In this study, we investigated the effects of multi-session tDCS on the adult rat brain in vivo, focusing on the cellular responses to stimulation. We demonstrate a pro-inflammatory effect of both cathodal and anodal tDCS that tends to be transient. In addition, only cathodal tDCS induced the recruitment of proliferating NSC to the stimulated hemisphere. Our data suggest a polarity-specific migratory effect on endogenous neural stem cells in vivo. tDCS is capable of attracting cells inflicted in reparative and regenerative responses to the site of ischemic stroke. Beneficial effects of tDCS may at least result partly from NSC activation and the modulation of neuroinflammation.
In our experimental paradigm, we took great care to ensure that tDCS would not cause any lesion to the brain tissue. In a recent pioneer study, Liebetanz et al. described the relationship between charge density and the occurrence and size of cortical lesions 
. While we confirmed that the charge density of 128571 C/m2
per single tDCS session was not associated with cortical lesions, it should be noted that the charge density used in our study is several orders of magnitude higher than the charge density usually applied in humans (up to 480 C/m2
). We chose this high charge density – just below the lesion threshold – in order to detect any putatively faint effect of tDCS on cellular processes. Using this high charge density in multi-session tDCS sessions allowed us to detect the expected effects at a magnitude that reached statistical significance. We hypothesize that lower charge densities as applied in humans may have the same effects, albeit to a potentially lesser extent. Focusing on a proof of principle design, we did not titrate the tDCS stimulation density in this study.
We found the numbers of Iba1-positive cells to be increased early after cathodal tDCS, decreasing again with time. Similar findings were obtained after anodal tDCS, albeit to a lesser extent. This rapid and transient activation of resident microglia is typical for an innate neuroinflammatory response to tDCS. ICAM1 expression is a prerequisite of recruitment of haematogenous cells in an antigen specific manner, categorized as adaptive immunity. ICAM-1 was used as an indicator of upcoming adaptive immune responses. It showed a typical later time-course with a maximum at 10 days following cathodal stimulation, while anodal stimulation caused a more transient, unspecific effect. However, those neuroinflammatory processes observed after both cathodal and anodal stimulation could – at least in part – also be caused by thermal effects that were sub-threshold to induce a lesion to the brain tissue.
Based upon work in focal cerebral ischemia, neuroinflammation has already been characterized as a ‘double-edged sword’ with both beneficial and detrimental effects on the prevention of secondary tissue damage, regeneration and recovery 
. Destructive effects of neuroinflammation include the damage caused by reactive oxygen species and excessive production of pro-inflammatory cytokines by immune cells, beneficial aspects are the containment of necrotic tissue and the induction of a strong regenerative response including the recruitment of endogenous NSC 
. Quality, extent and timing of neuroinflammatory processes determine whether manipulating that particular response will be deleterious or therapeutically beneficial. The activation of resident microglia can under some circumstances be neurotoxic 
, especially when the mode of activation triggers a defense-oriented reaction, such as the application of bacterial endotoxin lipopolysaccharide (LPS). But the engagement of microglia can also be neuroprotective, since ablation of microglial cells in stroke mice leads to a significant increase in infarct size 
. The difference in the role and function of microglia seems to depend on the activating conditions 
. Interestingly, differentially activated microglia also have opposing effects on NSC: microglia activated by LPS attenuate neurogenesis, while their activation by cytokines associated with T-helper cells promotes neurogenesis 
. In our model of non-lesional tDCS, we propose that cathodal tDCS induces a local immune response, predominantly involving innate immunity, potentially conferring neuroprotection and attracting endogenous NSC. However, further experiments are needed to characterize the complex neuroinflammatory pattern following tDCS and the putative interplay between NSC and inflammation.
The transcription factor Hes3 was recently identified as a biomarker for a widespread population of quiescent endogenous NSC that is dynamically upregulated upon their activation 
. Several recent reports have demonstrated the direct effect of electric fields on the migration of stem cells in culture. This has been shown in adult mouse NSC from the subventricular zone 
, adult rat NSC from the hippocampus 
, embryonic rat NSC from the lateral ganglionic eminence 
, human embryonic stem (hES) cells and human induced pluripotent stem (hiPS) cells 
, and hES cell-derived NSC 
. All of those cell types were reported to migrate towards the cathode of the electric field, except the hiPS cells that moved towards the anode 
. Accordingly, we here show an upregulation of proliferating NSC in the non-invasively stimulated brain after cathodal tDCS can also be explained by the migration of endogenous NSC towards the cathode. The extent of eNSC accumulation after cathodal tDCS was comparable to that induced pharmacalogically, e.g. via activation of the Notch receptor 
. Enhancing the (physiological) mobilization of eNSC after injury such as stroke was previously shown to remarkably improve neurological function and recovery 
. However, the differentiation of mobilized eNSC into mature neurons that functionally integrate into the damaged circuitry has rarely been observed, and the vast majority of newly generated migrating neuroblasts in ischemic stroke models die by the time they have reached the peri-infarct area 
. Recent studies have elucidated several mechanisms other than neurogenesis that convey the stem cells' beneficial effects. An important function of eNSC seems to be neuroprotection 
, with neuroprotective effects mediated through several neuroprotective cytokines such as GDNF, VEGF, and Shh 
. Other functions of eNSC in regeneration involve the suppression of inflammation in the damaged tissue, and clearance of debris in the injured area such as the peri-infarct tissue 
Although BrdU-positive proliferating cells could at least in part constitute microglia, the extent of their upregulation and their exclusive appearance after cathodal tDCS makes it likely that at least the majority of those proliferating cells are not microglia. Furthermore, we found the number of Hes3-positive cells to increase with the number of cathodal tDCS session, further confirming an effect on NSC. In addition to the direct effect of the electric field on NSC migration, a second mechanism for the activation of cortical NSC is conceivable: tDCS was recently shown to modulate cerebral blood flow in a polarity-specific way, with an increase after anodal and decrease after cathodal stimulation, lasting for at least 30 min after tDCS 
. Transient ischemic stress caused by a brief decrease in blood flow is well known to induce the proliferation of endogenous NSC, conferring a certain ischemic tolerance in the process 
This study provides evidence that electric stimulation modulates responses of non-neuronal cells in the brain, and relevantly contributes to our scarce knowledge about the neurobiological effects of tDCS. tDCS both attracts endogenous NSC and activates innate immune responses. From the clinical point of view, applying tDCS after stroke may help to locally augment endogenous NSC known to promote neuroprotection and repair.