Biodiesel Exhaust–Induced Cytotoxicity and Proinflammatory Mediator Production in Human Airway Epithelial Cells Benjamin J. Mullins,1,2* Anthony Kicic,3,4,5,6* Kak-Ming Ling,3 Ryan Mead-Hunter,1,2 Alexander N. Larcombe3 1
Fluid Dynamics Research Group, Curtin University, GPO Box U1987, Perth, Western Australia 6845, Australia
2
School of Public Health, Faculty of Health Sciences, Curtin University, GPO Box U1987, Perth, WA, Australia
3
Telethon Kids Institute, University of Western Australia, Subiaco, Western Australia, 6008, Australia
4
Department of Respiratory Medicine, Princess Margaret Hospital for Children, Perth, Western Australia, 6001, Australia
5
School of Paediatrics and Child Health, University of Western Australia, Nedlands, Western Australia, 6009, Australia
6
Centre for Cell Therapy and Regenerative Medicine, School of Medicine and Pharmacology, The University of Western Australia, Nedlands, 6009, Western Australia, Australia
Received 10 March 2014; revised 15 June 2014; accepted 17 June 2014 ABSTRACT: Increasing use of biodiesel has prompted research into the potential health effects of biodiesel exhaust exposure. Few studies directly compare the health consequences of mineral diesel, biodiesel, or blend exhaust exposures. Here, we exposed human epithelial cell cultures to diluted exhaust generated by the combustion of Australian ultralow-sulfur-diesel (ULSD), unprocessed canola oil, 100% canola biodiesel (B100), and a blend of 20% canola biodiesel mixed with 80% ULSD. The physicochemical characteristics of the exhaust were assessed and we compared cellular viability, apoptosis, and levels of interleukin (IL)-6, IL-8, and Regulated on Activation, Normal T cell Expressed and Secreted (RANTES) in exposed cultured cells. Different fuel types produced significantly different amounts of exhaust gases and different particle characteristics. All exposures resulted in significant apoptosis and loss of viability when compared with control, with an increasing proportion of biodiesel being correlated with a decrease in viability. In most cases, exposure to exhaust resulted in an increase in mediator production, with the greatest increases most often in response to B100. Exposure to pure canola oil (PCO) exhaust did not increase mediator production, but resulted in a significant decrease in IL-8 and RANTES in some cases. Our results show that canola biodiesel exhaust exposure elicits inflammation and reduces viability of human epithelial cell cultures in vitro when compared with ULSD exhaust exposure. This may be related to an increase in particle surface
Correspondence to: A. N. Larcombe; e-mail: [email protected] Contract grant sponsors: Thoracic Society of Australia and New Zealand (Maurice Blackburn Grant-In-Aid); Friends of the Institute for Child Health Research (Provision of Financial Support for Researchers and Research) *Benjamin J. Mullins and Anthony Kicic are joint first authors.
Published online 5 July 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/tox.22020 C 2014 Wiley Periodicals, Inc. V
44
BIODIESEL EXHAUST EXPOSURE OF EPITHELIAL CELLS
45
area and number in B100 exhaust when compared with ULSD exhaust. Exposure to PCO exhaust elicited the greatest loss of cellular viability, but virtually no inflammatory response, likely due to an overall increase C 2014 Wiley Periodicals, Inc. Environ Toxicol 31: 44–57, 2016. in average particle size. V Keywords: biodiesel exhaust exposure; canola oil; airway epithelial cell; cellular viability; apoptosis; inflammation; cytokines
INTRODUCTION As fossil fuel resources continue to become more difficult and expensive to extract, increasing urgency has been placed on finding suitable renewable fuel sources. One alternative that has recently received considerable attention is biodiesel, which is a generic term used for fuel made by reacting one of a wide variety of renewable plant- and animal-based oils with an alcohol to produce fatty acid esters. A key advantage of biodiesel is that it can directly replace (or combine with) mineral diesel in unmodified engines, and some oils such as canola may be used unmodified if preheated to a sufficient temperature (Fontaras et al., 2011). The most common renewable oil used to make biodiesel is rapeseed (Brassica sp.), which accounts for 60% of all biodiesel produced in the European Union (Flach et al., 2012). Since it first went into commercial production 30 years ago (International Energy Agency, 2004), many types of biodiesel have been extensively studied with a focus on their effects on engine performance and exhaust characteristics (Buyukkaya, 2010; Dwivedi et al., 2011). These studies generally show that biodiesel combustion reduces most gaseous exhaust emissions when compared with mineral diesel; however, the degree of reduction depends on a range of variables, including the type of renewable oil used, the blend ratio, engine type, and testing conditions (Krahl et al., 1996; Schr€oder et al., 1999; Swanson et al., 2006). Recently, there has been a shift toward trying to understand the potential health effects of exposure to biodiesel exhaust (B€ unger et al., 2012; Schr€oder et al., 2013); however, there is a paucity of human or animal exposure studies, which directly compare the health consequences of mineral diesel, biodiesel, or blend exhaust exposure in vivo (Brito et al., 2010; Tzamkiozis et al., 2010). There are, however, numerous in vitro studies investigating the mutagenic and cytotoxic effects of biodiesel exhaust exposure (Bagley et al., 1998; B€ unger et al., 1998, 2000a,b, 2007, 2012; Kado and Kuzmicky, 2003; Turrio-Baldassarri et al., 2004; Kisin et al., 2013). The majority of these studies use the Ames Test to assess the mutagenic properties of biodiesel exhaust, whereby tester strains of Salmonella typhimurium are exposed to particles captured on filters and extracted using solvents (and not the gaseous component of exhaust). These studies provide conflicting evidence as to the comparative mutagenicity or toxicity of exposure to biodiesel exhaust when compared with mineral diesel exhaust. For example, some show that biodiesel exhaust exposure reduces mutage-
nicity (B€unger et al., 1998; Kado and Kuzmicky, 2003), others suggest no discernible difference (Turrio-Baldassarri et al., 2004), and yet others show biodiesel blends or biodiesel fuels combusted under certain conditions are the most mutagenic (Kado et al., 2001; Kisin et al., 2013). The findings of the smaller number of in vitro studies investigating nonmutagenic effects are only slightly more convincing. For example, extracts of rapeseed methyl ester (RME) exhaust have been shown to induce stronger cytotoxic effects in mouse fibroblasts when compared with mineral diesel particle extract (B€unger et al., 1998, 2000a,b), whereas studies which expose transformed human airway epithelial cell cultures to extracted biodiesel exhaust particles indicated that 20% biodiesel blends yield stronger cytotoxicity than pure biodiesel or mineral diesel exhausts (Ackland et al., 2007; Liu et al., 2008). No previous studies have measured the production of proinflammatory mediators by cultured human airway epithelial cells exposed to biodiesel exhaust as an aerosol, as has been done in older studies of mineral diesel exposure (Bayram et al., 1998; Ohtoshi et al., 1998; Boland et al., 1999; Abe et al., 2000; Takizawa et al., 2000; Bonvallot et al., 2001). These previous studies typically exposed airway epithelial cell cultures to mineral diesel particles prepared in solution. This exposure methodology ignores the potential effects of the gaseous component of the exhaust and also does not provide a realistic particle size distribution. Mineral diesel particles in solution induce production of interleukin (IL)-8, soluble intercellular adhesion molecule-1, and granulocyte-macrophage colony-stimulating factor in nasal epithelial and bronchial explant cell cultures (Bayram et al., 1998; Ohtoshi et al., 1998). IL-8 mRNA expression was also induced by exposure to whole diesel exhaust (Abe et al., 2000) in a transformed human airway epithelial cell culture; however, the exhaust gas alone did not cause a sustained increase in IL-8 protein levels and showed no induction of IL-8 mRNA. This suggests that the particles themselves may be the most important trigger for a respiratory system inflammatory response. In this study, we exposed two human airway epithelial cell lines (10KT and NuLi-1) to diluted exhaust generated by a diesel engine under partial load. Fuel types used were standard Australian ultra-low-sulfur-diesel (ULSD), unprocessed canola oil (pure canola oil [PCO]), 100% canola biodiesel (B100), and a blend of 20% canola biodiesel mixed with 80% ULSD (B20). As distinct from most existing studies, cells were exposed to diluted exhaust containing both exhaust gases and particles. Post exposure, we compared
Environmental Toxicology DOI 10.1002/tox
46
MULLINS ET AL.
Fig. 1. Experimental setup of the cell exposure system (A) and close-up of the exposure chamber (B). Exhaust was cooled in the buffer chamber before a 1 lpm sample was extracted isokinetically. This sample was diluted with HEPA-filtered air before entering the exposure chamber. Exhaust gases were monitored in the exposure chamber, and particle characteristics were measured both prior to and after exiting the chamber. Exhaust was drawn through the system by a vacuum pump.
cellular viability, apoptosis, and levels of proinflammatory mediator production to the different types of exhaust. We hypothesized that there would be increased programmed cell death and greater production of proinflammatory mediators in cell cultures exposed to biodiesel when compared with ULSD because of the altered physicochemical properties of the exhaust mix.
MATERIALS AND METHODS Fuels and Exhaust Generation Fuels used were (i) commercial 100% ULSD (produced by BP Western Australia, with confirmation that it contained no biodiesel), (ii) B100 (fatty acid methyl ester [FAME]) produced from new canola (a variety of rapeseed), (iii) 20% biodiesel (B20 a blend of 20% FAME and 80% ULSD by volume), and (iv) PCO, all preheated to 70 C prior to use.
Environmental Toxicology DOI 10.1002/tox
The ULSD used conformed to the Australian Diesel Fuel Quality Standard (2009). The canola oil used was new Australian grown food-grade canola. B20 and B100 were produced from the same canola oil and high purity chemicals to conform to the Australian Biodiesel Quality Standard (2006). PCO was studied as it is commonly used as a direct substitute for diesel in many agricultural applications and passenger vehicles (Golimowski et al., 2012). Fuels were combusted using a light-medium duty diesel engine (Isuzu 4BD1-T, 3.9L) operated at a constant 1800 rpm and 20% load for all experiments. Load was applied using an enginedriven air compressor. This engine and conditions were chosen because of the relatively high particle number produced which would provide the highest probability of detecting differences between treatments. The engine and fuel lines were completely purged between experiments to ensure no contamination between fuels. Exhaust was prepared using standard methods (Burtscher et al., 1998). A 1 liter per minute
BIODIESEL EXHAUST EXPOSURE OF EPITHELIAL CELLS
(lpm) sample of whole exhaust was extracted isokinetically then diluted with 9 lpm of HEPA-filtered makeup air (Fig. 1). This air (10 lpm total) was then passed through the exposure chamber. The gaseous atmosphere in the chamber was monitored continuously throughout exposures using a GasAlert Micro 5 and a GasAlert Extreme (BW Technologies by Honeywell International, Calgary, Canada). Particulates were monitored using a DustTrak DRX (TSI, Shoreview, MN, USA), with detailed measurements performed using a TSI Scanning Mobility Particle Sizer (SMPS; TSI), consisting of a 3081 differential mobility analyzer and a 3775 condensation particle counter. The DustTrak was fitted with a 37-mm gravimetric filter. Total mass was calculated from gravimetric measurements and used to calibrate the DustTrak and SMPS measurements. Between five and seven replicate samples were taken for each exhaust type. SMPS data were processed to obtain mass data (Liu et al., 2012). Surface area was calculated based on previously published relationships (Maricq and Xu, 2004). Sampling during tests was conducted in the exit from the chamber to ensure that the sampling did not influence particle deposition.
Cells and Culturing Two telomere-shortened airway epithelial cell lines were used in this study. The NuLi-1 cell line (Zabner et al., 2003) was obtained from ATCC (Manassas, VA, USA). Cells were cultured in human placenta collagen (HPC) precoated tissue culture plasticware using specific growth media consisting of BEBM supplemented with 0.025 mg/mL BPE, 0.025 mg/ mL EGF, 0.5 mg/mL epinephrine, 0.125 mg/mL amphotericin B, 1% (v/v) gentamicin, 1% (v/v) penicillin/streptomycin, 0.5 mg/mL hydrocortisone, 5 mg/mL insulin, 0.1 ng/mL retinoic acid, 0.01 ng/mL transferrin, 6.5 ng/mL triiodothyronine, and 2% (v/v) Ultroser-G. The 10KT (Ramirez et al., 2004) cell line was provided by Prof. D. Knight (University of Newcastle, NSW, Australia) who originally obtained the line from Dr. John Minna (University of Texas Southwestern Medical Centre, Dallas, USA). Cells were cultured in HPC precoated T25-cm2 flasks using growth media consisting of KSFM supplemented with KSFM supplements (BPE and EGF), 1% (v/v) gentamicin, and 1% (v/v) penicillin/streptomycin. Both cell types were derived from normal lungs from separate adult patients via dual retroviral infection with HPV-16 E6/E7-LXN (Zabner et al., 2003). As a result, both cell types have the advantage of not undergoing growth arrest when in cell culture due to exogenous expression of the telomerase and HPV-16 E6/E7 genes. Furthermore, they are considered more characteristic of airway epithelia because of forming highly polarized differentiated epithelia, which exhibit transepithelial electrical resistance, and maintaining ion channel physiology, which may have been lost in many other commercially available epithelial cell lines. The two cell types were used to provide potential interperson variation in responses to fine particle exposure.
47
All cells were grown in a Panasonic CO2 incubator at 37 C in an atmosphere of 5% CO2/95% air under strict aseptic conditions.
DEP Exposure Protocol Airway epithelial cells were initially seeded at 10,000 cells per centimeter squared in 12-well plates in their respective growth media and grown to confluence. Twenty-four hours prior to exposure to exhaust, growth medium was replaced with their respective basal media, without any supplements to remove any factors that may influence cytokine production. Cell culture plates were loaded into the exposure chamber (30 cm 3 30 cm 3 30 cm; volume 9 L) and exposed to one of the different types of exhaust for 1 h. Control cell cultures were simultaneously exposed to HEPA-filtered air for 1 h. During exposures, plates were maintained at 37 C using a heating pad. Between exposures, the chamber was cleaned and flushed with HEPA-filtered air. Cultures were then incubated for a further 6, 12, or 24 h at 37 C in an atmosphere of 5% CO2/95% air. At each time point, media supernatants were collected for inflammatory mediator production, and cells were collected for induction of apoptosis and viability.
Apoptosis and Viability Cells were initially seeded into a 96-well plate at 10,000 cells per centimeter squared and maintained at 37 C in an atmosphere of 5% CO2/95% air. After reaching 100% confluency, cells were exposed to exhaust for 1 h, and then the plates were allowed to incubate for a further 24 h at 37 C under aseptic conditions. For assessment of apoptosis, cellular supernatants were aspirated off, and then the cells were fixed in 80% (v/v) methanol. Apoptosis was then determined using a single-stranded DNA Apoptosis ELISA (Millipore, Billerica, MA, USA). Cell viability was determined using a CellTiter 96 Aqueous assay (Promega, Madison, WI, USA). Briefly, cells were washed twice with tissue culture PBS, and a set volume of RPMI-1640 (without phenol red) was added to each well. A set volume of the MTS substrate was then added to each well, and then the plates were incubated for 2 h at 37 C. Absorbances were recorded at 490 nm using a Multiskan FC plate reader (ThermoFisher Scientific, Scoresby, Victoria, Australia).
Measurement of Cytokine Production Expression of IL-8 (Becton Dickinson, Biosciences, San Diego, CA, USA) and Regulated on Activation, Normal T cell Expressed and Secreted (RANTES; Bio-Scientific, Sydney, NSW, Australia) was measured by ELISA. IL-6 was measured using a time-resolved fluorometry detection system as previously described (Perkin-Elmer, Waltham, MA, USA; Taylor et al., 2007; Stevens et al., 2008; Sutanto et al.,
Environmental Toxicology DOI 10.1002/tox
48
MULLINS ET AL.
Fig. 2. The combustion of ULSD, B20, B100, or PCO fuel in a light-medium duty Euro 1 diesel engine produces exhaust gases and particles with significantly different physicochemical properties. Different letters and “*” indicate significant differences between exhaust types (p < 0.05).
2011). Cytokine production was normalized to the number of viable cells.
Statistics Analyses were conducted using one-way ANOVA. The Holm-Sidak post hoc test was used to identify significant differences between groups. Data were transformed, where necessary, to satisfy the assumptions of normality and homoscedasticity. Analyses were performed using SigmaStat software (v3.50; SPSS Science, Chicago, IL, USA). A
Environmental Toxicology DOI 10.1002/tox
p-value of 0.071 in both cases). Twelve hours after exposure, IL-8 levels were higher in ULSD-exposed cells when compared with all other exposures (p < 0.043 in all cases); however, at this time point, IL-8 was also higher in B100-exposed 10KT cells when compared with B20exposed cells. Twenty-four hours after exposure, IL-8 levels were significantly higher in all exposure groups when compared with controls (p < 0.001 in all cases) or PCO-exposed
Environmental Toxicology DOI 10.1002/tox
cells (p 5 0.128). At this time point, IL-8 was still significantly higher in ULSD-exposed 10KT cells when compared with B20-exposed cells (p 5 0.008). The pattern of response was similar in the NuLi-1 cell line at 6 h after exposure, but quite different 12 and 24 h after exposure. At all three time points, there was no difference in IL-8 between controls and PCO (p > 0.136 in all cases). Six hours after exposure, IL-8 was significantly lower in B20-exposed cells when compared with either ULSD- or B100-exposed cells (p < 0.018 in both cases); however, there was no difference between ULSD and B100 (p 5 0.951). Twelve hours after exposure, IL-8 was significantly higher in B100-exposed cells when compared with either ULSD- or B20-exposed cells (p < 0.006 in both cases). There was no difference between ULSD and B20 (p 5 0.160). This pattern was even more pronounced 24 h after exposure, whereby IL-8 in B100exposed cells was significantly higher than that in either B20- or ULSD-exposed cells (p < 0.001). Furthermore, IL-8 was higher in B20-exposed cells when compared with ULSD-exposed cells (p 5 0.009).
RANTES For the 10KT cell line at 6 h after exposure, B100 was not significantly different than ULSD (p 5 0.618), although RANTES was significantly higher in B100-exposed cells
BIODIESEL EXHAUST EXPOSURE OF EPITHELIAL CELLS
51
Fig. 6. Production of IL-6, IL-8, and RANTES by 10KT (left) and NuLi-1 (right) cells after exposure to control, ULSD, B20, B100, or PCO exhaust for 1 h. IL-8 and RANTES were measured using ELISA, and IL-6 was measured using time-resolved fluorometry. The greatest inflammatory response was most often induced by exposure to B100 exhaust. PCO exhaust exposure did not increase inflammatory cytokine production at any time point. Responses of the two cell types were similar. Bars are mean 6 standard deviation. Different letters indicate significant differences between different exhaust types within the same time point (p < 0.05).
when compared with B20-exposed cells at this time point (p 5 0.002). The effects of different exposures on RANTES levels in 10KT cells were similar at the 12- and 24-h time points. At both time points, RANTES was significantly higher in B100-exposed cells when compared with all other exposures (p < 0.001 in all cases). Furthermore, RANTES
was significantly higher in ULSD-exposed cells when compared with B20-exposed cells. RANTES production by NuLi-1 cells after exposure was slightly different to the pattern seen in 10KT cells. Six hours after exposure, cells exposed to B100 produced significantly more RANTES (approximately double) when compared with any other
Environmental Toxicology DOI 10.1002/tox
52
MULLINS ET AL.
TABLE I. Linear regression analyses between exhaust physiochemical characteristics and in vitro responses of 10KT and NuLi-1 cell lines exposed to different exhaust types
10KT CO2 CO NO NO2 SOx GMPD SA MMC NuLi-1 CO2 CO NO NO2 SOx GMPD SA MMC
Apoptosis
Viability
IL-6
IL-8
RANTES
0.318 0.143 0.029 0.424 0.59 0.016 0.069 0.109
Fluid Dynamics Research Group, Curtin University, GPO Box U1987, Perth, Western Australia 6845, Australia
2
School of Public Health, Faculty of Health Sciences, Curtin University, GPO Box U1987, Perth, WA, Australia
3
Telethon Kids Institute, University of Western Australia, Subiaco, Western Australia, 6008, Australia
4
Department of Respiratory Medicine, Princess Margaret Hospital for Children, Perth, Western Australia, 6001, Australia
5
School of Paediatrics and Child Health, University of Western Australia, Nedlands, Western Australia, 6009, Australia
6
Centre for Cell Therapy and Regenerative Medicine, School of Medicine and Pharmacology, The University of Western Australia, Nedlands, 6009, Western Australia, Australia
Received 10 March 2014; revised 15 June 2014; accepted 17 June 2014 ABSTRACT: Increasing use of biodiesel has prompted research into the potential health effects of biodiesel exhaust exposure. Few studies directly compare the health consequences of mineral diesel, biodiesel, or blend exhaust exposures. Here, we exposed human epithelial cell cultures to diluted exhaust generated by the combustion of Australian ultralow-sulfur-diesel (ULSD), unprocessed canola oil, 100% canola biodiesel (B100), and a blend of 20% canola biodiesel mixed with 80% ULSD. The physicochemical characteristics of the exhaust were assessed and we compared cellular viability, apoptosis, and levels of interleukin (IL)-6, IL-8, and Regulated on Activation, Normal T cell Expressed and Secreted (RANTES) in exposed cultured cells. Different fuel types produced significantly different amounts of exhaust gases and different particle characteristics. All exposures resulted in significant apoptosis and loss of viability when compared with control, with an increasing proportion of biodiesel being correlated with a decrease in viability. In most cases, exposure to exhaust resulted in an increase in mediator production, with the greatest increases most often in response to B100. Exposure to pure canola oil (PCO) exhaust did not increase mediator production, but resulted in a significant decrease in IL-8 and RANTES in some cases. Our results show that canola biodiesel exhaust exposure elicits inflammation and reduces viability of human epithelial cell cultures in vitro when compared with ULSD exhaust exposure. This may be related to an increase in particle surface
Correspondence to: A. N. Larcombe; e-mail: [email protected] Contract grant sponsors: Thoracic Society of Australia and New Zealand (Maurice Blackburn Grant-In-Aid); Friends of the Institute for Child Health Research (Provision of Financial Support for Researchers and Research) *Benjamin J. Mullins and Anthony Kicic are joint first authors.
Published online 5 July 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/tox.22020 C 2014 Wiley Periodicals, Inc. V
44
BIODIESEL EXHAUST EXPOSURE OF EPITHELIAL CELLS
45
area and number in B100 exhaust when compared with ULSD exhaust. Exposure to PCO exhaust elicited the greatest loss of cellular viability, but virtually no inflammatory response, likely due to an overall increase C 2014 Wiley Periodicals, Inc. Environ Toxicol 31: 44–57, 2016. in average particle size. V Keywords: biodiesel exhaust exposure; canola oil; airway epithelial cell; cellular viability; apoptosis; inflammation; cytokines
INTRODUCTION As fossil fuel resources continue to become more difficult and expensive to extract, increasing urgency has been placed on finding suitable renewable fuel sources. One alternative that has recently received considerable attention is biodiesel, which is a generic term used for fuel made by reacting one of a wide variety of renewable plant- and animal-based oils with an alcohol to produce fatty acid esters. A key advantage of biodiesel is that it can directly replace (or combine with) mineral diesel in unmodified engines, and some oils such as canola may be used unmodified if preheated to a sufficient temperature (Fontaras et al., 2011). The most common renewable oil used to make biodiesel is rapeseed (Brassica sp.), which accounts for 60% of all biodiesel produced in the European Union (Flach et al., 2012). Since it first went into commercial production 30 years ago (International Energy Agency, 2004), many types of biodiesel have been extensively studied with a focus on their effects on engine performance and exhaust characteristics (Buyukkaya, 2010; Dwivedi et al., 2011). These studies generally show that biodiesel combustion reduces most gaseous exhaust emissions when compared with mineral diesel; however, the degree of reduction depends on a range of variables, including the type of renewable oil used, the blend ratio, engine type, and testing conditions (Krahl et al., 1996; Schr€oder et al., 1999; Swanson et al., 2006). Recently, there has been a shift toward trying to understand the potential health effects of exposure to biodiesel exhaust (B€ unger et al., 2012; Schr€oder et al., 2013); however, there is a paucity of human or animal exposure studies, which directly compare the health consequences of mineral diesel, biodiesel, or blend exhaust exposure in vivo (Brito et al., 2010; Tzamkiozis et al., 2010). There are, however, numerous in vitro studies investigating the mutagenic and cytotoxic effects of biodiesel exhaust exposure (Bagley et al., 1998; B€ unger et al., 1998, 2000a,b, 2007, 2012; Kado and Kuzmicky, 2003; Turrio-Baldassarri et al., 2004; Kisin et al., 2013). The majority of these studies use the Ames Test to assess the mutagenic properties of biodiesel exhaust, whereby tester strains of Salmonella typhimurium are exposed to particles captured on filters and extracted using solvents (and not the gaseous component of exhaust). These studies provide conflicting evidence as to the comparative mutagenicity or toxicity of exposure to biodiesel exhaust when compared with mineral diesel exhaust. For example, some show that biodiesel exhaust exposure reduces mutage-
nicity (B€unger et al., 1998; Kado and Kuzmicky, 2003), others suggest no discernible difference (Turrio-Baldassarri et al., 2004), and yet others show biodiesel blends or biodiesel fuels combusted under certain conditions are the most mutagenic (Kado et al., 2001; Kisin et al., 2013). The findings of the smaller number of in vitro studies investigating nonmutagenic effects are only slightly more convincing. For example, extracts of rapeseed methyl ester (RME) exhaust have been shown to induce stronger cytotoxic effects in mouse fibroblasts when compared with mineral diesel particle extract (B€unger et al., 1998, 2000a,b), whereas studies which expose transformed human airway epithelial cell cultures to extracted biodiesel exhaust particles indicated that 20% biodiesel blends yield stronger cytotoxicity than pure biodiesel or mineral diesel exhausts (Ackland et al., 2007; Liu et al., 2008). No previous studies have measured the production of proinflammatory mediators by cultured human airway epithelial cells exposed to biodiesel exhaust as an aerosol, as has been done in older studies of mineral diesel exposure (Bayram et al., 1998; Ohtoshi et al., 1998; Boland et al., 1999; Abe et al., 2000; Takizawa et al., 2000; Bonvallot et al., 2001). These previous studies typically exposed airway epithelial cell cultures to mineral diesel particles prepared in solution. This exposure methodology ignores the potential effects of the gaseous component of the exhaust and also does not provide a realistic particle size distribution. Mineral diesel particles in solution induce production of interleukin (IL)-8, soluble intercellular adhesion molecule-1, and granulocyte-macrophage colony-stimulating factor in nasal epithelial and bronchial explant cell cultures (Bayram et al., 1998; Ohtoshi et al., 1998). IL-8 mRNA expression was also induced by exposure to whole diesel exhaust (Abe et al., 2000) in a transformed human airway epithelial cell culture; however, the exhaust gas alone did not cause a sustained increase in IL-8 protein levels and showed no induction of IL-8 mRNA. This suggests that the particles themselves may be the most important trigger for a respiratory system inflammatory response. In this study, we exposed two human airway epithelial cell lines (10KT and NuLi-1) to diluted exhaust generated by a diesel engine under partial load. Fuel types used were standard Australian ultra-low-sulfur-diesel (ULSD), unprocessed canola oil (pure canola oil [PCO]), 100% canola biodiesel (B100), and a blend of 20% canola biodiesel mixed with 80% ULSD (B20). As distinct from most existing studies, cells were exposed to diluted exhaust containing both exhaust gases and particles. Post exposure, we compared
Environmental Toxicology DOI 10.1002/tox
46
MULLINS ET AL.
Fig. 1. Experimental setup of the cell exposure system (A) and close-up of the exposure chamber (B). Exhaust was cooled in the buffer chamber before a 1 lpm sample was extracted isokinetically. This sample was diluted with HEPA-filtered air before entering the exposure chamber. Exhaust gases were monitored in the exposure chamber, and particle characteristics were measured both prior to and after exiting the chamber. Exhaust was drawn through the system by a vacuum pump.
cellular viability, apoptosis, and levels of proinflammatory mediator production to the different types of exhaust. We hypothesized that there would be increased programmed cell death and greater production of proinflammatory mediators in cell cultures exposed to biodiesel when compared with ULSD because of the altered physicochemical properties of the exhaust mix.
MATERIALS AND METHODS Fuels and Exhaust Generation Fuels used were (i) commercial 100% ULSD (produced by BP Western Australia, with confirmation that it contained no biodiesel), (ii) B100 (fatty acid methyl ester [FAME]) produced from new canola (a variety of rapeseed), (iii) 20% biodiesel (B20 a blend of 20% FAME and 80% ULSD by volume), and (iv) PCO, all preheated to 70 C prior to use.
Environmental Toxicology DOI 10.1002/tox
The ULSD used conformed to the Australian Diesel Fuel Quality Standard (2009). The canola oil used was new Australian grown food-grade canola. B20 and B100 were produced from the same canola oil and high purity chemicals to conform to the Australian Biodiesel Quality Standard (2006). PCO was studied as it is commonly used as a direct substitute for diesel in many agricultural applications and passenger vehicles (Golimowski et al., 2012). Fuels were combusted using a light-medium duty diesel engine (Isuzu 4BD1-T, 3.9L) operated at a constant 1800 rpm and 20% load for all experiments. Load was applied using an enginedriven air compressor. This engine and conditions were chosen because of the relatively high particle number produced which would provide the highest probability of detecting differences between treatments. The engine and fuel lines were completely purged between experiments to ensure no contamination between fuels. Exhaust was prepared using standard methods (Burtscher et al., 1998). A 1 liter per minute
BIODIESEL EXHAUST EXPOSURE OF EPITHELIAL CELLS
(lpm) sample of whole exhaust was extracted isokinetically then diluted with 9 lpm of HEPA-filtered makeup air (Fig. 1). This air (10 lpm total) was then passed through the exposure chamber. The gaseous atmosphere in the chamber was monitored continuously throughout exposures using a GasAlert Micro 5 and a GasAlert Extreme (BW Technologies by Honeywell International, Calgary, Canada). Particulates were monitored using a DustTrak DRX (TSI, Shoreview, MN, USA), with detailed measurements performed using a TSI Scanning Mobility Particle Sizer (SMPS; TSI), consisting of a 3081 differential mobility analyzer and a 3775 condensation particle counter. The DustTrak was fitted with a 37-mm gravimetric filter. Total mass was calculated from gravimetric measurements and used to calibrate the DustTrak and SMPS measurements. Between five and seven replicate samples were taken for each exhaust type. SMPS data were processed to obtain mass data (Liu et al., 2012). Surface area was calculated based on previously published relationships (Maricq and Xu, 2004). Sampling during tests was conducted in the exit from the chamber to ensure that the sampling did not influence particle deposition.
Cells and Culturing Two telomere-shortened airway epithelial cell lines were used in this study. The NuLi-1 cell line (Zabner et al., 2003) was obtained from ATCC (Manassas, VA, USA). Cells were cultured in human placenta collagen (HPC) precoated tissue culture plasticware using specific growth media consisting of BEBM supplemented with 0.025 mg/mL BPE, 0.025 mg/ mL EGF, 0.5 mg/mL epinephrine, 0.125 mg/mL amphotericin B, 1% (v/v) gentamicin, 1% (v/v) penicillin/streptomycin, 0.5 mg/mL hydrocortisone, 5 mg/mL insulin, 0.1 ng/mL retinoic acid, 0.01 ng/mL transferrin, 6.5 ng/mL triiodothyronine, and 2% (v/v) Ultroser-G. The 10KT (Ramirez et al., 2004) cell line was provided by Prof. D. Knight (University of Newcastle, NSW, Australia) who originally obtained the line from Dr. John Minna (University of Texas Southwestern Medical Centre, Dallas, USA). Cells were cultured in HPC precoated T25-cm2 flasks using growth media consisting of KSFM supplemented with KSFM supplements (BPE and EGF), 1% (v/v) gentamicin, and 1% (v/v) penicillin/streptomycin. Both cell types were derived from normal lungs from separate adult patients via dual retroviral infection with HPV-16 E6/E7-LXN (Zabner et al., 2003). As a result, both cell types have the advantage of not undergoing growth arrest when in cell culture due to exogenous expression of the telomerase and HPV-16 E6/E7 genes. Furthermore, they are considered more characteristic of airway epithelia because of forming highly polarized differentiated epithelia, which exhibit transepithelial electrical resistance, and maintaining ion channel physiology, which may have been lost in many other commercially available epithelial cell lines. The two cell types were used to provide potential interperson variation in responses to fine particle exposure.
47
All cells were grown in a Panasonic CO2 incubator at 37 C in an atmosphere of 5% CO2/95% air under strict aseptic conditions.
DEP Exposure Protocol Airway epithelial cells were initially seeded at 10,000 cells per centimeter squared in 12-well plates in their respective growth media and grown to confluence. Twenty-four hours prior to exposure to exhaust, growth medium was replaced with their respective basal media, without any supplements to remove any factors that may influence cytokine production. Cell culture plates were loaded into the exposure chamber (30 cm 3 30 cm 3 30 cm; volume 9 L) and exposed to one of the different types of exhaust for 1 h. Control cell cultures were simultaneously exposed to HEPA-filtered air for 1 h. During exposures, plates were maintained at 37 C using a heating pad. Between exposures, the chamber was cleaned and flushed with HEPA-filtered air. Cultures were then incubated for a further 6, 12, or 24 h at 37 C in an atmosphere of 5% CO2/95% air. At each time point, media supernatants were collected for inflammatory mediator production, and cells were collected for induction of apoptosis and viability.
Apoptosis and Viability Cells were initially seeded into a 96-well plate at 10,000 cells per centimeter squared and maintained at 37 C in an atmosphere of 5% CO2/95% air. After reaching 100% confluency, cells were exposed to exhaust for 1 h, and then the plates were allowed to incubate for a further 24 h at 37 C under aseptic conditions. For assessment of apoptosis, cellular supernatants were aspirated off, and then the cells were fixed in 80% (v/v) methanol. Apoptosis was then determined using a single-stranded DNA Apoptosis ELISA (Millipore, Billerica, MA, USA). Cell viability was determined using a CellTiter 96 Aqueous assay (Promega, Madison, WI, USA). Briefly, cells were washed twice with tissue culture PBS, and a set volume of RPMI-1640 (without phenol red) was added to each well. A set volume of the MTS substrate was then added to each well, and then the plates were incubated for 2 h at 37 C. Absorbances were recorded at 490 nm using a Multiskan FC plate reader (ThermoFisher Scientific, Scoresby, Victoria, Australia).
Measurement of Cytokine Production Expression of IL-8 (Becton Dickinson, Biosciences, San Diego, CA, USA) and Regulated on Activation, Normal T cell Expressed and Secreted (RANTES; Bio-Scientific, Sydney, NSW, Australia) was measured by ELISA. IL-6 was measured using a time-resolved fluorometry detection system as previously described (Perkin-Elmer, Waltham, MA, USA; Taylor et al., 2007; Stevens et al., 2008; Sutanto et al.,
Environmental Toxicology DOI 10.1002/tox
48
MULLINS ET AL.
Fig. 2. The combustion of ULSD, B20, B100, or PCO fuel in a light-medium duty Euro 1 diesel engine produces exhaust gases and particles with significantly different physicochemical properties. Different letters and “*” indicate significant differences between exhaust types (p < 0.05).
2011). Cytokine production was normalized to the number of viable cells.
Statistics Analyses were conducted using one-way ANOVA. The Holm-Sidak post hoc test was used to identify significant differences between groups. Data were transformed, where necessary, to satisfy the assumptions of normality and homoscedasticity. Analyses were performed using SigmaStat software (v3.50; SPSS Science, Chicago, IL, USA). A
Environmental Toxicology DOI 10.1002/tox
p-value of 0.071 in both cases). Twelve hours after exposure, IL-8 levels were higher in ULSD-exposed cells when compared with all other exposures (p < 0.043 in all cases); however, at this time point, IL-8 was also higher in B100-exposed 10KT cells when compared with B20exposed cells. Twenty-four hours after exposure, IL-8 levels were significantly higher in all exposure groups when compared with controls (p < 0.001 in all cases) or PCO-exposed
Environmental Toxicology DOI 10.1002/tox
cells (p 5 0.128). At this time point, IL-8 was still significantly higher in ULSD-exposed 10KT cells when compared with B20-exposed cells (p 5 0.008). The pattern of response was similar in the NuLi-1 cell line at 6 h after exposure, but quite different 12 and 24 h after exposure. At all three time points, there was no difference in IL-8 between controls and PCO (p > 0.136 in all cases). Six hours after exposure, IL-8 was significantly lower in B20-exposed cells when compared with either ULSD- or B100-exposed cells (p < 0.018 in both cases); however, there was no difference between ULSD and B100 (p 5 0.951). Twelve hours after exposure, IL-8 was significantly higher in B100-exposed cells when compared with either ULSD- or B20-exposed cells (p < 0.006 in both cases). There was no difference between ULSD and B20 (p 5 0.160). This pattern was even more pronounced 24 h after exposure, whereby IL-8 in B100exposed cells was significantly higher than that in either B20- or ULSD-exposed cells (p < 0.001). Furthermore, IL-8 was higher in B20-exposed cells when compared with ULSD-exposed cells (p 5 0.009).
RANTES For the 10KT cell line at 6 h after exposure, B100 was not significantly different than ULSD (p 5 0.618), although RANTES was significantly higher in B100-exposed cells
BIODIESEL EXHAUST EXPOSURE OF EPITHELIAL CELLS
51
Fig. 6. Production of IL-6, IL-8, and RANTES by 10KT (left) and NuLi-1 (right) cells after exposure to control, ULSD, B20, B100, or PCO exhaust for 1 h. IL-8 and RANTES were measured using ELISA, and IL-6 was measured using time-resolved fluorometry. The greatest inflammatory response was most often induced by exposure to B100 exhaust. PCO exhaust exposure did not increase inflammatory cytokine production at any time point. Responses of the two cell types were similar. Bars are mean 6 standard deviation. Different letters indicate significant differences between different exhaust types within the same time point (p < 0.05).
when compared with B20-exposed cells at this time point (p 5 0.002). The effects of different exposures on RANTES levels in 10KT cells were similar at the 12- and 24-h time points. At both time points, RANTES was significantly higher in B100-exposed cells when compared with all other exposures (p < 0.001 in all cases). Furthermore, RANTES
was significantly higher in ULSD-exposed cells when compared with B20-exposed cells. RANTES production by NuLi-1 cells after exposure was slightly different to the pattern seen in 10KT cells. Six hours after exposure, cells exposed to B100 produced significantly more RANTES (approximately double) when compared with any other
Environmental Toxicology DOI 10.1002/tox
52
MULLINS ET AL.
TABLE I. Linear regression analyses between exhaust physiochemical characteristics and in vitro responses of 10KT and NuLi-1 cell lines exposed to different exhaust types
10KT CO2 CO NO NO2 SOx GMPD SA MMC NuLi-1 CO2 CO NO NO2 SOx GMPD SA MMC
Apoptosis
Viability
IL-6
IL-8
RANTES
0.318 0.143 0.029 0.424 0.59 0.016 0.069 0.109
