Analysis of volatile organic compounds released from human lung cancer cells and from the urine of tumor-bearing mice

To extract the VOCs, upon reaching confluency the cell lines were grown for one, two and three weeks. After a one week incubation period, no dead cells were observed in any of the cell lines. After two week, some floating cells were observed and the culture medium had turned red (all cell lines). After the three week incubation period, many floating cells were observed and in all cell lines the culture medium was becoming amber. Thus, the culture conditions of the one and two week incubation periods ensured that the release of VOCs into the medium was mostly due to living cells. However, following three week incubation, cell viability diminished, indicating that released VOCs were not derived from the living cells. Typical total ion chromatograms (TICs) of VOCs measured in the medium samples of each cell line after one week incubation are shown in Figure 1. A sample of media obtained from the A549 cell lines showed a peak at 12.82 min that was clearly increased when compared with the control cell lines. However, the measured TICs were difficult to analyze because of multiple ion peaks, making it difficult to find the different EIC peaks in each of the cell lines. To analyze in more detail, comparative analysis between the A549 cell line and control cell lines was performed using XCMS software [33] and statistical analyses. Under our experimental conditions, an average of 751 ion peaks per sample was detected in the culture medium by XCMS software and statistical analyses. The significantly increased or decreased ion peaks (with a p-value < 0.10 and 1.5-fold higher or lower than average peak area of the normal control) were selected for identification. These ion peaks were summarized by deconvolution analysis of retention time and identified by spectral library match using the NIST'08 and Wiley library. However, identification of the VOCs was done only by means of spectral library matching (higher than 80% match) without confirmation of their retention times and comparison against a commercially available standard reagent. In this study, any peaks that could not be determined due to too low a library match were disregarded. The summary of determined VOCs by XCMS analysis are listed in Table 1, which also shows the rate of change with averaged peak area (using major m/z), comparing between the A549 and control cells. After a one week incubation period, there was a difference in the concentration of 31 compounds in the A549 cells compared with the control cells, whereby 18 compounds were found to be significant increased and 13 compounds were found to be significantly decreased (p < 0.10) (Table 1A). After a two week incubation period differences in the concentration of 45 compounds between the A549 and control cells culture medium were detected whereby 31 compounds were significantly increased, and 14 compounds were significantly decreased (p < 0.10) (Table 1B). After the three week incubation different concentrations were observed in 30 compounds, whereby 28 compounds were significant increased, and two compounds were significant decreased (p < 0.10) (Table 1C).

To observe the proliferation of the A549 cells implanted into the mice, tumor size was measured (Figure 2a), with measurement commencing at day 13 post implantation. The body weight of the mice was also recorded (Figure 2b). Although there was a slight decrease in body weight in the two days following transplantation, this then gradually increased over subsequent days. There was no significant difference in body weight between the tumor-bearing and control mice.

The pooled urine samples from the tumor-bearing (n = 24) and control (n = 16) mice were analyzed by headspace solid-phase microextraction (HS-SPME) and gas chromatography time-of-flight mass spectrometry (GC-TOF MS). Two typical TICs of the VOCs obtained from the urine samples are shown in Figure 3. The chromatographs obtained from urine samples were very similar to the cultured medium samples, making it difficult to identify differences between peaks in EICs of the tumor-bearing and control mice. However, the results obtained from detailed and comparative analyses of the two groups did reveal some differences. On average the chromatogram from each urine sample contained 1493 ion peaks and from this 68 ion peaks were significantly increased and 65 ion peaks were decreased relative to the control group (p < 0.10 and characterized by a 1.5-fold higher or lower than average peak area of the control). Deconvolution analysis of the retention time of each peak allowed for the isolation of 76 VOCs that were significantly increased and six VOCs that were decreased. The isolated VOCs were picked up by similarity searching using the mass spectral library (NIST'08 and Wiley). Any VOC that could not be determined because a library match is too low was disregarded. Table 2 shows the 43 VOCs that were found be highly similar following a search of the database. However, these VOCs were identified only by means of spectral library match (higher than 80% match) without confirmation of their retention times and mass spectral by comparison against a commercially available standard reagent. Moreover, Table 2 shows the rate of change with averaged peak area (using major m/z), comparing between the tumor-bearing and control mice. In the urine obtained from the tumor-bearing mice, the concentration of 43 compounds was different relative to the control group (p < 0.10), whereby 42 compounds were significantly increased, and one compound was significantly decreased.

We also investigated the compounds in more detail that were increased in both the one and two week culture medium samples (A549 cells), and urine of the tumor-bearing mice. Seven compounds were increased in all three samples. It was also discovered that dimethyl succinate, phenol, 2-methylpyrazine, 2-hexanone, 2-butanone, 2-pentanone and acetophenone could be identified not only by spectral library match using NIST'08 and Wiley library but also by matching with retention time and mass spectral of commercially available reagents (data not shown).

To determine the concentration of the seven VOCs, a standard curve was prepared with analysis by HS-SPME. The standard curves were induced using the peak area at most abundance m/z value. The calibration measurements were used to determine the limit of detection (LOD) and subsequent quantification (LOD) of seven VOCs. Results in Table 3 present the accuracy of the standard curve. The standard curve of each VOC indicated good linearity in the range of 0.02 μM to 10 μM (R² = 0.99). The concentration of each VOCs was then calculated using the respective standard curves. The quantitative findings, with accompanying statistical analysis, are summarized in Tables 4 and 5 (cell cultures) and Table 6 (mice urine). The significance of the results was calculated by ANOVA.

Summary of the concentration and significant difference of seven VOCs in culture medium were shown in Tables 4 and 5. After one week incubation, five VOCs in A549 cell medium were significantly increased (2-butanone, 2-pentanone, dimethyl succinate, acetophenone, phenol), 2-methylpyrazine was significantly decreased in comparison to the VOCs in the DMEM. Four VOCs in A549 cell medium were significantly increased in comparison to the VOCs in OUS-11 cell medium; 2-butanone (p = 0.001), 2-pentanone (p = 0.015), 2-methylpyrazine (p = 0.012), dimethyl succinate (p = 0.023). Five VOCs in A549 cell medium were significantly increased in comparison to the VOCs in WI-38 VA13 cell medium; 2-pentanone (p = 0.001), 2-hexanone (p = 0.002), dimethyl succinate (p = 0.016), acetophenone (p = 0.001), phenol (p = 0.011). After two week incubation, five VOCs in A549 cell medium were significantly increased (2-butanone, 2-pentanone, dimethyl succinate, acetophenone, phenol), 2-methylpyrazine was significantly decreased in comparison to VOCs in the DMEM. All seven VOCs in A549 cell medium were significantly increased in comparison to the VOCs in OUS-11 cell medium; 2-butanone (p < 0.001), 2-pentanone (p < 0.001), 2-hexanone (p = 0.008), 2-methylpyrazine (p < 0.001), dimethyl succinate (p < 0.001), acetphenone (p < 0.001) and phenol (p = 0.001). Five VOCs in A549 cell medium were significantly increased in comparison to the VOCs in WI-38 VA13 cell medium; 2-butanone (p = 0.001), 2-pentanone (p = 0.001), 2-hexanone (p = 0.039), dimethyl succinate (p < 0.001), acetphenone (p < 0.001).

Summary of the concentration and significant difference of all seven VOCs in mice urine was shown in Table 6. All VOCs were significantly increased in tumor-bearing mice urine in comparison to the VOCs in control mice urine; 2-butanone (p = 0.02), 2-pentanone (p = 0.08), 2-hexanone (p = 0.02), 2-methylpyrazine (p = 0.03), dimethyl succinate (p = 0.03), acetphenone (p = 0.02) and phenol (p = 0.01).

Three different human cell lines were used in the experiments: A549 (JCRB No. JCRB0076), OUS-11 (JCRB No. JCRB1034), WI-38 VA13 (JCRB No. JCRB9057). The A549 cells were originally isolated from a lung carcinoma of a 58-y-old man and are characterized by the presence of a mutated K-ras but a wild-type B-raf gene. The OUS-11 and WI-38 VA13 cells were used as normal control cells for the purposes of comparison and were derived from normal tissue of lung cancer patient and from a SV40 virus transformed lung fibroblast cell line, respectively. The cell lines were obtained from the Health Science Research Resource Bank (Osaka, Japan). All cells have grown in DMEM high-glucose culture medium (Sigma-Aldrich Japan, Tokyo) supplemented with 10% fetal bovine serum (HyClone, Logan, UT), penicillin (30 unit/mL, Meiji Seika, Tokyo, Japan) and streptomycin (30 μg/mL, Meiji Seika, Tokyo, Japan). The cells were cultured in two culture dishes (100 mm × 20 mm) in 10 mL culture medium and every three days fresh medium was applied. For all experiments, the cells were cultured under standard conditions at 37°C in humidified atmosphere containing 5% CO₂. For the VOC measurements, the A549, OUS-11 and WI-38 VA13 cells were incubated for one, two and three-weeks after all cell lines grew as monolayer adherent on the surface of culture dishes. After this incubation period, the culture medium was collected and stored at -80°C until analysis.

Five-week-old female C.B-17/lcr-scidJcl mice (CLEA Japan, Tokyo, Japan) were separated into two groups, those bearing tumors (n = 24) and those used as controls (n = 16). To establish the human tumor-bearing model, A549 cancer cells were prepared at a concentration of 1.0 × 10⁸ cells/mL in PBS, and implanted subcutaneously into the right ventral flank area by injecting 0.1 mL of the suspension. The control mice were injected only with 0.1 mL PBS on the same position as tumor-bearing mice. The day of inoculation was defined as day 0. The mice were kept in glass-shielded metabolic cages (Metabolica; Sugiyamagen, Tokyo, Japan) under sterile conditions and were supplied with radiation sterilized feed CE-2 (CLEA Japan, Tokyo Japan) and autoclaved drinking water, which was provided ad libitum. After transplantation, urine was collected from the mice daily, and stored at -80°C until time of analysis. Body weight and tumor size (the shortest and longest diameters) of each mouse was recorded every few days. Tumor size was calculated by multiplying the length of the shortest and longest diameters. The mice were sacrificed 30 days after transplantation. The urine samples were collected from 26 to 30 days after transplantation.

All animals were treated in accordance with the institutional guidelines for the care and use of laboratory animals.

The GC-TOF MS system was composed of a 7890a GC (Agilent, CA, USA) equipped with an auto sampler Combi-pal and GCT Premier (Waters, MA, USA) for time of flight mass spectrometry. The SPME fiber with absorbed volatile compounds was inserted into the injection port of the 7890a GC using the auto sampler and desorbed for 10 min at 240°C. The injection was pulsed splitless (closed 3 min) with a 0.75 mm liner. The GC-TOF MS system was equipped with an InertCap Pure-WAX T.L. column (60 m + 2 m transfer line, 0.25 mm i.d., 0.5 μm thick; GL science, Japan), which was used for separation and analysis of the desorbed volatiles. We employed the following chromatographic protocol for separation before MS analyses: 40°C for 5 min, then programmed at 3°C/min to 240°C with a 5 min hold at this final temperature. Column flow was constant at rate of 1 ml/min. The injection port was held at 240°C. Operating parameters for the mass spectrometer were as follows: the ion source temperature was 200°C, ionizing energy at 70 eV, scanning frequency was 0.2 s/spectrum from m/z 40 to m/z 500. Peak identification was accomplished through manual interpretation of spectra and matching against the mass spectral library (NIST/EPA/NIH mass spectral library (NIST 08); mass spectral library of drugs, poisons, pesticides, pollutants and their metabolites Wiley, USA) and comparison with commercially available standard samples, which were purchased from the Tokyo Chemical Industry (Tokyo, Japan). Each medium sample was analyzed in triplicate.

The chromatographic peak areas were integrated using the MassLynx 4.1 (Waters, MA, USA). Genesis peak detection was applied for a list of putatively annotated ions. Detection and integration of generated ion peak from electron ionization were also performed using a XCMS software package version 1.16.3 (http://​masspec.​scripps.​edu) [33], running under R version 2.10.1 (http://​cran.​r-project.​org/​). The significance levels of differences between groups were calculated using a student's t-test that come with XCMS software. We also determined that the "increased extracted ion peaks" had a p-value of < 0.10, and were 1.5 times higher than the averaged peak area of the other group. Of those peaks, a manual inspection of the EIC for each peak was made to validate the detected peak. Refine the VOC was performed by manual deconvolution based on retention time and peak shape of each increased extracted ion peak.