An Isuzu 4JA1 pre-Euro I, 2,5 L, direct injection, four-cylinder diesel engine was used, whose characteristics, operating conditions, nano-structural and morphological properties of the PM are discussed in [26,50]. The engine was started on clean low-sulfur fossil diesel, it was kept on stationary conditions until reaching 2410 rpm and the PM was collected at 43 and 95 Nm of torque. When these conditions were reached, 10% of the energy fraction of diesel was substituted with ethanol injected through a multipoint injection system (dual-injected ethanol/diesel fuel). In order to eliminate any residue after running the engine with the dual-injected ethanol/diesel fuel, the combustion system was drained for 15 min before being run again with clean low-sulfur fossil diesel.

Four undiluted PM samples were filtered and collected with stainless steel fibers positioned at 1,5 m after the exhaust manifold pipe. Gas temperature was kept below 200°C to avoid oxidative reactions in the PM, and the PM recovered by scraping the stainless-steel fibers was stored in sterile amber glass containers.

The collected PM was sonicated with dichloromethane (DCM) as a solvent [28,51-62].

Fifty milligrams of PM from each fuel was mixed with DCM in a 1:2 ratio, the mix was sonicated at 40Hz for one hour at room temperature [63]. Subsequently, it was centrifuged for 10 minutes at 2.000 rpm, the extracted suspension was collected, and the process was repeated three more times using the precipitate. The supernatant was filtered through Nylon membranes (0,22 µm), followed by evaporation under a gentle stream of nitrogen gas and resuspended in dimethyl sulfoxide (DMSO). Because we were not able to weigh the organic extract of each fuel, doses were expressed in amounts (µg) starting material of PM of each fuel, so, the maximum concentration of each extract was of 0,16 µg PM Equivalents/µL. The expression of concentration in this way has been used by Claxton when referring to NIST Standard Reference Material (Diesel SMR 1650) [64] and by Umbuzeiro et al for fractions of air particulate material [65].

Clean stainless-steel fibers with the same characteristics as those used to collect the PM were subjected to the previous organic extraction process to determine their influence on the toxicity of the PM extracts.

The mutagenic activity of the organic extracts was evaluated with the Salmonella/microsome test [36], the Kado micro-suspension method [38,66]. TA98 and TA100 Salmonella strains auxotrophs for histidine (his-) and sensitive to detect most environmental mutagens were used [42]. The TA98 strain with a frameshift (hisD3052, Δ(uvrB, bio), rfa, Ap^r (ampicillin resistance) (pKM101) and the TA100 strain with base substitution (hisG46, A(uvrB, bio), rfa, Ap^r (ampicillin resistance), (pKM101), both strain in the absence of (-S9) and the presence of (+S9) metabolic fraction from a male rat's liver (induction with Aroclor 1254, Moltox S.A., USA). For each strain, nine concentrations from each extract were used per assay: 0,002; 0,005; 0,01; 0,02; 0,04; 0,07; 0,15; 0,3 and 0,6 µg equiv, of PM/µL on strain TA100 and 0,0002; 0,0005; 0,001; 0,002; 0,005; 0,01; 0,02, 0,04, 0,07, 0,15 µg equiv, of PM/µL on strain TA98. Negative controls were dMsO and the extract of clean stainless-steel fibers, and positive controls were Daunomycin (6 µg/plate) for strain TA98 -S9, Sodium Azide (1,5 µg/plate) for strain TA100 -S9 and 2-Aminoanthracene (5 µg/plate) for both strains, TA98 +S9 and TA100 +S9. The results were expressed as average revertants/µl and potential mutagenic activity was considered when the treatments showed statistically significant differences with the negative control and when the concentration-response relationship was significative (p<0,05) in any of the used strains.

The liver epithelial cell line HepG2 ATCC® (CRL -10741 LOT: 61777384), competent for the expression of a great variety of cytochromes P450 and widely used for genotoxicity assays of complex mix [67] was cultivated in Dulbecco's Modified Eagle Medium (DMEM), supplemented with fetal bovine serum (FBS) at 10%, 50 IU/ml of penicillin, 50 ug/ml of streptomycin. Cells were incubated at 37°C in a humid atmosphere and 5% of CO2. Cells were used when reached a cell confluence of 80% and changed every 20 passages.

Cell viability in the HepG2 line was evaluated through the MTT assay by reduction of the compound 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazole bromide. The methodology suggested by [68] was followed, 5.000 cells/well were seeded in 96-well dishes and were exposed in darkness to 20 µl of each extract of organic PM, then, were incubated at 37°C in humid atmosphere and 5% of CO2 during 24 hours. Twenty 1:1 serial dilution was evaluated, ranging from 0,16 to 0,0005 µg equiv of PM/µl.

Later, 10 µl of MTT were added to the cells, placed in an orbital shaker at 37°C for 5 h and then, 100 µl of cold isopropanol were added to each well to dilute the formazan crystals. The culture plates were placed in an orbital shaker platform for 12-24 hours and absorbance was measure using a Multiskan-GO spectrophotometer (Thermo Scientific) at 570 nm.

Viability (%) was calculated according to the equation:

Where: OD TnT is the optical density of the treated cells and OD neg.con is the optical density of untreated cells or negative control (clean stainless-steel fibers extract).

Genotoxic activity of the organic PM extracts from fuels was determined by alkaline version of the comet assay modified by McNamee [69]. Thirty-five thousand HepG2 cells were seeded in 12-well dishes, cells were treated with 10 µl concentrations of 0,0007, 0,0013; 0,026; 0,005; 0,01; 0,02; 0,042, 0,08 and 0,16 µg equivalent of PM/µl for 24 hours at 37°C in a humid atmosphere and 5% of CO2. Afterwards, cells were embebed in low melting point agarose and placed on gelbond films, then lysed and subjected to alkaline electrophoresis (pH > 13). Finally, the cells were stained with GelRed^TM (Invitrogen Corp; CA, USA), observed and measured using a fluorescent microscope Nikon® Eclipse 55i (Tokyo, Japan) with a 20X objective. The length of the comet was randomly measured in 100 cells with an eyepiece micrometer (um) [70-72], the fraction of damaged cells (%DC) and the cut-off point were from the equation:

Where: D is the damaged DNA, avg. len. nc is the average length of DNA migration from the negative control or untreated cells (clean stainless-steel fibers extract) and SD is the standard deviation of the mean.

Three independent assays were performed, which included untreated cells, cells subjected to the clean stainless-steel fiber extract as negative controls, and 100 µM of H2O2 as positive control. A genotoxic effect was considered when the treatments exhibited statistically significant differences with the negative control and the concentration-response relationship was significant (p<0,05).

For gene expression analysis, HepG2 cells were grown as described above and exposed for 24 hours to 0,16 µg equivalents of PM/µl of the organic extract obtained from each fuel, fossil diesel and dual-injected ethanol/diesel. The experiment involved three technical replicates per organic extract, including negative controls (HepG2 cells exposed to organic extract from clean stainless-steel fibers). Total RNA extraction was performed to 1x 10⁶ cells from each group with Trizol (Invitrogen) following the manufacturer's recommendations. Quantity and quality of the RNA was determined with an UV light spectrophotometer on a NanoDrop^TM2000 (ThermoFisher Scientific) at an absorbance of 260 nm. The RIN (RNA Integrity Number) was determined by utilizing the Agilent RNA 6000 Pico kit (Agilent) and a Bioanalyzer (Agilent). The sequencing libraries were prepared with the TruSeq RNA library Prep kit (Ilumina INC, CA, USA) using poli-T beads for mRNA enrichment. Finally, paired-end libraries (2 x 150 nt) were sequenced in a HiSeq 4000 plataform (Ilumina Inc., CA, USA).

The raw reads underwent a cleaning process to completely remove reads or low-quality ends (Phred score <30), for which the TRIMMOMATIC software version 0.39 was implemented [73]. Subsequently, the clean reads were mapped to the reference human genome (GRCh38) with the STAR software version 2.6.0 [74] with all the default parameters. The reads count associated with each gene were generated with the HTSEQ-COUNT software version 0.11.1 utilizing all the default parameters [75].

The differential gene expression analysis was performed with the EDGER package version 3 [76]. Differentially expressed genes (DEGs) were identified by establishing a cut-off point of log2 FC (Fold Change) of 1,5 and a P value <0,05. The REACTOME database version 73 was implemented to group DEGs according biological processes [77] and enriched pathways reported by the database were selected considering a P value <0,05.

Data analysis was performed with the statistical package IBM SPSS Statistics version 23.0 IBM Corp. Mutagenicity and genotoxicity were considered positive when the ANOVA and the concentration-response curve were significant (p<0,05).

The organic extracts of PM obtained through both operation modes, 43 and 95 Nm of torque, from fossil diesel and from dual-injected ethanol/diesel were mutagenic (Fig. 1), especially in TA98 strain in absence of metabolic enzymes. Extract from dual-injected ethanol/diesel was more mutagenic than fossil diesel extract. Organic extracts of PM obtained at 43Nm were significantly more mutagenic than those obtained at 95Nm of torque (p<0,05).

Figure 1 Results represent the average of three independent experiments. Concentration - response significance obtained with the ANOVA was p<0,001. Source: authors.

Figure 2 Source: authors.

After exposing HepG2 cells to 20 concentrations of organic extracts from PM for 24 hours, there was no evidence of loss of mitochondrial functionality. Therefore, the extracts were considered non-toxic.

Independent of the fuel or the operation mode of the engine, all the extracts were genotoxic. The concentration at which most of the organic extracts caused damage to 50% of the cells oscillated from 0,02 to 0,04 µg equivalent of PM/µl. One-hundred percent of the cells exhibited DNA damage from 0,08 µg equivalent of PM/µl with all the extracts. Regarding the operational mode, a higher genotoxic effect was evident at 43Nm of torque (Fig. 2).

Considering the results of the comet assay, the 0.16 µg equivalent of PM/µl concentration of the organic extracts obtained at an engine load of 43Nm was selected to perform the gene expression analysis. After normalizing the data, the average count of reads was 17 with approximately 10.000 genes detected with at least 5 reads for each of the conditions analyzed. Fifty percent of the detected genes had a count of 39 reads. Differentially expressed genes were identified when the expression profiles of HepG2 cells exposed to organic extracts from both fuels, fossil diesel and dual-injected ethanol/diesel, were compared to the expression profile of HepG2 cells exposed to extracts from clean stainless-steel fibers.

Fig.3 shows the distribution of DEGs of each organic extract from the fuels compared to HepG2 cells exposed to clean stainless-steel fibers extract (negative control). Both treatments displayed a higher proportion of upregulated genes in contrast to downregulate genes. Organic extract of PM from fossil diesel affected fewer genes than the organic extract of PM from dual-injected ethanol/diesel. Non-coding RNA were detected, although the methodology involved a mRNA enrichment based on polyT beads, those transcripts correspond to pseudogenes, micro-RNAs, long non-coding RNAs (LncRNAs), and antisense genes.

Figure 3 Source: authors.

Figure 4 Source: authors.

An enrichment analysis of pathways in the Reactome database identified the most affected biological processes. In their respective order, the immune system, transduction signals and lipid-carbohydrate-protein metabolism (Fig. 4). The enrichment results evidence that the organic extracts modulated the expression of various pleiotropic genes, such as the CASTPERG (Cation channel sperm associated auxiliary subunit gamma) and CXCL8 gene (C-X-C motif chemokine ligand 8) for the diesel organic extract, and the FOS genes (Fos proto-oncogene, AP1 transcription factor) or CSF1R (Colony stimulating factor 1 receptor) for the dual-injected ethanol/diesel extract. It was also observed that several genes might affect a single pathway, for example, for the immune system, the interleukin signals, cytokines, and the JAK-STAT pathway through cytokines (Table 1).

Table 1 ^a

p<0,05

Among other enriched pathways, those related to the embryogenesis and fertility were found in cells treated with the extract from diesel, while the dual-injected ethanol/diesel extract affected carbohydrate metabolism pathways. It is important to highlight that some modulated genes from both PM organic extracts take part and link the complex network of the DNA damage repair (DDR) system with the immune system pathways through different mechanisms (Table 2).

Table 2

The arrows ↑, ↓ indicate gene regulation, up-regulated genes (Up) or down-regulated genes (Down), respectively more oxygenated aliphatic groups (alkenes, methylenes, and methyls) in the surface of the PM than fossil diesel, mainly when the engine operated at 43Nm of torque. Hydrocarbons, especially PAH, could be responsible for the mutagenicity with +S9 [81], while the aliphatic groups could cause mutagenicity without metabolic activation (-S9) [34]. Additionally, the increase of NOx in the combustion of diesel spread with alcohols could favor the formation of modified PAH as nitro-PAH, whose reactivity is greater in the TA98 strain without S9 [1,82]. Salmonella strains with the sensibility to specific chemical groups should be tested to identify the chemical groups in the extracts responsible for the mutagenicity, such as YG1041 (sensitive to nitroaromatics without S9 and amino aromatics with S9), YG5185 (sensitive to unmodified PAH), TA104 (sensitive to free radicals) or YG7108 (highly sensitive to alkylating agents), among others. Furthermore, the PM organic extract could be chemically characterized according to the differential mutagenic response of the strains.