Using the Matrix-Induced Ion Suppression Method for Concentration

Sep 10, 2015 - Studies of the cell metabolome greatly improve our understanding of cell biology. Currently, most cellular metabolomics studies control...
0 downloads 5 Views 4MB Size
Download PDF
Article pubs.acs.org/ac

Using the Matrix-Induced Ion Suppression Method for Concentration Normalization in Cellular Metabolomics Studies Guan-Yuan Chen,†,‡ Hsiao-Wei Liao,†,‡ I-Lin Tsai,†,‡ Yufeng Jane Tseng,†,‡,§,∥ and Ching-Hua Kuo*,†,‡,⊥ †

School of Pharmacy, College of Medicine, National Taiwan University, No. 33, Linsen S. Rd., Chongcheng Dist., Taipei, 10051 Taiwan ‡ The Metabolomics Core Laboratory, Center of Genomic Medicine, National Taiwan University, No.2, Xuzhou Rd., Zhongzheng Dist., Taipei 10055, Taiwan § Graduate Institute of Biomedical Electronics and Bioinformatics, National Taiwan University, No.1, Sec. 4, Roosevelt Rd., Zhongzheng Dist., Taipei 10090, Taiwan ∥ Department of Computer Science and Information Engineering, National Taiwan University, No.1, Sec. 4, Roosevelt Rd., Zhongzheng Dist., Taipei 10090, Taiwan ⊥ Department of Pharmacy, National Taiwan University Hospital, No.7, Zhongshan S. Rd., Zhongzheng Dist., Taipei 10002, Taiwan S Supporting Information *

ABSTRACT: Studies of the cell metabolome greatly improve our understanding of cell biology. Currently, most cellular metabolomics studies control only cell numbers or protein content without adjusting the total metabolite concentration, mainly because of the lack of an effective concentration normalization method for cell metabolites. This study proposed a matrix-induced ion suppression (MIIS) method to measure the total amount of cellular metabolites by utilizing flow injection analysis coupled with electrospray ionization mass spectrometry (FIA-ESI-MS).We used series dilutions of HL-60 cell extracts to establish the relationship between cellular metabolite concentration and the degree of ion suppression of the ion suppression indicator, and a good correlation was obtained between 2- and 12-fold dilutions of cell extracts (R2 = 0.999). Two lung cancer cells, CL1-0 and CL1-5, were selected as the model cell lines to evaluate the efficacy of the MIIS method and the importance of metabolite concentration normalization. Through MIIS analysis, CL1-0 cells were found to contain metabolites at a concentration 2.1 times higher than in CL1-5, and the metastatic properties of CL1-5 could only be observed after 2.1-fold dilution of CL1-0 before metabolomic analysis. Our results demonstrated that the MIIS method is an effective approach for metabolite concentration normalization and that controlling metabolite concentrations can improve data integrity in cellular metabolomics studies.

W

has deepened scientific understanding in cancer biology, stem cell research, and neuron degenerative diseases.9−14 In contrast to genomic and proteomic studies that control the total amount of nucleic acid or protein, most cellular metabolomics studies use controlled cell numbers.12,14−24 Few cellular metabolomics studies have normalized the protein content25−28 or DNA content29 but not the total metabolite concentrations, which is probably due to the lack of generally accepted metabolite normalization methods. Various statistical normalization methods have been developed for both mass-spectrometry-based and nuclearmagnetic-resonance-based metabolomics.30−32 Although normalization of metabolomic data by chemometric approaches

ith the growing importance of omic sciences, cell-based omics studies in a controlled environment have been widely used in drug discovery, cell growth and differentiation, cancer therapy, toxicology, and stem cell studies.1,2 For genomics and proteomics studies, it is common practice to adjust the total amount of nucleic acid or protein before data acquisition to provide unbiased comparisons between studied cells.3,4 For example, to better understand the molecular mechanisms underlying how the metastatic property of nonsmall cell lung cancer is affected at the protein level, the nonmetastatic CL1-0 and highly metastatic CL1-5 cell lines were compared using a fixed amount of proteins.5,6 Numerous methods have been proposed to measure the total concentration of proteins and nucleic acid.7,8 Metabolomics is the latest omic science, and it has emerged as a powerful tool to correlate phenotype with metabolic changes. The successful application of cell-based metabolomics © XXXX American Chemical Society

Received: May 19, 2015 Accepted: September 8, 2015

A

DOI: 10.1021/acs.analchem.5b01869 Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

■

can greatly improve the reproducibility of quality control samples and can be used to calibrate the total metabolite concentration in individual samples,33 the calibration results may not accurately reflect the true concentration of metabolites in samples. The dynamic ranges of metabolite concentrations are quite wide, and the signals of high-concentration metabolites may become saturated in the mass spectrometer detector. Moreover, polar metabolites are difficult to retain in reversed-phase (RP) columns, and their signal intensities are often affected by the matrix effect.34 With the matrix effect, the MS signal intensity may not reflect the true metabolite concentration in the samples. Therefore, it is important to have effective experimental methods to normalize metabolite concentrations. Experimental calibration methods in metabolomics are mostly designed for urinary metabolomics studies. In urinary metabolomics studies, creatinine is frequently used to normalize urine concentrations, as the excretion of creatinine is relatively constant under normal conditions.35 The osmolarity method has also been widely used for urine concentration normalization.36 Recently, Wu and Li used chemically labeled metabolites as a means of metabolome sample normalization and sample loading optimization in mass-spectrometry-based metabolomics.37 This method might be applicable for various types of samples, including cell extracts. However, it requires chemical labeling for each sample. Considering that there is no convenient method for metabolite concentration normalization in cellular metabolomics studies, this study proposes using the matrix-induced ion suppression (MIIS) method for estimation and adjustment of total cellular metabolites by flow injection analysis coupled with electrospray ionization mass spectrometry (FIA-ESI-MS). The MIIS method uses an ion suppression indicator (ISI) to evaluate the extent of MIIS and to estimate the relative concentration of metabolites in studied samples. In our previous study, we demonstrated that MIIS is an accurate method for concentration normalization in urinary metabolomics studies, and the results of MIIS analysis showed a high correlation with data obtained by the creatinine and osmolarity methods for the same urine samples.38 To resolve the normalization problem in cellular metabolomics studies, we applied the MIIS method to normalize cellular metabolite concentrations, using two human lung adenocarcinoma cell lines as our study model. CL1-0 and CL1-5 are both nonsmall cell lung cancer cells. The CL1-5 cell line was derived from the human lung adenocarcinoma CL1-0 cell line using a transwell invasion chamber to select progressively more invasive cancer cell populations.39,40 Because of their distinct invasion properties, their morphologies and sizes are quite different. Controlling cell numbers is the current standard approach for cell-based metabolomics studies. This study compared the differences in the metabolomics analysis results obtained by controlling cell numbers and after concentration normalization by the MIIS method. We optimized the FIA-ESI-MS conditions of the MIIS method using cell extracts. The MIIS method optimized for cellular metabolomics study was applied for concentration normalization of CL1-0 and CL1-5. The importance of concentration normalization for cellular metabolomics studies was discussed in reference to the compared results.

Article

EXPERIMENTAL SECTION

Chemicals. Hexakis (1H,1H,3H-perfluoropropoxy)phosphazene (HKP) was purchased from Apollo (Apollo, Graham, NC, USA). LC-MS grade acetonitrile and Tris-Cl were obtained from J.T. Baker (J.T. Baker, Phillipsburg, NJ, USA). MS-grade water and methanol were bought from Scharlau (Scharlau, Sentmenat, Barcelona, Spain). Formic acid, β-mercaptoethanol, Coomassie blue reagent, glycerol, and bromophenol blue were purchased from Sigma (Sigma, St. Louis, MO, USA) Cell Culture and Sample Preparation. The lung carcinoma cells CL1-0 and CL1-5 were kindly provided by Prof. Min-Liang Kuo’s lab (National Taiwan University, Taipei, Taiwan). CL1-0 was the parental cell, and CL1-5 was the subline selected from CL1-0. Both were cultured in RPMI-1640 medium (Gibco, Gaithersburg, MD, USA) with 10% fetal bovine serum (Gibco) and 1% penicillin/streptomycin (Invitrogen, Carlsbad, CA, USA) at 37 °C in 5% CO2 and were then subcultured upon reaching 80−90% confluence. The medium was renewed every 2 days. Malarvu hepatotic carcinoma cells, hepG2 cells, colon cancer HCT 16 cells, and suspensive leukemia HL-60 cells were also used to verify this model. The culture conditions for Malarvu hepatotic carcinoma cells (Malarvu), hepG2 cells, colon cancer HCT 16 cells, and HL-60 were the same as those described for CL1-0 and CL1-5. Cell were washed twice with pH 7.4 phosphate buffered saline (PBS) and then trypsinized after harvesting. Suspended cells were counted, and their size measured with a Cellometer (Nexcelom, Lawrence, MA, USA).Controlled cell numbers (1 × 106) were washed twice with PBS, which was then removed by centrifugation at 200× rcf. A volume of 500 μL of ice-cold methanol was added to the cell pellets, and the suspension was vortexed. The extracts were put on ice for 5 min. A volume of 500 μL of ice-cold DI water was further added to the cell extract and vortexed, followed by placing the samples on ice for 5 min. The supernatants were collected by centrifugation at 15 000× rcf for 5 min at 4 °C and filtered with a 0.22 μm PP membrane (RC-4, Sartorius, Göttingen, Germany). Serial dilutions of the cell extracts were prepared by diluting the cell extracts with 50% methanol into 2-, 4-, 6-, 8-, 10-, and 12fold diluted samples. Cell samples were mixed with 3000 ng/ mL HKP in 50% methanol (1:1, v/v) and then transferred into glass inserts for FIA-ESI-MS analysis. Protein Analysis. A total of 1 × 106 cells was counted, and samples were washed twice with PBS. The PBS was removed, and then, the cell pellets were lysed using 0.1% Triton X-100 for protein extraction. A 2-μL aliquot of 6-fold sample buffer (containing 60 mM Tris-Cl, pH 6.8, 2% SDS, 10% glycerol, 5% β-mercaptoethanol, and 0.01% bromophenol blue) was added to 10 μL of cell extract and mixed. The mixtures were heated in a water bath at 95 °C for 5 min to allow the proteins to denature into negatively charged linear structures and then quickly put on ice for 10 min to avoid protein annealing. The 10 μL protein extract samples was loaded onto 10% SDS-PAGE for separation. Electrophoresis was performed at 100 V for 180 min. The SDS-PAGE was carefully rinsed with deionized water and fixed with 7% acetic acid in 40% MeOH for 1 h. Gels were further stained with Coomassie blue reagent for 2 h and then destained with 10% acetic acid in 25% MeOH for 60 s. The gel was further rinsed with 25% MeOH until the background was clear for observation. The protein concentrations were B

DOI: 10.1021/acs.analchem.5b01869 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

nucleic acids. Total metabolites, which contain polar lipid and metabolites, could therefore be used to represent cellular matrix. In electrospray ionization, the level of ion suppression is proportional to the concentration of sample matrix, and the MIIS method uses this phenomenon to estimate relative metabolite concentrations. The workflow of the MIIS method is diagrammed in Figure 1. We used reference cells to correlate

determined using the Bradford protein assay (Biorad, Hercules, CA, US). FIA-ESI-MS Analysis. FIA-ESI-MS analysis was performed using an Agilent 1290UHPLC system coupled with an Agilent 6460 QqQ (Agilent Technologies, Santa Clara, CA, USA). A Jet Stream electrospray ionization source was used for sample ionization. The following parameters were used throughout the study: 100 °C gas temperature, 3 L min−1 gas flow, 10 psi nebulizer, 150 °C sheath gas temperature, 3 L min−1 sheath gas flow, 40 kV capillary voltage in positive mode, and 120 V fragmentation voltage. The mobile phase was composed of 50% A and 50% B at a flow rate of 0.2 mL min−1, where A was 0.1% formic acid and B was 0.1% formic acid in acetonitrile. A 2 μL sample volume was injected into ESI-MS by FIA. Selected ion monitoring (SIM) mode was used for monitoring the ISI signals. To estimate the sample matrix concentration by the FIA-ESIMS (the MIIS method), the ion suppression indicator (ISI) was spiked in the blank solution and the cell extracts. The signal reduction was calculated as the extent of ISI in blank solution subtracted by ISI in cell samples. The calibration curve was constructed by signal reduction and reference cell concentration. Metabolic Profiling. Metabolic profiling of CL1-0 and CL1-5 was performed using an Agilent 1290UHPLC system (Agilent Technologies, Santa Clara, CA, USA) coupled with a Burker maXis QTOF (Bruker Daltonics, Bremen, Germany). A 2 μL sample of cell extract was injected into an Acquity HSS T3 column (2.1 × 100 mm, 1.8 μm, Waters, Milford, MA, USA), and the analytical column was maintained at 40 °C. The mobile phase was composed of solvent A, water/0.1% formic acid, and solvent B, acetonitrile/0.1% formic acid. The gradient elution program was as follows: 0−1.5 min, 2% B; 1.5−9 min, linear gradient from 2% to 50% B; 9−14 min, linear gradient from 50% to 95% B; hold at 95% B for 3 min. The flow rate was 300 μL min−1. For sample ionization, an Ion Booster electrospray ionization source was employed with a capillary and end plate offset voltage of 1 k and 400 V in both positive and negative modes. The MS parameters were set as follows: 200 °C drying gas temperature, 4 L/min drying gas flow, 60 psi nebulizer flow, and 300 °C vaporizer temperature. The mass parameter selection aims to achieve better sensitivity for low concentration metabolites and to avoid signal saturation for high concentration metabolites. The mass spectrometer was calibrated with 10 mM sodium formate before daily use and a lockmass between runs. Metabolites were identified using the in-house library with accurate masses and retention times. Data Analysis. All of the HKP extracted ion chromatography (EIC) data were collected and exported to .XY files using Bruker Data Analysis software. Integration of peak areas and regression analysis were accomplished using Microsoft Excel 2007 (Albuquerque, NE). Data obtained using the Agilent triple quadruple were converted into comma-separated values (csv) format and processed using Microsoft Excel 2007. These data were further plotted using R programming software.41

Figure 1. Workflow of MIIS-based approach for adjusting cell concentrations in cellular metabolomics studies.

the cellular metabolite concentrations and level of ISI ion suppression by measuring the differences in peak area of the FIA chromatograms. The signal reduction of ISI in cell extract was calculated through reference to the measurement of ISI signal in blank solution. Reference cell extracts were diluted in series and spiked with a fixed amount of ISI. After analysis with FIA-ESI-MS, a calibration curve was constructed between the degree of ion suppression and the cell extract dilution factors, which can be used to estimate the relative metabolite concentrations in tested cells. Optimization of the FIA-ESI-MS Conditions and Deduction of Relative Concentrations of Cell Extracts by the MIIS Method. Hexakis (1H,1H,3H-perfluoropropoxy) phosphazene (HKP, MW: 921) was demonstrated to be an effective ISI in our previous urinary metabolomics study.38 In the current study, HKP was again selected as the ISI for cellular metabolomics experiments. For method optimization, 3 μg mL−1 HKP was prepared in 50% methanol and mixed at a 1:1 ratio with cell extracts or blank solution, followed by FIA-ESIMS analysis. Most cellular metabolomics studies use 1 × 106 cells as the study material. Compared to urine samples, cell extracts containing 1 × 106 cells are highly diluted. Because the MIIS method measures the extent of ion suppression caused by the sample matrix, the diluted character of cell extracts makes the ion suppression behavior less obvious if the MS parameters

■

RESULTS Description of the MIIS Method. The MIIS method measures the extent of matrix-effect-induced signal suppression of the ion suppression indicator (ISI) to estimate the sample matrix concentration. Polar lipid and small metabolites are the major components in the cell extracts that were pretreated with methanol to remove macromolecules such as proteins and C

DOI: 10.1021/acs.analchem.5b01869 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

formic acid in acetonitrile as the mobile phase, and the flow rate was set at 0.1 mL min−1. The cell extract sample injection volume was 2 μL. Under the optimal FIA-ESI-MS conditions, a linear signal reduction was observed at cell extract dilutions from 2- to 12-fold (R2 = 0.999; y = 17 287x2 − 433 429x + 3 603 192; Figure 3). To verify FIA-ESI-MS as a universal method to adjust the total metabolite concentration for various cells, this approach was used to establish regression curves on serial dilutions of CL1-0, HL-60, HCT-16, Malarvu, and hepG2 cells. The correlation coefficients for all tested cell extracts were all higher than 0.99 (Figure S1). To determine the total metabolite concentration in test samples, all test samples should be mixed with HKP standard solution, followed by FIA-ESI-MS analysis. The extent of HKP signal reduction in each test sample can be used to deduce the relative total metabolite concentration using the abovementioned calibration curve, which can also be used for concentration normalization before metabolomics studies. Physical Characteristics of the Cell Model for the Cellular Metabolomics Study. CL1-0 and CL1-5 are cells derived from nonsmall cell lung cancer lines with different metastatic properties.39,40 The CL1-5 cells are mesenchymallike with a spindle-like shape and are more malignant, whereas CL1-0 cells are epithelial-like with a round shape and are less malignant (Figure 4a).42 The distinctions in morphology, size, and malignancy between CL1-0 and CL1-5 provided an ideal model to demonstrate the effectiveness of the MIIS method for metabolite concentration normalization and to elucidate the importance of controlling total metabolite concentrations in cellular metabolomics studies. Figure 4b shows that the cellular diameter of CL1-5 cells is less than that of CL1-0 cells. The diameter of suspended CL1-5 is approximately 17 μm, whereas that of CL1-0 is approximately 21 μm, suggesting the volume of CL1-0 cells to be significantly larger than that of CL1-5 cells. To understand whether the differences in cell size resulted in different metabolite concentrations, we first measured the protein quantities of these two cell lines. Proteins were separately extracted in CL1-0 and CL1-5 samples with cell numbers of 1 × 106, and the protein quantities in the extracts were analyzed by SDS-PAGE (Figure 4c). The results indicated that the protein quantities in CL1-0 were 1.4 times higher than those in CL1-5. We further investigated the differences in metabolite concentration between the two cell lines. Using the MIIS Method for CL1-0 and CL1-5 Metabolite Concentration Normalization. Metabolites in CL1-0 and CL1-5 samples with cell numbers of 1 × 106 were extracted using MeOH and water to remove macromolecules (protein, nucleic acid, and cell debris), and the remained metabolites of extracts were mixed with HKP standard solution for MIIS measurement. The CL1-0 cell extract exhibited a greater inhibition of HKP signal, which suggested that the metabolite concentration in CL1-0 is higher than that in CL1-5 for the same number of cells (Figure 5a). The regression equation obtained from HL-60 cells (y = 17 287x2 − 433 429x + 3 603 192) was used to estimate the relative total metabolite concentration in CL1-0 and CL1-5. HL-60 is a cell line composed of suspension leukemia cells. The reason we used HL-60 as the reference cell is due to the convenience in operation of this cell line in the related cell culture procedures. The results of MIIS analysis indicated that the metabolite concentration in CL1-0 was 2.1 times higher than that in CL1-

were not optimal. The drying gas temperature showed the most significant effect on the MIIS performance. Drying gas temperatures were tested from 300 to 100 °C to evaluate their effect on matrix-induced ion suppression behavior in the ISI (HKP). The test range of MIIS used 2- to 8-fold dilutions of cell extract, and the results are shown in Figure 2. At drying

Figure 2. Effect of the drying gas temperature on MIIS performance. MIIS analysis of a series dilution of cell extracts at drying gas temperatures of (a) 300, (b) 200, and (c) 100 °C.

temperatures of 300 and 200 °C, the differences in the degree of ion suppression caused by 2- to 8-fold diluted cell extract was not obvious (Figure 2a,b). A linear correlation between the degree of dilution and extent of ion suppression was only obtained at a drying temperature of 100 °C (Figure 2c). In contrast to the drying gas temperature, decreasing the sheath gas temperature from 350 to 250 °C did not improve the correlation between the degree of dilution of the cell extracts and the extent of ion suppression. The effect of LC parameters on the performance of the MIIS method was less important than that of the ESI parameters. Finally, we selected 0.1% D

DOI: 10.1021/acs.analchem.5b01869 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 3. (a) Overlaid extracted ion chromatograms showing matrix-dependent signal reduction of HKP in HL-60 cell extracts with different degrees of dilution. (b) Quadratic regression curve of the degree of signal reduction versus dilution factor (R2 = 0.999; y = 17 287x2 − 433 429x + 3 603 192). H represents HKP.

Figure 4. Morphology and protein content comparison between CL1-0 and CL1-5. (a) Microscope images used for morphological analysis of cell shape; (b) distribution of cell diameters determined by Cellometer; (c) SDS-PAGE results showing the total protein content in a uniform number of cells (1 × 106).

5; therefore, CL1-0 was diluted 2.1-fold with 50% methanol before metabolomic analysis. To verify the accuracy of the estimation results, the extents of ISI signal suppression were measured again after 2.1-fold dilution of the CL1-0 extracts. The EIC of ISI revealed that the signal suppression was similar after concentration normalization of CL1-0 and CL1-5 (Figure

5), which indicated that their total metabolites were present in similar concentrations after adjustment. Five exogenous compounds including 15N2-theophylline, D8-phenylanaline, oxazepam, flunitrazepam, and amphetamine with different retention times were spiked into cell extracts of CL1-0 and CL1-5 after being normalized according to MIIS analytical results. The spiked samples were analyzed with LCE

DOI: 10.1021/acs.analchem.5b01869 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 5. FIA chromatograms of CL1-0 and CL1-5 extracts obtained before (a) and after (b) concentration normalization by the MIIS method.

MS. The results indicated that the peak areas of the five compounds were similar in the two cell extracts which demonstrated that the MIIS method is effective for concentration normalization in cellular metabolomics studies (Figure S2). Comparison of Metabolomics Study Results Using Controlled Cell Numbers or Controlled Total Metabolite Concentration. Controlling cell numbers is widely applied in cellular metabolomics studies. CL1-0 and CL1-5 samples with cell numbers of 1 × 106 were extracted using MeOH and water, followed by metabolic profiling by liquid chromatography timeof-flight mass spectrometry (LC-TOFMS) analysis. The base peak chromatogram (BPC) of metabolic profiles of CL1-0 and CL1-5 showed that the signals for CL1-0 were generally much larger than those of CL1-5. The differences in metabolite concentration between CL1-0 and CL1-5 were further investigated. The metabolites in CL1-0 and CL1-5 were identified using our in-house library containing 160 standards, and the identified signals were compared with their peak areas. A total of 73 metabolites were identified in these two cell extracts. More than 80% of the identified metabolites were found in higher quantities in CL1-0 than in CL1-5 (Figure 6a, Table S1), which is unreasonable considering the homeostasis of cell biology. A metabolomic study of CL1-0 and CL1-5 was also performed using the controlled metabolite concentration approach. Figure 6b shows the relative changes between CL1-0 and CL1-5 obtained after metabolite concentration normalization by the MIIS method. Compared to Figure 6a, the relative metabolite changes were quite different after adjusting the total metabolite concentration.

To maintain the homeostasis state in cells, the amount of upregulated and down-regulated metabolites should be similar. Therefore, the mean metabolites’ fold changes of CL1-0 to CL1-5 should theoretically be close to 1. To verify this concept, the mean fold changes of the metabolites were calculated using the data obtained before and after MIIS adjustment. We found that the average value of the fold changes of the identified metabolites calculated before adjustment is about 2.01, and the value became 1.01 when calculated using the fold change data obtained after MIIS adjustment. We additionally compared the results obtained by the protein normalization method (adjusted to total protein concentration), and the fold change average value was 1.44. The results suggested that the MIIS method could provide more unbiased metabolomics comparison when compared to controlling cell numbers and the protein normalization method. We further discuss the dysregulated metabolites in CL1-0 and CL1-5. Metabolites related to glycolysis (hexose, pentose, hexose-6 phosphate) and polyamine synthesis (N-acetylputrescine, 5′-methylthioadenosine and adenosine) had higher peak areas in CL1-5, whereas CL1-0 cells showed a higher concentration of TCA cycle intermediates (succinate, pyruvate, malate) and the antioxidant, glutathione. CL1-5 is characterized by its higher invasion property, malignancy, and proliferation compared to CL1-0. Previous studies have shown that the doubling time of CL1-5 is shorter than that of CL1-0.43 Our metabolomic results show that higher content of metabolites involved in glycolysis implies the Warburg effect is more prominent in CL1-5 supporting the proliferation of the cell. Increased polyamine levels have been associated with increased cell proliferation, tumor invasion, and metastasis.44 Elevation of polyamine metabolism in CL1-5 may facilitate metastasis and F

DOI: 10.1021/acs.analchem.5b01869 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 6. Fold change of the identified metabolites before (a) and after (b) concentration normalization by the MIIS method. Positive values stand for higher expression in CL1-0; negative values stand for higher expression in CL1-5, and red lines indicate 33% range of changes.

degree of ion suppression and degree of dilution of cell extracts. A linear correlation was only observed when we decreased the drying gas temperature. This phenomenon was not observed in urinary samples. As the cell extracts are much more diluted than urine samples, the nonvolatile components in cell extracts are assumed to be much lower. Therefore, the amount of nonvolatile components between cell extracts with different degrees of dilution did not differ greatly. Under high drying gas temperature, the ionization efficiency was high, and the small changes in ionization efficiency caused by the nonvolatile components were not obvious. When the drying gas temperature was decreased, the poor ionization efficiency was more readily affected by the nonvolatile components in cell extracts. Therefore, a linear relationship between the degree of ion

invasion in this line. The higher concentration of TCA metabolites in CL1-0 suggests that it produces energy by a more aerobic pathway which could be associated with its lower proliferation rate compared to CL1-5.

■

DISCUSSION Several mechanisms have been proposed as the causes of ion suppression in the Supporting Information. One theory recognized as one of the main causes of ion suppression suggests that nonvolatile components in the matrix precipitate with analytes and decrease the ionization efficiency by depressing droplet formation or evaporation.45 In this study, we observed that the ESI parameter is the most critical parameter that affects the linear relationship between the G

DOI: 10.1021/acs.analchem.5b01869 Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

■

suppression and the level of dilution can only be observed under lower drying gas temperature. The MIIS method was used to estimate and adjust the total metabolite concentration for CL1-0 and CL1-5. CL1-0 and CL1-5 are both nonsmall cell lung cancer cells but show distinct metastatic properties, and these different properties result in significant morphology and size differences (Figure 4). The cell diameter of CL1-0 was 1.3-fold that of CL1-5, and the protein content of CL1-0 was also found to be 1.4-fold higher for controlled cell numbers. The results of MIIS analysis indicated that the total metabolite concentration was 2.1-fold higher in CL1-0 than in CL1-5. Although controlling cell numbers is the most widely used approach for cellular metabolomics studies, we found that the metastatic properties of CL1-5 could only be observed when using the controlled total metabolite concentration approach. The relative change in the identified metabolites revealed that more than 80% of the metabolites concentrations were higher in CL1-0 compared to CL1-5 when using the controlled cell number approach (Figure 6a). The higher concentration in CL1-0 relative to CL1-5 might be partially due to the larger size of CL1-0 cells compared to CL1-5. We therefore concluded that typical cell-based metabolomics studies that only control cell numbers without considering the difference in total metabolite concentration between different cells may lead to a biased comparison. For proteomic studies, it is a common approach to normalize the protein amount in sample cell extracts instead of controlling cell numbers. Our results suggested that normalizing the metabolite concentration using cellular matrix concentration provided by the MIIS method could also improve the data integrity for metabolomics studies when compared to the controlling cell number approach. The MIIS method represents an efficient and effective method for the estimation and adjustment of total cellular metabolites before metabolomics studies, which could greatly improve the quality of cell-based metabolomic studies.

AUTHOR INFORMATION

Corresponding Author

*Tel: +886.2.33668766. Fax: +886.2.23919098. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This study was supported by the Ministry of Science and Technology of Taiwan (MOST 103-2321-B-002-094; MOST 103-2320-B-002-013). The authors thank the NTU Integrated Core Facility for Functional Genomics of National Research Program for Genomic Medicine of Taiwan for technical assistance.

■

REFERENCES

(1) Hu, Z. Z.; Huang, H.; Wu, C. H.; Jung, M.; Dritschilo, A.; Riegel, A. T.; Wellstein, A. Methods Mol. Biol. 2011, 719, 547−571. (2) Gao, H.; Dahlman-Wright, K. EMBO J. 2011, 30, 2516−2517. (3) Fan, G.; Wrzeszczynski, K. O.; Fu, C.; Su, G.; Pappin, D. J.; Lucito, R.; Tonks, N. K. Biochem. J. 2015, 465, 433−442. (4) Iorio, M. V.; Ferracin, M.; Liu, C. G.; Veronese, A.; Spizzo, R.; Sabbioni, S.; Magri, E.; Pedriali, M.; Fabbri, M.; Campiglio, M.; Menard, S.; Palazzo, J. P.; Rosenberg, A.; Musiani, P.; Volinia, S.; Nenci, I.; Calin, G. A.; Querzoli, P.; Negrini, M.; Croce, C. M. Cancer Res. 2005, 65, 7065−7070. (5) Tian, T.; Hao, J.; Xu, A.; Luo, C.; Liu, C.; Huang, L.; Xiao, X.; He, D. Cancer Sci. 2007, 98, 1265−1274. (6) Beckmann, T.; Thute, T.; Heinrich, C.; Buntemeyer, H.; Noll, T. BMC Proc. 2011, 5 (Suppl 8), P92. (7) Peterson, G. L. Anal. Biochem. 1977, 83, 346−356. (8) Munro, H. N.; Fleck, A. Methods Biochem. Anal. 1966, 14, 113− 176. (9) Kumar, K. K.; Goodwin, C. R.; Uhouse, M. A.; Bornhorst, J.; Schwerdtle, T.; Aschner, M.; McLean, J. A.; Bowman, A. B. Metallomics 2015, 7, 363−370. (10) McFadden, J. W.; Aja, S.; Li, Q.; Bandaru, V. V. R.; Kim, E. K.; Haughey, N. J.; Kuhajda, F. P.; Ronnett, G. V. PLoS One 2014, 9, e115642. (11) Folmes, C. D. L.; Nelson, T. J.; Martinez-Fernandez, A.; Arrell, D. K.; Lindor, J. Z.; Dzeja, P. P.; Ikeda, Y.; Perez-Terzic, C.; Terzic, A. Cell Metab. 2011, 14, 264−271. (12) Panopoulos, A. D.; Yanes, O.; Ruiz, S.; Kida, Y. S.; Diep, D.; Tautenhahn, R.; Herrerias, A.; Batchelder, E. M.; Plongthongkum, N.; Lutz, M.; Berggren, W. T.; Zhang, K.; Evans, R. M.; Siuzdak, G.; Belmonte, J. C. I. Cell Res. 2012, 22, 168−177. (13) Jain, M.; Nilsson, R.; Sharma, S.; Madhusudhan, N.; Kitami, T.; Souza, A. L.; Kafri, R.; Kirschner, M. W.; Clish, C. B.; Mootha, V. K. Science 2012, 336, 1040−1044. (14) Halama, A.; Möller, G.; Adamski, J. OMICS 2011, 15, 325−335. (15) Frezza, C.; Zheng, L.; Tennant, D. A.; Papkovsky, D. B.; Hedley, B. A.; Kalna, G.; Watson, D. G.; Gottlieb, E. PLoS One 2011, 6, e24411. (16) Lin, S.; Liu, N.; Yang, Z.; Song, W.; Wang, P.; Chen, H.; Lucio, M.; Schmitt-Kopplin, P.; Chen, G.; Cai, Z. Talanta 2010, 83, 262− 268. (17) Selvarasu, S.; Ho, Y. S.; Chong, W. P.; Wong, N. S.; Yusufi, F. N.; Lee, Y. Y.; Yap, M. G.; Lee, D. Y. Biotechnol. Bioeng. 2012, 109, 1415−1429. (18) Tiziani, S.; Lodi, A.; Khanim, F. L.; Viant, M. R.; Bunce, C. M.; Gunther, U. L. PLoS One 2009, 4, e4251. (19) Pirman, D. A.; Efuet, E.; Ding, X. P.; Pan, Y.; Tan, L.; Fischer, S. M.; DuBois, R. N.; Yang, P. PLoS One 2013, 8, e61379. (20) Rijk, J. C.; Peijnenburg, A. A.; Blokland, M. H.; Lommen, A.; Hoogenboom, R. L.; Bovee, T. F. Chem. Res. Toxicol. 2012, 25, 1720− 1731.

■

CONCLUSION Following genomics and proteomics, cell-based metabolomics is gaining growing importance in various fields such as cancer biology and the study of stem cells and neuron degenerative diseases.46 Similar to typical genomic and proteomic studies that control the total concentrations of study materials, we observed that it is also important in metabolomics studies to control the total metabolite concentration before investigation of differential metabolites. In our study of the lung cancer cell models CL1-0 and CL1-5, which have distinct morphologies, the metastatic properties of CL1-5 were only observed when controlling the total metabolite concentration. We also demonstrated that our proposed MIIS method represents a simple and effective method for the estimation and adjustment of total metabolite concentrations, which can improve data integrity for cellular metabolomics studies.

■

Article

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b01869. Supporting table, Table S-1; supporting figures, Figures S-1 and S-2 (PDF) H

DOI: 10.1021/acs.analchem.5b01869 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry (21) Dietmair, S.; Timmins, N. E.; Gray, P. P.; Nielsen, L. K.; Kromer, J. O. Anal. Biochem. 2010, 404, 155−164. (22) Lorenz, M. A.; Burant, C. F.; Kennedy, R. T. Anal. Chem. 2011, 83, 3406−3414. (23) Chong, W. P.; Reddy, S. G.; Yusufi, F. N.; Lee, D. Y.; Wong, N. S.; Heng, C. K.; Yap, M. G.; Ho, Y. S. J. Biotechnol. 2010, 147, 116− 121. (24) Hartmann, M.; Zimmermann, D.; Nolte, J. In Vitro Cell. Dev. Biol.: Anim. 2008, 44, 458−463. (25) Sheikh, K. D.; Khanna, S.; Byers, S. W.; Fornace, A., Jr.; Cheema, A. K. J. Biomol. Technol. 2011, 22, 1−4. (26) Finley, L. W.; Carracedo, A.; Lee, J.; Souza, A.; Egia, A.; Zhang, J.; Teruya-Feldstein, J.; Moreira, P. I.; Cardoso, S. M.; Clish, C. B.; Pandolfi, P. P.; Haigis, M. C. Cancer Cell 2011, 19, 416−428. (27) Spegel, P.; Malmgren, S.; Sharoyko, V. V.; Newsholme, P.; Koeck, T.; Mulder, H. Biochem. J. 2011, 435, 277−284. (28) Tran, M. Q.; Nygren, Y.; Lundin, C.; Naredi, P.; Bjorn, E. Anal. Biochem. 2010, 396, 76−82. (29) Silva, L. P.; Lorenzi, P. L.; Purwaha, P.; Yong, V.; Hawke, D. H.; Weinstein, J. N. Anal. Chem. 2013, 85, 9536−9542. (30) Dieterle, F.; Ross, A.; Schlotterbeck, G.; Senn, H. Anal. Chem. 2006, 78, 4281−4290. (31) Aberg, K. M.; Torgrip, R. J.; Kolmert, J.; Schuppe-Koistinen, I.; Lindberg, J. J. Chromatogr. A 2008, 1192, 139−146. (32) Craig, A.; Cloarec, O.; Holmes, E.; Nicholson, J. K.; Lindon, J. C. Anal. Chem. 2006, 78, 2262−2267. (33) Katajamaa, M.; Oresic, M. BMC Bioinf. 2005, 6, 179. (34) Matuszewski, B. K.; Constanzer, M. L.; Chavez-Eng, C. M. Anal. Chem. 2003, 75, 3019−3030. (35) Anker, R. M. J. Lab. Clin. Med. 1954, 43, 798−801. (36) Gyamlani, G. G.; Bergstralh, E. J.; Slezak, J. M.; Larson, T. S. Am. J. Kidney Dis. 2003, 42, 685−692. (37) Wu, Y.; Li, L. Anal. Chem. 2012, 84, 10723−10731. (38) Chen, G. Y.; Liao, H. W.; Tseng, Y. J.; Tsai, I. L.; Kuo, C. H. Anal. Chim. Acta 2015, 864, 21−29. (39) Yang, P. C.; Luh, K. T.; Wu, R.; Wu, C. W. Am. J. Respir. Cell Mol. Biol. 1992, 7, 161−171. (40) Chu, Y. W.; Yang, P. C.; Yang, S. C.; Shyu, Y. C.; Hendrix, M. J. C.; Wu, R.; Wu, C. W. Am. J. Respir. Cell Mol. Biol. 1997, 17, 353−360. (41) R Development Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2008. (42) Shih, J. Y.; Yang, P. C. Carcinogenesis 2011, 32, 1299−1304. (43) Chang, Y. H.; Lee, S. H.; Chang, H. C.; Tseng, Y. L.; Lai, W. W.; Liao, C. C.; Tsay, Y. G.; Liao, P. C. J. Proteome Res. 2012, 11, 5167− 5185. (44) Babbar, N.; Ignatenko, N. A.; Casero, R. A., Jr.; Gerner, E. W. J. Biol. Chem. 2003, 278, 47762−47775. (45) King, R.; Bonfiglio, R.; Fernandez-Metzler, C.; Miller-Stein, C.; Olah, T. J. Am. Soc. Mass Spectrom. 2000, 11, 942−950. (46) D’Alessandro, A.; Zolla, L. Drug Discovery Today 2012, 17, 3−9.

I

DOI: 10.1021/acs.analchem.5b01869 Anal. Chem. XXXX, XXX, XXX−XXX