (Fig. 1e).
In vivo antitumor activity of CUDC‑907
in a BxPC3‑derived xenograft mouse model

To test the in vivo antitumor activity of CUDC-907, we established a BxPC-3-derived xenograft model in nude mice. At the end of the animal trial (day 20 post the first CUDC-907 treatment), results showed that CUDC-907 treat- ment significantly delayed tumor growth compared to that in the vehicle control-treated group (p < 0.01), with a tumor inhibition rate [Treated/control (%)] of 52% (Fig. 2a). The body weight loss of the CUDC-907 treatment group was less than 10%, indicating that CUDC-907 treatment was well tolerated (Fig. 2b). To further investigate the in vivo effects of CUDC-907 treatment, tumors were analyzed by H&E staining, PCNA immunohistochemical staining, and TUNEL staining. CUDC-907 treatment resulted in increased tumor necrosis, as indicated by H&E staining (Fig. 2c); decreased proliferation, as indicated by PCNA staining and

Fig. 1 CUDC-907 impairs cell viability and induces cell death in pancreatic cancer cell lines. a BxPC-3, CFPAC-1, and MIAPaCa-2 cells were treated with vehicle control or 1 μM CUDC-907 (CUDC) for 1 h. Whole cell lysates were subjected to Western blotting and probed with the indicated antibodies. The experiment was repeated two times and representative blots are shown. b, c Pancreatic can- cer cell lines were cultured with variable concentrations of CUDC for 72 h, and then subjected to MTT assay. Data are presented as means ± standard errors from at least 3 independent experiments. d,
e BxPC-3, CFPAC-1, and MIAPaCa-2 cells were treated with vari- ous concentrations of CUDC for 48 h. Cells were subjected to trypan blue exclusion test or fixed, stained with PI, and then analyzed by flow cytometry. Dead cells are expressed as the percentage of cells with sub-G1 DNA content shown in panel D, or with blue staining shown in panel E. The experiments were repeated twice and data are presented as means of triplicates ± standard errors from one repre- sentative experiment. *Indicates p < 0.05, **indicates p < 0.005, and
***indicates p < 0.001 compared to no drug treatment control

proliferation index (Fig. 2d, e); and increased apoptosis, as measured by the TUNEL assay and calculation of the apop- totic index (Fig. 2f, g). Taken together, our results show that CUDC-907 has antitumor activity against pancreatic cancer in vivo.
Effect of CUDC‑907 on the PI3K/mTOR and MEK/ERK signaling pathways

Next, we investigated the effect of CUDC-907 on pancre- atic cancer cells at the molecular level. Treatment of the cell lines with 20–160 nM CUDC-907 for 48 h resulted in hyperacetylation of histone H4 but had no effect on α-tubulin acetylation. This indicates that CUDC-907 had no inhibi- tory activity against HDAC6 under this treatment condition. Notably, CUDC-907 increased the phosphorylation of S6 at Ser240/244 and ERK at Thr202/Tyr204, indicating acti- vation of the PI3K/mTOR and MEK/ERK pathways. As a
PI3K inhibitor, CUDC-907 treatment resulted in a modest decrease in p-AKT (T308) (Fig. 3). In contrast, CUDC-907 treatment for 48 h led to increased phosphorylation of AKT at S473 (Fig. 3). These results are contrary to the expected effect of CUDC-907 on the PI3K/mTOR and MEK/ERK pathways, potentially due to complex feedback loops and crosstalk among these pathways.
Effect of PI3K, mTOR, and/or ERK inhibition on CUDC‑907 antitumor activity

Results shown in Fig. 3 indicate that induction of mTOR and ERK activation by prolonged treatment with CUDC- 907 at lower concentrations may represent a mechanism of resistance to this agent in pancreatic cancer cells. To inves- tigate the roles of mTOR and ERK activation in the antitu- mor activity of CUDC-907 in pancreatic cancer, we used the mTOR inhibitor rapamycin [15], mTORC1/2 inhibitor

Fig. 2 CUDC-907 displays potent antitumor activity in a BxPC-3 xenograft mouse model. Tumor volumes were calculated according to the following formula:
m12 × m2 × 0.5236 (m1: short
diameter; m2: long diameter) (a). Body weights were meas- ured every 1–3 days (b). Tumor specimens were fixed in 10% formalin, embedded in paraffin, and cut into 4 μM-thick sections for H&E (c) and PCNA (d) staining. The proliferation index was graphed as means ± stand- ard errors (e). Apoptosis was measured using the TUNEL assay in tumor specimens
(f). The apoptosis index was graphed as means ± standard errors (g). **Indicates p < 0.005 when compared to vehicle control

KU-0063794 [16], PI3K/mTOR dual inhibitor VS-5584 [17], and ERK inhibitor SCH772984 [18], and used them in combination with CUDC-907 to treat MIAPaCa-2 cells. Interestingly, all of these inhibitors significantly enhanced CUDC-907-induced cell death, with VS-5584 causing maximum cell death (Fig. 4a). VS-5584 also significantly enhanced CUDC-907-induced cell death in BxPC-3 and CFPAC-1 cells (Fig. 4b). Trypan blue exclusion test was used to verify the combined effect of CUDC-907 and
VS-5584 on cell death (Fig. 4c). These results suggest that simultaneous targeting of PI3K and mTOR may significantly enhance the antitumor activity of CUDC-907 in pancreatic cancer cells.
To confirm that the effect of VS-5584 on the antitumor activity of CUDC-907 was on-target, we determined the effects of these compounds on the PI3K-AKT-mTOR and MEK/ERK pathways. Results show that VS-5584 alone and VS-5584 in combination with CUDC-907 substantially

Fig. 3 CUDC-907 treatment decreases p-AKT (T308) and increases p-AKT (S473) and p-S6 in pancreatic cancer cells. CFPAC-1 (a) and MIAPaCa-2 (b) cells were treated with 0–160 nM CUDC-907
for 48 h. Whole cell lysates were subjected to western blotting and probed with the indicated antibodies. The experiments were repeated at least two times and representative blots are shown

decreased the phosphorylation of AKT at S473 and T308 and completely abolished CUDC-907-induced phosphoryla- tion of S6. However, VS-5584 and VS-5584 in combina- tion with CUDC-907 induced phosphorylation of ERK at Thr202/Tyr204 to a much greater extent than did CUDC- 907 alone (Fig. 4d, e). These results suggest that abrogating CUDC-907-induced activation of the PI3K/mTOR pathway is responsible for VS-5584 enhancement on CUDC-907-in- duced cell death.

Discussion
Multi-target therapy is a powerful and promising strategy to treat pancreatic cancer. CUDC-907 inhibits/inactivates a variety of signaling pathways and molecular targets that play important roles in cancer cell survival, including PI3K/ mTOR, MEK/ERK, STAT3, MYC, CHK1, Wee1, and
RRM1 [10, 19–21]. Thus, CUDC-907 represents a potential candidate drug for the treatment of pancreatic cancer. In this study, we determined the antitumor activity of CUDC-907 in preclinical models of pancreatic cancer. Approximately
90% of pancreatic cancer patients harbour KRAS mutation [22]. BxPC-3 cells with KRAS wild-type showed moderate sensitive to CUDC-907 compared to other cell lines with KRAS mutation (Fig. 1b, c). These results suggest that the antitumor activity of CUDC-907 is irrelevant to the status of KRAS gene in pancreatic cancer, but more studies are needed to verify it. While our study was underway, Fu et al. reported the antitumor activity of CUDC-907 in pancreatic cancer [13]. We obtained results consistent with those pre- sented by Fu et al. [13] for both in vitro and in vivo models. However, the dose of CUDC-907 used in our in vivo experi- ments was far lower than that used by Fu et al. (10 mg/kg vs. 300 mg/kg).
CUDC-907 was designed to simultaneously inhibit HDACs and the PI3K/mTOR pathway [10]. We confirmed that short-term treatment with high concentrations of CUDC-907 (1 μM) results in the inhibition of HDACs and the PI3K/mTOR pathway, which is reflected by the increase of acetylated H4, acetylated α-tubulin, and phosphorylated S6 (Fig. 1a). Although long-term treatment with CUDC-907 at much lower concentrations (20–160 nM) resulted in a modest decrease in p-AKT (T308), it substantially increased

Fig. 4 VS-5584 enhances the antitumor activity of CUDC-907. a MIAPaCa-2 cells were treated with 40 nM CUDC-907 (CUDC) alone or in combination with 2 μM rapamycin (RAPA), 2 μM KU-0063794 (KU), 2 μM VS-5584 (VS), or 2 μM SCH772984 (SCH) for 48 h.
Cells were fixed, stained with PI, and subjected to flow cytometry analysis. Cells with sub-G1 DNA content were deemed dead cells. The experiment was repeated twice and data are presented as mean percent of cells ± standard errors of triplicates from one representa- tive experiment. b, c Pancreatic cancer cell lines were treated with 40 nM CUDC and 2 μM VS, alone or in combination, for 48 h. Cells were subjected to trypan blue exclusion test or fixed, stained with PI,
and then analyzed by flow cytometry. Dead cells are expressed as the percentage of cells with sub-G1 DNA content shown in panel B, or with blue staining shown in panel C. The experiment was repeated twice and data are presented as means of triplicates ± standard errors from one representative experiment. ### Indicates p < 0.001 when compared to single drug treatments. d, e CFPAC-1 and MIAPaCa-2 cells were treated with 40 nM CUDC and 2 μM VS alone or in com- bination for 48 h. Whole cell lysates were subjected to Western blot- ting and probed with the indicated antibodies. The experiment was repeated at least two times and representative blots are shown

the levels of phosphorylated AKT (S473), S6, and ERK. In a previous study, CUDC-907 treatment resulted in an increase in p-AKT (S473) in Capan-1 cells but a decrease in AsPC-1 and PANC-1 cells [13]. These results suggest that the effect of CUDC-907 on the PI3K-AKT pathway can be cell type- specific. In addition, it is worth noting that there exists func- tional compensation among the PI3K isoforms [23–26]. Thus, the activation of PI3K/mTOR pathway may be due to alternative mechanisms resulting from isoform selectivity of CUDC-907. Complex feedback loops exist within the PI3K/ mTOR pathway and/or crosstalk occurs between the PI3K/ mTOR pathway and other signaling pathways such as the
MEK/ERK pathway [27]. Ribosomal protein S6 kinase β (S6K2), which phosphorylates S6 at S240/244, is a common downstream target of both the PI3K/mTOR and MEK/ERK pathways [28, 29]. Thus, compensatory activation of ERK in response to PI3K inhibition [reflected by the decrease in p-AKT (T308)] is a potential mechanism responsible for the induction of p-AKT (S473) and p-S6 by CUDC-907. These results emphasize the complexity of the PI3K/mTOR path- way as well as the impact of drugs on this pathway in the complex genetic context of pancreatic cancer cells.
As shown in Fig. 4a, combining CUDC-907 with inhibi- tors that target the PI3K/mTOR or MER/ERK pathways can

augment CUDC-907-induced cell death. This suggests that inhibition of PI3K/mTOR or MEK/ERK pathways represents a reasonable way to enhance CUDC-907 antitumor activity in pancreatic cancer. As VS-5584 increased CUDC-907-induced cell death to the greatest extent among the inhibitors tested, we used VS-5584 for additional studies. Although VS-5584 and its combination with CUDC-907 treatment substantially decreased the levels of p-AKT (T308) and p-AKT (S473) and completely abolished the induction of p-S6 by CUDC-907, it caused high induction of p-ERK. This induction of ERK activation by VS-5584 is consistent with a previous study per- formed in acute myeloid leukemia [30]. These results suggest that the PI3K/mTOR pathway plays a critical role in resistance to CUDC-907 in pancreatic cancer cells. Besides that, com- bining the ERK selective inhibitor SCH772984 with CUDC- 907 significantly enhanced cell death induced by CUDC-907. Thus, we cannot rule out the contribution of ERK activation to CUDC-907 resistance.
In summary, through the use of cell lines and a mouse xenograft model, we confirmed that CUDC-907 has promis- ing preclinical antitumor activity against pancreatic cancer. Combining CUDC-907 with a PI3K/mTOR dual inhibitor may result in even better antitumor activity against pancreatic cancer. Overall, our study supports the clinical development of CUDC-907 for the treatment of pancreatic cancer, especially in combination with a PI3K/mTOR dual inhibitor.
Acknowledgements This study was supported by the Key Laboratory for Molecular Enzymology and Engineering, the Ministry of Educa- tion, National Engineering Laboratory for AIDS Vaccine, School of Life Sciences, Jilin University.
Author contributions Conceptualization, GW; in vitro experiments, SL, YD, and TW; in vivo experiments, LZ and XN; data analysis, SL, SZ, and GW; supervision, SZ and GW; draft preparation and funding acquisition, GW.

Funding This study was supported by the School of Life Sciences of Jilin University a grant from the National Natural Science Foundation of China (NSFC 81800154).

Availability of data and materials All data generated and analyzed dur- ing this study are included in this published article.

Compliance with ethical standards
Conflict of interest The authors declare no competing interests.
Ethics approval The animal study was conducted following internation- ally recognized guidelines and was approved by the Animal Research Committee of Norman Bethune College of Medicine, Jilin University.
References
⦁ Rahib L, Smith BD, Aizenberg R, Rosenzweig AB, Fleshman JM, Matrisian LM (2014) Projecting cancer incidence and deaths to 2030: the unexpected burden of thyroid, liver, and pancreas cancers in the United States. Cancer Res 74(11):2913–2921. ⦁ https
://doi.org/10.1158/0008-5472.can-14-0155
⦁ Kikuyama M, Kamisawa T, Kuruma S, Chiba K, Kawaguchi S (2018) Early diagnosis to improve the poor prognosis of pancre- atic cancer. Cancer (Basael) 10(2):48. ⦁ https://doi.org/10.3390/ ⦁ cance⦁ rs10020048
⦁ Zhang Y, Yang C, Cheng H, Fan Z, Huang Q, Lu Y, Fan K, Luo G, Jin K, Wang Z, Liu C, Yu X (2018) Novel agents for pancreatic ductal adenocarcinoma: emerging therapeutics and future direc- tions. J Hematol Oncol 11(1):14. ⦁ https://doi.org/10.1186/s1304 ⦁ 5-017-0551-7
⦁ Feng W, Zhang B, Cai D, Zou X (2014) Therapeutic potential of histone deacetylase inhibitors in pancreatic cancer. Cancer Lett 347(2):183–190. https://doi.org/10.1016/j.canlet.2014.02.012
⦁ Thorpe LM, Yuzugullu H, Zhao JJ (2015) PI3K in cancer: diver- gent roles of isoforms, modes of activation and therapeutic target- ing. Nat Rev Cancer 15(1):7–24. https://doi.org/10.1038/nrc3860
⦁ Polivka J Jr, Janku F (2014) Molecular targets for cancer therapy in the PI3K/AKT/mTOR pathway. Pharmacol Ther 142(2):164– 175. https://doi.org/10.1016/j.pharmthera.2013.12.004
⦁ Schlieman MG, Fahy BN, Ramsamooj R, Beckett L, Bold RJ (2003) Incidence, mechanism and prognostic value of activated AKT in pancreas cancer. Br J Cancer 89(11):2110–2115. ⦁ https:// ⦁ doi.org/10.1038/sj.bjc.6601396
⦁ Wu CY, Carpenter ES, Takeuchi KK, Halbrook CJ, Peverley LV, Bien H, Hall JC, DelGiorno KE, Pal D, Song Y, Shi C, Lin RZ, Crawford HC (2014) PI3K regulation of RAC1 is required for KRAS-induced pancreatic tumorigenesis in mice. Gastroenter- ology 147(6):1405-1416.e1407. ⦁ https://doi.org/10.1053/j.gastr ⦁ o.2014.08.032
⦁ Baer R, Cintas C, Dufresne M, Cassant-Sourdy S, Schönhuber N, Planque L, Lulka H, Couderc B, Bousquet C, Garmy-Susini B, Vanhaesebroeck B, Pyronnet S, Saur D, Guillermet-Guibert J (2014) Pancreatic cell plasticity and cancer initiation induced by oncogenic Kras is completely dependent on wild-type PI 3-kinase p110α. Genes Dev 28(23):2621–2635. ⦁ https://doi.org/10.1101/ ⦁ g⦁ ad.249409.114
⦁ Qian C, Lai CJ, Bao R, Wang DG, Wang J, Xu GX, Atoyan R, Qu H, Yin L, Samson M, Zifcak B, Ma AW, DellaRocca S, Borek M, Zhai HX, Cai X, Voi M (2012) Cancer network disruption by a single molecule inhibitor targeting both histone deacetylase activ- ity and phosphatidylinositol 3-kinase signaling. Clin Cancer Res: An Off J The Am Assoc Cancer Res 18(15):4104–4113. ⦁ https:// ⦁ doi.o⦁ rg/10.1158/1078-0432.ccr-12-0055
⦁ Oki Y, Kelly KR, Flinn I, Patel MR, Gharavi R, Ma A, Parker J, Hafeez A, Tuck D, Younes A (2017) CUDC-907 in relapsed/ refractory diffuse large B-cell lymphoma, including patients with MYC-alterations: results from an expanded phase I trial. Hae- matologica 102(11):1923–1930. ⦁ https://doi.org/10.3324/haema ⦁ t⦁ ol.2017.172882
⦁ Younes A, Berdeja JG, Patel MR, Flinn I, Gerecitano JF, Nee- lapu SS, Kelly KR, Copeland AR, Akins A, Clancy MS, Gong L, Wang J, Ma A, Viner JL, Oki Y (2016) Safety, tolerability, and preliminary activity of CUDC-907, a first-in-class, oral, dual inhibitor of HDAC and PI3K, in patients with relapsed or refractory lymphoma or multiple myeloma: an open-label, dose- escalation, phase 1 trial. Lancet Oncol 17(5):622–631. ⦁ https://doi. ⦁ o⦁ rg/10.1016/s1470-2045(15)00584-7
⦁ Fu XH, Zhang X, Yang H, Xu XW, Hu ZL, Yan J, Zheng XL, Wei RR, Zhang ZQ, Tang SR, Geng MY, Huang X (2019) CUDC-907

displays potent antitumor activity against human pancreatic ade- nocarcinoma in vitro and in vivo through inhibition of HDAC6 to downregulate c-Myc expression. Acta Pharmacol Sin 40(5):677– 688. https://doi.org/10.1038/s41401-018-0108-5
⦁ Wang G, Niu X, Zhang W, Caldwell JT, Edwards H, Chen W, Taub JW, Zhao L, Ge Y (2015) Synergistic antitumor interac- tions between MK-1775 and panobinostat in preclinical models of pancreatic cancer. Cancer Lett 356(2 Pt B):656–668. ⦁ https:// ⦁ doi.o⦁ rg/10.1016/j.canlet.2014.10.015
⦁ Cardenas ME, Cutler NS, Lorenz MC, Di Como CJ, Heitman J (1999) The TOR signaling cascade regulates gene expression in response to nutrients. Genes Dev 13(24):3271–3279. ⦁ https://doi. ⦁ o⦁ rg/10.1101/gad.13.24.3271
⦁ García-Martínez JM, Moran J, Clarke RG, Gray A, Cosulich SC, Chresta CM, Alessi DR (2009) Ku-0063794 is a specific inhibi- tor of the mammalian target of rapamycin (mTOR). Biochem J 421(1):29–42. https://doi.org/10.1042/bj20090489
⦁ Hart S, Novotny-Diermayr V, Goh KC, Williams M, Tan YC, Ong LC, Cheong A, Ng BK, Amalini C, Madan B, Nagaraj H, Jayara- man R, Pasha KM, Ethirajulu K, Chng WJ, Mustafa N, Goh BC, Benes C, McDermott U, Garnett M, Dymock B, Wood JM (2013) VS-5584, a novel and highly selective PI3K/mTOR kinase inhibi- tor for the treatment of cancer. Mol Cancer Ther 12(2):151–161. ⦁ https⦁ ://doi.org/10.1158/1535-7163.mct-12-0466
⦁ Morris EJ, Jha S, Restaino CR, Dayananth P, Zhu H, Cooper A, Carr D, Deng Y, Jin W, Black S, Long B, Liu J, Dinunzio E, Windsor W, Zhang R, Zhao S, Angagaw MH, Pinheiro EM, Desai J, Xiao L, Shipps G, Hruza A, Wang J, Kelly J, Paliwal S, Gao X, Babu BS, Zhu L, Daublain P, Zhang L, Lutterbach BA, Pel- letier MR, Philippar U, Siliphaivanh P, Witter D, Kirschmeier P, Bishop WR, Hicklin D, Gilliland DG, Jayaraman L, Zawel L, Fawell S, Samatar AA (2013) Discovery of a novel ERK inhibitor with activity in models of acquired resistance to BRAF and MEK inhibitors. Cancer Discov 3(7):742–750. ⦁ https://doi. ⦁ o⦁ rg/10.1158/2159-8290.cd-13-0070
⦁ Sun K, Atoyan R, Borek MA, Dellarocca S, Samson ME, Ma AW, Xu GX, Patterson T, Tuck DP, Viner JL, Fattaey A, Wang J (2017) Dual HDAC and PI3K inhibitor CUDC-907 downregulates MYC and suppresses growth of MYC-dependent cancers. Mol Cancer Ther 16(2):285–299. ⦁ https://doi.org/10.1158/1535-7163. ⦁ mct-16-0390
⦁ Kotian S, Zhang L, Boufraqech M, Gaskins K, Gara SK, Quez- ado M, Nilubol N, Kebebew E (2017) Dual inhibition of HDAC and tyrosine kinase signaling pathways with CUDC-907 inhib- its thyroid cancer growth and metastases. Clin Cancer Res: An Off J The Am Assoc Cancer Res 23(17):5044–5054. ⦁ https://doi. ⦁ o⦁ rg/10.1158/1078-0432.ccr-17-1043
⦁ Chen Y, Peubez C, Smith V, Xiong S, Kocsis-Fodor G, Kennedy B, Wagner S, Balotis C, Jayne S, Dyer MJS, Macip S (2019) CUDC-907 blocks multiple pro-survival signals and abrogates microenvironment protection in CLL. J Cell Mol Med 23(1):340– 348. https://doi.org/10.1111/jcmm.13935
⦁ Knudsen ES, O’Reilly EM, Brody JR, Witkiewicz AK (2016) Genetic diversity of pancreatic ductal adenocarcinoma and oppor- tunities for precision medicine. Gastroenterology 150(1):48–63. ⦁ https⦁ ://doi.org/10.1053/j.gastro.2015.08.056
⦁ Ramadani F, Bolland DJ, Garcon F, Emery JL, Vanhaesebroeck B, Corcoran AE, Okkenhaug K (2010) The PI3K isoforms p110alpha and p110delta are essential for pre-B cell receptor signaling and B cell development. Sci Signa 3(134):ra60. ⦁ https://doi.org/10.1126/ ⦁ scisignal.2001104
⦁ Iyengar S, Clear A, Bödör C, Maharaj L, Lee A, Calaminici M, Matthews J, Iqbal S, Auer R, Gribben J, Joel S (2013) P110α- mediated constitutive PI3K signaling limits the efficacy of p110δ-selective inhibition in mantle cell lymphoma, particularly with multiple relapse. Blood 121(12):2274–2284. ⦁ https://doi. ⦁ org/10.1182/blood-2012-10-460832
⦁ Schwartz S, Wongvipat J, Trigwell CB, Hancox U, Carver BS, Rodrik-Outmezguine V, Will M, Yellen P, de Stanchina E, Baselga J, Scher HI, Barry ST, Sawyers CL, Chandarlapaty S, Rosen N (2015) Feedback suppression of PI3Kα signaling in PTEN-mutated tumors is relieved by selective inhibition of PI3Kβ. Cancer Cell 27(1):109–122. ⦁ https://doi.org/10.1016/j.ccell
.2014.11.008
⦁ Xie S, Ni J, McFaline-Figueroa JR, Wang Y, Bronson RT, Ligon KL, Wen PY, Roberts TM, Zhao JJ (2020) Divergent roles of PI3K isoforms in PTEN-deficient glioblastomas. Cell reports 32(13):108196. https://doi.org/10.1016/j.celrep.2020.108196
⦁ Rozengurt E, Soares HP, Sinnet-Smith J (2014) Suppression of feedback loops mediated by PI3K/mTOR induces multiple over- activation of compensatory pathways: an unintended consequence leading to drug resistance. Mol Cancer Ther 13(11):2477–2488. ⦁ https⦁ ://doi.org/10.1158/1535-7163.mct-14-0330
⦁ Pardo OE, Arcaro A, Salerno G, Tetley TD, Valovka T, Gout I, Seckl MJ (2001) Novel cross talk between MEK and S6K2 in FGF-2 induced proliferation of SCLC cells. Oncogene 20(52):7658–7667. https://doi.org/10.1038/sj.onc.1204994
⦁ Rosner M, Hengstschlager M (2010) Evidence for cell cycle- dependent, rapamycin-resistant phosphorylation of ribosomal protein S6 at S240/244. Amino Acids 39(5):1487–1492. ⦁ https:// ⦁ doi.o⦁ rg/10.1007/s00726-010-0615-2
⦁ Su Y, Li X, Ma J, Zhao J, Liu S, Wang G, Edwards H, Taub JW, Lin H, Ge Y (2018) Targeting PI3K, mTOR, ERK, and Bcl-2 sign- aling network shows superior antileukemic activity against AML ex vivo. Biochem Pharmacol 148:13–26. ⦁ https://doi.org/10.1016/j. ⦁ bcp.2017.11.022
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.