Contribution of glial cells to the neuroprotective effects triggered by repetitive magnetic stimulation: a systematic review

Abstract

Repetitive transcranial magnetic stimulation has been increasingly studied in different neurological diseases, and although most studies focus on its effects on neuronal cells, the contribution of non-neuronal cells to the improvement triggered by repetitive transcranial magnetic stimulation in these diseases has been increasingly suggested. To systematically review the effects of repetitive magnetic stimulation on non-neuronal cells two online databases, Web of Science and PubMed were searched for the effects of high-frequency-repetitive transcranial magnetic stimulation, low-frequency-repetitive transcranial magnetic stimulation, intermittent theta-burst stimulation, continuous theta-burst stimulation, or repetitive magnetic stimulation on non-neuronal cells in models of disease and in unlesioned animals or cells. A total of 52 studies were included. The protocol more frequently used was high-frequency-repetitive magnetic stimulation, and in models of disease, most studies report that high-frequency-repetitive magnetic stimulation led to a decrease in astrocyte and microglial reactivity, a decrease in the release of pro-inflammatory cytokines, and an increase of oligodendrocyte proliferation. The trend towards decreased microglial and astrocyte reactivity as well as increased oligodendrocyte proliferation occurred with intermittent theta-burst stimulation and continuous theta-burst stimulation. Few papers analyzed the low-frequency-repetitive transcranial magnetic stimulation protocol, and the parameters evaluated were restricted to the study of astrocyte reactivity and release of pro-inflammatory cytokines, reporting the absence of effects on these parameters. In what concerns the use of magnetic stimulation in unlesioned animals or cells, most articles on all four types of stimulation reported a lack of effects. It is also important to point out that the studies were developed mostly in male rodents, not evaluating possible differential effects of repetitive transcranial magnetic stimulation between sexes. This systematic review supports that through modulation of glial cells repetitive magnetic stimulation contributes to the neuroprotection or repair in various neurological disease models. However, it should be noted that there are still few articles focusing on the impact of repetitive magnetic stimulation on non-neuronal cells and most studies did not perform in-depth analyses of the effects, emphasizing the need for more studies in this field.

Introduction

Transcranial magnetic stimulation (TMS) was introduced as a neurophysiological technique in 1985 by Anthony Barker and his team (Klomjai et al., 2015). This technique uses the time-varying magnetic field generated by one coil placed over the patient’s head which induces an electrical stimulus in adjacent conductive brain tissues, parallel to the coil, interacting with the neuronal activity of the targeted brain region (Zhong et al., 2018; Burke et al., 2019). The induced electrical field induces the flow of ions in the brain, altering the electrical charge on the cell membranes (Rossi et al., 2009).

There are two conventional types of repetitive TMS (rTMS), which are described as low-frequency rTMS (LF-rTMS) and high-frequency rTMS (HF-rTMS). While LF-rTMS uses stimulation frequencies lower than 1 Hertz (Hz) and is characterized by continuous trains of single pulses, HF-rTMS uses stimulation frequencies higher than 5 Hz and is usually characterized by trains of stimuli lasting 5–10 seconds (s) separated by pauses of 20 to 50 seconds (Rossi et al., 2009; Klomjai et al., 2015). The functional effects triggered by rTMS last longer than just the period of the treatment protocol. The most prevalent theory suggests that these effects are associated with long-term depression (LTD) induced by LF-rTMS or with the long-term potentiation (LTP) induced by HF-rTMS, given that LTD results in a decrease in synaptic strength, whereas LTP promotes increased synaptic strength (Klomjai et al., 2015). In addition to these conventional forms of rTMS, there are some other approaches, one of which is delivered in the form of a theta pattern, the theta-burst stimulation (TBS). This type of TMS consists of short bursts of three pulses at 50 Hz that occur every 200 milliseconds (i.e., 5 Hz, the theta frequency) (Rossi et al., 2009; Burke et al., 2019). There are two forms of TBS, intermittent (iTBS) and continuous (cTBS). While iTBS delivers 2-second stimuli every 10 seconds, for 600 seconds in total (around 180 seconds) and can result in motor-evoked potential facilitation, cTBS also delivers 600 pulses with a continuous stimulus for 40 seconds, which can result in motor-evoked potential inhibition (Huang et al., 2005).

The LTP and LTD processes are related to the electric current induced by TMS, apparently causing changes in synaptic plasticity and neurotransmitter release, as previously mentioned. However, several studies have shown that the intrinsic magnetic field is also responsible for several effects including macromolecular magnetic effects, magnetic spin effects, genetic magnetoreception, and quantum effects (Chervyakov et al., 2015). Furthermore, the combination of the induced electric current and the intrinsic magnetic field is seen as a possible explanation for the long-term effects caused by rTMS (Chervyakov et al., 2015).

Due to the ability to modulate specific brain areas in a focalized way and the possibility of adjusting the stimulation protocols to target specific effects, rTMS is becoming a promising technique to normalize and/or enhance the cortical activity on several neurological disorders, promoting the recovery of brain functions (Klomjai et al., 2015; Lefaucheur, 2019).

rTMS has been proposed as a therapeutic approach for neurological disorders such as major depressive disorder (MDD), ischemic stroke (IS), neurogenic pain, Parkinson’s disease (PD), Alzheimer’s disease (AD), and multiple sclerosis (MS) (Fregni and Pascual-Leone, 2005). In animal studies, it is not yet clear whether the therapeutic potential of rTMS is similar in the two sexes (Korb et al., 2015; Lee et al., 2017). A few studies suggest that there may be differences in physiological responses to rTMS between sexes (Adamson et al., 2022), which may be explained by the distance between scalp and cortex, gray matter density, and by estradiol/progesterone levels (Hanlon and McCalley, 2022). However, although studies comparing the action in males and females are scarce, two clinical studies have shown greater efficacy of high-frequency single pulse application and iTBS in females than in males suffering from major depression (Engelbertson, 2022; Kinback et al., 2022).

Although the beneficial effects of repetitive magnetic stimulation (rMS) are recognized, the cellular and molecular mechanisms triggered by rMS and TBS that are responsible for these improvements are not completely understood. The available data about how rTMS modulates cells is still scarce and focuses mainly on the effect of rTMS on neuronal cells disregarding the role that non-neuronal cells may have. In addition to neurons, the central nervous system is composed of glial cells (von Bernhardi et al., 2016) and vascular cells (Ross et al., 2020). Among the glial cells, astrocytes have the ability to actively participate in almost all processes that are associated with neuronal homeostasis, since they provide metabolic support to neurons, participate in the removal of metabolic waste by the glymphatic system, reduce excitotoxicity by removing glutamate from the synaptic cleft, regulate blood flow, control potassium concentrations, regulate synapse formation and communication, and regulate neuroinflammation (Sofroniew and Vinters, 2010; Colombo and Farina, 2016; Liu and Chopp, 2016). Oligodendrocytes are fundamental for neuronal communication, providing support to axons and facilitating action potential transmission through myelination (von Bernhardi et al., 2016; Kuhn et al., 2019), whereas microglia, participate in the immune defense of the central nervous system, but also play an important role in the regulation of neurogenesis, neuroinflammation, synaptic pruning, and synaptic plasticity (Nayak et al., 2014; Colonna and Butovsky, 2017). Glial cells respond to brain insults and have a fundamental role in the outcome after injury. These cells may have a dual action, either protecting and repairing or exacerbating the injury. Thus, studying how rMS modulates glial cells can be of major interest to understanding how this strategy can promote protection and repair after brain lesions. So far, the available data indicate that rMS can promote neuroprotection (Hausmann et al., 2001; Sasso et al., 2016), neuroplasticity (Cacace et al., 2017), neurogenesis (Caglayan et al., 2019), and prevent hyperexcitability (Vucic et al., 2021). Due to their properties and intrinsic characteristics, each subset of glial cells reacts differently to electrical activity and hypothetically all of these cells can be affected by rMS, however, the way this occurs, what mechanisms are involved, and the magnitude of this modulation is not fully understood (Cullen and Young, 2016).

Although the published work on the modulation of glial cells by rMS is not numerous, evidence that these cells may be an important target for the protective action exerted by rMS has been growing. Therefore, the major objective of this systematic review is to gather information about the modulatory effects of rMS on non-neuronal cells, promoting a better understanding of how this stimulation functions in this type of cells and what type of stimulation might be most effective in inducing neuromodulation.

Methods

Literature search

The literature search was performed using two databases (PubMed and Web of Science) on October 2, 2021. An update to the research was made on February 10, 2023. The literature search included the following search terms: (Glia* AND “magnetic stimulation”) OR (Glia AND “magnetic stimulation”) OR (Astrocyte* AND “magnetic stimulation”) OR (Astrocyte AND “magnetic stimulation”) OR (Microglia* AND “magnetic stimulation”) OR (Microglia AND “magnetic stimulation”) OR (Astroglia* AND “magnetic stimulation”) OR (Astroglia AND “magnetic stimulation”) OR (Oligodendrocyte* AND “magnetic stimulation”) OR (Oligodendrocyte AND “magnetic stimulation”) OR (“brain vascular cells” AND “magnetic stimulation”) OR (Pericyte AND “magnetic stimulation”) OR (“Smooth muscle cells” AND “magnetic stimulation”) OR (“Endothelial cells” AND brain AND “magnetic stimulation”) OR (Glia* AND “frequency magnetic field”). Not all of the above terms exist as MESH terms, so the search was carried out using only keywords. Only peer-reviewed articles published in English were selected. Duplicates were manually removed from the search results. Two authors conducted the literature search and screened the abstracts and the full text of the studies independently. Any disagreements were settled by a discussion with a third author. The literature search and exclusion criteria applied to select the studies are shown in Figure 1.

Figure 1.

Flow diagram representing the literature search and exclusion criteria applied to select the studies included in this systematic review.

rTMS: Repetitive transcranial magnetic stimulation.

Inclusion and exclusion criteria

The literature search and the exclusion criteria used are summarized in Figure 1. In consideration of our review question and following a PICO framework, we included only preclinical studies, in cell cultures or animals that always included a control group, evaluating the effect of various stimulation protocols (HF-rMS, LF-rMS, iTBS, and cTBS) on glial cells. In the results, we included motor and cognitive assessments, or evaluation of different cellular and inflammatory markers.

We excluded all review articles, letters to the editor, and congress presentations. We also excluded studies with TMS that only used TMS to make electrophysiological evaluations or that only evaluated its effect on neurons, and studies that only used electrical and not magnetic stimulation. We also excluded Helmholtz coils because they have very different characteristics from the commonly used coils that are indicated in clinical practice.

Data extraction

From the included articles, we extracted the following information: type of stimulation, stimulation frequency (high or low), stimulation intensity, protocol of stimulation, model (in vivo or in vitro), neurological models, and the major results that include parameters of astrocyte and microglial reactivity, oligodendrocyte proliferation, differentiation and enhancement of myelination, and release of pro-inflammatory cytokines. The astrocyte and microglial reactivity included counting the number of cells positive for glial fibrillary acidic protein (GFAP) or ionized calcium-binding adaptor molecule 1 (Iba-1), respectively, their expression levels, or the cellular proliferation. This information is summarized in Additional Tables 1–6.

Additional Table 1.

Studies focusing on the effects of HF-rTMS in unlesioned animals or cells

Studies	Stimulation	Frequency (Hz)	Intensity(T)	Stimulation protocol	Model	Release of inflammatory cytokines	Microglia reactivity	Astrocyte reactivity	Oligodendro cytes	Results
Chan et al., 1999	rMS	10	0.10;0.21; 0.42;0.63	Figure-of-eight commercial coil, 1 session	PCC-Astrocytes from cortex	-	-	↑ GFAP levels and returnto baseline	-	Observed 0, 1, 3, 5, 7, 12, or 20 days after stimulation: Increased GFAP levels on day 3 whichreturned to baseline after 5 days.
Clarke et al., 2021	rMS	10	0.018	Custom built round coil, 1 session	PCC-Astrocytes from cortex	↓ Cxcl2, Dusp1, Egr1,Stat3,Smad3, Clcf1, Hmox1 -	-	-	-	Observed 5 hours after stimulation: Reduced pro-inflammatory cytokines, reductionof genes involved in calcium signaling. There were changes inKcnmb4 gene expression, but no change in protein levels, and different protein levels of STIM1 and ORAI3 despite no significant changes in mRNA levels.
Clarke et al., 2017b	rMS	10	0.018	Custom built round coil, 1 session	PCC-Astrocytes from cortex	-	-	-	-	Observed when cells were confluent: No effects on calcium levels, no effect on astrocyte migration or proliferation.
Hausmann et al., 2001	rMS	20;50	50%;75% of max	Figure-of-eight commercial coil, 1 session on DIV 14	OC from parietalcortex	-	-	No effects in GFAP+ cells	-	Observed 0, 1.5, 3, 6, 14, or 24hours after stimulation: No alterations on GFAP levels.
Liu et al., 2020	rMS	10	2.4;4.8	Figure-of-eight commercial coil, 2 sessions per day, from DIV 12 to35	iPSCs line	-	-	No effects in GFAP+ cells	↑ Proliferation and differentiation	Observed on the last day of stimulation: Glial cell markers GFAP or Oligo2 were not detected, indicatingfew glial cells were generated. rMS significantly increased Olig2 transcription compared to the control group.
Cullen et al., 2019	rTMS	10	0.12	Custom built round coil, 28 sessions	Male and female transgenic C57BL/6Jmice (adult) -cre-lox lineage tracing	-	-	-	No effects on oligodendroc yte number	Observed 7, 14, or 28days after lesion: Did not affect oligodendrocyte number
Fujiki and Steward, 1997	rTMS	25	1.63	Commercial coil, 1 session	Male C57BLr6J mice (60–90 days)	-	-	-↑ GFAP levels and returnto baseline	-	. Observed 12, 24, 36, 48hours, 4 or 8 days after stimulation: Increased GFAP levels which returned tobaseline after 8 days. The animals thatreceived magnetic stimulation while anesthetized with ketamine did not exhibit increases inGFAP levels.
Hausmann et al., 2000	rTMS	20	75% of max	Figure-of-eight commercial coil, 14 sessions	Male SD rats (adult)	-	-	No effects in GFAP levels	-	Observed 12 hours after the last stimulation: No changes inGFAP.
Post et al., 1999	rTMS	20	4	Round commercial coil, 1 session or 5 sessions per week for 11 weeks with anesthesia	Male Wistar rats (3 months old)	-	-	↑ GFAP+ cells and returnto baseline	-	Observed after the last stimulation and after 11 weeks: Increased GFAP levels whichreturned to baseline after 11 weeks.
Zorzo et al., 2019	rTMS	100	0.33	Custom built round coil3 sessions	Male Wistar rats (young)	-	No effects on Iba-1+ cells	No effects in GFAP+ cells	-	Observed 90minutes after the last stimulation: No changes in astrocyte or microglia reactivity.

Clcf1: Cardiotrophin-like cytokine factor 1; CXCL2: C-X-C motif chemokine ligand 2; DIV: days in vitro; DUSP1: dual specificity phosphatase 1; EAAC1: glutamate transporter excitatory amino-acid carrier 1; EAAT4: excitatory aminoacid transporter 4; Egr1: early growth response 1; GAT2, GAT3, and GAT4: GABA transporters; GFAP: glial fibrillary acidic protein; GLAST: glial glutamate and aspartate transporter; GLT1: glial glutamate transporter; GLYT1 and GLYT2: glycine transporters; HF-rMS: high-frequency repetitive magnetic stimulation; HF-rTMS: high-frequency repetitive transcranial magnetic stimulation; Hmox1: heme oxygenase 1; Iba-1: ionized calcium binding adaptor molecule 1; iPSCs: induced pluripotent stem cells; Kcnmb4: calcium-activated potassium channel subunit beta-4; max: maximum output of the stimulator; OC: organotypic cultures; Olig2: oligodendrocyte transcription factor 2; OLIGO2: gene expressed by oligodendrocytes; ORAI3: calcium release-activated calcium modulator 3; PCC: primary cell culture; rMS: repetitive magnetic stimulation; rTMS: repetitive transcranial magnetic stimulation; SD: Sprague-Dawley; Smad3: SMAD family member 3; Stat3: signal transducer and activator of transcription 3; STIM1: stromal interaction molecule 1; TGF-β1: transforming growth factor beta 1. "-": Not assessed.

Additional Table 6.

Studies focusing on the effects of TBS in disease models

Studies	Stimulation	Main frequency (Hz)	Intensity (T)	Stimulation protocol	Model	Neurological conditions	Releaseof inflammatory cytokines	Microglia reactivity	Astrocyte reactivity	Oligodendro cytes	Results
Stanojevic et al., 2022	iTBS	5	33%of max	Figure-of-eight commercial coil, 2 sessions 2 days apart, starting6 days after surgical recovery (each session lasted 5 days)	Male Wistar rats (adult) – Intracerebroventr icularly injected streptozotocin model	AD	-	↓ Iba-1+ cells	↓ GFAP+ cells	-	Observed 6 days after the last stimulation: iTBS treatment attenuates cognitive impairment. After iTBS treatment, reduction inthe intensityof stainingcould be observed in all regions accompaniedby reverting microglial morphology towards restingphenotype. iTBS decreased GFAP+ cells.
Stekic et al., 2022	iTBS	5	33%of max	Figure-of-eight commercial coil, 15 sessions, 2 stimulations per day starting 3 days after lesion	Male Wistar rats (2 months old) – intraperitoneal injectionof trimethyltin	AD	↓ IL-1β ↑ IL-10	-	No effects	-	Observed 21 days after lesion: iTBS reduced hyperactivity, aggressive behavior, anxiety, and tremor and improve cognitive impairment, effect that could be mediated via PI3K/Akt/mTOR signalingpathway.
Cai et al., 2020	iTBS	5	20%of max	Figure-of-eight commercial coil, 14 sessions, starting 22 days after surgery	Male SD rat (4 weeks old) -Chronic Hypertension-induced CSVD	CSVD	-	↓ Iba-1+ cells	↓ GFAP+ cells	-	Observed on last day of stimulation: iTBS improved spatialcognitive function and NR2B, p-CaMKIIa, and GluR1 expression. iTBS decreased GFAP+ cells andIba-1+ cells.
Luo et al., 2017	iTBS	Information not available	1.44; 0.96	Figure-of-eight commercial coil, stimulation from day 3 after lesion to day 7 and from day 10 today 14	Male Wistar rats (adult) – MCAO model	IS	-	-	No effects in GFAP+ cells	-	Observed on day 7 or 14after lesion: Functional recovery, enhanced neurogenesis, and activation of BDNF/TrkB pathway. Reduced infarct volume. Observed 24 hours after the last lesion: Promoted motor functionalrecovery, noteworthy,
Luo et al., 2022b	iTBS	5	25%of max	Roundcommercial coil, for 8 days starting day1 after lesion	Male C57BL/6J mice(adult) – MCAO model	IS	↓ IL-1β, IL-17a, TNF-α, IFN-γ, ↑ IL-10	↓ CD86, iNOS ↑ CD206, Arg1	-	-	depletion of microglia eliminated themotor functional improvementsafter iTBS treatment. iTBS reverse M1 polarization of microglia reducingmicroglia-associated neuroinflammation.
Dragic et al., 2020	iTBS and cTBS	5	30%of max	Figure-of-eight commercial coil, 10 sessions starting 14 days after lesion	Female DA rats (10-14weeks old) – EAE model	MS	-	↓ Iba-1+ cells	↓ GFAP+ cells	↓ Demyelinatio n	Observed 24 hours after the last lesion: TBS treatmentattenuated reactive gliosis and restored myelin sheet.
Dragic et al., 2021	cTBS	Information not available	30%of max	Figure-of-eight commercial coil, 10 sessions starting 14 days after lesion	Female DA rats (2 months old) – EAE model	MS	↓ IL-1β, CD73↑ IL-10	↓ Iba-1+ cells	↓ GFAP+ cells	-	Observed on the last day of stimulation: Lowered the number of reactive microglial cells and hypertrophied astrocytes.
Stevanovic et al., 2019	iTBS and cTBS	5	30%of max	Figure-of-eight commercial coil, 10 sessions starting 14 days after lesion	Female DA rats (10-14weeks old) – EAE model	MS	-	↓ Iba-1 levels	↓ ki67and GFAP levels	-	Observed on the last day of stimulation: Both protocols of TBS increased BDNF and decreased GFAP, ki67 and Iba-1.
Cacace et al., 2017	iTBS	5	30%of max	Figure-of-eight commercial coil, 1 session 8 days after lesion	Male Wistar rats -6-OHDA model	PD	-	↓ Iba-1+ cells	↓ GFAP+ cells	-	Observed 20 or 80minutes after stimulation: Reduction in the total number of GFAP and Iba-1 cells 80 minutes after iTBS.
Muri et al.,2020	iTBS and cTBS	5	90% of max	Commercialcoil, 2 sessions starting 5 days after infection	PCC-Astrocytes from cortex & maleWistar rat (11days old) -injectionof living S. pneumoniae	PM	↑ IL-1β, IL-10, TNF-α, IL-6	↑ Iba-1, Cd14, Cd45, Cd68, Cd84, Cd86, F4/80 ↓ CD163, Arg1, Cd206, Ym1, Fizz1	↑ GFAP, S100b, S100a4	-	Observed 24 hours after the last stimulation: iTBS and cTBS induced polarization of microglia andastrocytes towards an inflammatory phenotype. cTBS induced the release of pro-inflammatory cytokines.
Sasso et al., 2016	iTBS	5	30%of max	Figure-of-eight commercial coil, 7 sessions starting 1 hour after surgery	Male Wistar rats (young) -removal of the rightcerebellar hemisphere	TBI	-	↓ Iba-1 levels	↓ GFAP levels	-	Observed 1 hour after the last stimulation: Reduced neuronal deathand improved functional recovery.

AD: Alzheimer’s disease; Arg1: arginase 1; BBB: blood-brain barrier; BDNF: brain-derived neurotrophic factor; CNTF: ciliary neurotrophic factor; CSVD: cerebral small vessel disease; cTBS: continuous theta-burst stimulation; CX3CL1: C-X3-C motif chemokine ligand 1; DA: Dark Agouti; EAE: experimental autoimmune encephalomyelitis; Fizz1: resistin-like molecule alpha1; GFAP: glial fibrillary acidic protein; GluR1: glutamate receptor 1; HIF-1α: hypoxia-inducible factor 1-alpha; Iba-1: ionized calcium binding adaptor molecule 1; IFN-γ: interferon gamma; IGFBP1: insulin like growth factor binding protein 1; iNOS: inducible nitric oxide synthase; IS: ischemic stroke; iTBS: intermittent theta-burst stimulation; max: maximum output of the stimulator; MCAO: middle cerebral artery occlusion; MS: multiple sclerosis; NR2B: N-methyl D-aspartate receptor subtype 2B; p-CaMKIIa: Ca2+/calmodulin-dependent protein kinase II; PCC: primary cell culture; PD: Parkinson’s disease; PM: pneumococcal meningitis; SD: Sprague-Dawley; TBI: traumatic brain injury; TGF-β: transforming growth factor beta; TNF-β: tumor necrosis factor-α; TrkB: tropomycin receptor kinase B; Ym1: chitinase-like protein. “-”: Not assessed.

The organization of the extracted data was done considering the various stimulation protocols (HF-rMS, LF-rMS, iTBS, and cTBS) and each of these has been divided into two sections (unlesioned cells/animals and models of disease). Finally, it was also analyzed if the published works were performed in males, females, or both sexes.

With the literature search, we identified 205 articles, which were screened. After assessing their eligibility, only 116 seemed to be suitable. However, after a more rigorous evaluation of these articles, only 52 fit our criteria. None of the studies include information regarding randomization and blinding of the experimental design, so the risk of bias has not been assessed.

Results

Effects of high-frequency repetitive transcranial magnetic stimulation on non-neuronal cells

In addition to the several effects that have already been described in neuronal cells, there is strong evidence that HF-rTMS has the ability to affect non-neuronal cells. The majority of the effects described on non-neuronal cells are associated with the modulation of astrocyte and microglial reactivity, as well as a decrease in pro-inflammatory cytokines (Figure 2; Kim et al., 2013; Caglayan et al., 2019; Zorzo et al., 2019).

Figure 2.

Effects of different magnetic stimulation protocols on astrocyte and microglial reactivity, oligodendrocyte proliferation, differentiation and enhancement of myelination, and release of pro-inflammatory cytokines, as described throughout this systematic review.

AD: Alzheimer’s disease; CP: chronic pain; CSVD: cerebral small vessel disease; cTBS: continuous theta-burst stimulation; HF-rTMS: high-frequency repetitive transcranial magnetic stimulation; HS: hemorrhagic stroke; IS: ischemic stroke; iTBS: intermittent theta-burst stimulation; LF-rTMS: low-frequency repetitive transcranial magnetic stimulation; MDD: major depressive disorder; MS: multiple sclerosis; NI: non-injured animals or cells; PD: Parkinson’s disease; PM: pneumococcal meningitis; SCI: spinal cord injury; TBI: traumatic brain injury.

In this review, 36 articles that used an HF-rTMS protocol were analyzed. Ten studies were performed in animals or cells without any lesions (Additional Table 1), but the impact of stimulation in different disease models was also analyzed (Additional Table 2). However, for each disease model, there are a small number of studies evaluating the impact of HF-rTMS on non-neuronal cells.

Additional Table 2.

Studies focusing on the effects of HF-rTMS in disease models

Studies	Stimulation	Frequency (Hz)	Intensity(T)	Stimulation protocol	Model	Neurological conditions	Releaseof inflammatory cytokines	Microglia reactivity	Astrocyte reactivity	Oligodendr ocytes	Results
Cao et al., 2022	rTMS	25	60% of max	Figure-of-eight commercialcoil, 21 sessions	Male and female 3xTg-AD model mice (6-8 month-old)	AD	↓ IL-6, IL-1β, TNF-α	↓ Iba-1+ cells	--	-	Observed 28 days after lesion: Improve cognitive function, ameliorate oxidative stress, and improve glucose metabolism. rTMS alleviated neuroinflammatory response, enhanced synaptic plasticity, and reduced neuronal loss and cell apoptosis by activation of the PI3K/Akt/GLT-1 pathway.
Li et al., 2021	rTMS	20	90% of motor threshold	Figure-of-eight commercialcoil, 14 sessions	Male and female double transgenic 5xFAD mice and wild-type mice of either gender (2-month-old)	AD	↓ IL-6, TNF-α	↓ Iba-1+ cells	--	-	Observed on thelast day of stimulation: rTMS reduced Aβ levels, Iba-1, IL-6, and TNF-α.
Lin et al., 2021	rTMS	20	1.38	Roundcommercial coil, 14sessions	Male and female 5xFAD mice (4 to5 months)	AD	-	↓ Iba-1+ cells	↓ GFAP+ cells -	-	Observed 10 days after the last stimulation: rTMS reduced Iba-1+ and GFAP+ cells.
Shiiba et al., 2012	rTMS	83M; 2k	2x10-7	Custom built round coil, 7 sessions starting7 days after lesion	Male SD rats -CCI model	CP	-	-	↓ GFAP+ cells -	-	Observed on thelast day of stimulation: rTMS ameliorated chronicpain and increased 5-HT levels. ReducedGFAP + cells
Yang et al., 2018	rTMS	20	40% of max	Does not indicate the coil, 10sessions starting3 days after lesion	Male SD rats -sciatic nerve ligation	CP	-	-	↓ BrdU/ GFAP+ cells -	-	Observed on theday after stimulation: rTMS decreased GFAP + cells
Cui et al., 2019	rTMS	10	Information not available	Figure-of-eight commercialcoil, 5 sessions, startingthe day after lesion	Male C57BL/6 J mice (7 to 8 weeks) -ICH model	HS	-	-	↓ GFAP+ cells -	-	Observed 24, 72, or 120 hours after stimulation: Glial differentiation was reduced, decreasing GFAP+ cells
Caglayan et al., 2019	rTMS	20	26% of max	Figure-of-eight commercialcoil, stimulation on the day of reperfusion (a), within3 days starting on the day of reperfusion (b), or 28 days starting 3 days after reperfusion (c) withanesthesia	Male Balb/c mice (10 to12 weeks) – MCAO model	IS	↓ IL-1β, TNF-α, TGF-β, MMP9	↓ BrdU/ Iba-1+ cells	No effects in BrdU/ GFAP+ cells -	-	Observed 1 day (a, b) or 10days (c) after the last stimulation: HF-rTMS increased CD31+ cells, decreased IL-1β, TNF-α, TGF-β, and MMP9, and upregulated VEGF-A and VEGF-B. Decreased glial scar formation and reduced Iba-1+ cells
Chen et al., 2023	rTMS	10; 20	33% of max	Custom built round coil, from the 7thto 28thday after lesion	Male SD rats (7 weeks) – MCAO model	IS	↓ TNF-α, IL-1β, iNOS ↑ Arg-1, IL-10, TGF-β	↓ Number of CD68-and CD16/32-positive microglia;↑ number of CD206-positive microglia	-	-	Observed 33 days after lesion: 20 Hz HF-rTMS more markedly improved the cognitive function atday 28 after operation. HF-rTMS attenuates white matter lesion, decreased proinflammatory cytokine levels, and increased anti-inflammatory cytokine levels. It also decreased the number of CD68-and CD16/32-positive microglia and increased the number of CD206- positive microglia.
Hong et al., 2020	rTMS	5; 10	1.9	Roundcommercial coil, 2 sessions in vitro and 7 sessions in vivo startingthe day after lesion	PCC -Astrocytes and neurons from cortex & male SD rats (adult) – MCAO model	IS	↓ TNF-α ↑ IL-10	-	↓ C3; iNOS; BrdU/ GFAP+ cells; ↑ S100A10; Arg1	-	Observed the day after the last stimulation: A1 markers, C3 and iNOS were inhibited and S100A10 and Arg1 were significantly increased. Decreased pro-inflammatory mediator TNF-α and increased anti-inflammatory
Hong et al., 2022	rTMS	10	1.9	Roundcommercial coil, 2 sessions in vitro and 7 sessions in vivo startingthe day after lesion	PCC -Astrocytes and neurons from cortex & microglial CL & male SD rats (adult) – MCAO model	IS	↓ TNF-α ↑ IL-10	↑ CD206 ↓ iNOS	-	-	mediator IL-10. Observed 1, 3,5,7, and14 days after lesion:Reverse M1 polarization of microglia reducing microglia-associated neuroinflammation.
Gava-Junior et al., 2022	rTMS	10	60% of max	Figure-of-eight commercialcoil, stimulation immediately after OGD	PCC -Astrocytes and neurons from cortex	IS	-	-	No effects in GFAP+ cells	-	Observed 1 day after stimulation: HF-rMS prevented neurite degeneration, increase the number of cells expressing ERK1/2,c-Fos, modulates the release of growth factors byastrocytes, namely GDNF thatprotects neurons from ischemia- induced injury.
Luo et al., 2017	rTMS	20	1.44; 0.96	Figure-of-eight commercialcoil, stimulation from day 3 after lesion today 7 and from day 10 to day 14	Male Wistar rats (adult) – MCAO model	IS	-	-	No effects in GFAP+ cells	-	Observed on day7 or 14 after lesion: Promoted functional recovery. Enhanced neurogenesis and activationof BDNF/TrkB signaling pathways. Reduced infarct volume.
Luo et al., 2022a	rTMS	10	30% of max	Figure-of-eight commercialcoil for rats, 7 and 28 days (2 times per day) beginning2 days after lesion in vivo and 2 days in vitro	PCC – neural stem cells and microglia & male Wistar rats (adult) – MCAO model	IS	↓ TNF-α, IL-1β ↑ IL-4, IL-10	↑ CD206	-	-	Observed at day 7 or 28after lesion in vivo and 24hours after the last stimulation in vitro: improvement in neurocognitive behavioral function. Long-term rTMS reverse M1 polarization of microglia reducing microglia-associated neuroinflammation.
Roque et al., 2021	rMS	10	60% of max	Figure-of-eight commercialcoil, 1 session, starting after lesion	PCC -Astrocytes and neurons from cortex	IS	-	-	-	-	Observed 1 day after stimulation: HF-rMS prevented neuronal death and the neurite degeneration, increase the number cells expressing ERK1/2, c-Fos, increased the number of synapticpuncta as well as their intensity.
Jiang et al., 2022	rTMS	10	100% of max	Circular animal commercialcoil, three courses of stimulation thatlasted 5 days each and were separated by 2 days	Male SD rats (adult) – CUS model	MDD	-	-	↓ GFAP levels -	-	Observed 5 days after the last stimulation: Ameliorated depressive likebehaviors and reduce GFAP levels.
Peng et al., 2018	rTMS	5; 10	0.84; 1.26	Figure-of-eight commercialcoil, 2 sessions in vitro at DIV 10/14and 7 sessions in vivo starting4 weeks after lesion	PCC -Astrocytes from prefrontal cortex & male SD rats (adult) -CUS model	MDD	-	-	--	-	Observed after stimulation in vitro and two days after the last stimulation in vivo: Ameliorated depressive like behaviors, increased 5-HT, dopamine, and NE levels, decreased the 5-HIAA level and Sirt1and MAO-A activity.
Yan et al., 2022	rTMS	10	Information not available	Information of coil not available, for 3 weeks, starting5 weeks after the start of the injury model	Male SD rats (adult) – CUS model	MDD	-	-	↑ GFAP+ cells -	-	Observed 8 weeks after the startof the injury model: Ameliorated depressive like behaviors. Improved the survivalrate of astrocytes.
Zuo et al., 2022	rTMS	15	80% of max	Figure-of-eight commercialcoil, stimulation for 4 weeks starting 4 weeks after lesion	Male C57BL/6J mice – CUS model	MDD	↓ IL-6, IL-1β and TNF-α	↓ CD68, CD16/32 Iba-1+ cells ↑ CD206	↑ GFAP+ cells -	-	Observed 1 week after the last stimulation: Ameliorated depression-like behaviors. rTMS not only markedly alleviated the activationof microglia but induced a switch of microglia polarization from pro-inflammatory M1 phenotype to anti-inflammatory M2 phenotype. rTMS reversed the downregulation of astrocytes and inhibitedhigh levels of interleukin IL-6, IL-1β and TNF-α.
Wang et al., 2021	rTMS	40	< 0.002	Does not indicate the coil, 5 sessions per week for 2 or 4 weeks starting12 weeks after lesion	Female C57BL/6 mice (8 weeks old) -cuprizone model	MS	↑ TGF-β -		-	↑ Myelination	Observed on thelast day of stimulation: Improved cognitionand depression-like behavior. Promoted myelin repair. Improved TGF-β which is importantin the myelination process and in inflammatory responses in MS.
Kang et al., 2022	rTMS	10	20% of max	Circular commercial coilfor rodent use, daily startingat 1 week post injury for 4 weeks	Male SD rats (adult) – 6-OHDA model	PD	↓ TNF-α, IL-1β, IL-6 -		↓ GFAP levels	-	Observed after the last stimulation: suppressed the elevation of GFAP and astrocyticCB2R. The study presents evidencethat the modulationof CB2R is a potential mechanism for the inhibitionof astrogliosis byrTMS. Improve motor symptoms.
Chalfouh et al., 2020	rTSMS	10	0.4	Figure-of-eight commercialcoil, 14 sessions with anesthesia startingthe day after lesion	Male and female C57BL/6 mice and 18hFoxJ1-CreERT2 mice (juvenile, adult, and aged) -surgical spinal cord transection	SCI -	-	↓ Iba-1+ cells	↑ GFAP+ cells	↓ Demyelinati on	Observed on thelast day of stimulation or 76days after the last stimulation: Decreased glial scar formation, decreasing fibrosis, increased proliferation of spinalcord stem cells, decreased demyelination, and improved motor function. Decreased Iba-1+ cells and induced ependymal cells todifferentiate into astrocytes and oligodendrocytes.
Delarue et al., 2021	rTSMS	10	0.4	Figure-of-eight commercialcoil, 14 sessions, startingthe day after lesion with anesthesia	Female C57BL/6 mice and Luciferase mice -surgical spinal cord transection	SCI -	-	↓ Iba-1+ cells	↑ GFAP+ cells	↓ Demyelinati on	Observed 45 days after the last stimulation: Reduced fibrosis, demyelination, and Iba-1+ cells. Increased the astroglial component of the scar and enhanced functional recovery.
Feng et al., 2021	rTMS	10	Information not available	Commercialcoil, 5 sessions per week for 4 weeks starting the day after lesion	Female SD rats (8 weeks old) -spinal cord contusion injury bythe weight-drop method	SCI -	-	-	↓ GFAP levels	-	Observed on thelast day of stimulation: Enhanced locomotor functional recovery and inhibited astrocyte reactivity.
Kim et al., 2013	rTMS	25	0.2	Custom built round coil, 5 sessions per week for 8 weeks, starting4 days after lesion	Female SD rats (adult) -spinal cord contusion injury by the weight-drop method	SCI -	-	↓ Iba-1+ cells	↓ GFAP+ cells	-	Observed on thelast day of stimulation: Reduced expression of GFAP and reduced Iba-1 immunoreactive cells. rTMS ameliorated cold allodynia and produced motor recovery.
Li et al., 2010	rTMS	5; 10;20	1.52	Figure-of-eight commercialcoil, 14 sessions starting the day after lesion	Male SD rats (adult) – EB -induced spinal cord injury	SCI -	-	-	↑ GFAP levels	-	Observed on thelast day of stimulation: Decreased lesion volumes and increased GFAP and ERK1/2expression.
Robac et al., 2021	rTSMS	10	0.4	Figure-of-eight commercialcoil, 14 sessions with anesthesia startingthe day after lesion	Female SD rats (8 – 10weeks) -surgical spinal cord transection	SCI -	-	↓ Iba-1+ cells	↓ GFAP+ cells	-	Observed on thelast day of stimulation or 45days after the last stimulation: Enhanced locomotor functional recovery Decreased glial scar formation, decreasing fibrosis. Decreased Iba-1+ cells.

5-HIAA: 5-Hydroxyindole-3-acetic acid; 5-HT: 5-hydroxytryptamine; AD: Alzheimer’s disease; Arg1: arginase 1; Aβ: amyloid beta; BDNF: brain-derived neurotrophic factor; CB1R: cannabinoid type 1 receptor; CCI: chronic constriction injury; CP: chronic pain; CUS: chronic unpredictable stress; DA: Dark Agouti; DAGLα: diacylglycerol lipase alfa; EAE: experimental autoimmune encephalomyelitis; EB: ethidium bromide; ERK: extracellular signal-regulated kinase; GDNF: glial cell line-derived neurotrophic factor; GFAP: glial fibrillary acidic protein; HF-rMS: high-frequency repetitive magnetic stimulation; HF-rTMS: high-frequency repetitive transcranial magnetic stimulation; HS: hemorrhagic stroke; Iba-1: ionized calcium binding adaptor molecule 1; ICH: intracerebral hemorrhage; iNOS: inducible nitric oxide synthase; IS: ischemic stroke; MAO-A: monoamine oxidase A; max: maximum output of the stimulator; MCAO: middle cerebral artery occlusion; MDD: major depressive disorder; MMP9: matrix metallopeptidase 9; MS: multiple sclerosis; NE: norepinephrine; NSCs: neural stem cells; OC: organotypic cultures; OGD: oxygen and glucose deprivation; OS: oxidative stress; PCC: primary cell culture; PD: Parkinson’s disease; rMS: repetitive magnetic stimulation; rTMS: repetitive transcranial magnetic stimulation; rTSMS: repetitive trans-spinal magnetic stimulation; SCI: spinal cord injury; SD: Sprague-Dawley; Sirt1: sirtuin 1; TGF-β: transforming growth factor beta; TNF-α: tumor necrosis factor-α; TrkB: tropomycin receptor kinase B; VEGF: vascular endothelial growth factor. “-”: Not assessed.

Effects of high-frequency repetitive transcranial magnetic stimulation in unlesioned animals or cells

Throughout this section, we analyzed the publications using HF-rTMS in animals or cells without any lesions. Ten articles were included, among which five used in vitro models (Chan et al., 1999; Hausmann et al., 2001; Clarke et al., 2017b, 2021; Liu et al., 2020), and five used in vivo models (Fujiki and Steward, 1997; Post et al., 1999; Hausmann et al., 2000; Cullen et al., 2019; Zorzo et al., 2019). Three reports showed that HF-rTMS induces an increase in GFAP expression both in vivo (Fujiki and Steward, 1997; Post et al., 1999) and in vitro (Chan et al., 1999). Interestingly, it was observed that the increase in GFAP expression was transient, increasing a few days after stimulation, and returning to basal levels in the coming weeks, which indicated that rTMS can transiently activate GFAP expression. However, only GFAP expression was evaluated, and no detailed evaluation was made of other markers of astrocyte reactivity. Interestingly, when male C57BL/6J mice were exposed to HF-rTMS while anesthetized with ketamine, the GFAP levels remained unaltered, suggesting that ketamine blocks the effects of HF-rTMS (Fujiki and Steward, 1997). However, contrary to these results, no changes in GFAP⁺ cells or levels were induced by HF-rTMS in in vivo and in vitro models of organotypic cultures, and in an induced pluripotent stem cells (iPSCs) line (Hausmann et al., 2000, 2001; Zorzo et al., 2019; Liu et al., 2020). The ability of HF-rMS to affect astrocyte migration and proliferation was also studied, but no effects were observed (Clarke et al., 2017b). These data suggest that HF-rMS does not induce long-term astrocyte reactivity in unlesioned animals or cells.

Intracellular calcium is a major regulator of astrocytes, so, the capacity of HF-rMS to alter calcium levels of astrocytes was studied as a measure of astrocyte activity (Clarke et al., 2017b, 2021). Although Clarke et al. (2017b) did not observe differences in intracellular calcium levels, the same group reported that HF-rMS decreased the expression of stromal interaction molecule 1 (STIM1), calcium release-activated calcium modulator 3 (ORAI3), and calcium-activated potassium channel subunit beta-4, proteins involved in the regulation of intracellular calcium (Clarke et al., 2021).

Regarding the impact of HF-rTMS on microglia, Zorzo and collaborators using Wistar rats reported that HF-rTMS did not alter the expression of the microglia marker Iba-1 (Zorzo et al., 2019).

Concerning oligodendrocytes, exposure of iPSC cultures to HF-rMS resulted in an increase in oligodendrocyte proliferation demonstrated by increased oligodendrocyte transcription factor 2 transcription (Liu et al., 2020). On the contrary, Cullen et al. (2019) using Pdgfrα-CreERT2 transgenic mice did not observe any effect of HF-rTMS on the number of oligodendrocytes.

In summary, few articles explored the impact of HF-rTMS on non-neuronal cells in unlesioned models and the existing articles evaluated a small number of parameters (Figure 3), so no clear conclusions can be drawn. Astrocyte reactivity was the most studied effect, and although three articles noted an increase in GFAP in the first few days after stimulation, its levels returned to baseline, so it is considered to have transitory effects (Fujiki and Steward, 1997; Chan et al., 1999; Post et al., 1999). Therefore, the seven articles that evaluated this parameter found no long-term effects (Fujiki and Steward, 1997; Chan et al., 1999; Post et al., 1999; Hausmann et al., 2000, 2001; Zorzo et al., 2019; Liu et al., 2020). As for microglial reactivity (Zorzo et al., 2019) and release of pro-inflammatory cytokines (Clarke et al., 2021), only one article observed each of these variables, impeding any conclusion to be drawn. Two articles evaluated oligodendrocyte proliferation, with one article reporting increased oligodendrocyte proliferation (Liu et al., 2020) while the other article reported the absence of effects (Cullen et al., 2019). Additionally, the induction of oligodendrocyte progenitor cell differentiation and increased myelination by HF-rTMS were also reported (Liu et al., 2020).

Figure 3.

Distribution of the main results of studies that used HF-rTMS in unlesioned animals or cells and by a neurological condition.

Evaluation of long-lasting effects of astrocyte reactivity, microglia reactivity, myelin breakdown, and oligodendrocyte loss, and release of pro-inflammatory cytokines. AD: Alzheimer’s disease; CP: chronic pain; HF-rTMS: high-frequency repetitive transcranial magnetic stimulation; HS: hemorrhagic stroke; IS: ischemic stroke; MS: multiple sclerosis; NI: non-injured animals or cells; PD: Parkinson’s disease; SCI: spinal cord injury; TBI: traumatic brain injury.

Preclinical studies using high-frequency repetitive transcranial magnetic stimulation

This section includes 26 articles that evaluated the effects of HF-rTMS in different models of neurologic conditions. The effects of HF-rTMS were studied in different models of neurological diseases to explore the impact of this stimulation protocol on the markers or deficits characteristic of these diseases.

Neurodegenerative diseases

HF-rTMS improved cognitive deficits present in AD models (Li et al., 2021; Lin et al., 2021; Cao et al., 2022). In 5×FAD mice, the reduction of astrocytic reactivity induced by HF-rMS was demonstrated by a reduction in GFAP⁺ cells. It is known that in AD there is an accumulation of amyloid beta (Aβ) in the brain that induces glial reactivity. HF-rTMS treatment significantly increased the drainage efficiency of brain clearance pathways, including the glymphatic system in the brain parenchyma and the meningeal lymphatics, which lead to a reduction in Aβ deposits. This data suggests that by reducing Aβ, HF-rTMS also reduced astrocyte reactivity (Lin et al., 2021). Also, in 5×FAD and 3×Tg-AD model mice, HF-rTMS reduced microglial reactivity, observed by a decrease in Iba-1⁺ cells (Li et al., 2021; Lin et al., 2021; Cao et al., 2022). This effect was coupled with a decrease in Aβ deposits (Lin et al., 2021) and the activation of the PI3K/Akt signaling pathway, which plays an important role in the modulation of microglia in neurodegenerative diseases (Li et al., 2021; Cao et al., 2022). Moreover, the application of HF-rTMS in this murine model of AD promoted a reduction in the levels of pro-inflammatory cytokines including interleukin (IL)-6, IL-1β and tumor necrosis factor-alpha (TNF-α) (Li et al., 2021; Cao et al., 2022). In addition, rTMS promoted synaptic plasticity and reduce a neuronal loss (Cao et al., 2022). The authors observed that the diminished microglial reactivity and the decrease in the release of pro-inflammatory cytokines resulted in lessened cognitive impairment, as assessed by behavioral tests such as novel object recognition and the Morris Water Maze test (Li et al., 2021; Lin et al., 2021; Cao et al., 2022).

Effects on the levels of inflammatory cytokines and astrocyte reactivity were assessed in the 6-OHDA PD model. HF-rTMS reduced the pro-inflammatory cytokines IL-6, IL-1β, and TNF-α, and decreased GFAP levels. In addition, HF-rTMS was able to improve the motor symptoms (Kang et al., 2022).

In C57BL/6 mice, after cuprizone model HF-rTMS improved cognition and depressive-like behavior. On the other hand, HF-rTMS promotes an increase in transforming growth factor-β expression. HF-rTMS was also shown to prevent myelin breakdown and oligodendrocyte loss, facilitating axonal regrowth, neuronal survival, and locomotor recovery (Wang et al., 2021).

Ischemic and traumatic injuries

The impact of HF-rTMS on HS was only evaluated by one study in a male C57BL/6J mice intracerebral hemorrhage model. HF-rTMS reduced brain edema and alleviated functional neurological deficits, measured, respectively by the determination of brain water content and by the modified neurological severity score (MNSS), which assesses sensory, balance, reflexes, and movement (Cui et al., 2019). The authors showed that HF-rTMS reduced astrocyte GFAP⁺ cells, which indicates a reduction in astrocyte reactivity.

HF-rTMS applied to Sprague-Dawley rats that suffered middle cerebral artery occlusion (MCAO) reduced the infarct volume, and improved cognitive functions, assessed by the MNSS (Luo et al., 2017, 2022a; Caglayan et al., 2019; Hong et al., 2020; Chen et al., 2023). HF-rMS treatment of primary cortical cultures exposed to oxygen and glucose deprivation and Sprague-Dawley rats that suffered MCAO induced a decrease in C3 and inducible nitric oxide synthase levels and an increase of S100A10 and arginase 1 (Arg1), suggesting a conversion of astrocytes presenting A1 phenotype to the A2 phenotype (Hong et al., 2020). Regarding GFAP⁺ cells, three studies reported the absence of alterations, whereas Hong et al. (2020) reported a decrease in GFAP⁺ cells induced by HF-rTMS after MCAO model (Luo et al., 2017; Caglayan et al., 2019; Hong et al., 2020; Gava-Junior et al., 2022). Although HF-rMS prevents the neuronal death and neurite degeneration induced by oxygen and glucose deprivation and increased the number of cells expressing extracellular signal-regulated kinase 1/2 and c-Fos (Roque et al., 2021; Gava-Junior et al., 2022), and the number and intensity of synaptic puncta, these effects were only observed in neuron-astrocyte cultures, and not in neuron-enriched cultures (Roque et al., 2021). In addition, HF-rMS modulated the release of growth factors by astrocytes, namely glial cell line-derived neurotrophic factor, which protected neurons from ischemic injury (Gava-Junior et al., 2022), suggesting that astrocytes are essential for the beneficial effects induced by HF-rMS after ischemia (Roque et al., 2021; Gava-Junior et al., 2022). Hong et al. (2020) also reported that astrocytes were crucial for the beneficial effects of HF-rMS after ischemia and demonstrated that the culture medium from astrocytes that underwent HF-rMS was sufficient to decrease neuronal apoptosis. Therefore, these results demonstrate that HF-rMS can stimulate astrocytes to secrete mediators that promote neuronal recovery after ischemia. This study also reported that HF-rTMS decreased the levels of the pro-inflammatory mediator TNF-α and increased the anti-inflammatory mediator IL-10 in the medium, which may contribute to the observed beneficial effects (Hong et al., 2020). Moreover, using the MCAO model, four studies observed that HF-rTMS reduced the release of pro-inflammatory cytokines, as showed by decreased expression of pro-inflammatory cytokines such as IL-1β, TNF-α, and inducible nitric oxide synthase and decreased expression of pro-inflammatory enzyme matrix metallopeptidase 9, and a decrease in the number of Iba-1⁺ cells (Caglayan et al., 2019; Hong et al., 2022; Luo et al., 2022a; Chen et al., 2023), and also an increase of anti-inflammatory cytokines Arg-1, IL-4, IL-10, and transforming growth factor-β (Hong et al., 2022; Luo et al., 2022a; Chen et al., 2023). In addition, HF-rTMS significantly reduced the expression of CD16/32⁺ Iba-1⁺ cells (M1) and increased the expression of CD206⁺ Iba-1⁺ cells (M2), indicating that HF-rTMS may modulate the conversion of microglia from the M1 phenotype to the M2 phenotype (Hong et al., 2022; Luo et al., 2022a; Chen et al., 2023). Besides the effects already mentioned, HF-rTMS caused an upregulation of the vascular endothelial growth factor-A and -B, indicating angiogenesis, that in turn protected capillary integrity, observed by an increase of CD31⁺ cells (Luo et al., 2017; Caglayan et al., 2019).

In the spinal cord injury (SCI) models, HF-rTMS was also shown to decrease the lesion volume (Li et al., 2010), to improve motor recovery evidenced by higher Basso, Beattie, and Bresnahan scores (Feng et al., 2021) and by the locotronic test (Delarue et al., 2021; Robac et al., 2021) and to ameliorated cold-induced allodynia (Kim et al., 2013). The effects of HF-rTMS on GFAP levels reported in SCI models were not consistent. While three studies observed a reduction in GFAP⁺ cells or levels (Kim et al., 2013; Feng et al., 2021; Robac et al., 2021), three studies reported an increase (Li et al., 2010; Chalfouh et al., 2020; Delarue et al., 2021). However, both the increase and the decrease in GFAP⁺ cells or levels can be associated with beneficial effects. The increase in GFAP was correlated with increased glial scar formation and decreased fibrosis, which is beneficial since glial scar limits axonal regrowth and functional recovery after injury (Chalfouh et al., 2020). Furthermore, the increase in GFAP levels was related to increased migration of astrocytes to the spinal cord lesion area which may promote a rapid repair by decreasing the formation of the cavity after spinal cord injury as far as possible (Li et al., 2010; Delarue et al., 2021). Concerning microglia, a decrease in the number of Iba-1⁺ cells was reported (Kim et al., 2013; Chalfouh et al., 2020; Delarue et al., 2021; Robac et al., 2021). Interestingly, the attenuation of astrocyte and microglia reactivity by HF-rTMS was associated with pain reduction (Kim et al., 2013). Moreover, HF-rTMS was also reported to prevent myelin breakdown and oligodendrocyte loss, a common problem in SCI (Chalfouh et al., 2020; Delarue et al., 2021). Additionally, HF-rTMS reduced fibrosis, increased the proliferation of spinal cord stem cells, and induced their differentiation into astrocytes and oligodendrocytes, which increased the astrocytic component of the glial scar and reduced the histopathological injury (Chalfouh et al., 2020; Delarue et al., 2021; Feng et al., 2021). Raf/MEK/ERK signaling is involved in neuronal apoptosis, cellular proliferation and differentiation and is upregulated in SCI model rats, and HF-rTMS suppressed the activation of this signaling pathway (Feng et al., 2021).

Other neurological diseases

Chronic pain was studied in Sprague-Dawley rats using two different models, the chronic constriction injury model, and the sciatic nerve ligation. In both models, HF-rTMS induced a decrease in GFAP⁺ cells, suggestive of a reduction in astrocyte reactivity and proliferation (Shiiba et al., 2012; Yang et al., 2018). In addition, these studies described a decrease in chronic pain after HF-rTMS application (Shiiba et al., 2012; Yang et al., 2018). Interestingly, the alleviation of spontaneous pain and brush-evoked pain was negatively correlated with the number of GFAP/BrdU co-labeled astrocytes, suggesting that reactive astrocytes and their proliferation are involved in the maintenance of chronic pain (Yang et al., 2018). HF-rTMS increase 5-hydroxyanisole-3-acetic acid, which is the main metabolite of 5-hydroxytryptamine (5-HT). Therefore, this study suggests that HF-rTMS acts on the 5-HT pathway and may induce an analgesic effect by activating the descending inhibitory system. Moreover, chronic pain (CP) can often lead to depressive conditions and an increase in 5-HT release may prevent this outcome (Shiiba et al., 2012).

In a Sprague-Dawley rat chronic unpredictable stress model HF-rTMS reduced the depressive-like behaviors (Jiang et al., 2022; Yan et al., 2022; Zuo et al., 2022), increased 5-HT, dopamine, and norepinephrine levels, decreased the 5-hydroxyanisole-3-acetic acid and Sirtuin 1 (Sirt1) levels and monoamine oxidase A (MAO-A) activity (Peng et al., 2018). The authors hypothesized that the decreased expression of Sirt1/MAO-A in astrocytes is related to the antidepressant effects of HF-rMS since Sirt1 can interfere with serotonin levels in the brain and with deacetylated NHLH2, which is a transcription factor regulating MAO-A expression. Accordingly, overexpression of Sirt1 activates MAO-A and reduces 5-HT levels (Peng et al., 2018). In addition to improving depressive-like behaviors, HF-rTMS was also found to promote decreased astrocyte reactivity, showed by lowered GFAP levels (Jiang et al., 2022), and to improve the survival of astrocytes (Yan et al., 2022; Zuo et al., 2022). Furthermore, HF-rTMS reduced the levels of interleukin IL-6, IL-1β, and TNF-α, and not only markedly alleviated the activation of microglia, but induced a switch of microglia polarization from a pro-inflammatory M1 phenotype to an anti-inflammatory M2 phenotype as showed by the decrease of CD68, CD16/32, Iba-1 and an increase of CD206 (Zuo et al., 2022).

In summary, as can be seen in Figure 3, HF-rTMS seems to impact glial cells by decreasing markers of astrocyte reactivity (GFAP), a fact that was observed in seven different disease models. However, in IS three articles observed no effect (Luo et al., 2017; Caglayan et al., 2019; Gava-Junior et al., 2022), and only one observed a decrease in astrocyte reactivity (Hong et al., 2020). However, this same article besides noting a decrease in GFAP also observed modulation of the astrocytic phenotype from A1 to A2 (Hong et al., 2020). All articles that evaluated microglial reactivity and release of pro-inflammatory cytokines, observed a decrease in these parameters (Kim et al., 2013; Caglayan et al., 2019; Chalfouh et al., 2020; Delarue et al., 2021; Li et al., 2021; Lin et al., 2021; Robac et al., 2021; Cao et al., 2022; Hong et al., 2022; Luo et al., 2022a; Zuo et al., 2022; Chen et al., 2023). However, the reported decrease in microglial reactivity is based only on the evaluation of Iba-1 labeling and the authors did not use other markers that can elucidate which microglial phenotype was predominant after HF-rTMS. On the other hand, a decrease in the release of pro-inflammatory cytokines was observed after HF-rTMS, namely a decrease in pro-inflammatory cytokines and an increase in anti-inflammatory cytokines (Caglayan et al., 2019; Hong et al., 2020, 2022; Li et al., 2021; Wang et al., 2021; Cao et al., 2022; Kang et al., 2022; Luo et al., 2022a; Zuo et al., 2022; Chen et al., 2023). Oligodendrocyte proliferation and myelination were studied in two disease models, with two articles in SCI (Chalfouh et al., 2020; Delarue et al., 2021) and one article in MS (Wang et al., 2021) observing an increase.

Effects of low-frequency repetitive transcranial magnetic stimulation on non-neuronal cells

LF-rTMS is considered an inhibitory stimulation and has been associated with functional improvements in several neurological disorders, such as IS (Caglayan et al., 2019; Hong et al., 2020), SCI (Fang et al., 2010; Li et al., 2010), CP (Yang et al., 2018), PD (Kang et al., 2022), MDD (Peng et al., 2018; Xue et al., 2019) and traumatic brain injury (TBI) (Clarke et al., 2017a).

Fifteen articles that used an LF-rTMS protocol and analyzed its impact on non-neuronal cells were included in this section. Six studies used unlesioned animals or cells (described in Additional Table 3) and the impact of stimulation in different disease models was also analyzed (described in Additional Table 4).

Additional Table 3.

Studies focusing on the effects of LF-rTMS in unlesioned animals or cells

Studies	Stimulat ion	Frequency (Hz)	Intensity(T)	Stimulation protocol	Model	Releaseof inflammatory cytokines	Microglia reactivity	Astrocyte reactivity	Oligodendro cytes		Results
Clarke et al., 2021	rMS	1	0.018	Custom built round coil, 1 session	PCC -Astrocytes from cortex	↓ Atf3, Ednra, ICAM-1 -	-	-	-		Observed 5 hours after stimulation: Reduced the release of pro-inflammatory cytokines, observed by the reduction on Atf3, Ednra,and ICAM-1. Increased STIM1 and ORAI3 proteinlevels.
Clarke et al., 2017b	rMS	1	0.018	Custom built round coil, 1 session	PCC -Astrocytes from cortex	--	-	↓ BrdU/ GFAP+ cells	-		Observed when cells were confluent: Rise in intracellular calcium.
Hausmann et al., 2001	rMS	1	50%;75% of max	Figure-of-eight commercial coil, 1 session in DIV 14	OC from parietal cortex	--	-	No effects in GFAP+ cells	-		Observed 0, 1.5, 3,6, 14or 24 hours after stimulation: No change in GFAP levels.
Liu et al., 2020	rMS	1	4.8	Figure-of-eight commercial coil, 2 sessions per day, from DIV 12 to35	iPSCs line	--	-	No effects in GFAP+ cells	No effects in oligodendroc yte number		Observed on the last day of stimulation: iPSCs differentiated into neurons while glial cell markers (GFAP or Oligo2) were not detected,indicating that few glia cells were generated.
Fujiki and Steward,1997)	rTMS	0.1	1.63	Commercial coil, 1 session	Male C57BLr6J mice (60–90 days) -	-	-	No effects in GFAP levels		-	Observed 12, 24, 36, 48hours, 4 or 8 days after stimulation: No effects in GFAP levels..
Liebetanz et al., 2003	rTMS	1	7	Round commercial coil, 5 sessions	Male Wistar rats -	-	No effects in Iba-1+ cells	No effects in GFAP+ cells		-	Observed 48hours after the last session of stimulation: No changes in microglialor astrocyticreactivity after rTMS

Atf3: Activating transcription factor 3; Ednra: endothelin receptor type A; GFAP: glial fibrillary acidic protein; Iba-1: ionized calcium binding adaptor molecule 1; ICAM-1: intercellular adhesion molecule-1; iPSCs: induced pluripotent stem cells; LF-rMS: low-frequency repetitive magnetic stimulation; LF-rTMS: low-frequency repetitive transcranial magnetic stimulation; max: maximum output of the stimulator; OC: organotypic cultures; ORAI3: calcium releaseactivated calcium modulator 3; PCC: primary cell culture; rMS: repetitive magnetic stimulation; rTMS: repetitive transcranial magnetic stimulation; rTSMS: repetitive trans-spinal magnetic stimulation; STIM1: stromal interaction molecule 1. "-": Not assessed.

Additional Table 4.

Studies focusing on the effects of LF-rTMS in disease models

Studies	Frequency (Hz)	Intensity(T)	Stimulation protocol	Model	Neurolog ical condition s	Release of inflammatory cytokines	Microglia reactivity	Astrocyte reactivity	Oligodendro cytes	Results
Yang et al., 2018	1	40% of max	Does not indicate the coil, 10sessions starting 3 days after lesion	Male SD rats -sciatic nerve ligation	CP	-	-	↑ BrdU/ GFAP+ cells	-	Observed on the day after stimulation: Increased BrdU/ GFAP+ cells.
Caglayan et al., 2019	1	26% of max	Figure-of-eight commercial coil, stimulation on the day of reperfusion (a), within 3 days starting on the day of reperfusion (b), or 28 days starting3 days after reperfusion (c) with anesthesia	Male Balb/c mice (10 to12 weeks) -MCAO model	IS	No effects	-	Noeffects in BrdU/ GFAP+ cells	-	Observed 1 day (a, b) or 10 days (c) after the last stimulation: No effect on pro-inflammatory cytokines levels or astrocyte reactivity.
Hong et al.,2020	1	1.9	Roundcommercial coil, 2 sessions in vitro and 7 sessions in vivo starting the day after lesion	PCC-Astrocytes and neurons from cortex & male SD rats (adult) – MCAO model	IS	No effects	-	Noeffects in BrdU/ GFAP+ cells	-	Observed on the day after the last stimulation: No effect of LF-rTMS on astrocyte regulation.
Peng et al.,2018	1	0.84; 1.26	Figure-of-eight commercial coil, 2 sessions in vitro at DIV 10/14and 7 sessions in vivo starting 4 weeks after lesion	PCC-Astrocytes from prefrontalcortex & male SD rats (adult) -CUS model	MDD	-	-	-	-	Observed after stimulation in vitro and two days after the last stimulation in vivo: No effects on depressive like behaviors.
Xue et al., 2019	1	0.84; 1.26	Figure-of-eight commercial coil, 2 sessions in vitro at DIV 10/14and 7 sessions in vivo starting 4 weeks after lesion	PCC-Astrocytes from prefrontalcortex & male SD rats (adult) -CUS model	MDD	-	-	-	-	Observed after stimulation in vitro and two days after the last stimulation in vivo: No effects on depressive like behaviors.
Kang et al., 2022	1	20% of max	Circular commercialcoil for rodent use, daily startingat 1 week post injury for 4 weeks	Male SD rats (adult) – 6-OHDA model	PD	No effects	-	Noeffects in GFAP levels	-	Observed after thelast stimulation: improve motor symptoms.
Fang et al.,2010	1	0;0.76; 1.52; 1.9	Roundcommercial coil, 14 sessions starting the day after lesion	Male SD rats (adult) -EB-induced spinal cord injury	SCI	-	-	↑ GFAP levels	-	Observed on the last day of stimulation: Decreased lesion volume, increase number of GFAP and ERK1/2 suggestingincreased migratory capacity.
Li et al., 2010	1	1.52	Figure-of-eight commercial coil, 14 sessions starting the day after lesion	Male SD rats (adult) - EBinduced spinal cord injury	SCI	-	-	↑ GFAP levels	-	Observed on the last day of stimulation: Decreased lesion volume, the number of GFAP and ERK1/2 increased which suggests astrocyte migration and promotes the formation of the glial scar.
Clarke et al., 2017a	1	0.0054	Custom built roundcoil, stimulation for 2 weeks starting24 hours after lesion	Male and female C57BL/6 J mice (3 months old) -unilateral penetrating corticalstabinjury model	TBI	-	↑ and ↓ Iba-1+ cells	↑ and↓ GFAP+ cells	-	Observed 2 hours after last stimulation: Reduceddensityof GFAP+ and Iba-1+ in adultand aged females but increased density inadult and aged males. Greater densities of GFAP+ astrocytes, andIba-1+ microglia, in 18-month mice compared to 3-monthmice.

CP: Chronic pain; CUS: chronic unpredictable stress; DIV: days in vitro; EB: ethidium bromide; ERK: extracellular signal-regulated kinase; GFAP: glial fibrillary acidic protein; Iba-1: ionized calcium binding adaptor molecule 1; IS: ischemic stroke; LF-rTMS: low-frequency repetitive transcranial magnetic stimulation; max: maximum output of the stimulator; MCAO: middle cerebral artery occlusion; MDD: major depressive disorder; PCC: primary cell culture; rTMS: repetitive transcranial magnetic stimulation; SCI: spinal cord injury; SD: Sprague-Dawley; TBI: traumatic brain injury. "-": Not assessed.

Effects of low-frequency repetitive transcranial magnetic stimulation in unlesioned animals or cells

The ability of LF-rTMS to modulate the astrocyte phenotype was investigated in an astrocyte-enriched primary culture from the cortex (Clarke et al., 2017b), an iPSCs line (Liu et al., 2020), an organotypic culture from the parietal cortex (Hausmann et al., 2001), and in rats (Fujiki and Steward, 1997; Liebetanz et al., 2003). While data from astrocyte cultures demonstrated a decrease in GFAP⁺ cells exposed to LF-rMS (Clarke et al., 2017b), no effects were observed in the other studies (Fujiki and Steward, 1997; Hausmann et al., 2001; Liebetanz et al., 2003; Liu et al., 2020). Besides, LF-rMS lead to an increase in intracellular calcium levels which is indicative of increased astrocyte activity (Clarke et al., 2017b, 2021). Clarke et al. (2021) also reported that LF-rMS increased STIM1 and ORAI3 protein levels, proteins related to calcium signaling, involved in sensing, and restoring astrocytic endoplasmic reticulum calcium stores following depletion. Besides this, LF-rMS reduced the levels of pro-inflammatory molecules such as intercellular adhesion molecule-1, activating transcription factor 3, and endothelin receptor type A (Clarke et al., 2021).

In addition, the ability of LF-rMS to modulate microglial reactivity and the release of pro-inflammatory cytokines was also evaluated, and while it was observed that LF-rMS could decrease pro-inflammatory cytokines (Clarke et al., 2021), no effects were observed in microglial reactivity (Liebetanz et al., 2003). The proliferation of oligodendrocytes was also investigated, but oligodendrocyte transcription factor 2 was not detected indicating low oligodendrocytes proliferation (Liu et al., 2020).

The few articles that applied LF-rMS to glial cells in a basal state did not observe significative changes in these cells (Figure 4).

Figure 4.

Distribution of the main results of studies that used LF-rTMS in unlesioned animals or cells and by a neurological condition.

Evaluation of long-lasting effects of astrocyte reactivity, microglia reactivity, myelin breakdown and oligodendrocyte loss, and release of pro-inflammatory cytokines. CP: Chronic pain; IS: ischemic stroke; LF-rTMS: low-frequency repetitive transcranial magnetic stimulation; NI: non-injured animals or cells; PD: Parkinson’s disease.

Preclinical studies using low-frequency repetitive transcranial magnetic stimulation

Only nine articles analyzed the impact of LF-rTMS in non-neuronal components in different preclinical models. One article used a CP model (Yang et al., 2018), two articles used an MDD model (Peng et al., 2018; Xue et al., 2019), two articles used an IS model (Caglayan et al., 2019; Hong et al., 2020), one article used a PD model (Kang et al., 2022), two articles used an SCI model (Fang et al., 2010; Li et al., 2010), and one article used a TBI model (Clarke et al., 2017a).

The only study that assessed the action of LF-rTMS on the 6-OHDA PD model did not observe any changes in inflammatory cytokines or astrocyte reactivity. Nevertheless, the application of LF-rTMS was able to improve motor symptoms (Kang et al., 2022). Regarding ischemic and traumatic injuries, the modulation of astrocyte reactivity by LF-rTMS was assessed in two studies that used the MCAO model and both reported the absence of changes in GFAP⁺ cells (Caglayan et al., 2019; Hong et al., 2020). The release of pro-inflammatory cytokines was also evaluated, by the measurement of IL-1β, IL-10, TNF-α, transforming growth factor-β pro-inflammatory cytokines, and matrix metallopeptidase 9 pro-inflammatory enzyme after application of LF-rTMS both in vivo and in vitro in models of IS, but no significant changes were detected (Caglayan et al., 2019; Hong et al., 2020).

Using the ethidium bromide model of spinal cord demyelination, two studies reported histological recovery. The application of LF-rTMS after SCI was associated with a reduction in lesion volume. LF-rTMS also increased the GFAP levels, increases that were associated with the migration of astrocytes to the lesioned area (Fang et al., 2010; Li et al., 2010).

A unilateral penetrating cortical stab injury model of TBI showed both an increase or decrease in GFAP levels depending on the gender and age of the mice after LF-rTMS. An increase in GFAP⁺ cells and Iba-1⁺ cells induced by LF-rTMS was observed in male rats, whereas in female rats there was a decrease, suggesting that sex differences need to be taken into consideration in therapeutic rTMS protocols. Additionally, there was a greater density of GFAP and Iba-1 labeling surrounding the injury in 18-month male and female mice when compared to 3-month male and female mice which indicates that the effects of the protocols are dependent on the age of the animals (Clarke et al., 2017a).

Regarding other neurological diseases, in Sprague-Dawley rats with sciatic nerve ligation, LF-rTMS promoted an increase in GFAP⁺ cells, but evaluation of spontaneous pain and brush-evoked pain showed no effects of LF-rTMS on CP (Yang et al., 2018).

The effect of LF-rTMS was analyzed in the Sprague-Dawley chronic unpredictable stress model. However, no changes were detected in depressive-like behaviors, the 5-HT, dopamine, or norepinephrine levels, or in the endocannabinoid system (Peng et al., 2018; Xue et al., 2019).

In summary, few studies evaluated the effect of LF-rTMS in models of disease in which the participation of glial cells was assessed and only astrocyte reactivity and the release of pro-inflammatory cytokines have been analyzed. In CP an increase in astrocyte reactivity was reported (Yang et al., 2018) and in IS no effects were observed after this stimulation protocol (Caglayan et al., 2019; Hong et al., 2020). Interestingly, all articles that evaluated astrocyte reactivity after LF-rTMS used males, except one that also evaluated this parameter in females (Clarke et al., 2017a). This study noted a decrease in astrocyte reactivity after LF-rTMS only in females, which may be an indication that the effects of LF-rTMS are dependent on the sex of the animal (Clarke et al., 2017a). However, the number of studies per disease present in this section is not sufficient to draw any conclusions. No effects were observed regarding the release of pro-inflammatory cytokines.

Effect of theta-burst stimulation on non-neuronal cells

TBS is an rTMS protocol capable of inducing several physiological changes in short periods (Liu et al., 2020; Dragic et al., 2021). While cTBS suppress excitatory synaptic transmission, inducing LTD (Stevanovic et al., 2019), iTBS potentiates excitatory synaptic transmission inducing LTP (Mancic et al., 2016; Cacace et al., 2017; Stevanovic et al., 2019). The application of this technique has been associated with several functional improvements, both motor and cognitive, raising the hypothesis that it can be used as a therapeutic approach in several neurological disorders such as PD, AD, IS, and MDD (Mancic et al., 2016; Dragic et al., 2021). Sixteen articles using TBS protocols analyzed the effects on non-neuronal cells. Five studies were done in control situations, in unlesioned animals or cells (described in Additional Table 5), and eleven studies examined the impact of stimulation in different disease models, being MS models the most studied (described in Additional Table 6).

Additional Table 5.

Studies focusing on the effects of TBS in unlesioned animals or cells

Studies	Stimulation	Main Frequency (Hz)	Intensit y (T)	Stimulation protocol	Model	Releaseof inflammator y cytokines	Microglia reactivity	Astrocyte reactivity	Oligodendrocytes	Results
Clarke et al., 2017b	cTBS	5	0.018	Custom built round coil, 1 session	PCC -Astrocytes from cortex	-	-	-	-	Observed when cells were confluent: No effect on astrocyte migrationor proliferation.
Liu et al., 2020	iTBS	5	2.4	Figure-of-eight commercialcoil, 2 sessions per day, from DIV 12 to35	iPSCs line	-	-	Noeffects in GFAP+ cells	-	Observed on the last day of stimulation: Increased formation of mature neurons and synapseformation. No changes inGFAP levels.
Cullen et al., 2021	iTBS	Information not available	0.120	Custom built round coil, 7, 14, and 28 sessions	Male andfemale transgenic mice C57BL/6J (adult) -cre-lox lineage tracing	-	-	-	↓ Demyelination	Observed 7, 14, or 28 days after lesion: Shortens nodes of Ranvier and increased the periaxonal space. Observed 7, 14, or 28 days after lesion: iTBS
Cullen et al., 2019	iTBS and cTBS	5	0.120	Custom built round coil28 sessions	Male andfemale transgenic mice C57BL/6J (adult) -cre-lox lineage tracing	-	-	-	↑ Myelination ↑ proliferation	increased myelin internodelength, increased number of newborn oligodendrocytes, and oligodendrocyte survival. Decreased apoptoticcells and enhanced myelination.
Mancic et al., 2016	iTBS and cTBS	5	33% of max	Figure-of-eight commercialcoil, 1 session (a) or 10 sessions (b) in2 weeks	Male Wistar rats (4 weeks old)	-	-	Noeffects in GFAP levels	-	Observed 24hours (a) or 48hours (b) after the last stimulation: Increased expression of GLT-1 induced byiTBS and decreased expression of vGluT1 after cTBS. Unchanged GFAP expression.

cTBS: Continuous theta-burst stimulation; GFAP: glial fibrillary acidic protein; GLT-1: glutamate transporter-1; iPSCs: induced pluripotent stem cells; iTBS: intermittent theta-burst stimulation; max: maximum output of the stimulator; PCC: primary cell culture; vGluT1: type I vesicular glutamate transporter. “-”: Not assessed.

Effects of theta-burst stimulation in unlesioned animals or cells

Only five articles focused on the effect of TBS in unlesioned animals or cells, with two using in vitro models (Clarke et al., 2017b; Liu et al., 2020) and three using an in vivo model (Mancic et al., 2016; Cullen et al., 2019, 2021). Of these, two papers used the iTBS protocol (Liu et al., 2020; Cullen et al., 2021), one used the cTBS protocol (Clarke et al., 2017b), and two used both protocols (Mancic et al., 2016; Cullen et al., 2019).

Data from an iPSCs line demonstrated that iTBS promotes the formation of mature neurons from iPSCs and synapse formation, however, iTBS did not affect astrocyte proliferation or altered GFAP⁺ cells (Liu et al., 2020).

Concerning cTBS, its effect was studied in a primary astrocyte culture and no effect on astrocyte migration or proliferation was observed (Clarke et al., 2017b). Mancic et al. (2016) observed, in Wistar rats, that a single session of iTBS increases type I glutamate transporter expression, which could indicate a greater efficiency in glutamate uptake leading to a reduction in excitatory transmission, whereas multiple sessions of cTBS cause a decrease in type I vesicular glutamate transporter expression which suggests reduced glutamate levels at excitatory synaptic terminals. The same study reported that none of the TBS protocols altered the expression of GFAP, suggesting an absence of effects on astrocytic reactivity (Mancic et al., 2016).

The effects of both TBS protocols on oligodendrocytes were studied in Plp-CreER:Tau-mGFP (Cullen et al., 2021) and Pdgfrα-CreERT2 transgenic mice (Cullen et al., 2019). iTBS, but no cTBS, increased the number of newly formed oligodendrocytes (Cullen et al., 2019). iTBS also reduced the number of apoptotic cells, having a more pronounced effect under the circumference of the coil, where the induction is strongest (Cullen et al., 2019). iTBS also increased the length of the internodes and enhanced myelination (Cullen et al., 2019). The same group showed that iTBS reduced the length of the nodes of Ranvier and increased the periaxonal space, which affects the transmission of action potentials. This effect was not associated with a pathological event but rather with a physiological adaptation to altered neuronal activity (Cullen et al., 2021). Based on the reported results, the authors argued that iTBS could be used to noninvasively promote myelin formation, having the potential to be used for the treatment of demyelinating diseases such as MS (Cullen et al., 2019).

In summary, none of the reports that analyzed the effect of cTBS or iTBS on astrocytes in basal conditions reported effects on astrocyte reactivity, although an in vivo study detected changes in glutamate transporters (Mancic et al., 2016). Microglial reactivity, the release of pro-inflammatory cytokines, and two studies demonstrated enhanced myelination (Cullen et al., 2019, 2021).

Preclinical studies using theta-burst stimulation

Eleven articles related to preclinical models of several neurological conditions such as AD (Stanojevic et al., 2022; Stekic et al., 2022), cerebral small vessel disease (Cai et al., 2020), IS (Luo et al., 2017, 2022b), MS (Stevanovic et al., 2019; Dragic et al., 2020, 2021), PD (Cacace et al., 2017), pneumococcal meningitis (PM) (Muri et al., 2020), and TBI (Sasso et al., 2016) analyzed the impact of TBS on non-neuronal cells. One article used cTBS (Dragic et al., 2021), seven articles used iTBS (Sasso et al., 2016; Cacace et al., 2017; Luo et al., 2017, 2022b; Cai et al., 2020; Stanojevic et al., 2022; Stekic et al., 2022), and three articles used both protocols (Stevanovic et al., 2019; Dragic et al., 2020; Muri et al., 2020).

Neurodegenerative diseases

In the intracerebroventricularly injected streptozotocin model (Stanojevic et al., 2022) and in the trimethyltin model of AD (Stekic et al., 2022), the authors found that iTBS treatment attenuates cognitive impairment, assessed by radial arm maze test and the object recognition test, an effect that could be mediated via PI3K/Akt/mTOR signaling pathway. iTBS also reduced hyperactivity, aggressive behavior, anxiety, and tremor (Stanojevic et al., 2022; Stekic et al., 2022). After iTBS treatment, a reduction in the intensity of Iba-1 staining could be observed in all brain regions accompanied by reversion of the microglia morphology towards resting phenotype. Furthermore, one study demonstrated that iTBS decreased pro-inflammatory cytokines IL-1β and increased anti-inflammatory cytokines IL-10, which was related to the reversal of the M1 phenotype to the M2 phenotype (Stekic et al., 2022). In addition, a significant reduction in GFAP⁺ cells was observed in one study (Stanojevic et al., 2022) while other study did not observe any changes (Stekic et al., 2022).

In the 6-hydroxydopamine model of PD, iTBS promoted functional recovery by increasing the levels of neurotransmitters such as dopamine in the brain (Cacace et al., 2017). In addition to the mentioned effects, a significant reduction in GFAP⁺ cells and Iba-1⁺ cells was reported 80 minutes post-iTBS (Cacace et al., 2017).

In the experimental autoimmune encephalomyelitis model, both protocols of TBS decreased GFAP⁺ cells or levels (Stevanovic et al., 2019; Dragic et al., 2020, 2021), which correlated with decreased expression of brain-derived neurotrophic factor and Ki67 (Stevanovic et al., 2019). Moreover, three studies reported that both protocols of TBS reduced the number of Iba-1⁺ cells (Stevanovic et al., 2019; Dragic et al., 2020, 2021). cTBS also reduced CD73 and IL-1β levels and increased IL-10 in the experimental autoimmune encephalomyelitis model, which is indicative of reduced neuroinflammation (Dragic et al., 2021). Dragic et al. (2020) also reported that both TBS protocols lead to a decrease in myelin loss. Furthermore, both protocols of TBS could increase the levels of neurotransmitters in the brain such as GABA (Stevanovic et al., 2019).

Ischemic and traumatic injuries

iTBS promoted functional recovery in a rat MCAO model by enhancing neurogenesis and activating brain-derived neurotrophic factor/tropomyosin receptor kinase B signaling pathway. This functional recovery was demonstrated by the improvements in the neurological severity score (NSS) test and by the MNSS. The authors also demonstrated that iTBS reduced the infarct volume; however, there were no reported changes in GFAP⁺ cells (Luo et al., 2017). Another study demonstrated, also in the MCAO model, that iTBS decreased pro-inflammatory cytokines IL-1β, IL-17a, TNF-α, and interferon-γ and increased anti-inflammatory cytokines IL-10, a change that can be associated with a switch of the M1 phenotype to the M2 phenotype. This change in the microglial phenotype was also supported by a decrease in cells labeled for CD86 and inducible nitric oxide synthase and an increase in cells labeled for CD206 and Arg1. Furthermore, this study demonstrated that the functional improvements promoted by iTBS were microglia-dependent (Luo et al., 2022b).

iTBS promoted functional recovery in this TBI model, demonstrated by improvements in the NSS test and the MNSS. The reduction of GFAP and Iba-1 levels was observed in a model of TBI after iTBS induced by the removal of the right cerebellar hemisphere (Sasso et al., 2016).

Other neurological diseases

In a Sprague-Dawley rat chronic hypertension-induced cerebral small vessel disease, iTBS caused a reduction in the number of GFAP and Iba-1⁺ cells in the hippocampus. iTBS also promoted the expression of N-methyl D-aspartate receptor subtype 2B, Ca²⁺/calmodulin-dependent protein kinase II, and glutamate receptor 1, changes that were associated with an improvement in spatial memory assessed by the Morris Water Maze test (Cai et al., 2020).

Muri et al. (2020), using an inflammatory model of PM induced by the infection with S. pneumoniae both in vivo and in vitro, reported an increase in S100b, S100a4, and in GFAP genes after both protocols of TBS. Contrary to the majority of results that have been described so far, Muri et al. (2020) reported that both protocols of TBS induce the polarization of microglia towards an inflammatory phenotype, as there was an increase in the expression of genes associated with microglial reactivity (Iba-1, CD14, CD45, CD68, CD84, CD86, F4/80) and a reduction in the expression of genes associated with resting microglia (CD163, Arg1, Cd206, Ym1, Fizz1). This was accompanied by increased release of pro-inflammatory cytokines such as IL-1β, IL-6, IL-10, and TNF-α after TBS treatment. Together, the data indicate that cTBS intensified neuroinflammation after PM (Muri et al., 2020).

iTBS, as can be seen in Figure 5A, decreased astrocyte and microglial reactivity in cerebral small vessel disease, MS, PD, and TBI models, with opposite effects in PM, which is probably related to the inflammatory characteristics of this model. Similarly, cTBS, as can be seen in Figure 5B, promoted gliosis and the release of pro-inflammatory cytokines in PM (Muri et al., 2020), whereas it led to a decrease of these parameters in the MS model (Dragic et al., 2021).

Figure 5.

Distribution of the main results of studies that used iTBS (A) and cTBS (B) in unlesioned animals or cells and under a neurological condition.

Evaluation of long-lasting effects of astrocyte reactivity, microglia reactivity, myelin breakdown and oligodendrocyte loss, and release of pro-inflammatory cytokines. AD: Alzheimer’s disease; CSVD: cerebral small vessel disease; cTBS: continuous theta-burst stimulation; IS: ischemic stroke; iTBS: intermittent theta-burst stimulation; MS: multiple sclerosis; NI: non-injured animals or cells; PD: Parkinson’s disease; PM: pneumococcal meningitis; TBI: traumatic brain injury.

The importance of male-specific or female-specific factors is increasingly being recognized as crucial in biomedical research

As both the lesion and the effects of rTMS could be influenced by sex, it is important to assess whether the studies analyzed the effects in males or females. Regarding the HF-rTMS protocol, of the twenty-four studies that used animals, fifteen used males, five used females, and four used both sexes. As for the LF-rMS protocol, nine studies used animals in their experimental method, eight studies used males and only one study used both sexes (Additional Table 7). Regarding the iTBS protocol, of the ten studies that used animals, eight used males and two used females. As for the cTBS protocol, four studies used animals, one study used males and three used females. It is also noteworthy that of all the protocols, the studies that only used females were just performed in the MS or SCI model. However, although there were studies that used both sexes, they did not discriminate the effects in each sex except for one study (Clarke et al., 2017a). Thus, an important limitation in the existing studies is that they do not consider the possible differences in the effects of rMS between genders.

Additional Table 7.

Studies of all protocols discriminating the sex of the animals used

Stimulation	Article	Neurological conditions	Male	Female	Separate analysis
HF-rMS	Li et al., 2021; Lin et al., 2021; Caoet al., 2022	AD	✔	✔	-
	Shiibaet al., 2012; Yanget al., 2018	CP	✔	-	-
	Cui et al., 2019	HS	✔	-	-
	Luo et al.,2017, 2022a; Caglayan et al., 2019;Hong et al., 2020;Hong et al., 2022;Chen et al., 2023	IS	✔	-	-
	Peng et al., 2018; Jiang et al., 2022; Yan et al., 2022; Zuo et al., 2022	MDD	✔	-	-
	Wang et al., 2021	MS	-	✔	-
	Kang et al., 2022	PD	✔	-	-
	Chalfouhet al., 2020	SCI	✔	✔	-
	Kim et al., 2013;Delarue et al., 2021, Feng et al., 2021; Robac et al., 2021	SCI	-	✔	-
	Li et al., 2010	SCI	✔	-	-
LF-rMS	Yang et al., 2018	CP	✔	-	-
	Caglayan et al., 2019;Hong et al., 2020	IS	✔	-	-
	Peng et al., 2018; Xue et al.,2019	MDD	✔	-	-
	Kang et al., 2022	PD	✔	-	-
	Fang et al., 2010; Li et al., 2010	SCI	✔	-	-
	Clarke et al., 2017a	TBI	✔	✔	✔
iTBS	Stanojevic et al., 2022; Stekicet al., 2022	AD	✔	-	-
	Cai et al., 2020	CSVD	✔	-	-
	Luo et al.,2017, 2022b	IS	✔	-	-
	Stevanovicet al., 2019; Dragic et al., 2020	MS	-	✔	-
	Cacace et al.,2017	PD	✔	-	-
	Muri et al., 2020	PM	✔	-	-
	Sassoet al., 2016	TBI	✔	-	-
cTBS	Stevanovicet al., 2019; Dragic et al., 2020,2021	MS	-	✔	-
	Muri et al., 2020	PM	✔	-	-

AD: Alzheimer’s disease; CP: chronic pain; CSVD: cerebral small vessel disease; cTBS: continuous theta-burst stimulation; HF-rMS: high-frequency repetitive magnetic stimulation; HS: hemorrhagic stroke; IS: ischemic stroke; iTBS: intermittent theta-burst stimulation; MDD: major depressive disorder; MS: multiple sclerosis; PD: Parkinson’s disease; PM: pneumococcal meningitis; SCI: spinal cord injury; TBI: traumatic brain injury. “-”: Not assessed.

Discussion

In this systematic review, the effects of several rTMS protocols on non-neuronal cells reported in 39 articles were analyzed. All the articles included in this analysis used at least one magnetic stimulation protocol (HF-rTMS, LF-rTMS, cTBS, iTBS), and twelve articles used more than one protocol.

Not only are the protocols varied, but they were also applied in different situations, animals, or cells without injury, and to various models of neurological diseases. Regarding the effects of rTMS in unlesioned animals or cells, one study reported that LF-rTMS induced a decrease in GFAP levels (Clarke et al., 2017b), HF-rTMS and LF-rTMS led to a reduction in the release of pro-inflammatory cytokines (Clarke et al., 2021), and HF-rTMS promoted myelination (Liu et al., 2020). However, it should be emphasized that most studies report the absence of rTMS effects in glial cells that were not subjected to adverse stimuli. Nevertheless, this lack of effect is expected since the cells and/or animals are in basal conditions and changes of astrocyte and microglial reactivity, on the levels of inflammation or oligodendrocyte differentiation/proliferation are not expected to occur on a significant level.

The HF-rMS protocol was the most widely used and was the one with the most consensual results, leading in the majority of the disease models to lessened disease markers or deficits, whether by decreasing astrocyte and/or microglial reactivity or by increasing myelination and oligodendrocyte proliferation. Therefore, by targeting glial cells, the HF-rTMS protocol seems to be a promising option for the control of glial cell dysfunction occurring in several neurological disorders. The LF-rMS protocol showed the least effect, leading only to decreased GFAP levels in CP and SCI models, however, this protocol was also less used and in fewer disease models. It is also worth noting that a study with the IS model that used both the HF-rMS and LF-rMS protocols observed no effects, but this is most likely due to the injury extent (60%) which can be too large for the therapy to induce significant effects. Furthermore, the intensity of the stimulation used was relatively low (26% of the maximum capacity of the stimulator) which may also have led to the lack of effects (Caglayan et al., 2019). Both TBS protocols led to decreased astrocyte and/or microglial reactivity or increased myelination in all diseases except in the PM model. In this model, which is characterized by an excessive inflammatory reaction associated with blood-brain barrier breakdown, increased intracranial pressure, hydrocephalus, and cerebral ischemia (Mook-Kanamori et al., 2011), both TBS protocols increased astrocyte and microglial reactivity and increased pro-inflammatory cytokine levels, which is probably related to the specificities of this model and its highly inflammatory characteristics (Muri et al., 2020).

The articles analyzed showed great variability both in the protocols applied and in the assessment of the effects. Regarding the coils used, of the 52 articles reviewed, 49.3% used commercial figure-of-eight coils, 15.5% used commercial circular coils, 18.3% used custom-built round coils, and 16.9% did not indicate which coil was used. The number of stimulation sessions applied was another parameter where there is a large variability. Of the articles reviewed, 19.4% performed only one stimulation session and 80.6% performed multiple stimulation sessions (2–28 sessions). Moreover, the onset of the sessions ranged from 1 day to 12 weeks. 28.6% of the studies perform rMS within 1 day of injury, 44.3% of the studies perform rMS between 1 day and 12 weeks after injury, and 27.1% of the studies did not indicate when the stimulation was initiated. In what concerns to the evaluation of the rMS effects, 67.3% of the studies performed this evaluation up to 1 day after the last rMS protocol, and 32.7% of the studies performed this evaluation between 1 day and 11 weeks after the last rMS. These data show that most studies do not assess the actions of rMS in the long term. This analysis highlights the wide diversity of stimulation protocols used, which associated with the reduced number of studies, significantly limits the conclusions that can be drawn.

Concerning the use of models, although animal model studies have the great advantage of including the interactions between the different cells and tissues, cell culture studies permit the evaluation of the direct effects of rMS on each cell type. From the analysis of the publications included in this review, 77.9% of the studies were carried out in animal models, and only 22.1% used cell cultures. It is not possible to conclude whether the effects observed in vivo result from a direct action on cells, or interactions between various components of the tissues since there are very few studies in cellular models. It should also be noted that, to our knowledge, no studies analyze the direct effect of rMS in cellular cultures of microglia, oligodendrocytes, or vascular cells, so there is no data on the direct impact of rMS protocols in these cells.

On the other hand, the vast majority of published studies on all protocols use males, or those that use both sexes do not analyze the effects on each sex separately, not allowing us to conclude whether there are differential effects of rTMS regarding sex, which can be a limitation on these studies.

In summary, of the articles included in this review, most studies were conducted in male rodents, and the most frequently used protocol, with the most effects, was HF-rTMS. The most studied disease models were IS (8 HF-rMS, 2 LF-rMS, 2 iTBS) and SCI (6 HF-rMS, 2 LF-rMS) with LF-rMS being the protocol that seems to have the least effects. Nevertheless, it is not possible to draw conclusions about which protocol is better because these are too variable. However, it is noteworthy that magnetic stimulation had beneficial effects on different diseases although they have different causes since some are acute lesions and others are degenerative diseases associated with glial changes. It is also remarkable that the total number of articles that have analyzed the impact of rMS on non-neuronal cells is small and that most of these studies did not perform an in-depth analysis of the impact of the protocols on non-neuronal cells, evaluating only a small number of parameters, leading to insufficient knowledge on the response of these cells to magnetic stimulation. Furthermore, the high diversity of protocols makes the comparison between the various studies challenging.

Nonetheless, the existing data support a relevant contribution of glial and vascular cells and indicate that these cells are important intermediaries to the improvements observed after rTMS in models of neurological conditions. These results are of particular interest for the treatment of diseases in which astrocyte and microglial reactivity, neuroinflammation, or loss of oligodendrocytes lead to the worsening of the patient’s condition.

Additional files:

Additional file 1: Open peer review reports 1 ^{(86.9KB, pdf)} and 2 ^{(86.8KB, pdf)} .

Additional Table 1: Studies focusing on the effects of HF-rTMS in unlesioned animals or cells.

Additional Table 2: Studies focusing on the effects of HF-rTMS in disease models.

Additional Table 3: Studies focusing on the effects of LF-rTMS in unlesioned animals or cells.

Additional Table 4: Studies focusing on the effects of LF-rTMS in disease models.

Additional Table 5: Studies focusing on the effects of TBS in unlesioned animals or cells.

Additional Table 6: Studies focusing on the effects of TBS in disease models.

Additional Table 7: Studies of all protocols discriminating the sex of the animals used.