Fatty acid synthase, a novel poor prognostic factor for acute lymphoblastic leukemia which can be targeted by ginger extract

Altered metabolism of fatty acid synthesis is considered a hallmark characteristic of several malignancies, including acute lymphoblastic leukemia (ALL). To evaluate the impact of fatty acid synthase (FASN) on drug resistant ALL, bone marrow samples were collected from 65 pediatric ALLs, including 40 de novo and 25 relapsed patients. 22 non-cancer individuals were chosen as controls. Quantitative RT-PCR showed increased expression levels of FASN in drug resistant patients compared with the therapy responders. Single and combined treatment of malignant cells were analyzed using Annexin-V/PI double staining and MTT assays. Incubation of resistant primary cells with ginger showed simultaneous increased apoptosis rates and reduced FASN expression levels. Furthermore, docking studies demonstrated high affinity bindings between ginger derivatives and FASN thioesterase and ketosynthase domains, compared with their known inhibitors, fenofibrate and morin, respectively. Finally, combined treatment of in-house multidrug resistant T-ALL subline with ginger and dexamethasone induced drug sensitivity and down regulation of FASN expression, accordingly. To the best of our knowledge, this is the first study that introduces FASN upregulation as a poor prognostic factor for drug resistant childhood ALL. Moreover, it was revealed that FASN inhibition may be applied by ginger phytochemicals and overcome dexamethasone resistance, subsequently.

. (a) Ginger extract components (according to the manufacturer), (b) patient characteristics. USP United States Pharmacopeia, NMT not more than, ppm parts per million, ppb parts per billion, HPLC highperformance liquid chromatography, ICP-MS inductively coupled plasma mass spectrometry, MRD minimal residual disease in newly diagnosed patients, determined 1 year after the onset of treatment.  www.nature.com/scientificreports/ Relative expression levels of FASN in de novo patients. In order to characterize the expression pattern of FASN in children with ALL, quantitative reverse transcriptase polymerase chain reaction was used to determine the expression levels of this gene in the bone marrow mononuclear cell samples of 40 children with newly diagnosed ALL and 22 non-cancer control cases. Results indicated no significant difference in the expression levels of FASN in de novo patients compared to the control group [1.123 ± 0.1228 vs. 0.9596 ± 0.05020 vs. mean ± SEM, P = 0.3432] (Fig. 1a).
To determine whether there was any significant difference between two subtypes of ALL, the mRNA expression levels of FASN was analyzed in 36 B-ALL and 4 T-ALL samples. As shown in Fig. 1b, there was no significant difference in FASN expression levels in these two groups [1.069 ± 0.1197 vs. 1.372 ± 0.4133, mean ± SEM, P = 0.3884].
FASN expression levels in drug sensitive vs. resistant patients. The relative gene expression levels of FASN in MRD+ and MRD− patients are presented in Fig. 2a. A significantly higher mRNA expression level of FASN was determined in MRD+ patients compared with the MRD− patients [1.841 ± 0.3311, n = 9 vs. 0.9242 ± 0.1134, n = 31, P = 0.0021]. Moreover, ROC curve analysis introduced the mRNA FASN level as a prognostic biomarker which may distinguish MRD+ from MRD− ALL patients. The total area under the curve (AUC) was 0.82, confirming the ability and accuracy of this measurement to classify the innate drug resistant patients from the sensitive group (95% CI 0.675-0.978, P = 0.0039) (Fig. 2a,b).
In order to determine the association between FASN and adaptive drug resistance, the expression levels of FASN was measured in 25 relapsed patients. It was revealed that FASN was significantly upregulated in ALL relapsed group compared with the MRD− patients (1.169 ± 0.15 vs. 0.7669 ± 0.09448, mean ± SEM, P = 0.0180). Moreover, ROC curve analysis revealed that FASN expression levels could discriminate between the relapsed and MRD− patients (AUC = 0.7023, 95% CI 0.545-0.891, P = 0.0187) (Fig. 2c,d).
Effect of ginger extract on FASN expression in primary ALL cells. The anti-leukemic effect of ginger extract was previously introduced by our group. Moreover, it was shown that this effect was not attributed to the expression levels of ABC transporters 14 . In order to identify the possible mechanism through which ginger could conquer ALL multidrug resistance in patients primary cells, fresh samples were collected from 7 children with relapsed ALL and 1 non-cancer control, treated with 167 μg/ml ginger extract for 48 h, and analyzed for any post-treatment alteration of the FASN expression levels. Cell death was measured using Annexin V/PI double staining and flow cytometry analysis. Results supported our previous data considering the significantly increased cell death in ginger treated patient samples compared with the untreated cells [39.11 ± 9.089% vs. 18.80 ± 7.433%, mean ± SEM, P = 0.0023]. In addition normal mononuclear cells (MNCs) were not significantly sensitive to proliferation inhibition of ginger extract (Fig. 3a). On the other hand, RT-PCR showed that FASN expression in relapsed patients was decreased upon cells exposure to ginger in comparison with the untreated samples (0.5894 ± 0.08593, mean ± SEM, P = 0.0031) but in normal MNCs were not significantly decreased (Fig. 3b).  www.nature.com/scientificreports/ Effect of ginger extract on CCRF-CEM and dexamethasone resistant CCRF-CEM/MVCD subline. The cytotoxic effect of ginger extract on the in house multidrug resistant CCRF-CEM/MVCD subline was previously defined 14 , and RT-PCR showed overexpression of FASN in this cell line compared with normal MNCs and its parental cell (P = 0.0119 and P = 0.0241, respectively) (Fig. 4a). Considering our previous data regarding CCRF-CEM/MVCD, among diverse examined chemotherapy drugs, dexamethasone showed the highest half maximal concentration (IC50) for inhibiting cell growth (Table 2). To investigate the possible impact of ginger in generating CCRF-CEM/MVCD sensitivity to dexamethasone, cells were treated with ginger extract, alone and in combination with dexamethasone. MTT assay was performed in addition to RT-PCR in order to determine the expression levels of FASN. Results showed that the cytotoxic effect of dexamethasone/ ginger extract was significantly more than that of the dexamethasone alone. In other words, cell viability was www.nature.com/scientificreports/ reduced down to 15.4 ± 0.821% in the presence of a combination of 1,000 μM dexamethasone with 167 μg/ ml ginger, compared with dexamethasone alone (56.794 ± 0.808%, P = 0.0008) (Fig. 4b).
Subsequently, RT-PCR demonstrated decrease in FASN expression followed by combination therapy compared with single drug treatment [1.791 ± 0.043 vs. 3.2 ± 0.210 (mean ± SEM; n = 2), P = 0.0225] (Fig. 4c). Interestingly, FASN expression level was increased in response to incubation with dexamethasone alone (P = 0.0116).  domains were illustrated using in silico studies ( Fig S1). The related interacting residues and binding energy of each docked ligand to the active site of the TE domain were calculated and demonstrated in Table 3. As shown, among diverse ginger phytochemicals (Fig S2), gingerenone family molecules had the highest affinity to the substrate binding site of the TE domain. Interestingly, the binding energies of these molecules were as low as fenofibrate, an experimentally-proved TE inhibitor 15 , and their binding energies were markedly lower than orlistat; which is another known TE inhibitor. The interactions of fenofibrate and other gingerenones with the TE domain are illustrated in Fig. 5 and Fig. S3. As shown, gingerenone C covers both, the interface cavity and specificity channel, generating the highest affinity towards the TE domain (see Fig. S4 for more details). Other gingerenone family members, not only lie on the  www.nature.com/scientificreports/ specificity channel and block the substrate binding site, but also extend to the catalytic site and form several hydrogen bonds with important residues for catalytic activity such as Ser2308 and Tyr2343 16 , through which the enzyme could be suppressed.

Docking results of the ketosynthase (KS) domain.
Binding energies of docked ligands to the activesite cavity of the KS domain and interacting residues of each KS-ligand complex are shown in Table 4. In the present study, none of our ligands displayed a significant affinity to the distal substrate binding site. By contrast, several ginger compounds, including quercetin and gingerenone family molecules, showed high affinity to the active-site cavity. Through the occupation of the cavity volume, these ligands may block accessibility to the active site residues Cys161, His293 and His331 and inhibit FASN activity. Morin, the well-known KS domain inhibitor, shares similar molecular scaffold and binding modes to the KS domain with quercetin. However, none of them show direct interaction with the main KS active site residues, implying that their inhibitory function might be due to the blocking of the substrate entry (Fig. 6, Fig. S5). In contrast, the gingerenone family molecules, may not only block the substrate entry, but also enter deep inside the cavity and form several hydrogen bonds with some of the main active site residues such as His293 and His331. (Fig. 6, Figs. S5, S6). Moreover, gingerenones form pi stacking interactions with His293 (data not shown). Cerulenin and C75, the two other docked known inhibitors, showed lower affinity to the KS domain in comparison to the majority of docked ginger phytochemicals. www.nature.com/scientificreports/

Discussion
Despite remarkable advances in the treatment strategies, drug resistance is still a major cause of chemotherapy failure leading to relapse in pediatric acute lymphoblastic leukemia 17 . Multiple studies suggested that metabolic rearrangements, newly recognized as prominent features of cancer 18 , are associated with the development of drug resistance in cancer cells 19 . Changes in lipid metabolism, in particular increased synthesis of fatty acids, are recognized as one of the key hallmarks in several cancer cells. Besides, FASN overexpression has shown to be associated with poor prognosis and resistance to chemotherapy 20 .
In the current study, the expression profile of FASN was determined in children with ALL. Although our findings showed no increase in FASN mRNA levels in de novo ALL patients compared with the control group, FASN showed significant upregulation in positive MRD patients known as drug resistant group compared with the drug sensitive or MRD− group. These data supported the hypothesis that FASN up-regulation contributes to poor response to chemotherapy. We also examined the potential prognostic value of FASN dysregulation in both intrinsic and adaptive drug resistance using ROC curve analysis. Results showed AUCs of 0.82 and 0.7 in discriminating MRD+ from MRD− (P = 0.0039) new case samples with one year follow up, and relapsed patients from MRD− individuals (P = 0.0187), respectively. Accordingly, it can be hypothesized that FASN might be served as a potential prognostic biomarker in pediatric ALL. The increased expression levels of FASN in a relapsed patient, compared with the expression levels of this gene at the time of diagnosis (0.075 vs. 1.14, respectively) was another interesting support for this hypothesis (data not shown). Investigating larger populations of ALL paired samples in prospective cohort studies may help intensify the validity of these results.
An increasing number of studies have examined the anti-cancer activity of ginger and its bioactive compounds in drug resistant cancer cells 21 . Our group previously reported the anti-leukemic effect of ginger extract. On the other hand, It was shown that fresh normal peripheral MNCs were not significantly sensitive to proliferation inhibition induced by 50% inhibition concentration of the ginger extract 14 . Furthermore, it was shown that this effect was not ascribed to the expression levels of ABC transporters. To identify the possible targets through which ginger may exert its cytotoxic effect on drug-resistant cells, the expression profile of the FASN was analyzed after treating the relapsed ALL patients primary cells with ginger extract. Results revealed that cell death was significantly increased in ginger treated samples compared with the untreated cells. On top of that,  22 .
Considering the possible contribution of FASN in drug resistance, and the negative impact of glucocorticoids on FASN expression levels in B-ALL cell lines 12 , we further investigated the effect of dexamethasone on the sensitive and resistant T-leukemic cells, using RT-PCR followed by MTT assays. The rationale for selecting T cells in these examinations was the aggressive behavior of this phenotype and the low survival rate of T-ALL patients compared with those with B-ALL 23 . Results showed that resistance to dexamethasone was associated with failure to FASN downregulation (Fig. 4b). Considering the cytotoxicity of ginger extract on resistant patient primary cells, combination treatments were designed to determine whether ginger extract may overcome resistance to dexamethasone. Intriguingly, results showed that ginger/dexamethasone combined therapy was associated with decreased expression levels of FASN and cell growth inhibition (Fig. 4b,c). These data may open up new avenues for improved combination therapies against leukemia drug resistance.
In cancer cells dysregulation of de novo FA synthesis and upregulation of enzymes involved in this pathway occur largely at the transcriptional levels through the activation of sterol regulatory element-binding proteins (SREBPs) 24 . The activity of SREBP is regulated by mTORC1, one of the crucial downstream effector of AKT 25 . In both B-cell and T-cell ALL primary bone marrow samples, AKT hyperactivation has been observed 11 . Similar to this data was our comparison between the transcripts levels of FASN in B-ALL and T-ALL patients which revealed no marked difference between these two groups. www.nature.com/scientificreports/ Cancer cells are extremely dependent on de novo lipogenesis for proliferation and survival. Therefore, FASN inhibitors seem to play promising role in cancer treatment. It is shown that FASN inhibitors can induce tumor cell apoptosis and sensitize breast cancer cells to chemotherapies 7 . However, off-target activities and detrimental systemic side effects of such components have prevented their clinical development. On the other hand, anticancer activity of the plant components is currently undergoing preclinical evaluation. Therefore, in the next step of this research, molecular docking was used to determine which one of the ginger phytochemicals can interfere with FASN activity. From the six different catalytic domains of FASN, KS (β-ketoacyl synthase) and TE (thioesterase), known as the first and the last catalytic domains of this enzyme, were chosen 26 . Subsequently, we retrieved twenty ginger phytochemicals from published literatures and investigated their inhibitory interactions with FASN.
Since FASN thioesterase domain is involved in palmitate synthesis termination and also in maintaining of the length of fatty acid chain, it is particularly a promising target to inhibit the enzyme activity (TE dom). Orlistat and fenofibrate are the FASN inhibitors which can prevent tumor growth and induce malignant cell death through blocking the TE domain 15 . In order to predict the inhibitory effects of ginger phytochemicals on FASN activity, they were docked with the crystal structure of TE domain and their binding energies were compared to those of orlistat and fenofibrate, in order to prioritize these ligands. Docking results revealed the binding energies of fenofibrate and orlistat to be − 7.4 and − 6.7 kcal/mol, respectively. Of all docked ginger phytochemicals, gingernone family molecules showed the highest binding affinity to FAS-TE domain (C form: − 7.5 kcal/mol, B form: − 7.4 kcal/mol and A form: − 7.1 kcal/mol) even higher than orlistat which is a US FDA-approved and marketed drug for management of obesity, acting through FAS-TE inhibition 26 (Fig. 5, Fig. S3). The first catalytic domain of FASN is KS 27 . Concerning the KS domain, several inhibitors such as morin, cerulenin and C75 have been reported 28 . The binding free energy of the 20 selected compounds against FAS-KS domain showed the highest binding affinity to the KS in quercetin (− 8.9 kcal/mol), which was even more than morin inhibitor (− 8.5), followed by gingernones (B form: − 7.7 kcal/mol, A form: − 7.7 kcal/mol and C form: − 7.5 kcal/mol) and Isogingernone B (− 7.5 kcal/mol) (Fig. 6), showing much higher binding affinities than the known KS inhibitors, cerulenin and C75 (− 6.5 kcal/mol). Interestingly, it can be noted that gingernones elicit the greater inhibitory effects on both TE and KS domains among all tested ginger compounds. Collectively, these results suggest that ginger may suppress FASN activity and overcome drug resistance through its gingernones. Additional cell-based studies and test tube experiments including dual luciferase (Firefly-Renila) reporter assays are required to confirm the molecular docking data.
In conclusion, our findings emphasize the significance of fatty acid synthesis as a potential target for leukemia treatment. Moreover, ginger constituents are introduced as promising agents able to effectively overcome drug resistance by possibly reducing FASN expression level and inhibiting its activity. Followed by additional FASN overexpression and knockdown studies confirming the causative role of ginger ingredients in FASN downregulation and drug sensitivity, this laboratory investigation could be taken into consideration for the design of animal model studies followed by clinical trials to evaluate the effect of combined treatment of ginger constituents and chemotherapeutic drugs in multidrug resistant leukemia.

Materials and methods
in vitro studies. Patients and control samples. 65 children with ALL and 22 non-cancer controls were included in the present study. Individuals were referred to Sayed-ol-Shohada Hospital, Isfahan, Iran in 2014-2017 for bone marrow evaluation. The project was performed in accordance with the Declaration of Helsinki and permitted by the Ethics Committee of the University of Isfahan (agreement number 94/31540). All Samples of children with ALL and non-cancer controls were collected with full written informed parents' consents in compliance with the ethical protocols and standards of Sayed-ol-Shohada Hospital. Two to five milliliters of bone marrow heparinized sample was collected from cALL patients and controls and sent on ice to the Cellular and Molecular Biology laboratory of University of Isfahan. Mononuclear cells (MNCs) were isolated by sedimentation on lymphoprep density gradients (Axis Shailed Diagnostics Ltd., Oslo, Norway), according to the manufacturer recommended protocol.
Cell lines and patient primary cells. CCRF-CEM (derived from a 4-year-old girl with T-ALL) human cell lines was purchased from Pasteur Institute (Tehran, Iran). Multidrug resistant CCRF-CEM/MVCD subline was generated in-house. Briefly, CCRF-CEM cells sequentially exposed to stepwise concentrations of Methotrexate (MTX) from 5 nM to 1.2 μM. In order to allow cells reaching regular growth rate, cells were kept in the same concentration of MTX for two or three passages. After full growth recovery, the concentration of MTX was increased by twofold each time. Finally, it was revealed that CCRF-CEM/MVCD subline had developed cross-resistance to a number of other chemotherapy drugs including dexamethasone. For further experiments, parental and resistant cell lines were cultured in RPMI1640 containing 10% (v/v) heat-inactivated FBS and 100 μg/ml streptomycin Scientific RepoRtS | (2020) 10:14072 | https://doi.org/10.1038/s41598-020-70839-9 www.nature.com/scientificreports/ and 1% (v/v) 100 IU/ml penicillin. Freshly collected patient samples were grown in RPMI-1640 supplemented with 20% FBS and 1% l-glutamine.
Cell treatment. Cell lines were seeded in 96-well cell culture plates at a density of 15 × 10 4 cells per well. Cells were suspended in 100 μl supplemented media and treated with 50 μl of freshly made ginger extract (167 μg/ml) for 72 h. Combination treatment was carried out by the addition of 25 μl ginger extract to the same volumes of dexamethasone (1,000 μM), followed by the same incubation time. MTT assay was initiated by the addition of MTT dye to each well. After 3 h incubation at 37 °C, in order to dissolve the formazan crystals in each well, the supernatant was removed and replaced with 100 μl of DMSO. The absorbance was measured at a wavelength of 492 nm using a Stat Fax-2100 microtiter plate reader (Palm City, FL). Cell viability ratio was evaluated as mentioned before 14 .
Flow cytometry analysis. Mononuclear cells isolated from relapsed patients samples and normal MNCs were seeded at a density of 25 × 10 4 cells per well and treated with 167 μg/ml ginger extract for 48 h. At the end of the treatment period, cells were harvested, washed with PBS supplemented by 0.5% FBS, resuspended in 100 μl of cold 1 × Annexin-V-binding buffer after centrifugation, and incubated with 5 μl of FITC conjugated Annexin-V and 10 μl of PI at room temperature for 15 min. The quantitative analysis of cell death was conducted by BD FACSCalibur Flow Cytometer (London, UK). Data was acquired and analyzed using Cell Quest Pro (BD Biosciences, San Jose, CA) and FlowJo softwares (Tree Star Inc., Ashland, OR).
RNA extraction and cDNA synthesis. In accordance with the manufacturer's protocol, total RNA isolation was performed from treated cell lines as well as MNCs of patient and control samples using TRIzol reagent. Extracted RNA was transformed to cDNA in accordance with the instructions provided by PrimeScript RT reagent kit (Takara, Japan) utilizing random hexamers and oligo dT primers. The obtained cDNA was preserved at − 20 °C for further analyses.
Real-time PCR analysis. Gene expression assessment was conducted utilizing ExiLENT SYBR Green master mix and Chromo4 system (Bio-Rad, Foster City, CA), according to the manufacturer's instructions. Data normalization was carried out utilizing glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as the internal control gene. qRT-PCR was carried out in duplicate according to the following cycling conditions: 5 min pre-incubation at 95 °C followed by 95 °C denaturation for 20 s, 60 °C annealing for 30 s, and 72 °C product expansion for 30 s. All relative expression levels were evaluated and reported using the 2 −ΔΔCt method. The forward and reverse primers sequences for FASN were CCG CTT CCG AGA TTC CAT CCT ACG C and GGA TGG CAG TCA GGC TCA CAA ACG; and for GAPDH were GCC CCA GCA AGA GCA CAA GAG GAA GA and CAT GGC AAC TGT GAG GAG GGG AGA TT, respectively.
Response to chemotherapy. ALL patients were treated based on the Australian and New Zealand Children's Cancer Study Group ALL study (https ://www.anzct r.org.au/trial _view.aspx?ID=1568). For evaluation of the treatment response, at the end of the first year, the presence of minimal residual disease (MRD) in new patients was assessed utilizing PCR-SSCP (Polymerase chain reaction coupled single-strand conformation polymorphism) analyses for T-cell receptor gamma (TcRγ) and immunoglobulin heavy chain (IgH) gene rearrangements. MRD provides evidence for the presence of post-therapeutic leukemia cells within the bone marrow and more rarely in peripheral blood circulation. Persistent monoclonality, referred to as MRD+, was considered as drug resistance and non-response chemo-treatment. MRD− individuals were appointed as drug sensitive patients.
Statistical analysis. SPSS23.0 and GraphPad Prism8.0.2 softwares were used to analyze the data of each experiment statistically. The Kolmogorov-Smirnov normality (KS test) and Shapiro-Wilk tests were applied to evaluate the normality of data distribution. Kruskal-Wallis and Mann-Whitney two tailed tests were carried out to compare the difference of continuous variables between two groups. The statistical significance of differences between two sets of data were estimated using unpaired nonparametric t test. Receiver operating characteristic (ROC) curves and the area under the ROC (AUC) were depicted, using GraphPad Prism, to evaluate the specificity and sensitivity of FASN as a prognostic biomarker for ALL patients. The greater the area under the curve, the more accurate the test. All data were expressed as mean ± standard error of mean (SEM). Data were results of 2 to 3 independent experiments which were performed in triplicates for cell line analyses, and in duplicates for patient samples. P < 0.05 was considered significant, statistically.
in silico studies. In order to determine the inhibitory impact of ginger phytochemicals on hFASN through in silico studies, various biologically active ginger compounds were selected from literatures [29][30][31][32] . Two domains of the enzyme were selected to perform docking simulation study; the ketosynthase (KS) and thioesterase (TE) domains. These domains are particularly important in targeting and inhibiting the FASN activity since the KS domain initiates the fatty acid synthesis cycle 27 and the TE domain terminates the cycle by hydrolyzing the thioester bond, which results in releasing the 16-carbon fatty acid, palmitate 33 . To determine and prioritize the significance of the docking results of the chosen phytochemicals, some known inhibitors of these two domains were selected and their binding affinities were determined as well 26 .
Docking simulation. Methods for selection and preparation of ligands and receptors were described in the supplementary methods. All molecular docking simulations were performed using AutoDock Vina 1. www.nature.com/scientificreports/ windows platform. Compounds and the protein preparations as well as results analyses were done using UCSF Chimera 1.13.1 35 , AutoDockTools 1.5.6 36 and LIGPLOT + 2.1 37 . AutoDock Vina was used to perform semi-flexible docking simulations in which proteins were considered as rigid, while ligands were allowed to be completely rotatable. Since the majority of our ligands were conformationally-flexible with large numbers of rotatable bonds, exhaustiveness was set to 24 for each docking simulation 38 . Other parameters were set to default. Grid boxes were defined for each domain using AutoDockTools. For TE domain, chain B of the pdb file (ID: 2PX6) was used. Subsequently, the Grid box was adjusted around the catalytic triad (Ser2308, Asp2338, His2481), specificity channel and interface cavity, all of which are important for substrate binding as described by John et al. 16 . For KS domain, chain A of the pdb file (ID: 3HHD) was utilized and since it was comprised of KS-MAT didomain, the KS domain was isolated by removing the sequence from Pro410 to Pro824 as described by Pappenberger et al. 39 . The Grid box was then adjusted around the active site residues (Cys161, His293, His331) which are deep inside the active-site cavity and the distal substrate binding sites 40 . Docking simulation was done for each ligand and results were compared and analyzed.