Analytical and Bioanalytical Chemistry

Electronic supplementary material

Catalytic diesel particulate filters reduce the in vitro estrogenic activity of diesel exhaust

Daniela Wenger, Andreas C. Gerecke (*), Norbert V. Heeb, Hanspeter Naegeli, and Renato Zenobi


ISO 8178/4 C1 test cycle

The eight-stage ISO 8178/4 C1 cycle is a steady-state emission cycle for construction site engines. It is composed of a sequence of constant engine speed and load modes with different weighting factors and simulates eight engine operating situations that cover typical field conditions [1]. The cycle is driven in the sequence described by Fig. S1. The cycle starts with four load-states at a maximum of 2,000 rounds per minute (RPM), followed by three load-states at an intermediate RPM (60% of max. RPM) and ends with an idling phase. The dwell time of each load per RPM point is 10-15 min, resulting in a total cycle time of 100 min. These 100 min also include the transition from one load-stage to the next.

In our study, exhaust sampling was carried out on the test stands of the University of Applied Sciences in Biel (BFH-TI Biel, Switzerland). The ISO 8178/4 C1 cycle (Fig. S1) was driven two times without interruptions. Thus, the duration of a sampling arrangement was 200 min. Mass flow proportional aliquots of undiluted diesel exhaust were taken during the two consecutive runs [1, 2], yielding 4.4–6.7 m3 of exhaust per sample. After each fuel change, engine and diesel particulate filters (DPFs) were preconditioned for one hour at the first stage of the ISO 8178/4 C1 cycle (max. torque, full load).

All-glass sampling devices

Exhaust sampling was performed according to the filter/condenser method described in European standard EN-1948-1 [3]. We used all-glass sampling devices to collect samples of diesel exhaust for bioassay analysis. As illustrated by Fig. S2, a sampling device consisted of a sampling probe, a quartz fiber filter, a cooler, a condensate separator, and a two-stage adsorber unit (XAD-2). With this sampling device particle-bound compounds and semivolatile compounds were obtained. Samples were collected in individual sampling devices, which all had been cleaned (water, detergent, solvents) and heated up to 450 °C for several hours to remove any traces of organics.

European standard EN-1948-1 has been validated for collection of PCDD/Fs, both adsorbed on particles and in the gas phase, from the hot, humid exhaust of stationary sources [3]. Validation trials were performed on emissions of municipal waste incinerators at a level of about 0.1 ng I-TEQ m-3 and a dust loading of 1–15 mg m-3. Although EN-1948-1 has been developed and validated primarily for gaseous streams emitted by waste incinerators, practical experience showed that it can be applied for a wide range of concentrations and for various emission sources, for example, for diesel engine emissions [1]. Besides for the analysis of PCDD/Fs, the sampling methods described in EN-1948-1 are suitable for the determination of other low- and semivolatile compounds.

Culture of T47D.Luc cells

Genetically modified human breast adenocarcinoma T47D cells (T47D.Luc) were purchased from BioDetection Systems (BDS, Amsterdam, The Netherlands) and cultured according to the protocols of BDS, as described elsewhere [4]. All media and supplements used for cell culture were from Gibco (Paisley, Scotland). Cells were grown in a 1:1 mixture of Dulbecco’s modified Eagle’s medium and Ham’s F12 nutrient mixture (D-MEM/F-12) with phenol red and supplemented with NaHCO3 (1.25 g L-1), 1% (v/v) non-essential amino acids, and 7.5% (v/v) fetal bovine serum (FBS). Cells were incubated (37 °C, 5% CO2, 100% relative humidity) until a confluency of 85–95% was reached. Cells were collected by trypsinization and used for cell culture or assay analysis.

Quality assurance/quality control for assay analysis

Extracts of diesel exhaust were analyzed in triplicate on 96-well microtiter plates. Results with relative standard deviations above 15% (based on triplicate measurements) were discarded. In addition to triplicate analysis, E2-CEQ concentrations measured after 24 h of cell exposure were determined on 3–4 independent microtiter plates. The relative standard deviations of interplate measurements were between 3–19%.

The limit of detection (LOD) was defined as 3 times the standard deviation of the solvent control measured in RLUs. The limit of quantification (LOQ) was set at 10 times the standard deviation of the solvent control measured in RLUs. The LOD and LOQ were calculated for each individual plate and converted to pM E2 using the fitted curve of the E2 dilution series. The average LOD and LOQ of all plates measured after 24 h of exposure were 0.5±0.2 pM E2 (15 fg E2 per well) and 1.1±0.5 pM E2 (31 fg E2 per well), respectively.

The protocols of BioDetection Systems (BDS, Amsterdam, The Netherlands) set the range for quantification of luciferase activity between the LOQ and the 50% effect concentration (EC50) of the dose-response curve of E2. It has been shown that CALUX determinations performed on the lower part of the curve of the reference compound result in more precise and less variable values, and thus give more reliable results [5]. We conducted range finding experiments to choose sample dilutions that induced an activity close to the LOQ. The selected sample dilutions did not exhibit luciferase activity above 20% of the maximal induction level of E2 (i.e., activity at the EC20 of the fitted curve of E2).

Expression of the results as E2-CEQs was allowed if the fitted curve of E2 had a coefficient of determination (r2) ≥0.98, an EC50 value of 2–8 pM E2, and a maximum induction factor (IF) ≥6 (IF = [a0 + RLUDMSO]/RLUDMSO; a0, maximum activity in RLUs; RLUDMSO, activity of solvent alone). For the 12, 48, and 72 h experiments, the curve fit had to fulfill exclusively the criterion of r2 ≥ 0.98.

A plate internal reference compound (EE2, Fluka, Switzerland) and the 3 pM concentration point of E2 were used to control assay performance. If their assay response was within two standard deviations of the long-term mean, the results of a microtiter plate were considered as reliable.

An ER antagonist, ICI 182,780 (Tocris Bioscience, Avonmouth, U.K.) [6], was used in a concentration of 10 nM [7] to test whether observed luciferase activity was ER-mediated or induced by other pathways.

E2-CEQ concentrations

ER-mediated activity of diesel exhaust generated by a heavy duty diesel engine (Liebherr, 914 T, 6.11 L, 105 kW) was determined from exhaust samples collected during two consecutive runs of the ISO 8178/4 C1 test cycle. Variations of engine load, consumed fuel, and exhaust gas flow were small (<3%) [2]. On average, 25.7±0.2 L (21.4±0.1 kg) of fuel were consumed during one pass of the ISO 8178/4 C1 cycle (100 min) at a mean engine load of 55.3±0.1 kW resulting in 531.7±12.1 m3 of dry exhaust [2]. With these mean parameters, ER-mediated activity per m3 of diesel exhaust can be converted to ER-mediated activity per kWh or per L and vice versa. In Table S1, the overall ER-mediated activity of the collected exhaust samples is reported in E2-CEQ m-3, E2-CEQ kWh-1, and E2-CEQ L-1.

Fig. S1 Torque-RPM diagram and weighting factors of the eight load-states of the ISO 8178/4 C1 cycle valid for construction site engines (RPM: rounds per minute). The holding time at each load-stage corresponds to the weighting factor in minutes. Emission factors for regulated pollutants and CO2 were calculated based on the weighting factors and are reported by Heeb et al. [2].

Fig. S2 All-glass sampling device used to collect samples of unfiltered diesel exhaust and samples of exhaust treated by a diesel particulate filter (DPF). The flow of the exhaust was ≤2 m3 h-1.

Table S1 Overall ER-mediated activity of the exhaust
ER-mediated activitya
Sampleb / ng E2-CEQ m-3 / ng E2-CEQ kWh-1 / ng E2-CEQ L-1
Ref / 1.63 ± 0.31 / 9.39 ± 1.78 / 33.7 ± 6.4
Fe / 1.31 ± 0.13 / 7.56 ± 0.78 / 27.1 ± 2.8
FeF / 0.74 ± 0.07 / 4.29 ± 0.40 / 15.4 ± 1.5
Cu / 1.05 ± 0.07 / 6.04 ± 0.41 / 21.7 ± 1.5
CuF / 0.55 ± 0.09 / 3.16 ± 0.50 / 11.3 ± 1.8
Cl / 1.62 ± 0.24 / 9.34 ± 1.38 / 33.5 ± 5.0
ClFe / 1.83 ± 0.17 / 10.53 ± 0.97 / 37.8 ± 3.5
ClFeF / 0.72 ± 0.09 / 4.16 ± 0.50 / 14.9 ± 1.8
ClCuF / 0.58 ± 0.04 / 3.32 ± 0.21 / 11.9 ± 0.8
xClCuF / 0.60 ± 0.02 / 3.46 ± 0.12 / 12.4 ± 0.4
Avg. of all tests without DPFs / 1.49 ± 0.31 / 8.57 ± 1.77 / 30.7 ± 6.3
Avg. of all tests with DPFsc / 0.64 ± 0.09 / 3.68 ± 0.51 / 13.2 ± 1.8
a ER-mediated activity of the exhaust was determined using the ER-CALUX assay (24 h of cell exposure). Data show the mean ± standard deviation of 3–4 independent exposure experiments. b Exhaust samples were generated using a heavy-duty diesel engine (Liebherr) and collected during two consecutive runs of the ISO 8178/4 C1 test cycle. c Two new, uncoated, cordierite-based, wall-flow diesel particulate filters (DPFs; Greentop) were used. Avg., average; E2-CEQ, 17β-estradiol CALUX equivalents; Ref, reference fuel; Fe, reference fuel with iron-based additive; Cu, reference fuel with copper/iron-based additive; Cl, xCl, chlorine-enriched reference fuel; F, exhaust treatment by DPF.

References

1. Heeb NV, Ulrich A, Emmenegger L (2005) Society of Automotive Engineers 2005-26-014:328-338

2. Heeb NV, Zennegg M, Gujer E, Honegger P, Zeyer K, Gfeller U, Wichser A, Kohler M, Schmid P, Emmenegger L, Ulrich A, Wenger D, Petermann J-L, Czerwinski J, Mosimann T, Kaspar M, Mayer A (2007) Environ Sci Technol 41:5789-5794

3. European Committee for Standardization, Brussels (1996) European Standard EN 1948-1

4. Wenger D, Gerecke AC, Heeb NV, Zenobi R (2006) Organohal Comp 68:1295-1298

5. Van Wouwe N, Windal I, Vanderperren H, Eppe G, Xhrouet C, Massart A-C, Debacker N, Sasse A, Baeyens W, De Pauw E, Sartor F, Van Oyen H, Goeyens L (2004) Talanta 63:1157-1167

6. Howell A, Osborne KC, Morris C, Wakeling AE (2000) American Cancer Society 817-825

7. Legler J, van den Brink CE, Brouwer A, Murk AJ, Van der Saag PT, Vethaak DA, Van der Burg B (1999) Toxicol Sci 48:55-66