Journal Archive

To conduct the DTT assay, most researchers employ the major methods developed by Cho et al. (2005) or Li et al. (2009), who simplified the methodology presented by Cho et al. (2005) (Lin and Yu, 2019; Fang et al., 2016; Kramer et al., 2016; Verma et al., 2014; Charrier and Anastasio, 2012; Lin and Yu, 2011; Li et al., 2009; Cho et al., 2005). There are several improved methods of DTT assay from Cho et al. (2005) or Li et al. (2009) (Lin and Yu, 2019; Fang et al., 2016; Kramer et al., 2016; Verma et al., 2014; Charrier and Anastasio, 2012; Lin and Yu, 2011). Thus, the DTT assay by Cho et al. (2005) or Li et al. (2009) can be considered the representative method.

The experimental DTT assays employed by Cho et al. (2005) or Li et al. (2009) are divided into three stages, as illustrated in Figs. 1 and 2 (Li et al., 2009; Cho et al., 2005). The stages consist of the collection and extraction of materials for the analysis, the development of a system for the DTT reaction, and the measurement of 2-nitro- 5-thiobenzoic (TNB) through the reaction of DTT solution with 5,5′-dithio-bis (2-nitrobenzoic acid) (DTNB). The extractions from collective filters conducted by Cho et al. (2005) or Li et al. (2009) were carried out using methanol or water; these utilized the ultrasonic extraction process. In addition, the researchers analyzed the extracted samples after measuring the concentrations according to the concentration of OP. In this stage, the solution used for extraction is important. Methanol or water was used in the experiments conducted by Cho et al. (2005) or Li et al. (2009), wherein DTT assay was organized by a single extraction. Thereafter, dichloromethane and simulated lung fluid (SLF) are added in the modified experiments to measure OP; in these experiments, water in the first stage and methanol in the second stage were used for the extraction (Patel et al., 2018; Pietrogrande et al., 2018; Calas et al., 2017; Gao et al., 2017). The selection of solvent and the processing of several stages for extraction are chosen based on the presence of hydrophilic or hydrophobic sub-fractions in the samples, and SLF can reflect the OP resulting from in vivo reactions (Patel et al., 2018; Pietrogrande et al., 2018; Calas et al., 2017; Gao et al., 2017). Relevant details will be discussed in Section 3.

In the second stage of the DTT assay, the DTT depletion is occurred, while the third stage involves measurement of the TNB created by the reaction with DTNB and residual DTT. The oxidation-reduction active quinone, 9,10-phenanthraquinone (9,10-PQ), delivers electrons to oxygen through DTT, thereby inducing the creation of superoxide (O₂^-) (Li et al., 2009). That is, an oxidation-reduction active compound like quinone catalyzes the reduction of O₂ to H₂O₂ by DTT, wherein the DTT is oxidized to DTT-disulfide (Li et al., 2009). Thereafter, in the third stage of DTT analysis, the thiol that remains after DTT reaction reacts with DTNB to create a mixture of bisulfide and TNB, wherein the TNB absorbs UV 412 nm (Cho et al., 2005). The reaction processes in the second and third stages of DTT analysis are similar to the oxidation-reduction of quinone that creates ROS in the human body (Fig. 3). Thus, the results of the DTT assay can be used to measure the biotoxicity of PM, wherein the loss ratio of DTT in the DTT assay is exploited as the measure of OP.

Fig. 1. 
Schematic illustration of experimental procedures used in the DTT assays by Cho et al. (2005) and Li et al. (2009).

Fig. 2. 
Chemical reactions in the DTT assay with PM (Cho et al., 2005).

Fig. 3. 
Redox cycling of quinones generating ROS in vivo and in the DTT assay (Li et al., 2009).

In the first stage of extraction for the OP analysis, the results vary according to the types of filters and solvents that are used (Table 1). As presented in Table 1, the collection of samples for the OP analysis in previous studies employed quartz, Teflon, and Teflon-coated glass-fiber filters; quartz filters were the dominant choice. However, when using a quartz filter, the amount of injected solvent needed for extraction is relatively high due to the higher rate at which moisture is absorbed by a quartz filter, which results in dilution of the extracted solution. This necessitates concentration of the extracted suspension for the OP analysis (Yang et al., 2014). According to the results of previous studies, quartz fiber filters can be damaged during the extraction process. As a result, OP may be underestimated. In a comparison between Teflon and quartz filters, quartz filters produced lower OP readings than Teflon filters (Yang et al., 2014). This is attributable to the lower efficiency of ultrasonic extraction of OP-related compounds by quartz vs. Teflon filters. The choice of a quartz or Teflon filter affects the absolute value of OP, but does not influence the tendency of the concentration gradient of OP. Thus, some researchers solely presented the trends in OP, or conducted an analysis of OP without removing PM in the suspension to quantify the absolute value of OP collected through the quartz filter (Patel and Rastogi, 2018). Therefore, a Teflon filter is frequently employed as a collection filter for the analysis of OP rather than a quartz filter to attain a higher extraction efficiency and to simplify the experimental procedure.

The solvents employed for the extraction of PM collected by filters are distinguished into three types. The first includes water, methanol, dichloromethane, etc. (Patel and Rastogi, 2018; Wang et al., 2018; Janssen et al., 2014; Yang et al., 2014; Eiguren-Fernandez et al., 2010). The efficiency of extraction of each solvent is determined by the polarity of the compounds in the collected PM (Patel and Rastogi, 2018). That is, the samples containing abundant water-soluble particles exhibit higher efficiency of extraction by water, whereas samples containing abundant non-water-soluble particles exhibit higher efficiency of extraction by dichloromethane. According to the results of previous studies, the OP readings when water was used in the extraction were higher than those when non-polar dichloromethane was used, while measurements taken using methanol were higher than those extracted with water (Patel and Rastogi, 2018; Yang et al., 2014; Eiguren-Fernandez et al., 2010). This indirectly suggests that OP-related compounds are more likely to be water-soluble materials than non-water-soluble materials. In addition, the characteristics of methanol extraction are similar to those of dichloromethane extraction, but the polarity of methanol is greater than that of dichloromethane, so some hydrophilic and hydrophobic compounds can be extracted by the former. Therefore, methanol is frequently employed as an extraction solvent to measure the maximum OP in a single extraction process (Calas et al., 2018).

Meanwhile, extraction with water, methanol, or dichloromethane is selectively used to examine the OP resulting from a specific component. That is, water and dichloromethane are applied to identical samples separately for the extraction so that the relative concentrations of either water-soluble compounds or non-water-soluble compounds can be measured (Eiguren-Fernandez et al., 2010). Furthermore, sequential extinctions using methanol and dichloromethane are exploited in stages to quantify OP after water extraction to calculate the absolute value of OP (Gao et al., 2017; Yang et al., 2014).

The second type of solvent used for filtered extraction is buffer solution (Romano et al., 2020; Pietrogrande et al., 2018; Charrier and Anastasio, 2012). The reason behind the employment of a buffer solution is to reduce blank oxidation. Blank oxidation is mainly triggered by trace metals included in the extraction solution; thus, trace metals should be removed from the extraction solution to avoid blank oxidation (Romano et al., 2020; Charrier and Anastasio, 2012). In previous studies, a buffer solution (pH 7.4) of phosphate and cation-exchange resin was used to remove trace metals. Pretreatment with extraction solution and employment of buffer solution reportedly do not lead to a higher rate of decrease of DTT in the blank solution, and increase the reproducibility of the experiments (Romano et al., 2020; Charrier and Anastasio, 2012).

The third type of solvent used for filtered extraction is SLF. The purpose of using SLF is to attain an accurate assessment of OP in the human body. To accurately measure the level of hazardousness to the human body, the bioaccessibility of compounds in PM needs to be taken into account. The bioaccessibility in the lungs can be represented by the degree of dissolution of PM compounds in SLF. The physiological conditions that induce reactions of PM in the lungs differ between water, methanol, and DCM, and SLF can be employed to simulate the pertinent physiological conditions (Calas et al., 2018).

When measuring OP using the DTT assay, the measurement is based on the decreasing rate of DTT. The chemical compounds in PM which exhibit strong correlations with a decreasing rate of DTT include elemental carbon, water-soluble organic carbon, and PAHs (Verma et al., 2009; Hu et al., 2008; Ntziachristos et al., 2007; Cho et al., 2005; Li et al., 2003). These materials do not directly interact with DTT in chemical reactions. For example, PAHs include compounds that are not involved in redox reactions, and are therefore regarded as irrelevant to direct loss of DTT. However, PAHs can indirectly decrease the concentration of DTT through chemical reactions with relationships such as quinone (Charrier and Anastasio, 2012). For example, phenanthrenequinone (PQN) and 1,4-naphthoquinone (1,4-NQN) oxidize DTT (Li et al., 2009; Kumagai et al., 2002).

On the other hand, the metals in PM may trigger in vivo oxidation damage, though metals are generally not thought to be active in the DTT assay (Chio et al., 2012; Leonard et al., 2004). In early research, an increase in OP was not observed with the addition of Fe and Cu to the DTT analysis (Charrier and Anastasio, 2012). However, the correlation between an increased concentration of metals and a decrease in DTT appeared thereafter, which was attributed to the association between metals and organic compounds (Hu et al., 2008; Ntziachristos et al., 2007). In addition, according to recent studies, the concentrations of heavy metals were found to remain constant even though that of DTT appeared to decrease by approximately 40-100% when the PM was heated. This indicates that semi-volatile organic compounds are responsible for the decrease in DTT (Biswas et al., 2009; Verma et al., 2009). However, Cu(II) and Zn(II) influenced the decrease in DTT, and the presence of metals increased the level of uncertainty in the DTT measurement due to the association between organic compounds (Gao et al., 2017).

Therefore, the most effective way to appraise the level of hazard to the human body is to measure OP by employing SLF, which is similar to compounds in the human body. Three main types of SLF were employed in previous studies. These are Gamble’s solution, artificial lysosomal fluid (ALF), and a mixture of Gamble’s solution and dipalmitoylphosphatidylcholine (DPPC). The compounds in Gamble’s solution and ALF are presented in Table 2 (Midander et al., 2007; Stopford et al., 2003). Gamble’s solution represents the interstitial fluid deep within the lungs. Lung macrophages are one of the major producers of ROS in the initial stages of inflammation (Calas et al., 2017; Marques et al., 2011). Macrophages use the antioxidant defense to control the discharge of ROS (Hamilton et al., 2008; May et al., 2005). ALF simulates the body fluid that is produced after phagocytosis by macrophages residing in the pulmonary alveoli and interstitial tissue; thus, it imitates the in vivo inflammatory state (Lacroze et al., 2009; Colombo et al., 2008). The mixture of Gamble’s solution and DPPC is a major phospholipid found in lung surfactant that reduces the surface tension at the air/water interface in the terminal respiratory tract and simulates the defensive function of the lungs (Lacroze et al., 2009; Colombo et al., 2008).

The correlations between DTT levels in the extractions using the three types of SLF and water, which were employed in previous studies, are shown in Table 3 (Calas et al., 2017). When comparing the extracted solutions resulting from the use of water and three kinds of SLF, the correlations between the concentrations of ionic compounds (NH₄⁺, SO₄²⁻ etc.) and quinone compounds (1,4-NQN etc.) all appear significant and similar to each other. However, with respect to the decreasing concentration of DTT, the extraction solution using SLF showed relatively higher relationships, in contrast to the lower correlations of water extraction for the trace metal compounds frequently represented by Mn, Cu, Zn, etc. DTT content decreased in the extraction solutions in the following order: Gamble’s solution, mixture of Gamble’s solution & DPPC, water, and ALF. That is, water extraction may result in a higher OP reading than ALF extraction, and water extraction may also result in an underestimation of OP compared to Gamble’s solution and Gamble’s solution & DPPC. OP measurements taken using each extraction solvent vary according to the particular chemical compounds in the samples. The available data on the effects of the complex of lung lining fluid and ligands on the measurement of air pollutants in PM is insufficient, so it would be unreasonable to conclude that SLF solution extraction is more accurate than water extraction (Calas et al., 2017). However, given the results of previous studies, we can assume that extraction solution containing SLF would produce better results than that containing plain water for the appraisal of hazardousness to the human body with respect to trace metals and PAHs.

The analysis of OP differs depending on the configuration of the DTT reaction system and the measurement tools. In short, the first-stage PM extracts catalyze the reduction of O₂ to H₂O₂ in oxidation-reduction compounds like quinine by DTT in the second stage, thereby oxidizing DTT to DTT-disulfide. In the third stage, the remaining DTT is reacted with DTNB to create TNB, enabling measurement using the light absorption characteristics of TNB. It would be difficult to strictly distinguish the second stage from the third stage because they constitute a continuous integrated process without preservation of separate samples.

The factors involved in the DTT reaction and measurement include the pH of the reaction, solution for the reaction, composition of the reaction stages, and TNB absorbance measurement. The pH affects the speed of the catalyzed oxidation-reduction reaction and the molar absorptivity of TNB upon completion of the reaction. As presented in Fig. 4, the absorbance increases as the pH increases from 3.8 to 5.0 in the reaction between DTT and DTNB, where oxidation-reduction reaction is absent. This phenomenon is attributable to the presence of mercaptide ion (TNB^2-), the absorbance of which is higher than that of the conjugate acid reacted with DTT and DTNB (Riddles et al., 1979). In addition, similar absorbance is seen in the pH interval from 3.8 to 5.0 when comparing reactions with DTT+DTNB and DTT+ Anthrone+DTNB, DTT+1,4-NQ+DTNB, and DTT+ 9,10-PQ+DTNB. This indicates that absorbance remains unchanged despite the presence of materials that trigger the reaction that decreases DTT, whereas the DTT reaction is almost absent at pH values below 5.0 (Li et al., 2009). In the pH range from 5.5-8.0, TNB appears in the form of TNB^2-, and no changes in the absolute values for the DTT+DTNB system are observed. However, the absorbance values of DTT+1,4-NQ+DTNB and DTT+9,10-PQ+DTNB decrease significantly. This indicates that the speed of the catalyzed oxidation-reduction reaction increases in accordance with the varying level of pH, while the DTT concentration decreases (Li et al., 2009). At pH values over 8.0, the absorbance in all systems increases rapidly due to alkaline decomposition of DTNB (Danehy et al., 1971). Thus, the range from pH 5.5 to 8.0 is superior for the DTT reaction. More specifically, 100% of DTT was converted into DTT-disulfide by 9,10-PQ at a pH of 7.3, while approximately 50% of 1,4- NQ was converted into DTT-disulfide. In addition, pH 7.3 is estimated to be optimal for the DTT reaction as it enables the simulation of in vivo reaction in the human body since it is almost identical to the pH of human blood (Li et al., 2009).

Fig. 4. 
Effect of pH on DTT reactions and TBN molar extinction coefficient (Li et al., 2009).

The selection of times for the DTT reaction depends on the specific experimental methods that are used; there is no single set of optimal experimental conditions in this regard. Cho et al. (2005) measured the rate at which DTT decreased according to various reaction times. The authors induced the DTT reaction for a predetermined period of time and then injected trichloracetic acid (TCA) to terminate the DTT reaction coercively. Thereafter, each sample was injected with DTNB to complete the DTT reaction at pre-set time intervals to measure the decreasing concentration of DTT corresponding to each reaction time (Eq. 1). OP measurements corresponding to each reaction time were obtained to finally calculate the rate of DTT decrease using the slope of a linear regression equation. On the other hand, Li et al. (2009) measured the final rate of DTT decrease of a sample by injecting DTNB after inducing the DTT reaction for a period of over 30 minutes. The method employed by Cho et al. (2005) has the advantage of reducing experimental errors by measuring the OP according to the time in culture, and the disadvantage of being more labor-intensive. However, the method employed by Li et al. (2009) may show greater experimental uncertainty, but it makes it possible to analyze many samples at the same time.

Here, A₀ represents the absorbance (blank) at t=0, while A and C_DTT represent the absorbance at a specified time t in culture and the initial concentration of DTT in the reaction solution, respectively.

The DTT assay to reduce errors triggered by a small amount of heavy metals has been used, as discussed previously. The solution that is added to remove a small amount of heavy metals in the reaction solution consists of a cation-exchange resin and buffer solution, as mentioned in the previous section. For the case of organic matter, previous studies reported that the concentration of organic matter and DTT decrease linearly, whereas that of DTT in a reaction with heavy metals varies according to the types of heavy metals (Lin and Yu, 2019; Fujitani et al., 2017; Charrier and Anastasio, 2012).

For example, as illustrated in Fig. 5, the decreasing concentration of DTT exhibits a linear relationship with the concentration of water-soluble compounds in PM, whereas the concentrations of Cu and Mn, which affect the DTT content, appear to have a non-linear relationship (Lin and Yu, 2019). Thus, it is additionally necessary to control the initial concentration of DTT in the DTT reaction in cases containing relatively high heavy metal concentrations that would affect the decrease in DTT in the samples (Lin and Yu, 2019). However, studies that definitively distinguish the decrease in DTT according to heavy metal content to determine the initial concentration of DTT have yet to be reported. Therefore, a small amount of heavy metals needs to be removed to reduce the uncertainty in the OP measurement. As explained in the previous section, heavy metals cause in vivo inflammation in the human body, and the relationship between a tiny amount of heavy metals and the decreasing rate of DTT appears as explained in the SLF extraction section. Thus, the limitations of each OP measurement method need to be clarified by expressing the processes of the OP assay, whether it be based on the use of an extractive solvent that removes heavy metals.

Fig. 5. 
DTT loss rates in the DTT20 and DTT100 assays by (a) water soluble compounds, (b) Cu, and (c) Mn (Lin and Yu, 2019).

The decreasing rate of DTT is calculated by measuring the TNB created through the DTT reaction. Absorption at 412 nm is frequently used to measure the TNB concentration. This is because TNB has a higher molar extinction coefficient in the visual range (Li et al., 2009). It should be remembered that, as illustrated in Fig. 6, the DTNB employed for the DTT reaction manifests its peak absorbance at approximately 315 nm, while it also exhibits a bit of absorbance in the neighborhood of 400 nm. Thus, the absorbance of DTNB needs to be removed from the TNB measurement results by taking the conditions of the measuring equipment into account.

Fig. 6. 
UV-VIS absorbance by DTNB and TNB (Li et al., 2009).