Introduction

Epidemiological studies support the notion that chronic exposure to organochlorine and related pesticides that are resistant to metabolism increases the risk factor for developing inflammatory cardiovascular and neurodegenerative diseases and cancer [22]. Although the actual mechanism responsible for the toxicity of organic pesticides is not completely understood, increased systemic oxidative stress triggered by elevated levels of reactive oxygen and nitrogen species (ROS/RNS) reportedly plays a key role [19]. The two major sources of ROS proposed to be responsible for the toxicity observed are mitochondria and NADPH oxidase enzymes (Nox) [9]. In this article, we focus mainly on Nox enzymes, especially Nox2, due to their role in the molecular mechanisms of pesticide toxicity [18].

ROS/RNS Cascade in Inflammatory Microenvironment

Figure 1 summarizes the cascade of oxidizing and nitrating species triggered by chlorinated-pesticide-induced generation of the superoxide radical anion (O2 ·–). O2 ·–, which is predominantly released upon Nox activation, can dismutate to form hydrogen peroxide (H2O2) or react with nitric oxide (NO) at a diffusion-controlled rate to form peroxynitrite (ONOO) [3, 28]. In the presence of bicarbonate (HCO3 ), which is ubiquitously present in cells, ONOO forms another transient intermediate, nitrosoperoxycarbonate (ONOOCO2 ), that decomposes to form nitrogen dioxide (NO2) and the carbonate radical anion (CO3 ·–), another highly potent one-electron oxidizing species. In addition, the formation of a chlorinating oxidant, hypochlorous acid (HOCl) from myeloperoxidase (MPO)/H2O2/chloride anion (Cl) oxidation, and hydroxyl radical (OH) from the redox-metal ion (e.g., reduced iron or copper) catalyzed reduction of H2O2 is also likely. MPO may also catalyze oxidation of the nitrite anion by H2O2 to produce nitrogen dioxide, NO2. Generation of other radicals derived from glutathione (GSH) such as the glutathiyl radical (GS) and their reaction with oxygen to form oxidizing radicals (GSOO) is also a distinct possibility. Because of this ROS/RNS cascade in the intracellular milieu, detection and assessment of the roles of different species using a single or several redox probes are nearly impossible without understanding their redox chemistry (reaction kinetics and product analyses) [47, 49]. Lack of progress in this area has so far stymied our understanding of Nox involvement in many areas of research, including pesticide toxicology.

Fig. 1
figure 1

ROS/RNS-induced radical cascade induced by superoxide and nitric oxide

Organochlorine Pesticides and NADPH Oxidase Activation

Organic chlorinated compounds (e.g., dieldrin, a metabolite of dichlorodiphenyltrichloroethane [DDT], dichlorodiphenyldichloroethylene [DDE], polychlorinated biphenyls [PCBs]) are mostly resistant to metabolism and biodegradation. Consequently, these chemicals tend to bioaccumulate in fatty tissues and release slowly with time, causing oxidative stress. Recent reports suggest that these chemicals activate Nox complex through activation of phospholipases A2/arachidonic acid (PLA2/AA) in monocyte/macrophages [18]. Monocytes treated with organochlorinated compounds enhanced Nox assembly and activation [18].

Unlike other redox-active enzymes in mitochondria and cytosolic compartments from which generation of ROS is an “accidental” byproduct of their primary catalytic function, the only known function of Nox enzymes is generation of O2 ·– and H2O2 [15, 24]. Several Nox isoforms including Nox2 form both O2 ·– and H2O2 (via dismutation of O2 ·–), with the exception of Nox4 that generates primarily H2O2 with little or no detectable O2 ·– [24, 30, 51]. High levels of O2 ·– generated from Nox2 are essential for bacterial cell killing and host defense, and low levels of ROS are chronically generated from Nox2 in response to stimulation (e.g., phorbol myristate ester [PMA]). PMA activates protein kinase C, leading to the phosphorylation of the p47phox cytosolic subunit, which in turn binds to the p22phox membrane protein [14]. After the assembly of all cytosolic and membrane components, NADPH is oxidized and electrons are transferred to oxygen, forming O2 ·–. Exogenously added compounds can activate or inhibit Nox expression, ligand receptor binding, trafficking of Nox components to cell membrane, activation and assembly of Nox complex, and/or affect NADPH binding, and electron transfer from the active site of the enzyme [2].

Recent studies have shown that organochlorine insecticides (trans-nonachlor, dieldrin, and DDE) induced enhanced expression of phospho-p47phox and enhanced its membrane localization [18]. Mechanistically, this was attributed to (PLA2) activation, leading to increased arachidonic acid and eicosanoid production in monocytes treated with organochlorinated compounds [18]. Chronic activation of monocytes by environmental toxicants could induce Nox activation through enhanced phosphorylation of p47phox mediated by protein kinase C activation and arachidonic acid release and subsequent translocation of p47phox to cell membranes [18]. Other xenobiotics such as dieldrin, lindane, paraquat, and rotenone activate microglial Nox, stimulating Nox-dependent ROS formation. Chemicals like 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) induce Nox2 expression and oxidative and nitrative stress in the substantia nigra of mice [8]. The oxidative metabolite 1-methyl-4-phenylpyridinium cation (MPP+) was shown to be the ultimate toxic metabolite. Activation of Nox2 and inducible nitric oxide synthase (iNOS) in glial cells is thought to be a major mechanism of toxicity. Reports suggest that organochlorine pesticides increase intracellular superoxide levels through activation of the PLA2/AA/Nox signaling, thereby posing a major risk factor for the onset of metabolic and cardiovascular diseases [18].

Probes for Superoxide Detection

More than a decade ago, we showed that hydroethidine (HE) or dihydroethidium (DHE) reacts with O2 ·– to form exclusively a highly diagnostic marker product, 2-hydroxyethidium (2-OH-E+) ([41, 42]; Fig. 2; Table 1). This finding negated the previous notion that ethidium (E+) is the product of oxidation of HE by O2 ·– [5]. We also showed that both 2-OH-E+ and E+ have very similar fluorescence characteristics, and thus, fluorescence microscopy is not a viable and reliable option to monitor intracellular O2 ·– formation [42, 45]. However, HPLC or ultra-high-performance liquid chromatography (UHPLC) and liquid chromatography-mass spectrometry (LC-MS/MS) approaches were used to separate and quantify 2-OH-E+ [45, 51]. Extensive research on the oxidation chemistry of HE revealed formation of both one- and two-electron oxidation products [46]. These include ethidium and several dimeric products that are all detectable by UHPLC [12, 51]. An additional benefit from HPLC (or LC-MS)-based detection and quantification of different HE oxidation products is the ability to monitor HOCl formation by following the formation of 2-chloroethidium (2-Cl-E+, Table 1), in case of H2O2/MPO/Cl system [10]. In contrast to other redox-sensitive fluorophores (e.g., dichlorodihydrofluorescein [DCFH], dihydrorhodamine [DHR]), and chemiluminescent probes (lucigenin, luminol, L-012)—which form radicals that react with oxygen to form superoxide—the HE-derived radical does not react with oxygen to form superoxide [47].

Fig. 2
figure 2

Chemical structures and oxidant-dependence of the products formed from hydroethidine (HE), MitoSOX Red (Mito-HE), hydropropidine (HPr+), 18F-labeled hydromethidine ([18F]-HMe), and tritium (3H)-labeled hydromethidine ([C3H3]-HMe) probes

Table 1 Structures of probes, marker products, and species detected

Another related cell-impermeable analog of HE is hydropropidine (HPr+), formed from a two-electron reduction of propidium (Pr++) ([20]; Fig. 2; Table 1). We showed that the oxidation chemistry of HPr+ is very similar to that of HE. Briefly, the HPr+/O2 ·– reaction formed 2-hydroxypropidium (2-OH-Pr++), and Pr++ is not formed in the HPr+/O2 ·– reaction. However, in the presence of other oxidants (e.g., ONOO, hydroxyl radical, or peroxidatic activity), other oxidation products (e.g., Pr++ and dimeric products, Pr++-Pr++, HPr+-HPr+, HPr+-Pr++) are formed (Fig. 2). HPLC or UHPLC and LC-MS/MS techniques were used to separate and identify these products [20]. Like the HE-derived radical, the HPr+-derived radical also does not reduce oxygen to superoxide. The HPr+ fluorescent probe is suitable for detecting extracellularly generated O2 ·– [20, 51].

Probes for Hydrogen Peroxide Detection

Boronates react with H2O2 stoichiometrically to form the corresponding hydroxyl derivative ([32, 49]; Fig. 3a). However, this reaction is very slow (rate constant of ca. 1 M−1s−1); therefore, in a cellular milieu, it is unlikely that a boronate probe is the preferred target for H2O2. In addition, other oxidants including ONOO and HOCl react with boronates at rates higher than does H2O2 (with the rate constants in the range of 103–106 M−1s−1, Fig. 3a), and therefore, boronate probes are not suitable for detecting H2O2 in cells or tissues under conditions generating ONOO or HOCl [32, 52, 53]. Under extracellular conditions, we monitored H2O2 formation using the coumarin boronate acid (CBA) probe (Fig. 3b; Table 1) with and without the catalase enzyme [51]. Peroxynitrite-mediated oxidation of boronates is not sensitive to catalase [32, 48, 49]. In addition, it will be necessary to rule out myeloperoxidase/H2O2/Cl-dependent oxidation of boronate via HOCl, as catalase could also inhibit this reaction.

Fig. 3
figure 3

Oxidation of boronate probes by hydrogen peroxide, hypochlorite, and peroxynitrite. a Comparison of the rate constants for different oxidants. be Examples of the boronate-based probes for in vitro and in vivo applications

The Amplex Red assay is widely used for H2O2 detection and quantification because of its high sensitivity. This assay is based on a horseradish peroxidase (HRP)/H2O2-dependent oxidation of Amplex Red probe to resorufin. Resorufin has a high extinction coefficient in the visible absorption region and can be conveniently monitored to measure extracellularly generated H2O2 (Table 1). Interference from photosensitized oxidation of Amplex Red by the analyzing (excitation) light, as well as other, H2O2-independent pathways of conversion of Amplex Red to resorufin, should be considered [23, 43]. Due to the requirement of HRP catalysis, the use of Amplex Red probe is limited to cell-free and extracellularly released H2O2.

Probes for Peroxynitrite Detection

Although over the last 5 years, an array of fluorogenic probes for the detection of ONOO has been reported (e.g., [16, 26, 36, 39, 44]), boronate-based probes seem to be best suited for that purpose. We have previously shown that boronate-based compounds react with peroxynitrite either added as a bolus or formed in situ from simultaneous generation of ·NO and O2 ·– nearly six orders of magnitude (106-fold) faster than H2O2 [32, 48]. The major product of this reaction is the corresponding hydroxyl derivative (phenol or alcohol), and the minor product (10 %) was from an intermediate radical [32, 33, 49, 52]. Indeed, the minor pathway led to nitrated benzene derivatives as the most diagnostic product [32, 33, 52]. Formation of minor products from the CBA probe (e.g., 7-nitrocoumarin, CNO2, Fig. 4a) allowed us to unequivocally demonstrate the formation of ONOO during the reaction of HNO with O2 [34]. Recently, we have characterized a new cyclic product, formed during the reaction with ortho-substituted boronate (o-MitoPhB(OH)2) via the phenyl radical addition to the phenyl ring of the triphenylphosphonium moiety (Fig. 4b, Table 1; [53]). By choosing the appropriate boronate, its reaction with ONOO can be monitored in real time using fluorescence or bioluminescence techniques [49, 50]. Figure 3b shows the reaction between ONOO and coumarin boronate. Whereas coumarin boronate CBA is not fluorescent, the product (7-hydroxycoumarin [COH]) is fluorescent, and ONOO can be monitored conveniently by monitoring COH by fluorescence in the presence of added catalase to exclude the minor reaction with H2O2 ([48, 50]; Table 1).

Fig. 4
figure 4

Chemical structures of the major and minor products formed during the reaction of ONOO with a CBA and b o-MitoPhB(OH)2 probes

By modifying the chemical structure, the fluorescence parameters can be altered. If the product absorbs in the red region, the applicability to in vivo situation is feasible. For in vivo applications, because of the poor tissue penetration of short-wavelength light, it is important to have boronates that yield products absorbing in the red or infrared regions [7, 40].

Mitochondrial Superoxide Detection: Problems with Mito-SOX

Recently, using a mitochondria-targeted hydroethidine probe (Mito-HE or Mito-SOX), superoxide generated in mitochondria was monitored [29]. In every aspect, the reaction chemistry between HE or Mito-SOX with superoxide and other oxidants is identical [46, 47]. For example, Mito-hydroethidine is oxidized by superoxide to form the characteristic 2-hydroxy-Mito-ethidium ([47]; Fig. 2). As with red fluorescence derived from HE oxidation, Mito-SOX/ROS-derived red fluorescence cannot be equated to superoxide detection and measurement [47], and it is essential to identify the product 2-hydroxy-Mito-ethidium (2-OH-Mito-E+) by HPLC or LC-MS before implicating superoxide involvement [12, 47]. These and other pitfalls of using a Mito-SOX probe to measure mitochondrial O2 ·– were elegantly described in a recent review [27]. Thus, nearly all of the studies that used Mito-SOX-red fluorescence as a measure of O2 ·– levels need to be repeated and reevaluated with rigorous methodologies using the LC-MS or HPLC techniques [25].

New Probes Developed in Other Laboratories: Assessment and Reinterpretation of Results

In Vivo Detection of Superoxide: Radiolabeled Probes

Recently, alternate sensitive and noninvasive approaches (e.g., positron emission tomography [PET] and radionucleotide imaging) for detecting superoxide, suitable for in vivo conditions, were developed [1, 6, 35]. These approaches are based on the intracellular trapping of the oxidation products of HE and its analogs. In PET detection, an 18F-labeled HE analog (Fig. 2) ([18F]-HMe) was used as a PET tracer. This is an interesting imaging modality that is easily translatable to humans, as there exist numerous PET probes (e.g., fluorodeoxyglucose [FDG]) that are currently being used in the clinic to track glycolytic metabolism in humans.

Although this imaging modality is very sensitive, the chemistry between an 18F-labeled HMe probe and ROS is the same as that of the unlabeled DHE and ROS (Fig. 2). Superoxide oxidizes [18F]-HMe to [18F]-2-OH-Me+ (Fig. 2). Other one-electron oxidants will oxidize this probe to the corresponding 18F-labeled ethidium analog and 18F-labeled dimeric oxidation products (Fig. 2). All products having the 18F tracer and trapped intracellularly will be imaged, and there is no way to distinguish between the superoxide-derived product and other nonspecific one-electron oxidation products. In essence, all of the limitations that we have previously described for fluorescence-based imaging are applicable to PET imaging as well [12, 47]. In addition, as with the unlabeled DHE that undergoes oxidation in the presence of heme (or hemoglobin), this PET tracer is also subject to nonspecific heme-catalyzed oxidation [47]. Any claims for noninvasive imaging of superoxide using this probe (in vivo or in vitro) should be reexamined. However, this probe may be used to investigate oxidative stress or oxidants formed in diseased and normal brains using PET imaging because of the likelihood of the 18F-labeled analog of DHE, and not the positively charged ethidium analog, crossing the blood–brain barrier [6].

Another radiolabeled HE analog (3H-hydromethidine, [C3H3]-HMe) (Fig. 2) containing the radiotracer tritium was recently developed to probe oxyradical formation in the brain [35]. Again, the radical chemistry of 3H-hydromethidine should be very similar to that of HE. Superoxide oxidizes [C3H3]-HMe to [C3H3]-2-OH-Me+ and nonspecific one-electron oxidation products include [C3H3]-Me+ and 3H-labeled dimers of hydromethidine (Fig. 2). A claim that O2 ·– reacts with 3H-hydromethidine to form 3H-methidium rather than a hydroxylated cation has not been substantiated. The authors cite a previous publication by Hall et al. [11] wherein O2 ·– reacts with hydroethidine under in vivo conditions to form ethidium and not 2-hydroxyethidium. That O2 ·– reacts with hydroethidine to form ethidium under low oxygen tension (but not at normal oxygen tension) was recently challenged by us [21]. We showed that irrespective of the superoxide flux, the major product of HE/O2 ·– reaction is 2-hydroxyethidium and not ethidium [21]. Simply measuring the extent of radioactivity in tissues is not sufficient for determining the identity of ROS; the products must be separated using the HPLC-radiolabeled detection method and the retention time compared with that of the appropriate standard. Despite the fact that the use of 3H-hydromethidine radiotracer is unlikely to yield definite information regarding the nature of an oxidant(s) formed in tissues or cells, the ability of the parent tracer to cross the blood–brain barrier is an advantage. The contribution of oxidative stress in the brain under pathological conditions can be qualitatively assessed.

In Vivo Targeting of Hydrogen Peroxide: Cell-Penetrating Peptides

Recently, in vivo detection of hydrogen peroxide was reported using a newly developed probe consisting of a polycationic cell-penetrating peptide and a polyanionic fragment connected through a boronate linker [37]. Fluorescent labeling of both of its peptide domains resulted in the fluorescence resonance energy transfer (FRET) signal (Fig. 3c). Reaction with H2O2 caused a disruption of FRET which was used to measure H2O2. Using the 40-fold ratio change in FRET, H2O2 generated by activated macrophages and neutrophils in a lipopolysaccharide (LPS) mouse model of inflammation was monitored [37]. However, several caveats with the use of this probe were not discussed in that study [37]. Boronates react very slowly with H2O2; in an intracellular milieu, this reaction probability is very low. We reported that ONOO reacts with boronates at least a million times faster than with H2O2 [32]. As discussed earlier, reports indicate that ONOO is generated during LPS treatment. Thus, additional experiments with NOS and/or Nox and MPO inhibitors (to rule out contribution from HOCl) are necessary for proper interpretation of the data reported in this study [37] as well as in another study using a lysosome-targeted boronate-based probe [13].

Bioluminescence and PET Imaging of ROS In Vivo

One of the most convenient modes of in vivo animal imaging is based on bioluminescence. Thus, a new probe has been synthesized, peroxy-caged luciferin-1 (PCL-1), which upon reaction with ROS/RNS forms luciferin in situ that is rapidly oxidized in luciferase-transfected cells generating green bioluminescence [31, 38]. This reaction uses adenosine triphosphate (ATP) as a cofactor (Fig. 3d).

A PET probe for detecting H2O2 was recently developed [4]. This ROS-specific PET agent is a thymidine analog, peroxy-caged-[18F]fluorodeoxy thymidine [18F]PC-FLT that is transported rapidly into cells via the nucleoside transporter. Unlike thymidine, [18F]PC-FLT is not phosphorylated by thymidine kinase and does not accumulate inside the cell [4]. Upon reaction with H2O2, PC-FLT-1 generated [18F]FLT in situ (Fig. 3e) that is phosphorylated, trapped intracellularly, and imaged by PET.

As discussed for other boronate probes [31, 32, 49], the PCL-1 and PC-FLT probes react with H2O2 rather slowly to be considered as effective H2O2 detectors in cells. Under conditions where ONOO and/or HOCl are generated, these probes will undoubtedly react with these species as opposed to H2O2.

Conclusion and Future Perspectives

With the advent of new and sensitive probes with relatively lower toxicity and better spatial resolution developed primarily in Chang’s laboratory [17], we are in a position to perform relevant preclinical imaging that can be translated to the clinical setting [4]. [18F]FLT is used in the clinic, and boronates have been administered to cancer patients for many years. Thus, it is conceivable that the peroxy-caged probe, [18F]PC-FLT, containing the boronate moiety will be tested in the clinic for imaging RNS. Equally significant and promising are the boronate-based bioluminescence probes. For example, the newly synthesized peroxy-caged luciferin-1 (PCL-1), upon reaction with ROS/RNS, forms luciferin in situ that is rapidly oxidized in luciferase-transfected cells generating green bioluminescence [31, 38]. This reaction uses ATP as a cofactor. As with other boronates, PCL-1 reacts with ONOO nearly a million times faster than with H2O2 and thus could be used to image ONOO formation in an inflammatory microenvironment in various toxicology models. With the proper experimental setup, we should be able to monitor the effects of pesticides on multiple cellular parameters including H2O2/ONOO generation and, for example, cellular bioenergetic status (ATP level), with a single detection modality (e.g., bioluminescence), as exemplified in Fig. 5.

Fig. 5
figure 5

Proposed approach for luciferin-based bioluminescence imaging of ROS and bioenergetic status in vivo