Discussion
It is well established that oxidative stress may cause profound alterations in biomolecules, including lipids, proteins and nucleic acids, and consequently may alter various cellular functions and structures (32,33). This explains why oxidative stress has been implicated in ageing and in a number of age-associated pathologies, including cancer, Alzheimer's disease, diabetes and cardiovascular diseases through genetic and/or epigenetic mechanisms (34). Concerning more particularly the role of oxidative stress in tumorigenesis, a general free radical theory was recently proposed linking oxidative stress to direct genetic toxicity and DNA mutagenesis, and indirectly to epigenetic alterations through free radical-induced protein epimutations (35). The present study reports for the first time, to the best of our knowledge, that ~80% of so-called EHS self-reporting patients present with oxidative stress, and thus may be considered to be bearing a truly objective pathological disorder, as is the case for cancer, Alzheimer's disease, or other diseases or pathological conditions. In the present study, the term EMFIS was preferred to EHS since, according to the clinical criteria used, it was not possible to clearly assess whether the patients exhibited a decreased tolerance threshold when exposed to EMFs. Furthermore, the term EMFIS was used preferentially to IEI-EMF, which has been proposed by the WHO, as all patients included in the present study clearly reported EMF-associated clinical symptoms (4).
During oxidative stress, among the reactive oxygen species (ROS) are the superoxide anion (O2°−), hydrogen peroxide (H2O2), the hydroxyl radical (OH°) and the hydroperoxyl radical (O2H°). Further, in redox cycling, transition metals including Fe, Cu, Ni and Co serve an important role in ROS formation (36).
Fe is the most commonly involved transition metal, and there are three classical reaction types. In the first step, namely the Haber-Weiss reaction, the superoxide anion reduces ferric ion into ferrous ion: Fe3+ + O2°− → Fe2+ + O2 [A]; while in a second step, namely the Fenton reaction, ferrous ion reacts with hydrogen peroxide to generate hydroxyl radicals and hydroxide ions: Fe2+ + H2O2 → Fe3+ + OH° + OH− [B].
Finally, in a third reaction, ferric ion is reduced to ferrous ion by reacting with a second hydrogen peroxide molecule, to recycle ferrous ion and form a hydroperoxyl radical and a proton: Fe3+ + H2O2 → Fe2+ + O2H° + H+ [C].
The net effect of [B] and [C] is the generation of the two ROS, OH° and O2H°, with H2O as a byproduct.
ROS are, however, difficult to measure directly due to their very short half-life. This explains why measurement of the products resulting from the molecular damage induced by ROS is the usual way to assess and measure oxidative stress. To that end, different oxidative stress and antioxidative response biomarkers were selected in the present study, representative of the different biochemical pathways and biological structural alterations that may occur when the organism is subjected to environmental stressors (37).
TBARs, which reflect the overall damage induced by non-enzymatic ROS-associated lipid peroxidation, among which MDA is the most prevalent byproduct (38), are commonly used biomarkers of lipoxidative stress (21). The reactive aldehyde MDA is a major indicator of the tissue damage resulting from the peroxidation of polyunsaturated fatty acids (PUFAs) induced by the two most prevalent ROS involved in lipoxidative stress: OH° and O2H° (39).
In fact, lipid peroxidation leads to the formation of numerous aldehydes, among which certain of them are highly reactive and may be considered as secondary messengers, which disseminate and amplify the initial oxidative stress. This is particularly the case for MDA, which is a bi-functional electrophile that is able to react strongly with nucleophiles, including amino acid residues in proteins (39). MDA adducts are thus biologically highly toxic, since they induce profound alterations in the structure and function of biomolecules by creating intramolecular or intermolecular protein/DNA cross-links (40,41).
This may explain why the majority of assays that have been developed to measure MDA on the basis of its derivatization with TBA have been challenged for their relative lack of specificity (42). This is due to the fact that TBA (in addition to MDA) is able to react spontaneously with numerous other molecules present in the test tube; by using a high temperature (90–100°C) to obtain the TBA/MDA spectrophotometrically-measurable condensation product, the process is able to generate in vitro further oxidation (43). In the present study, the method developed by Londero and Lo Greco (24) was used, which is considered to minimize the biases due to the procedure itself, and thus may increase specificity. Moreover, TBARs and MDA were measured simultaneously in the same sample and the values obtained for these two biomarkers were compared with the normal-range values obtained in healthy controls. Using this procedure, it was demonstrated that 40–50% of the patients had statistically significantly increased TBAR and MDA mean plasmatic values relative to normal values, a finding which strongly suggests that these patients present with an increased lipid peroxidation state detectable in their peripheral blood. In addition, these data were confirmed in the overall sample of 32 patients studied for TBARs, and tended to be significant for MDA.
Indeed these data may not be restricted to the peripheral blood, since cellular and nuclear membranes are primarily composed of fatty acids, including PUFAs. In the past 20 years MDA has been recognized as a reliable lipid peroxidation marker in a number of diseases, including cancer (44–47), type 2 diabetes (48), cardiovascular diseases (49,50) and Alzheimer's disease (51). On the basis of the present data, this is also the case for EMFIS, and this result is unsurprising since oxidative stress, including lipid peroxidation, has also been evidenced in similar recognized pathological conditions, including chronic fatigue syndrome (CFS) (52–56) and MCS (57).
However, as previously outlined, oxidative stress is an extremely complex redox cycling process resulting in various oxidizing/nitrosating free radical and molecular species attacks that exceed natural defense mechanisms; thus, it may not be measured by only one biomarker. In addition to TBARs and MDA, the present study measured GSH and, more specifically, GSSG and NTT as oxidative stress biomarkers. Glutathione is the primary compound that determines the redox state of a cell. It is a prototype antioxidant involved in cellular protection from the noxious effects of oxidative stress, directly and as cofactor of GPx. This thiol-containing tripeptide exists in an oxidized (GSSG) and reduced (GSH) form, and thus is a nucleophile and a reducing agent that is able to react with electrophilic and oxidizing species, allowing cells to escape the interaction of ROS with critical molecular targets, including proteins or nucleic acids (58). The ratio of GSH to the GSSG is a well-known marker of the redox state of a cell (59). Consequently GSH and GSSG were measured and their ratio (GSH/GSSG) and sum (GluT) were determined for the analysis of oxidative stress and the antioxidative stress response.
During oxidative stress, GSSG results from the oxidation of two GSH molecules by one hydrogen peroxide molecule, according to the following formula: 2GSH + H2O2 GSSG + 2H2O [D]; while the two GSH molecules are usually recycled from the reduction of GSSG, according to a reaction involving the coenzyme NADPH: GSSG + NADPH 2GSH + NADP+ [E].
It is notable that the oxidation of GSH according to [D] is catalyzed by GPx, while the reduction of GSSG according to [E] is catalyzed by GR; thus, according to [D] and [E], the activity of these two key enzymes in RBCs and plasma was measured.
In the present study, it was demonstrated that in comparison with normal-range values, the GSSG plasmatic mean value was statistically significantly increased in 40% of the patients, meaning that these patients presented with an oxidative redox state detectable in their peripheral blood. However, these data were not confirmed for the overall 32-patient sample for which the mean GSSG plasmatic mean value was not statistically significantly increased.
According to [D], it is assumed that increased GSSG may result in GSH depletion and consequently contribute to a decrease in antioxidant defenses. This may explain the result that in the overall patient sample studied, mean plasmatic values of GSH, the GSH/GSSG ratio and the GSH/GluT ratio were all statistically significantly decreased in comparison with normal control values; and that 20–40% of the patients with values of these biomarkers below the lower normal limits presented with statistically significantly decreased mean values in comparison with normal values, a finding which also confirmed that EHS self-reporting patients present with oxidative stress. Similar data were obtained in RBCs for GSH, although not for the GSSG/GluT ratio, in the De Luca et al (57) study, suggesting that the plasmatic measurement of GSSG, the GSH/GSSG ratio or the GSH/GluT ratio may be more informative compared with the measurement of the GSSG/GluT ratio in RBCs for the assessment of oxidative stress in EHS self-reporting patients.
Since the GSSG increase may be caused by an increase in GPx activity and/or by a decrease in GR activity, as indicated above, the present study measured the activity of these two key enzymes in RBCs and plasma. The overall mean GPx activity in the all 32 cases studied was not statistically significantly increased in RBCs and in plasma; with the exception of two cases, mean GR activity in RBCs and plasma was normal in all sample cases studied. However when considering the 18.75 and 9.28% of patients with increased GPx activity in RBCs and in plasma, respectively, a statistically significant difference was identified. Accordingly, it was suggested that the GSSG mean level increase in the peripheral blood may be associated with the increased GPx activity in ~19% of the patients, and/or by lower or normal GR activity; according to [E], the activity of this latter inducible enzyme is insufficient to recycle GSH from GSSG.
In the redox process GPx is an important enzyme as, by acting as a peroxynitrite reductase, it is able to efficiently reduce peroxynitrite/peroxynitrous acid (ONOO−/ONOOH) into nitrite (NO), thereby protecting cells against oxidative and nitrative reactions (60). The present data was not completely concordant with that obtained in the De Luca et al (57) study, which exhibited an overall statistically significant increase in GPx activity in RBCs in comparison with normal values. In addition, since GR activity was not measured in this latter study, it is not possible to confirm the accuracy of the present data demonstrating a normal level of GR activity in RBCs and plasma. In the present study, the evidence of oxidative stress in EHS self-reporting patients was, however, considerably reinforced by the evidence that, relative to normal values, the SOD1 mean value in RBCs was observed to be statistically significantly increased when considering the overall patient sample and the near 60% of the patients with mean values above the upper normal limit.
Such results were confirmed in patients with MCS in the De Luca et al (57) study, although not in patients with EHS; however, in these patients there was a strong tendency towards an increase in the SOD1 mean value in RBCs. The reason for such differences in comparison with the present data are not clear, and may be due to different inclusion criteria, since these criteria were not clearly detailed in the De Luca et al study, and/or to the use of a different dosage techniques.
SOD1 catalyzes the detoxification of the superoxide anion by dismuting this anion into hydrogen peroxide and molecular oxygen. O2°− + O2°− + 2H+ → H2O2 + O2 [F].
A further consideration to be made concerning the increased SOD1 activity in EMFIS-bearing patients is that according to [B] and [C], excessive production of H2O2 may provide excessive OH° and O2H° free radical and OH− ion production, and thus may amplify oxidative stress-induced detrimental health effects. Such a hypothesis is plausible since, in the present study, GPx activity was observed to be increased in only 10–18% of cases, thereby limiting its H2O2 detoxifying capacity. Another possibility for detoxifying H2O2 is catalase. However, catalase activity was not measured in the present study, although it was reported that in comparison with normal controls catalase activity tends to be decreased in EHS self-reporting patients (57), meaning that the H2O2 detoxifying capacity of catalase may be not sufficient in these patients.
It is possible, according to [D] and [F] by providing H2O2 in excess, that SOD1 may also indirectly contribute to the increased formation of GSSG, since as reported above it was demonstrated that GPx activity was normal or even increased in EHS self-reporting patients. In fact, whatever the resulting effects of the SOD1 increase in EHS self-reporting patients, it is notable that a similar increased level of SOD1 activity was reported in patients with Alzheimer's disease, with this increased level having been considered for the early diagnosis and therapeutic monitoring of this disease (61). This may also be the case for EMFIS-bearing patients.
During the oxidative process, peroxynitrite (ONOO−) may also be generated from the reaction of the superoxide anion with nitric oxide in the framework of an oxidative/nitrosative stress process, according to the following formula: O2−° + NO → ONOO− [G].
In this in vivo reaction, the radical coupling of NO with O2−° to form the non-free radical anion ONOO− is fast enough to outcompete the protective endogenous effect of SOD1. Although NO is regarded as a physiological cellular regulating agent, due to its rapid intra-tissue diffusion, it is also considered to be a crucial mediator of cellular damage occurring in different inflammation-associated pathological conditions, more particularly in neurodegenerative diseases, including Alzheimer's disease (62). Since NO is produced in large quantities in the brain, it is thought that it may serve a major contributing role in amplifying the peroxynitrite-induced toxicity in the central nervous system, thereby accounting for the fact that Alzheimer's disease is associated with peroxynitrite-associated oxidative stress. In fact, contrary to what was believed in the past, that the majority of oxidative stress-associated toxic effects may be attributed to NO, it is known that in vitro NO may inhibit lipid peroxidation (63); and it is now clearly established that due to the almost instantaneous formation of peroxynitrite each time NO and superoxide collide, peroxynitrite is the true toxic tissue damaging agent; peroxynitrite is a powerful oxidant that has been proven to cause pathogenic damage by interacting at a relatively slow rate and diffusion-limiting capacity with intracellular lipids, proteins and DNA (62).
A good example of such a selective reaction is the nitration of tyrosine residues in proteins and the formation of NTT, which thereby serves as a marker of peroxynitrite formation (64) and is a marker of oxidative/nitrosative stress (65).
The present study therefore included NTT in the battery of oxidative stress biomarkers used. Considering the overall sample of 60 EHS self-reporting patients investigated, wit was possible to define three categories of patients according to their NTT values: One-third of the patients exhibited values within the normal-range values, and another one-third exhibited values above the upper normal limits, while a further one-third of the patients presented with values below the lower normal limits. Notably, these findings were corroborated by the fact that in the latter two last categories, patients with abnormal mean values, these values were statistically significantly increased or decreased relative to normal values.
These data strongly suggested that one-third of the patients studied, those with statistically significantly increased NTT mean values, presented with detectable oxidative/nitrosative stress in the peripheral blood; overall, these data called into question why two-thirds of the patients presented with normal or statistically significantly decreased NTT values. The increase in NTT values in one-third of the patients may be easily explained by the increased peroxynitrite formation, and also by the fact that a decrease in GSH, usually an efficient scavenger of peroxynitrite (60), may result in a decrease in peroxynitrite detoxification. Furthermore, since SOD1 may also catalyze peroxynitrite-mediated tyrosine nitration (24), it may be hypothesized that the increased SOD1 activity that was identified in a number of patients may also contribute to the increased NTT detection in these patients.
Explaining the normal or decreased NTT values in two-thirds of the patients, however, is more problematic. A plausible hypothesis may be that according to [F], the increased SOD1 activity may strongly detoxify the organism of superoxide anions, thus decreasing the formation of peroxynitrite to such a level that the NTT may have been normalized or even decreased. If such hypothesis were to be validated, it may further confirm the existence of oxidative stress in these patients, since this detoxification process would involve increased SOD1 activity. A second hypothesis may be associated with the scavenging of peroxynitrite by GSH since, as reported above, peroxynitrite is able to directly oxidize low molecular weight thiols, including GSH; this hypothesis is in agreement with the decrease in GSH bioavailability that was observed in the patients. Finally, a third hypothesis may involve the inactivating effects of peroxynitrite on certain enzymes, by inducing the nitration of tyrosine and the oxidation of cysteine. In agreement with the present data, this may be the case for GR in a majority of patients, and for GPx and SOD1 in a number of them, and it may be the case for catalase, according to the data obtained by De Luca et al (57). Tyrosine nitration may indeed affect the structure and function of selective proteins (66), and consequently must be considered to be a central process of peroxynitrite-mediated toxicity. It is notable that tyrosine nitration and, more particularly, GSH depletion, in association with increased peroxynitrite toxicity, has been proposed to contribute to the occurrence and progression of a number of inflammation-associated diseases, particularly the neuro-degenerative diseases Parkinson's disease (67), Alzheimer's disease (68) and amyotrophic lateral sclerosis (69). A key molecular mechanism that may account for the occurrence of these pathological disorders may involve the activation of the proinflammatory transcription factor nuclear factor (NF)-κB by hydrogen peroxide (70) and/or peroxynitrite, possibly through the classical inhibitor of NF-κB kinase-dependent cell-type specific pathway (71). Further research in this field of molecular biology is required in order to elucidate the molecular causal role of oxidative stress in the onset of inflammation and, more particularly, inflammation-associated disease.
Whatever the precise molecular mechanism to be considered, the present data strongly suggest that EHS self-reporting patients, more precisely EMFIS-bearing patients, present with oxidative/nitrosative stress. This has been evidenced by measuring TBARs, MDA, GSSG/GSH and NTT in the plasma, and the inducible enzymes SOD1 in RBCs and GPx in RBCs and plasma. The search for a correlation between these different parameters confirmed the coherence of the present molecular dissection analysis.
A major finding of the present study was that by using a limited number of oxidative stress biomarkers, 70–80% of EHS self-reporting patients were able to be characterized by the existence of oxidative stress. Consequently, as is the case for numerous chronic pathological disorders, including cancer (44–47), diabetes (48), cardiovascular diseases (49,50), neurodegenerative diseases (51), and similar pathological syndromes including CFS (51–55) and MCS (57), the present data strongly suggested that EMFIS may be characterized by some degree of chronic inflammation (3,4) in addition to oxidative stress. This means that EMFIS (as for MCS and CFS) is a novel pathological disorder which merits recognition by the international biomedical community and classification as such by the WHO.
There remains no clear explanation as to the causal origin of oxidative stress in EHS self-reporting patients. The hypothesis that a nocebo effect may have been initially causal in the onset of oxidative stress is unlikely, since this is unable to explain the molecular abnormalities that were evident in the present study (4). The hypothesis that certain environmental stressors may be causally implicated in the onset of this pathology requires further investigation. Since it was previously demonstrated that MCS is frequently associated with EHS in EHS self-reporting patients (3), man-made chemicals may theoretically be these environmental stressors. However, in the present study, all patients with MCS that may have been associated with EHS were excluded, thus EMF exposure, as reported by the patients, may be an environmental stressor. This hypothesis merits consideration since numerous in vitro and animal experimental studies have reported that extremely low frequencies (ELF) radiation exposure (72,73), and more importantly radio-frequencies (RF) EMF exposure (74–79), are associated with oxidative stress occurrence, with the resulting biological effects including alterations in differentiation (72,73), inflammatory responses and DNA damage (77,80); all these detrimental effects occur more frequently in the brain (74,76–79).
Finally, it may be concluded that regardless of its causal origin, EMFIS may be biologically characterized as a novel pathological disorder, and thus may be diagnosed in medical practice on the basis of clinical symptoms, and more objectively by measuring: Inflammation-associated biomarkers, including histamine, protein S100B and the cellular stress chaperone proteins Hsp70 and Hsp27 (3); oxidative stress biomarkers, including TBARs, MDA, GSS and NTT in plasma; and anti-oxidative defense biomarkers, including SOD in RBCs, and GSH and GPx in plasma.