GP‐2250, a novel anticancer agent, inhibits the energy metabolism, activates AMP‐Kinase and impairs the NF‐kB pathway in pancreatic cancer cells

Abstract GP‐2250, a novel anticancer agent, severely limits the energy metabolism, as demonstrated by the inhibition of hexokinase 2 and glyceraldehyde‐3‐phosphate dehydrogenase and a decrease of ATP. Rescue experiments with supplementary pyruvate or oxaloacetate demonstrated that a TCA cycle deficit largely contributed to cytotoxicity. Activation of the energy‐deficit sensor, AMP‐dependent protein kinase, was associated with increased phosphorylation of acetyl‐CoA carboxylase and Raptor, pointing to a possible deficit in the synthesis of fatty acids and proteins as essential cell components. Binding of p65 to DNA was dose‐dependently reduced in nuclear lysates. A transcriptional deficit of NF‐κB (nuclear factor kappa‐light‐chain‐enhancer of activated B cells) was substantiated by the downregulation of cyclin D1 and of the anti‐apoptotic Bcl2, in line with reduction in tumour cell proliferation and induction of apoptosis, respectively. The upregulation of p53 concomitant with an excess of ROS supported apoptosis. Thus, the anticancer activity of GP‐2250 is a result of disruption of energy metabolism and inhibition of tumour promotion by NF‐κB.

hexokinase (HK) was a potential target for GP-2250, with HK2 being highly expressed in cancer cells. HK2 couples metabolic and proliferative effects by rewiring metabolism to aerobic glycolysis and protects against pro-apoptotic stimuli by docking to mitochondria. 8 Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is the ratelimiting enzyme of aerobic glycolysis thereby exerting control over glycolytic energy metabolism. 9 Tumour cells contain functional mitochondria 5,7 and inhibition of the trichloroacetic acid cycle (TCA cycle) can likewise contribute to an energy deficit either as a downstream consequence of the impaired glycolysis or by direct inhibition of TCA cycle enzymes such as α-ketoglutarate dehydrogenase (αKGDH).
Inhibition of glycolytic enzymes and TCA cycle by GP-2250 would be expected to reduce ATP and trigger an energy crisis.
Activation of the energy-deficit sensor, adenosine monophosphatedependent protein kinase (AMPK), suppresses tumour growth 10 by introducing a metabolic austerity programme via inhibition of downstream ATP-consuming biosynthetic pathways. 11,12 They include the synthesis of fatty acids, proteins and nucleotides, key building blocks for cell growth and proliferation. Tumour growth can also be limited by inhibition of the transcription factor NF-κB (nuclear factor kappalight-chain-enhancer of activated B cells), a major driver of tumour progression with a key impact on cell cycle regulation and apoptosis.
Its inhibition would be expected to downregulate the expression of cyclin D1, the driver of mitosis, and Bcl2, an anti-apoptotic survival factor. 13 The drug-induced oxidative reactive oxygen species (ROS) stress can be expected to upregulate the transcription factor p53, which provides additional anti-tumour activity. Thus, by targeting energy metabolism and the NF-κB pathway of tumour cells, GP-2250 would be expected to display a dual approach to cytotoxicity.

| Cell lines and culture conditions
Two human pancreatic cancer cell lines were used: BxPC3 14 (ATCC- LGC Standards GmbH) and Panc TuI (ATCC-LGC Standards GmbH).
Authentication was analysed by STR analysis. Panc TuI cells 15 were cultured in Dulbecco's Modified Eagle Medium (DMEM 25 mM Glucose).
BxPC3 were maintained in RPMI 1640. Both cultures were supplemented with the antibiotics penicillin (100 U/mL), streptomycin (100 U/ mL) and 2 mM L-glutamine. Cells were grown as monolayer and cultured in 25 cm 2 flasks at 37°C and 5% CO 2 in a humidified atmosphere.

| Western blot analysis
Protein isolation for western blotting 16

| ATP and cell viability determination
The ATP level in the cell lines was measured using the Luminescent ATP Detection Assay Kit (Abcam; ab1113849) following the manufacturer's instructions. The viability of cells was determined with the MTT assay as described earlier. 1 The assay was performed in at least three independent experiments with consecutive passages.

| HK2 assay
A HK activity assay kit (Merck, D, MAK091-1KT) 17 was used following the manufacturer's instructions. Cells were incubated with different concentrations of GP-2250 for 24 h before lysis. To analyse HK2 activity, lysates were partly incubated for 1 h at 45°C to eliminate HK2, which is sensitive to temperature. The difference between total HK activity (lysates on ice) and HK 1 activity (lysates for 1 h at 45°C) reveals the activity of HK2.

| rhHK2 enzyme activity assay
Recombinant human hexokinase 2 (rhHK2) enzyme activity was tested in vitro using HK2 from a HK2 inhibitor assay kit (Abcam, ab211114) following the manufacturer's instructions.

| GAPDH enzyme activity assay
Following incubation with GP-2250 for 24 h, cell lysates were prepared in lysis buffer [20 mM Tris (tris [hydroxymethyl] aminomethane) pH 7.8, 100 mM NaCL, 1% triton, protease inhibitors] followed by centrifugation. Enzyme activity was measured from 10 μL of cell lysate as described by Kornberg et al. 18 Assays were performed in 10 mM sodium pyrophosphate buffer (pH 8.5) in 96-well plates. Cell lysates were incubated with 20 mM sodium arsenate (made fresh on the day of experiment), 1 mM NAD+, and 2.88 mM glyceraldehyde-3-phosphate (G3P). Enzyme activity was measured using a microplate reader-spectrophotometer (Tecan) as the increase in absorbance at 340 nm is due to the reduction of NAD+. The assay was performed at 37°C. The lysate was first diluted into sodium pyrophosphate buffer to a volume of 100 μL. An additional 100 μL of reaction mix containing the sodium arsenate, NAD+ and G3P was added. Absorbance was measured every 5 min for 60 min.
2 μg of rhGAPDH was transferred to HEN buffer (250 mM Hepes-NaOH, 1 mM EDTA 0 and 1 mM neocuproine) and incubated with PBS or GP-2250 (100 and 250 μM) at 37°C. After 0, 30 and 60 min of incubation, an aliquot was taken for the enzyme assay. The assay was conducted using the GAPDH Activity Assay Kit (Abcam Cambridge, UK, ab204732) following the manufacturer's instructions. 19
Cells were incubated with different concentrations of GP-2250 for 24 h before lysis.

| ROS assay
ROS was analysed using the Cellular ROS/Superoxide Detection Assay Kit (ab139476, Abcam) 21 following the manufacturer's instructions.

| NFκ B transcription factor binding assay
The transcription factor NF-κB was assayed with the NFκB p65 Transcription Factor Assay Kit (Abcam, ab133112) 22 following the manufacturer's instructions. Lysates were prepared either from drug-treated cells or, alternatively, from untreated cells with the lysates being directly incubated with GP-2250. The nuclear extracts were prepared with the Nuclear Extraction Kit (Abcam, ab113474) according to the manufacturer's instructions.

| Statistics and calculations
Results of MTT test, ATP assay, ROS assay, HK2 assay, GAPDH assay, PDH assay, NF-κB assay, western blots and rescue experiments are expressed as mean ± SD. Comparison between experimental groups with normal distribution was performed using one-way anova followed by Tukey's post hoc test and pairwise tests performed using t-tests. p ≤ 0.05 were considered statistically significant and indicated in the figures as follows: *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001.

| Induction of energy deficiency
The hypothesis that GP-2250 induces an energy deficit was tested by measuring the level of ATP in the pancreatic tumour cell lines BxPC3 and Panc Tul. At the lowest dose tested (250 μM GP-2250), a significant decrease in ATP was already apparent at 6 h in both BxPC3 and Panc Tul cells (12.4% ± 2.4% and 19.8% ± 2.3%, respectively), with no loss of cell viability ( Figure 1B,C). A further decrease in ATP was apparent at the same time point with 500 μM GP-2250 in both cell lines (32.5% ± 3.1% and 48.9% ± 4.8%, respectively), at which dose only a minor decrease of cell viability of 10%-12% was observed ( Figure 1B,C). These findings of a druginduced deficit of ATP pointed to an inhibition of energy metabolism by GP-2250.

| Generation of ROS
In both Panc Tul and BxPC3 cells, a strong rise in ROS was already apparent after 90 min of incubation with higher concentrations of GP-2250 ( Figure 4A,B). Addition of the reducing agent NAC (n-acetylcysteine) abrogated the rise in ROS as shown by the negative control (NC). The ROS level was slightly higher in the untreated control (UC), which was devoid of NAC, a finding which presumably reflects the ROS base level in the cells.

| Activation of AMPK
The energy deficit induced by GP-2250 was expected to activate the metabolic regulator AMPK through phosphorylation at threonine 172 (Thr 172) by an upstream kinase. 11,12 In the presence of GP-2250, the phosphorylation of AMPK at Thr 172 was increased in a time-dependent manner (6-24 h) as tested in PancTul cells ( Figure 5A). The dose-dependency in the range of 250-1000 μM was determined in BxPC3 cells ( Figure 5B). To test whether the activation of AMPK was functionally significant, two downstream targets were analysed, acetyl-CoA carboxylase (ACC) and Raptor.

| Phosphorylation of the rate-limiting enzyme of fatty acid synthesis and Raptor
The enzyme ACC, the rate-limiting step of fatty acid synthesis, is susceptible to AMPK-induced inhibition through phosphorylation at serine 79 (Ser79). In the presence of GP-2250, ACC-1 was increasingly phosphorylated at Ser79 over time, which was apparent for both Panc TuI (250 μM) and BxPC3 (500 μM), as tested by western  blotting ( Figure 6). The phosphorylation was somewhat more prominent at 12 h than at 24 h. The phosphorylation of ACC-1 points to a potential impact of GP-2250 on fatty acid synthesis.
The mTOR complex 1 (mTORC1) drives cellular growth 23,24 and can be directly inhibited by AMPK through phosphorylation of its constituent protein Raptor at Ser792. 11,[24][25][26] In the presence of GP-2250 (500 μM), Raptor phosphorylation at Ser792 was time-dependently increased as shown in western blots with Panc TuI cells responding at a lower drug concentration (500 μM) than BxPC3 (1000 μM) ( Figure 6). In addition, the mTOR protein level was time-dependently reduced in both cell lines by GP-2250 (1000 μM) ( Figure 6). The protein level of the upstream serine/ threonine-kinase Akt, a driver of tumour proliferation, was likewise time-dependently decreased by GP-2250 (1000 μM) in both cell lines ( Figure 6). These results point to a possible inhibition of cell growth by GP-2250.

| Inhibition of the NFκ B and its transcriptional activity
To test whether GP-2250 might inhibit the NF-kB pathway, p65/DNA binding was analysed in nuclear lysates prepared from Panc Tul and

| Upregulation of p53
As a response to metabolic and oxidative stress, cells frequently respond by arresting cell division or by inducing apoptosis, triggered by the transcription factor and tumour repressor p53. 27,28 In the presence of GP-2250 (500 μM), p53 protein levels were timedependently increased, serving as a sign of activation of the stress response ( Figure 6).

| Inhibition of energy metabolism causes anticancer activity
As outlined in a schematic overview (Figure 7 Both HK2 and GAPDH are upregulated in many human tumours and serve as drug targets. 8 Inhibitors of HK2 such as benitrobenrazide show anticancer activity as demonstrated in tumour xenograft models 29 and clinical trials have started with the HK2 inhibitor lonidamine. 8 Apart from limiting glycolysis, HK2 inhibition is expected to contribute to the rise of ROS by reducing the synthesis of the reducing agent NADPH through inhibition of the pentose phosphate pathway. 5,30 The selective GAPDH inhibitors koningic acid 9

| Activation of AMPK
The impairment of energy metabolism by GP-2250 was most clearly apparent by the decrease of ATP ( Figure 1B,C), which represents a threat to survival of the tumour cells. As a countermeasure F I G U R E 7 Proposed scheme of the alteration of cancer cell metabolism and NF-κB inhibition by GP-2250. Bold black arrows indicate metabolic pathways; red and green arrows indicate drug-induced changes. By limiting the energy metabolism through the inhibition of hexokinase 2 and GAPDH, GP-2250 induces an energy deficit in line with an impairment of the TCA cycle. The reduction of ATP triggers the activation of the energy-deficit sensor AMPK. Its downstream events include the inhibition of mTOR, a major driver of tumour cell growth and potential impairment of fatty acid synthesis (FAS) through ACC-1 inhibition. The inhibition of NF-κB by GP-2250 limits the rate of tumour cell proliferation through cyclin D1 downregulation. It also promotes apoptosis through downregulation of the anti-apoptotic Bcl2. ROS contributes to the upregulation of the transcription factor p53, which supports apoptosis. AMPK, adenosine monophosphatedependent protein kinase; FAS, fatty acid synthesis; NF-κB, nuclear factor kappa-light-chain-enhancer of activated B cells; PDH, pyruvate dehydrogenase.
to preserve ATP, the energy-deficit sensor AMPK was activated as shown by the time-dependent phosphorylation of the AMPK alpha subunit at threonine 172 in Panc Tul cells ( Figure 5A) and the dose dependency of the phosphorylation of AMPK in BxPC3 cells ( Figure 5B). AMPK activation is known to inhibit major downstream ATP-consuming biosynthetic pathways required for cell growth and proliferation, including the synthesis of proteins, fatty acids, cholesterol, nucleotides, ribosomal RNA and glycogen. 11,12 Due to this austerity programme, AMPK activation suppresses tumour growth. 10 The activation of AMPK is therefore a hallmark of the anticancer activity of GP-2250.

| Inhibition of ACC
AMPK inhibits lipogenic transcriptional programmes including the synthesis of fatty acids. This is in keeping with the phosphorylation of ACC-1 in the presence of GP-2250 ( Figure 6), pointing to a potential deficit of fatty acid synthesis in Panc Tul and BxPC3 cells.
The power of ACC inhibition in suppressing the growth of cancers is underlined by the finding that drug-induced ACC phosphorylation or a specific ACC inhibitor was effective in suppressing hepatocellular carcinoma and pancreatic cancer. 39,40 Thus, the cytotoxicity of GP-2250 may be attributed, at least in part, to a lack of fatty acid synthesis. The ability of supplementary OAA to rescue Panc Tul and BxPC3 cells from GP-2250-induced cytotoxicity ( Figure 3A,B) might be attributed, at least in part, to a possible restoration of fatty acid synthesis. An excess of OAA is likely to override the ACC inhibition by increasing the transport capacity to provide mitochondrial Acetyl-CoA as substrate for the cytosolic ACC.

| Inhibition of RAPTOR/mTOR
AMPK regulates protein synthesis in large part by inhibition of the mTORC1 complex, a key driver of tumour growth. 24,41 The timedependent increase of RAPTOR phosphorylation at Ser792 by GP-2250 ( Figure 6), suggests that GP-2250 is able to inhibit protein synthesis. There are reciprocal interactions between mTOR and Akt.
The protein kinase Akt also drives tumour growth and proliferation.
At high concentrations of GP-2250, the protein levels of both mTOR and Akt were reduced, supporting the view that GP-2250 inhibits protein synthesis ( Figure 6). Besides inhibiting the activity of mTOR through RAPTOR phosphorylation, AMPK can also limit protein synthesis indirectly via TSC2 and directly by blocking RNA synthesis and protein elongation. 11,12

| A lack of nucleotides?
As part of the downstream events of AMPK activation, the synthesis of nucleotides can be switched off by inhibition of PRPP (phosphoribosyl-1-pyrophosphate)-synthase. 11 Although not tested experimentally, an inhibition of nucleotide synthesis by GP-2250 would be in keeping with the rescue experiment performed with OAA ( Figure 3A,B). The synthesis of both pyrimidine-and purinenucleotides requires aspartate. 5,7 As a precursor to aspartate, supplementary OAA may be able to restore the synthesis of nucleotides by replenishing aspartate and override PRPP inhibition. As nucleotides are essential building blocks for the formation of RNA and DNA, it cannot be excluded that impaired DNA and RNA synthesis is part of the cytotoxic effects of GP-2250.  Figure 4E,F). It is of note, that the dose dependency for p65/DNA binding in treated lysates was comparable to that from the lysates of treated cells. This finding suggests that GP-2250 is likely to directly interfere with the binding of p65 to DNA and not with an upstream event.
A reduction of the transcriptional activity of NF-κB was apparent in both Panc Tul and BxPC3 cell lines by the change in expression of cyclin D1 and Bcl2. This is testimony to the functional relevance of NF-kB inhibition by GP-2250. The expression of cyclin D1, the driver of cell cycle progression, was reduced by GP-2250 ( Figure 5). This finding is in line with the previously shown reduction of the rate of cell proliferation by GP-2250. 1 The expression of the anti-apoptotic protein Bcl2 was likewise reduced by GP-2250 ( Figure 6). Thus, by inhibiting NF-κB, GP-2250 is able to disrupt tumour progression in a two-pronged manner, by reducing the rate of cell proliferation and promoting apoptosis.
Several limitations of this study must be taken into account.
The present study is exclusively an in vitro analysis using pancreatic cancer cell line models. The impact of GP-2250 on growth control via lipid and protein synthesis requires further study. Whether the present findings can be confirmed in an analysis of pancreatic cancer tissue in mouse PDX models remains to be seen. Further studies are also required to test whether the mechanism of action described here extends to other types of tumours.

| CON CLUS ION
The

ACK N O WLE D G E M ENTS
The substance GP-2250 was kindly provided by Geistlich Pharma

DATA AVA I L A B I L I T Y S TAT E M E N T
The data presented in this study are available on request from the corresponding author.