Quantification of acetaldehyde and carbon dioxide in the headspace of malignant and non-malignant lung cellsin vitro by SIFT-MS

Abstract

Previous studies using selected ion flow tube mass spectrometry, SIFT-MS, showed that CALU-1 lung cancer cell cultures emit acetaldehyde in proportion to the number of cells in the culture medium. However, studies in another laboratory led to conflicting results, so these SIFT-MS studies have been repeated and extended to include NL20 normal lung epithelial cells and 35FL121 Tel+ telomerase positive lung fibroblast cells. Thus, SIFT-MS has been used to quantify acetaldehyde and, additionally, carbon dioxide, acetone and ethanol in the headspace of the cell culture medium alone and the headspace of the appropriate medium containing 50 × 10⁶ or 80 × 10⁶cells following incubation for 16 h at 37 °C. Three independent experiments were carried out for each cell type for both cell numbers and for each of the two culture media used. The results showed that acetone and ethanol were only released by the culture medium, specifically from the foetal calf serum contained in the medium, and not by the cells. Acetaldehyde was also released by the medium, but at relatively lower levels than the other three compounds, and was also generated by the CALU-1 and NL20 cell cultures in proportions to the number of cells in the medium. However, following incubation, the acetaldehyde levels in the headspace of the 35FL121 Tel+ cell cultures were much lower than those present in the headspace of the medium alone. Carbon dioxide was clearly generated by the CALU-1 and 35FL121 Tel+ cells indicating that they were respiring normally, but much less was produced by the NL20 cells, presumably indicating that normal metabolism was being inhibited.

---

Aim of investigation

Breath analysis has a significantly long history. An early authoritative discussion of the methodology of breath analysis and its possible value in clinical diagnosis and therapeutic monitoring was given 25 years ago by Manolis.¹ At that time the major diagnostic method was gas chromatography with mass spectrometry, GC/MS, by which several trace compounds were detected in the breath of healthy individuals and this remains an important analytical method in this research area.^2,3 In the interim there have been important developments in analytical methods that have increased the armoury available to the research scientists and clinicians. In particular, the development of proton transfer reaction mass spectrometry, PTR-MS,⁴ and selected ion flow tube mass spectrometry, SIFT-MS,^5,6 have added a new dimension to breath research in that on-line, real time analyses of exhaled breath can now be carried out. The introduction of spectroscopic and electrochemical analytical techniques, best exemplified by their use in detecting and quantifying nitric oxide in exhaled breath,^7,8 are also being shown to be useful for the detection of small molecules. The more recent developments are reported in detail in a recent book on breath research.³

Accurate analyses of several trace compounds (metabolites) can be obtained in single breath exhalations in real time using SIFT-MS.⁵ This allows non-invasive, painless breath analysis that can be used by frail patients and even children, the results being immediately available to the health professional. This is a real step forward in the anticipated introduction of breath analysis into the clinical setting. Using SIFT-MS, several compounds have been accurately quantified in the exhaled breath of many healthy adult volunteers and concentration/population distributions have been constructed that provide the reference ranges for these compounds in the exhaled breath of healthy people.^9,10 These data provide the baseline levels for analogous studies of patients in diseased states and are an important step towards the “holy grail” of detecting minimal, (early stage) disease via breath analysis.¹¹

Although progress in the recognition of new volatile biomarkers of disease and infection by breath analysis is slow, there have been some developments, notable examples being the GC/MS work by Phillips et al. in comparing the levels of branched chain hydrocarbons in the breath of cancer patients and healthy volunteers¹² and the detection ofPseudomonas infection in the airways of patients with cystic fibrosis by the presence of enhanced levels of hydrogen cyanide in their exhaled breath.¹³ For obvious reasons, there is a great desire to find biomarkers in breath for the presence of malignant tumours in the body. This might aid early detection, which is known to greatly improve the prognosis, especially so for lung cancer.^14,15

Current progress in the development of a diagnostic test for lung cancer through the analysis of breath volatiles, including acetaldehyde, has been recently reviewed by Mazzone.^16,17 In support of this, attention has been directed towards the study of the volatile compounds emitted by cancer cell lines in vitro in the hope that this would provide a focus for the breath analysis investigations. Several years ago, we used SIFT-MS to investigate the volatile compounds released by CALU-1 and SK-MES lung cancer cell lines cultured in complete media.¹⁸ The notable result of this study was that acetaldehyde was generated by both cell lines in close proportion to the number of cells in the culture medium, which were typically (50–80) × 10⁶, the generation rate of acetaldehyde molecules being about 10⁶ per cell per minute. Since that time this work has been taken up by others, notably by Amann and colleagues in Innsbruck, who have reported conflicting results.¹⁹ A recent study in the United States²⁰ has reported that acetaldehyde is released by HL60 leukaemia cells at approximately the same rate as that seen for the CALU-1 and SK-MES cells in the SIFT-MS study. But a recent report, as yet unpublished, indicates that a smaller number (5 × 10⁶) of SK-MES cells actually consume acetaldehyde when this is deliberately added to the support medium.²¹ It is also seen that both acetaldehyde and ethanol are present in the headspace of the media used in the absence of the cells and so it has been suggested that the increase in acetaldehyde seen on introducing the cells is due to the partial conversion of the ethanol to acetaldehyde by the cells. This certainly needs to be investigated; it is commented on later in this paper.

On this basis of the total evidence we decided to revisit this work first to more thoroughly investigate the source of ethanol and acetaldehyde emission from the medium used for these studies, then to check our earlier work on CALU-1 lung cancer cell lines and to extend this study by also investigating volatile emissions from two other cell lines, namely the normal lung epithelial cells NL20, and the non-malignant telomerase positive lung fibroblast cells 35FL121 Tel+ that can both be cultured to large numbers. We again focused on measurement of acetaldehyde, but by exploiting a very recent development of SIFT-MS²² we are able to measure simultaneously the production of carbon dioxide by the cells, which turns out to be a very significant addition.

Experimental

SIFT-MS

Selected ion flow tube mass spectrometry, SIFT-MS, has been developed principally for the rapid, real time analysis of trace gases in air and exhaled breath,⁵ a major focus being on the use of breath analysis for clinical diagnosis and therapeutic monitoring.³ Its value as an analytical tool for breath analysis and other areas has been demonstrated by detailed pilot studies in our laboratories^5,11 and in other laboratories.^23–25 This technique has been described in many previous papers^5,6,26,27 and only a very brief description is required here. The present experiments were carried out at Keele using a Profile 3SIFT-MS instrument (Instrument Science Limited, UK). The precursor ions are formed in a microwave discharge source and are selected according to their mass-to-charge ratio, m/z, by a mass filter and injected into flowing helium carrier gas where they are convected as a thermalised swarm along a flow tube. In the present experiments, H₃O⁺ precursor ions were used since these are the most appropriate ions to analyse the species of interest i.e.acetaldehyde, carbon dioxide, ethanol and acetone.^22,28 Thus, the count rates of the characteristic product ions at m/z 45 and 81 were used to calculate the absolute acetaldehyde concentrations avoiding any overlap due to the presence of carbon dioxide,²⁸ and the count rates of the product ion at m/z 63 was used to calculate the concentration of CO₂ taking the influence of acetaldehyde into the account.²² The product ion count rates at m/z 59 and 77 were used to calculate the acetone concentration and m/z 83 only was used to calculate ethanol concentration. Air/breath or liquid headspace samples are introduced at known flow rates into the carrier gas, and the precursor ions and the product ions of the reactions of the H₃O⁺ ions with the trace gases in the sample are detected and counted by a downstream analytical mass spectrometer system that can be scanned over predetermined m/z ranges for a given time period. An on-line computer immediately calculates the partial pressures of the trace gases in the air sample from the ion count rates.^29,30 This is the SIFT-MS full-scan mode (FSM) that is used when the trace compounds present in the sample need to be identified. Since we focused on known species in the present study we used the alternative mode of operation, the multi-ion monitoring mode (MIM) by which the downstream analytical mass spectrometer is rapidly switched between selected m/z values to target selected trace gas species. By this means, temporal changes in trace gas concentrations can be tracked and this is particularly valuable for the analysis of trace gases present in single exhalations of breath.⁶ Also, when sampling the headspace above a liquid in a sealed bottle the pressure within it should fall and this can be seen as the analysis of the headspace proceeds, which acts as a vital check on the integrity of the seal (the septum; see below). From the data obtained, together with the rate coefficients of the reactions as included in an on-board kinetics database, the partial pressures of the trace compounds in the sample are obtained.

To determine the concentrations of trace gases above aqueous liquids, including urine and cell cultures, we followed the procedure that we developed and successfully used previously.^18,31 Briefly, glass bottles typically of volume 150 mL containing a total volume of 50 mL of liquid are sealed with septa and placed in a temperature controlled water bath. After allowing the bottle plus contents to reach the water bath temperature, the septum is punctured by a hypodermic needle connected directly to the sample inlet line of the SIFT-MS instrument, whence the air/vapour headspace automatically flows into the helium carrier gas and analysis of the contents proceeds. The medium/cell cultures were prepared as described below and the measurements were carried out at a water bath temperature around 37 °C.

Studying the DMEM medium

The medium used for our previous studies of CALU-1 and SKMES cancer cell lines and which was used for the bulk of the present studies is DMEM culture supplemented with 10% foetal calf serum (FCS), antibiotics and glucose at 4.5 g L⁻¹. This medium is seen to release ethanol, acetaldehyde and acetone into its headspace at easily measurable levels. In an attempt to discover the origin of these compounds and to investigate if they were produced by glucose degradation, experiments were carried out on the DMEM culture supplied glucose free (Biosera Ltd, U. K.) and then at added glucose concentrations of 3 g L⁻¹ and the regular 4.5 g L⁻¹. These media were held at 37 °C for various periods ranging from a few hours to 16 h (typical of the overnight incubation period) after which the headspace levels of the above compounds were measured. It must be said immediately that on these relatively short time scales there was no measurable change in the levels of any of the headspace compounds. Further to these media-alone studies, the media at the given glucose levels were each cultured with 50 × 10⁶CALU-1 cells according to the protocol outlined below. In each case the headspace levels of the targeted compounds were measured by SIFT-MS, as explained above. These studies were carried out in duplicate. Additionally, and crucially, the headspace of the foetal calf serum alone was studied for the above compounds. The results obtained are clear and unequivocal, as is explained later.

Cell culture protocol

The non-small cell lung cancer cell line CALU-1 was obtained from the European Collection of Cell Cultures (Salisbury, UK) and the lung epithelial cell line NL20 from ATTCC (UK). The telomerase positive lung fibroblast 35FL121 Tel+ cell line is based on cells obtained from a 35 years old female lung.³² This cell line has been retrovirally infected to express telomerase in order to render them immortal. The CALU-1 and 35FL121 Tel+ cell lines were kept in the DMEM culture referred to above. The NL20 cell line was kept in Ham's F12 culture supplemented with FCS (4%), antibiotics, glucose (2.7 g L⁻¹) and growth factors according to the provider's instructions (American Type Culture Collection, Teddington, UK). Cell growth was carried out in tissue culture flasks (Sarstedt, UK) at 37 °C in an atmosphere containing 5% CO₂. The culture medium was changed every 3–4 days.

The CALU-1 and 35FL121 Tel+ cells were detached from the tissue culture flasks before reaching confluence using trypsin/EDTA. The NL20 cells were detached before reaching confluence using HBSS supplemented with FCS and EDTA according to the provider's instructions. The corresponding complete medium was then added and the cells/medium combination was spun in a centrifuge at 1200 rpm for 7 min. The supernatant was removed and complete medium was again added. The cell numbers were counted and the total was apportioned into the 150 ml glass bottles which were then sealed with the septa. Typically, the bottles contained either 50 × 10⁶ or 80 × 10⁶cells in 50 mL of medium. The air above the liquid was replaced with CO₂ free dry cylinder air and the bottles were resealed and placed in an incubator at 37 °C overnight for a total time of about 16 h together with a bottle containing the same volume of media only (no cells). This incubation period was chosen for convenience and to conform to our previous studies¹⁸ in which we showed that this was sufficient time to allow the trace compounds in the headspace to develop sufficiently to allow easy analysis. Thus, the headspaces above the medium-only and the cell/medium cultures were allowed to develop overnight after which they were sampled directly into the SIFT-MS instrument, as described above. It is important to recognise that during the sampling procedure the pressure in the sealed bottle decreases somewhat from atmospheric pressure P to about 0.8P. This is taken into account when calculating the absolute concentrations of the trace metabolites in the headspace.³¹Cell survival following the SIFT-MS analysis was carried out using the trypan blue exclusion method.

Results and discussion

Medium studies

The data obtained for two repeat studies are shown in Table 1. They clearly show that the baseline levels of acetaldehyde, ethanol and acetone in the medium headspace without the cells are not dependent on the glucose concentration. So these compounds must originate from another component of the medium and surely the major source is the foetal calf serum, which comprises 10% of the medium by volume. This deduction is supported when it is seen that the headspace levels of the ethanol and acetone above the neat serum (see Table 1) are about ten times greater than the levels in the medium headspace. The relative level of acetaldehyde above the serum is only about 3 times higher than that above the medium, but it must be noted that the absolute levels of this compound are much lower than those of the ethanol and acetone, and acetaldehyde is much more volatile than these two compounds and will thus be more readily lost from the concentrated (serum) solution. Thus, the conclusion must be that the sole source of the base line levels of ethanol and acetone, and probably acetaldehyde also, is the foetal calf serum. With hindsight, it is no surprise that ethanol and acetone are present in the calf serum, since these compounds, as well as acetaldehyde at a much lower level, are endogenously produced in most mammals, including human beings. Also, it should be expected that the levels of these endogenous compounds will be quite variable as they are in the exhaled breath (and hence in the blood stream) of people.⁹ It should also be noted that carbon dioxide also appears in the headspace of the medium alone at a level of about 1%. This might originate from the sodium bicarbonate that is present in the DMEM medium at the relatively high concentration of 3.7 g L⁻¹. This premise is given credence by the fact that there is less CO₂ into the headspace of the Ham's F12 culture medium that contains less bicarbonate (1.2 g L⁻¹), as we show later.

Table 1 The levels of acetaldehyde, ethanol and acetone in parts-per-billion, ppb, measured in two separate experiments in the headspace of glucose-free DMEM medium and after glucose has been added resulting in concentrations of 3 and 4.5 g L⁻¹, as indicated. Also shown in separate columns are the corresponding levels of the compounds in the headspace of the DMEM medium when cultured for 16 h together with 50 × 10⁶CALU-1 cells. Also given is the percentage survival of the cells following culture, determined immediately after the SIFT-MS analyses

glucose	acetaldehyde, ppb		ethanol, ppb		acetone, ppb		survival
	medium	cells	medium	cells	medium	cells
a Levels of acetaldehyde, ethanol and acetone in the headspace of the neat foetal calf serum, which is present at 10% by volume in the DMEM medium.
0 g	66	117	628	629	215	202	76%
0 g	70	144	633	523	156	116	69%
3 g	66	185	608	601	217	197	82%
3 g	84	183	520	439	193	121	76%
4.5 g	63	194	580	641	207	206	91%
4.5 g	65	218	822	803	223	221	85%
serum^a	166	—	6391	—	1569	—	—

---

Also clear in the data (Table 1) is that when 50 million CALU-1 cells are cultured in the medium, the headspace ethanol and acetone levels are not changed from the levels in the medium alone. However, the headspace acetaldehyde levels are markedly increased, even for the medium/cells combination in which there is no added glucose, being marginally highest for the highest glucose concentration. The fact that acetaldehyde levels are increased even though the ethanol levels are not, excludes the possibility suggested by Shan²¹ that the headspace acetaldehyde is formed from the headspace ethanol. Thus, the cells are producing the acetaldehyde and a minimum level of glucose is required for normal functioning of the cells. This is supported by the greater percentage survival rate of the cells at the higher glucose levels, as is indicated in Table 1, which is about 90% and only 70% for zero added glucose. That the majority of the cells survive in ostensibly zero glucose medium might mean that they are utilizing nutrients from the serum, presumably including some glucose. It appears that the safe glucose level in the DMEM medium is indeed the higher level (4.5 g L⁻¹) routinely used in the commercially available media and which was used in our previous experiments¹⁸ and in the following experiments.

Further cell studies

Identical experiments were carried out for lung cancer CALU-1 cells, telomerase positive lung fibroblast cells 35FL121 Tel+ and lung epithelial cells contained in 50 mL of the commercially supplied medium, as described above. Three independent experiments for each cell line for both 50 × 10⁶ and 80 × 10⁶cells in the appropriate medium were analysed over a total period of 3-months. Following overnight incubation, analyses of the headspace of each medium/cells combination and the medium alone was performed for acetaldehyde, carbon dioxide, ethanol and acetone simultaneously using the MIM mode of operation of the SIFT-MS instrument. Following these analyses, and following the venting of the bottles with dry cylinder air by simply puncturing the septa with a needle connected to the air supply, FSM spectral analyses were obtained to investigate if other trace compounds were present in the headspace at significant levels, using both H₃O⁺ and NO⁺ precursor ions for the analysis. Clearly, other compounds were present in the headspaces of all three cell cultures, but they were present at much lower levels than acetaldehyde, ethanol and acetone and the levels were close to the limit of detection and quantification of the current SIFT-MS instrument. So these trace compounds have not been investigated further, but we recognise that they might carry important information that only further research would reveal.

The headspace concentrations of the four targeted compounds obtained for each separate cell/medium mixture and the parallel values for the medium samples alone, incubated at the same time, are shown in Table 2. Also shown are the percentage survival rates of each of the cell collections, as determined a few minutes after the completion of the SIFT-MS analyses. These show that the CALU-1 cells had the highest mean survival rate at 94% closely followed by the fibroblast cells at 87% with the lung epithelial cells at 80% survival.

Table 2 The concentrations of acetaldehyde (parts-per-billion, ppb) and carbon dioxide (percent, %) measured in the headspace of the cell line cultures indicated. Note that three independent measurements were made for each cell line at 50 × 10⁶ and 80 × 10⁶cells in the commercial DMEM medium containing 4.5 g L⁻¹ of glucose for the CALU-1 and 35FL121 Tel+ cells and Ham's F12 medium containing 2.7 g L⁻¹ of glucose for the NL20 cells. Also given in separate columns are the concentrations of these compounds above the respective culture medium alone and the percentage survival of the cells in each experiment

Cell line	Number of cells, ×10^6	Survival factor, %	Acetaldehyde, ppb^a		CO2, %
			medium	cells	medium	cells
a Acetone and ethanol were also present in the headspace of the cell cultures at the following mean levels given in ppb (levels above medium without cells are given in parentheses): CALU-1: acetone 296 (286) ppb; ethanol 3312 (2380) ppb. 35FL121 Tel+: acetone 256 (248) ppb; ethanol 2787 (2514) ppb. NL20: acetone 120 (125) ppb; ethanol 1184 (1053) ppb. b n.d. not done.
CALU-1	50	95	245	426	1.2%	3.7%
CALU-1	50	91	265	419	1.4%	4.5%
CALU-1	50	93	278	387	1.6%	4.7%
CALU-1	80	95	245	533	1.2%	5.6%
CALU-1	80	94	247	545	1.5%	6.3%
CALU-1	80	94	251	492	1.6%	7.1%
35FL121 Tel+	50	91	233	83	1.8%	5.0%
35FL121 Tel+	50	84	297	28	1.1%	3.4%
35FL121 Tel+	50	89	248	49	1.5%	3.6%
35FL121 Tel+	80	n.d.^b	227	54	1.5%	5.5%
35FL121 Tel+	80	85	281	67	1.2%	6.4%
35FL121 Tel+	80	83	248	39	1.5%	4.6%
NL20	50	82	101	309	0.7%	1.2%
NL20	50	76	121	368	0.5%	1.5%
NL20	50	84	121	287	0.5%	1.8%
NL20	80	80	90	417	0.7%	2.0%
NL20	80	77	103	370	0.9%	2.5%
NL20	80	77	121	381	0.5%	2.8%

---

These analyses show that there is no significant increase in either acetone or ethanol due to the presence of the cells, these compounds clearly originating in the medium alone. The acetone concentration in the headspace of the commercially supplied DMEM total medium is comparable to that measured in the DMEM medium supplied without the glucose (typically 250 parts-per-billion, ppb). But the acetone level in the headspace of the Ham's F12 medium is typically 120 ppb and this mirrors the lower percentage of serum which is 4% compared to 10% in the DMEM medium. However, the ethanol level is about four times greater (typically 2400 ppb compared to 600 ppb) in the commercially supplied DMEM medium and this is also the case for the acetaldehyde (240 ppb compared to 60 ppb). Thus, the medium headspace levels of acetaldehyde and ethanol are linked suggesting that they have a common origin, which surely is the serum. Note again that the levels of these two compounds in the Ham's F12 medium are about 40% of those in the DMEM medium implying that they originate in the serum. However, it cannot totally be ruled out that the acetaldehyde increase is, in part, due to the partial metabolism of glucose and the oxidation of the ethanol.²¹ It is worth reiterating that over time scales of 16 h incubation the headspace levels of ethanol and acetone did not measurably increase.

However, very significantly, the concentrations of acetaldehyde and carbon dioxide are obviously increased above the cells/medium culture compared to the medium alone, as can be seen in Table 2. First notice that carbon dioxide is at a mean level of 1.4% in the headspace of the DMEM medium used for both the CALU-1 and fibroblast cells and lower at a mean level of 0.7% above the Ham's F12 medium used for the epithelial cells. This is probably because the DMEM medium has a higher concentration of sodium bicarbonate (3.7%) than the Ham's F12 medium (1.2%). But very interesting differences are clear in the relative concentrations of these two compounds above the respective cells/medium mixtures, and these are best seen in the plots shown in Fig. 1, as summarised here:

Fig. 1 Plots of the data values given in Table 1 for the headspace concentrations of acetaldehyde (left column) and carbon dioxide (right column), in parts-per-billion, ppb and percent, %, respectively, of the appropriate culture medium alone (no cells) and when 50 × 10⁶ and 80 × 10⁶cells of the named cell lines are present in the medium. The coefficients of determination, R², are also given. An assessment of these data is given in the text.

CALU-1 . When these cells are present, the acetaldehyde and carbon dioxide levels are both increased in the headspace in close proportion to the number of cells in the DMEM medium. The consistency of the measurements is confirmed by the three repeats for both 50 × 10⁶cells and 80 × 10⁶cells. The acetaldehyde increases above the “background” or baseline level (medium alone) by about 300 ppb and the carbon dioxide by about 3.5% when 80 × 10⁶cells are present in the medium.

35FL121 Tel+. Remarkably, the headspace acetaldehyde level is dramatically reduced due to the presence of these modified fibroblast cells from the medium alone level of about 250 ppb to about 50 ppb for both cell numbers, whereas the carbon dioxide level is increased in proportion to the cell number and to levels very similar to those for the CALU-1 cells.

NL20. The acetaldehyde in the headspace above these normal lung epithelial cells is increased in proportion to the number of cells from the medium alone level of about 110 ppb to about 360 ppb for the 80 × 10⁶cells, i.e. by 250 ppb, but for these cells the increase in carbon dioxide is only about 1.5%, which is much less than for the CALU-1 and, the 35FL121 Tel+ cells.

How can these results be explained? Firstly, it must be said that the increase in the acetaldehyde level in the presence of the CALU-1 cells seen in the present experiments is consistent with the results we obtained in our first study of several years ago,¹⁸ which revealed that these cells are producing acetaldehyde at a rate of about 10⁶ molecules (about 2 × 10⁻¹⁸ mol) per cell per minute. Additionally, in the present studies we have the carbon dioxide data, which indicates that these CALU-1 cells are respiring normally, the rate of CO₂ molecule production roughly approximating to that expected on the basis of CO₂ molecule production in the human body cell burden. This normal respiration is given further credence by the high survival rate of these cells (94%). It is worthy of note that glucose concentrations below some threshold level will surely lead to lower respiration rates and hence to lower headspace levels of CO₂.

The modified 35FL121 Tel+ fibroblast cells also produce a very similar amount of carbon dioxide, but in this case the unexpected happens in that acetaldehyde is efficiently removed from the medium/headspace and hence from the liquid medium. Is this the behaviour of healthy non-malignant cells that remove the toxic (carcinogenic) acetaldehyde from the mixture? Do these cells use acetaldehyde as a feed back for their increased proliferation activity and metabolism? The somewhat lower survival rate (86%) might indicate that the cells are inhibited in their growth by the acetaldehyde. We tentatively suggest that the acetaldehyde is being metabolised to acetic acid by these cells, but unfortunately acetic acid was not accurately monitored as part of this study. However, it is also possible that no build up of acetic acid occurs; rather that the terminating metabolites CO₂ and H₂O are released.

As for the normal lung epithelial NL20 cells they also produce acetaldehyde, but the carbon dioxide production is much lower, perhaps suggesting that normal cell metabolism is inhibited, which is why their survival rate is significantly lower than the other two cell types at only 79% (see Table 1). But it is important to note again that the glucose level in the medium is only at a concentration of 2.7 g L⁻¹ compared to that of 4.5 g L⁻¹ for the medium used for the other two cell lines, which might be marginal in the light of the data reported above. Clearly, we now need to extend the gas analysis to include acetic acid measurements and attempt to trace the metabolic pathways for the production of acetaldehyde by cellsin vitro.

Concluding remarks

The results of this limited study provided further information on the phenomenon of acetaldehyde production and consumption by cancer cells and non-malignant cellsin vitro. The initial motivation for the present study was to follow up our previous SIFT-MS study of CALU-1 cancer line lines, the results of which have recently been questioned by others. Now the previous results have been vindicated by the present experiments and the extension of the study to include carbon dioxide production by the cells has strengthened our conviction that the cell preparation and sampling methodology is sound. However, the true value of this study is the revelation that the NL20 non-malignant cell type also produces acetaldehyde and the 35FL121 Tel+ type actually removes acetaldehyde from the supporting medium, even though the respiration appears to be reasonably normal as viewed by the carbon dioxide production. Follow up studies will need to focus on the quantification of other volatile compounds that are produced and to include other cell types.

The formation mechanism of acetaldehyde by these cells is not yet clearly understood. It is known that yeast cells metabolise glucose to acetaldehyde, which is quickly reduced to ethanol by the fermentation process such that only a relatively small fraction of the acetaldehyde remains in the ferment.^33,34 It is also known that acetaldehyde is the first intermediate of alcohol metabolismin vivo and that it is a highly toxic compound.^35,36 There is a large literature on this topic, in particular with respect to cell apoptosis in the liver and brain that also indicates acetaldehyde to be a possible human carcinogen.^37–39Ethanol is not produced in measurable quantities by the cell types included in this study; the observed ethanol originates from the serum contained in the media used. We noted above that the modified 35FL121 Tel+ fibroblast cells might be converting acetaldehyde to acetic acid, so it is interesting to note that it has been reported⁴⁰ that in the metabolism of glucose by the fungus Trichoderma reesei, the acetaldehyde produced may be channelled into acetic acid.

It has been our hypothesis that should acetaldehyde appear in exhaled breath above physiological levels this could be an indicator of the presence of tumours in the body, but the fact that the lung epithelial cells also produce acetaldehydein vitro has cast serious doubt on this. As noted previously, acetaldehyde occurs in the breath of healthy individuals at low levels, typically around 10–20 ppb^9,11 and increases somewhat following the ingestion of alcohol,³⁶ so additional production by cancer cellsin vivo would need to appear in exhaled breath at levels greater than about 20 ppb. We have now begun pilot investigations of the metabolites in the breath of patients suffering from lung cancer using SIFT-MS, a topic that is receiving considerable attention worldwide.^18–21,41

Acknowledgements

We are grateful to Dr N.R. Forsyth, Keele University, for kindly donating the 35FL121 Tel+ cell line. We are grateful to the North Staffordshire Medical Institute and the Grant Agency of the Czech Republic (project numbers 202/09/0800 and 203/09/0256) for partial funding of this work.

References

1. A. Manolis, Clin Chem., 1983, 29, 5–15.
2. M. Phillips, J. Herrera, S. Krishnan, M. Zain, J. Greenberg and R. N. Cataneo, J. Chromatogr., B: Biomed. Sci. Appl., 1999, 729, 75–88.
3. Breath Analysis for Clinical Diagnosis and Therapeutic Monitoring, A. Amann and D. Smith, ed., World Scientific, Singapore, 2005.
4. W. Lindinger, A. Hansel and A. Jordan, Int. J. Mass Spectrom. Ion Processes, 1998, 173, 191–241.
5. D. Smith and P. Španěl, Mass Spectrom. Rev., 2005, 24, 661–700.
6. P. Španěl and D. Smith, Med. Biol. Eng. Comput., 1996, 34, 409–419.
7. L. E. Gustaffson, A. M. Leone, M. G. Persson, N. P. Wiklund and S. Moncada, Biochem. Biophys. Res. Commun., 1991, 181, 852–857.
8. C. Roller, K. Namjou, J. D. Jeffers, M. Camp, A. Mock, P. J. McCann and J. Grego, Appl. Opt., 2002, 41, 6018–6029.
9. D. Smith, C. Turner and P. Španěl, J. Breath Res., 2007, 1, 014004.
10. D. Smith and P. Španěl, Analyst, 2007, 132, 390–396.
11. T. Wang, A. Pysanenko, K. Dryahina, P. Španěl and D. Smith, J. Breath Res., 2008, 2, 037013.
12. M. Phillips, N. Altorki, J. H. M. Austin, R. B. Cameron, R. N. Cataneo, R. Kloss, R. A. Maxfield, M. I. Munawar, H. I. Pass, A. Rashid, W. N. Rom, P. Schmitt and J. Wai, Clin. Chim. Acta, 2008, 393, 76–84.
13. B. Enderby, D. Smith, W. Carrol and W. Lenney, Pediatr. Pulmonol., 2009, 44, 142–147.
14. A. Wehinger, A. Schmid, S. Mechtcheriakov, M. Ledochowski, C. Grabmer, G. A. Gastl and A. Amann, Int. J. Mass Spectrom., 2007, 265, 49–59.
15. B. J. Schneider, Cancer Imaging, 2008, 8, 181–185.
16. P. J. Mazzone, J. Thorac. Oncol., 2008, 3, 774–780.
17. P. J. Mazzone, J. Breath Res., 2008, 2, 037014.
18. D. Smith, T. S. Wang, J. Sule-Suso, P. Španěl and A. El Haj, Rapid Commun. Mass Spectrom., 2003, 7, 845–50.
19. W. Filipiak, A. Sponring, T. Mikoviny, C. Ager, J. Schubert, W. Miekisch, A. Amann and J. Troppmair, Cancer Cell Int., 2008, 8, 17.
20. H. W. Shin, B. J. Umber, S. Meinardi, S. Y. Leu, F. Zaldivar, D. R. Blake and D. M. Cooper, J. Transl. Med., 2009, 7, 31.
21. Detection of aldehydes in lung cancer cell culture by gas chromatography/mass spectrometry and solid-phase microextraction with on-fiber derivatization. G. Shan, Master of Science thesis. Texas A&M University, 2006.
22. D. Smith, A. Pysanenko and P. Španěl, Rapid Commun. Mass Spectrom., 2009, 23, 1419–1425.
23. C. Turner, P. Španěl and D. Smith, Physiol. Meas., 2006, 27, 321–337.
24. B. M. Ross and N. Vermeulen, Rapid Commun. Mass Spectrom., 2007, 21, 3608–3612.
25. P. F. Wilson, C. G. Freeman, M. J. McEwan, D. B. Milligan, R. A. Allardyce and G. M. Shaw, Rapid Commun. Mass Spectrom., 2002, 16, 427–432.
26. P. Španěl and D. Smith. In, Amann A, Smith D, editors. Breath Analysis for Clinical Diagnosis and Therapeutic Monitoring. Singapore, World Scientific, 2005, 3–34.
27. D. Smith and P. Španěl, Rapid Commun. Mass Spectrom., 1996, 10, 1183–1198.
28. P. Španěl and D. Smith, J. Breath Res., 2008, 2, 046003.
29. P. Španěl, K. Dryahina and D. Smith, Int. J. Mass Spectrom., 2006, 249/250, 230–239.
30. D. Smith, A. Pysanenko and P. Španěl, Int. J. Mass Spectrom., 2009, 281, 15–23.
31. P. Španěl, A. M. Diskin, S. M. Abbott, T. S. Wang and D. Smith, Rapid Commun. Mass Spectrom., 2002, 16, 2148–2153.
32. N. R. Forsyth, A. P. Evans, J. W. Shay and W. E. Wright, Aging Cell, 2003, 2, 235–243.
33. Biochemistry, G. L. Zubay, WCB/McGraw-Hill, Columbus, 1998.
34. D. Smith, T. S. Wang and P. Španěl, Rapid Commun. Mass Spectrom., 2002, 16, 69–76.
35. J. F. Brien and C. W. Loomis, Can. J. Physiol. Pharmacol., 1983, 61, 1–22.
36. D. Smith, T. S. Wang and P. Španěl, Physiol. Meas., 2002, 23, 477–489.
37. P. Boffetta and M. Hashibe, Lancet Oncology, 2006, 7, 149–156.
38. M. Salaspuro, Addiction, 2009, 104, 551–553.
39. N. Homann, J. Tillonen, J. H. Meurman, H. Rintamaki, C. Lindqvist, M. Rautio, H. Jousimies-Somer and M. Salaspuro, Carcinogenesis, 2000, 21, 663–668.
40. F. S. Chambergo, E. D. Bonaccorsi, A. J. Ferreira, A. S. Ramos, J. J. Ferreira, Jr., J. Abrahao-Neto, J. P. Farah and H. El Dorry, J. Biol. Chem., 2002, 277, 13983–13988.
41. X. Chen, F. Xu, Y. Wang, Y. Pan, D. Lu, P. Wang, K. Ying, E. Chen and W. Zhang, Cancer, 2007, 110, 835–844.

---