Effects of transcranial direct current stimulation (tDCS) on pro-inflammatory cytokines: a systematic review

Transcranial direct current stimulation (tDCS) is a neuromodulation technique that causes alterations in the synthesis of several proteins, including cytokines (e.g., Interleukins). Pro-inflammatory cytokines are associated with the presence of pain and their reduction occurs in several pathologies. The aim of this study was to investigate the effects of tDCS on the variation of tissue and serum blood levels of pro-inflammatory cytokines and its relationship with behavioral changes, through a systematic review. PubMed, Embase and Lilacs database searches were performed for articles published in all languages before October 1, 2020. The search was based on the keywords "Transcranial Direct Current Stimulation" or "tDCS" and "IL-1alpha " or "IL-1Beta" or "IL-6" or "IL-8" or "IL-17" or "Tumoral necrosis factor alpha" or "TNF-alpha". The systematic review protocol was registered in PROSPERO (CRD42021283417). Initially, 416 studies were identified in the electronic databases, of which 40 were eliminated because they were duplicates. Of the remaining 376, 358 were excluded after analyzing the title and abstract (selection stage) and 09 were excluded after a complete reading. Nine studies were considered for evaluation. The results demonstrate that tDCS can alter the levels of pro-inflammatory cytokines and modify behaviors in animals, however these findings are variable. Still, the cause and effect relationship between cytokine levels Five studies found a reduction in IL-1  levels in the cerebral cortex (De Oliveira et al., 2019), cerebral cortex, spinal cord, and brainstem (Cioato et al., 2016; Lopes et al., 2020); or hippocampus (Guo et al., 2020; Regner et al., 2020). However, one study  levels stable (Spezia Adachi et al., 2012). Another article brainstem et For one and in there was IL-1 α increased and after application tDCS et al., 2020).


Introduction
Transcranial Direct Current Stimulation (tDCS) applies electrical energy to the cerebral cortex through electrodes over different brain regions. This neurostimulation technique uses sustained direct current , being a noninvasive alternative, and easy to apply. In addition, this treatment involves modulation of brain areas associated with nociceptive processing (Fregni et al., 2006b;Liebetanz et al., 2009). Therefore, tDCS is a promising tool for treating pathologies that involve the central nervous system (CNS) (e.g., chronic pain). TDCS can use anodic or cathodic electrical currents depending on the treatment required. The anodal current causes a depolarization of the neuronal membrane, increasing the excitability of neurons (Nitsche et al., 2008). When using cathodal current, there is hyperpolarization of the cell membrane, inhibiting neuronal activity (Nitsche & Paulus, 2000;Nitsche et al., 2008). Due to the action of tDCS altering cortical excitability, the regions of interest vary according to the pathology/dysfunction. For example, patients with painful disorders can receive treatment with an electrical current directed to the prefrontal cortex or the primary motor cortex (M1) Lopes et al., 2020), cortical areas involved in pain processing and perception (Quevedo & Coghill, 2007). In addition to the direct effects on brain tissues, remote areas of the CNS can be affected by tDCS through descending pathways (i.e., top-down effect), which may contribute to pain control (Lefaucheur, 2006).
Each family member of interleukins may play different roles in the inflammatory cascade. For example, IL-1 ( and ß) induces the inflammatory response related to the acute phase of infections (Dinarello, 2011). IL-6 is considered a mixed cytokine, as it has pro-inflammatory (i.e., stimulating the immune response) (Lin et al, 2000) and anti-inflammatory effects (Oliveira, 2011). IL-8 is one of the primary mediators of the immune response to intracellular microorganisms. Furthermore, IL-8 is strongly associated with neutrophil chemotaxis and the activation of polymorphonuclear neutrophils (Baggiolini & Clark-Lewis, 1992;Curfs et al, 1997). On the other hand, IL-17 is associated with extracellular bacterial infections (Jaiswal, 2014).
The effectiveness of tDCS in pain management has been documented using different protocols (Nitsche et al., 2008).
However, the ideal parameters for optimizing the use of tDCS (i.e., cortical areas, treatment time, and current intensity) are the subject of investigations (Knotkova et al., 2019). Furthermore, the pathophysiologies of several diseases, such as chronic pain, are not fully understood. However, it has been characterized that blocking pro-inflammatory cytokines in the CNS can effectively control pain (Wieseler-Frank et al.;2005). In addition, the use of tDCS is effective in reversing behaviors related to pathological processes (e.g., pain, epilepsy, anxiety) (Cioato & Torres, 2014;De Oliveira et al., 2019;Regner et al., 2020;Santos et al., 2020) and improvement in cognitive performance (Guo et al., 2020). Therefore, through a systematic review, the present study aimed to investigate the effects of tDCS on pro-inflammatory cytokine levels in the CNS and serum and behavioral changes.

Methodology
The review protocol was registered with PROSPERO (registration number CRD42021283417) and followed PRISMA guidelines (Page et al., 2021)

Literature search
This systematic review searched three databases: Embase, PubMed (Public/Publisher MEDLINE), and Lilacs (Latin American and Caribbean Literature in Health Sciences). Embase was chosen because it is the most complete database. PubMed is the main reference database in the health area. Likewise, Lilacs is also widely known in this area. All searches were based on the same criteria. The strategy employed included the following keywords ("Mesh Terms") with the Booleans arranged as follows: "tDCS" or "Transcranial Direct Current Stimulation" and "IL-1alfa" or "IL-1Beta" or "IL-6" or "IL-8" or "IL-17" or "Tumor necrosis factor-alpha" or "TNF-alpha". These keywords and booleans must be contained in the title or abstract of the articles. The search found 43 studies on the PubMed platform, 373 on Embase, and 0 (zero) on Lilacs.

Study selection
The collection and analysis of the articles took place from 10/01/2020 to 12/10/2020. The search and inclusion of studies in this review were performed by two researchers independently (LEG and INA), and no discrepancies were found between them.
Initially, the duplication of articles was verified through the management of the Microsoft Office Access 2013 database, where forty duplicate articles were observed. Afterward, the selected articles were evaluated separately by the two evaluators using the title and abstract according to the eligibility criteria, excluding articles irrelevant to the objective of the present review. Next, the selected articles were read in full by the two evaluators individually for deliberation on their inclusion. Then, the researchers extracted the data separately, and after collection, a new consensus meeting was held to verify the degree of agreement between the authors. If no agreement was reached between the authors, a third party would be asked to reach an agreement. As there was 100% agreement between the two evaluators, requesting any evaluation from this third researcher was unnecessary.

Inclusion criteria
All publication types, except gray literature, were accepted. In addition, studies published in all languages, performed in animals, and using the tDCS technique involving at least one proinflammatory cytokine were included. Therefore, the eligible studies met the following criteria:

1.
Experimental studies in animals that received tDCS as an intervention.

2.
The primary or exploratory objective evaluates tDCS effects on the production or release of proinflammatory cytokines.

3.
Studies that include behavioral assessment in addition to biochemical tests.

6.
Activated electrode over the cortical areas.

Exclusion criteria
Studies that did not meet the inclusion criteria, articles that used other stimulation techniques, evaluated in humans, or only considered anti-inflammatory interleukins, systematic reviews, clinical trials, abstracts, congress data, theses, dissertations, in vitro studies or even studies that associate some drug therapy with tDCS were excluded. Excluded studies were selected first by title, then by abstract, and then by full text.

Analysis of the quality of studies
The detection of risk of bias was performed using the SYRCLE RoB tool, an adapted version of the Cochrane RoB tool for animal studies (Hooijmans et al., 2014)

Results
According to the defined search strategy, the search results of 416 studies found in the electronic databases were identified in the first step. Forty were excluded because they were duplicates. Of the remaining 376 articles, 358 were excluded after analyzing the title and abstract in the exhibition phase. Finally, nine articles were excluded after the complete reading. In addition, articles were excluded because they did not achieve the objectives of this study, used another electrotherapy, analyzed only anti-inflammatory cytokines, and studied humans. Therefore, nine studies were selected at the eligibility stage.  Seven studies used bimodal stimulation (Table 1)    Most studies used the current intensity of 0.5mA and 20 minutes of application time. Several models were applied to study the effects of tDCS, including nerve and vascular manipulation. Likewise, the effects of tDCS were evaluated by different behavioral tests such as von Frey, Hot plate, and Open field. Statistical analysis was performed according to experimental design (i.e., intra-vs. inter-group analysis) or data distribution (i.e., parametric vs. non-parametric tests).
Finally, other animal models were used, such as Bilateral Carotid Artery Occlusion (Guo et al., 2020) Maze (Guo et al., 2020) and Racine Scale (Regner et al., 2020) were used during post-tDCS treatment. Two studies reported that the application of tDCS reduced mechanical and thermal hyperalgesia, causing an antiallodynic and analgesic effect, respectively (Table 3)  One study reported that bimodal tDCS, aerobic exercise, or both treatments combined promoted analgesic effects for neuropathic pain (Lopes et al., 2020). Cathodal tDCS did not change the convulsive behavior of rats (epileptic seizure occurrence or latency), while Anodal tDCS increased the number of epileptic seizures (Regner et al., 2020). A study stated that animals with induced pain (sciatic nerve constriction) showed less pain behavior from thermal and mechanical stimulation; however, there was no correlation between pain intensity (increase or reduction) and tDCS application (Cioato et al., 2016).
One study performed several measurements in a water maze test and observed that animals treated with tDCS performed statistically significantly better when compared to untreated animals (Guo et al., 2020).
For the analysis of pro and anti-inflammatory mediators (Table 3) (Guo et al., 2020). Several cytokines have been associated with tDCS effects (Table 3)  The effects of tDCS were assessed by behavioral changes (e.g., hyperalgesia, obesity, and allodynia). tDCS was able to alter the levels of TNF-α and several interleucins (e.g., IL-1α, IL-1β, and IL-6) . In addition to cytokines, tDCS causes biochemical changes in levels of nerve growth factor (NGF), lactate dehydrogenase (LDH), and brain-derived neurotrophic factor (BDNF).
To investigate changes in the cytokine levels after tDCS, studies used the cerebral cortex  and hippocampus (Guo et al., 2020;Spezia Adachi et al., 2012), (Table 3). After applying tDCS, three studies claimed a reduction in TNF-α levels Guo et al., 2020;Spezia Adachi et al., 2012). One study stated that TNF-α remained stable in the brainstem (Callai et al., 2019), and other studies indicated that there was an increase in this cytokine in the cerebral cortex, spinal cord, brainstem (Cioato et al., 2016), and hippocampus (Regner et al., 2020). Neuropathic pain groups exhibited thermal hyperalgesia. There was a reduction in hyperalgesia in the neuropathic pain + alcohol + tDCS group One-way ANOVA Posthoc: SNK test. IL-1α: ⇧ neuropathic pain + tDCS group (cerebral cortex) IL-1α: ⇧ neuropathic pain + alcohol + tDCS group (cerebral cortex) IL-1β: ⇧ neuropathic pain + tDCS group (cerebral cortex) IL-1β: ⇧ neuropathic pain + alcohol + tDCS group (cerebral cortex) Five studies found a reduction in IL-1 levels in the cerebral cortex , cerebral cortex, spinal cord, and brainstem (Cioato et al., 2016;Lopes et al., 2020); or hippocampus (Guo et al., 2020;Regner et al., 2020). However, one study indicated that IL-1 levels remained stable in the hippocampus (Spezia Adachi et al., 2012). Another article observed an increase in IL-1 levels in the cerebral cortex and brainstem (Santos et al., 2020). For IL-6, one study observed a reduction in the hippocampus (Guo et al., 2020), and in another study, there was an increase in this cytokine in the brainstem . In addition, the IL-1α levels were increased in the cerebral cortex and brainstem after the application of tDCS (Santos et al., 2020).

Discussion
Pro-inflammatory cytokines are involved in pain processing, neuronal membrane depolarization, and hyperalgesia (Campos Kraychete et al, 2006). TDCS has been investigated as a modulatory factor of the CNS (Nitsche et al., 2008), which may help to inhibit or increase the synthesis/release of these cytokines. This systematic review analyzed the effects of tDCS on the pro-inflammatory cytokine levels in CNS and serum. Furthermore, behavioral changes associated with the variation in cytokine levels were investigated. Although there are methodological differences, studies show that tDCS may alter the synthesis of cytokines.
Bimodal stimulation was the most used method (Callai et al., 2019;De Oliveira et al., 2019;Scarabelot et al., 2019;Cioato et al., 2016, Santos et al., 2020Lopes et al., 2020). This tDCS protocol was cited in the first time by Spezia and collaborators (Spezia Adachi et al., 2012) and has been reproduced by other authors (Callai et al., 2019;De Oliveira et al., 2019;Scarabelot et al., 2019;Cioato et al., 2016;Lopes et al., 2020;Regner et al., 2020;Santos et al., 2020). Spezia also determined tDCS parameters such as the application time, the current intensity, and the number of tDCS sessions (Spezia Adachi et al., 2012) When nerves are damaged, microglia and astrocytes are activated and release pro-inflammatory cytokines, which might play an essential role in developing neuropathic pain (Dimming & Algorithm, 2014). In addition, these mediators generate a warning to the body about the potential risk through the activation of nociceptive fibers (Rocha et al., 2007, Cho et al., 2018Kotani et al., 1999;Ouyang et al., 2011). The literature suggests that the initial response (protective), mainly proinflammatory, has harmful properties if it becomes a chronic process. Furthermore, there is a correlation between this inflammatory phase and the occurrence of pain after tissue damage (Thelin et al., 2020).
Besides being associated with allodynia/hyperalgesia, the presence of pro-inflammatory cytokines also can cause fever, increase protein synthesis by the liver, increase the release of corticosteroids, and reduce appetite. These changes occur to accelerate defensive enzymatic reactions, reducing the replication of pathogens and increasing the proliferation of immune cells to immobilize the injured area and conserve energy (Campos Kraychete et al., 2006). In addition, the presence of allodynia/hyperalgesia causes neuronal excitability. Those clinical findings are related to the mechanisms of maladaptation and chronicity (Ashmawi et al., 2016). In chronic pain, there are alterations in the nociceptive system, such as changes in information processing, neuroplastic rearrangement, and apoptosis of interneurons (Moore et al., 2002;Raghavendra et al, 2003). Accordingly, the presence of chronic pain causes a reduction in the patient's physical/mental condition and quality of life (Langley et al., 2013). The increase in pro-inflammatory cytokines among immune, neural or glial cells is essential for the development of pain, as these are the cytokines that are responsible for the establishment of chronic pain (Vanderwall et al., 2019).
One of the possible mechanisms involved in tDCS-induced analgesia is the modulation of levels of pro-inflammatory cytokines such as TNF-α (De Oliveira et al., 2019). This cytokine plays an essential role in inflammatory hyperalgesia and neuropathic pain. For example, in a model of induced pain in rats, there was a significant increase in the levels of TNF-α and IL-1β (Woolf et al., 1997). Furthermore, Cunha and collaborators demonstrated that the injection of TNF-α caused mechanical and thermal hyperalgesia (Cunha, 1992). Likewise, the injection of TNF-α directly into nerves induces Wallerian degeneration, a condition found in painful nerve injuries. Those data indicate a close relationship between increased TNF-α and pain. During the literature review, six studies evaluated the effect of tDCS on TNF levels (Spezia Adachi et al., 2015;Cioato et al., 2016;De Oliveira et al., 2019;Callai et al., 2019;Guo et al., 2020;Regner et al., 2020). Three articles showed decreased cytokine levels after using tDCS (Spezia Adachi et al., 2015;De Oliveira et al., 2019;Guo et al., 2020). One study stated that TNF-α remained stable (Callai et al., 2019), and two other studies indicated that there was an increase in this cytokine (Cioato et al., 2016;Regner et al., 2020).
Using a neuropathic pain model, Yana and collaborators demonstrated that IL-1β and IL-6 levels increase after peripheral nerve constriction (Yana et al., 1992). The increase in IL-1β is also associated with prostaglandins and substance P production in neurons and glial cells reinforcing the association of this interleukin with the presence of pain (Schweizer et al., 1988). Out of the seven studies that investigated the action of tDCS on IL-1β levels (Spezia Adachi et al., 2015;Cioato et al., 2016;De Oliveira et al., 2019;Guo et al., 2020;Regner et al., 2020;Lopes et al., 2020;Santos et al., 2020), five showed reduced levels of this cytokine after the treatment (Cioato et al., 2016;De Oliveira et al., 2019;Guo et al., 2020;Regner et al., 2020;Lopes et al., 2020).
IL-6 is involved in microglial and astrocytic activation (Jones et al., 1997), and its increase contributes to the development of neuropathic pain after peripheral nerve injury (Ramer et al., 1998). In a study using a model of neuropathic pain caused by freezing the sciatic nerve, an increase in IL-6 in the spinal cord was observed, demonstrating the association of this interleukin with the presence of pain (Joyce, 1996). One study observed a reduction of IL-6 after the application of tDCS (Guo et al., 2020), while the other study observed increased levels of this interleukin . As for IL-1, which was analyzed in only one study (Santos et al., 2020), the level increased after treatment with tDCS. IL-17, considered a pro-inflammatory cytokine (Jaiswal, 2014), has not yet been analyzed in the studies selected for the review.
Behavioral tests are used to investigate the presence of allodynia or hyperalgesia. In addition, these procedures allow for the evaluation of the non-communicating subjects (Deuis et al, 2017). In addition, pain can change several behaviors, such as locomotion, and may cause depression and anxiety (Guo et al., 2020;Regner et al., 2020;Santos et al., 2020). There are controversies in the literature concerning the effects of tDCS on pain behavior, especially using von Frey and Hot Plate test.
On the other hand, some authors observed that the tDCS application efficiently reduced mechanical and thermal pain (Cioato et al., 2016;Scarabelot et al., 2019;Spezia Adachi et al., 2012;Lopes et al., 2020). Other study demonstrated that the association of tDCS and alcohol caused analgesia, but tDCS alone did not reduce pain in rats (Santos et al., 2020). Moreover, a study failed to find analgesic effect of tDCS in neuropathic pain rat model (Callai et al., 2019). For anticonvulsant control, the cathodal tDCS also did not benefit the animals (Regner et al., 2020). Likewise, cathodal tDCS had no anticonvulsant action in an animal model. However, the use of tDCS improved cognitive impairment and stress effects in rats submitted to a rat model of vascular dementia (Guo et al., 2020).

Conclusion
The present literature review supports the idea that the application of tDCS can effectively modulate the levels of proinflammatory cytokines. However, even though there was a decrease in these cytokine levels in most studies, these results are not unanimous. For example, some studies found that tDCS might increase cytokine levels, and others were not able to find any effects of tDCS on these biochemical factors. One of the possible explanations for the lack of effect is the methodological variability applied by different studies. Furthermore, it is worth noting that the duration of these effects of tDCS on cytokine levels is limited to a few hours (Liebetanz at al., 2009).
In animal models, it has been seen that tDCS may decrease cytokine levels and pain measurements. However, it is difficult to make a direct relationship (cause/effect) between the induced analgesia and variation in cytokine levels because these results may be independent. Future studies should use new approaches (e.g., different brain areas stimulated and new technologies to deliver varying intensities of currents). Therefore, it will be possible to investigate further the mechanisms involved in the effect of tDCS on the levels of pro-inflammatory cytokines and their consequences on pain.