Electrical stimulation of the dorsal motor nucleus of the vagus in male mice can regulate inflammation without affecting the heart rate

Abstract

Background:

The vagus nerve plays an important role in neuroimmune interactions and in the regulation of inflammation. A major source of efferent vagus nerve fibers that contribute to the regulation of inflammation is the brainstem dorsal motor nucleus of the vagus (DMN) as recently shown using optogenetics. In contrast to optogenetics, electrical neuromodulation has broad therapeutic implications. However, the anti-inflammatory effectiveness of electrical stimulation of the DMN (eDMNS) and the possible heart rate (HR) alterations associated with this approach have not been investigated. Here, we examined the effects of eDMNS on HR and cytokine levels in mice administered with lipopolysaccharide (LPS, endotoxin) and in mice subjected to cecal ligation and puncture (CLP) sepsis.

Methods:

Anesthetized male 8–10-week-old C57BL/6 mice on a stereotaxic frame were subjected to eDMNS using a concentric bipolar electrode inserted into the left or right DMN or sham stimulation. eDMNS (500, 250 or 50 μA at 30 Hz, for 1 min) was performed and HR recorded. In endotoxemia experiments, sham or eDMNS utilizing 250 μA or 50 μA was performed for 5 mins and was followed by LPS (0.5 mg/kg) i.p. administration. eDMNS was also applied in mice with cervical unilateral vagotomy or sham operation. In CLP experiments sham or left eDMNS was performed immediately post CLP. Cytokines and corticosterone were analyzed 90 mins after LPS administration or 24h after CLP. CLP survival was monitored for 14 days.

Results:

Either left or right eDMNS at 500 μA and 250 μA decreased HR, compared with baseline pre-stimulation. This effect was not observed at 50 μA. Left side eDMNS at 50 μA, compared with sham stimulation, significantly decreased serum and splenic levels of the pro-inflammatory cytokine TNF and increased serum levels of the anti-inflammatory cytokine IL-10 during endotoxemia. The anti-inflammatory effect of eDMNS was abrogated in mice with unilateral vagotomy and was not associated with serum corticosterone alterations. Right side eDMNS in endotoxemic mice suppressed serum TNF and increased serum IL-10 levels but had no effects on splenic cytokines. In mice with CLP, left side eDMNS suppressed serum IL-6, as well as splenic IL-6 and increased splenic IL-10 and significantly improved the survival rate of CLP mice.

Conclusions:

For the first time we show that a regimen of eDMNS which does not cause bradycardia alleviates LPS-induced inflammation. These eDMNS anti-inflammatory effects require an intact vagus nerve and are not associated with corticosteroid alterations. eDMNS also decreases inflammation and improves survival in a model of polymicrobial sepsis. These findings are of interest for further studies exploring bioelectronic anti-inflammatory approaches targeting the brainstem DMN.

1. Introduction

Inflammation is a complex, protective response that is triggered by a variety of stimuli and conditions, including infection and injury (1–3). Precisely controlled, localized and timely-resolved inflammation results in restoration of physiological homeostasis (2–4). However, dysregulated inflammatory responses and excessive or non-resolved inflammation are implicated in the pathogenesis of a broad spectrum of diseases (5, 6). Discoveries during the last 20 years have revealed the key role of the vagus nerve and the vagus nerve based inflammatory reflex in the regulation of inflammation (1, 7, 8). The inflammatory reflex is a physiological mechanism in which afferent (sensory) vagus neurons detecting increased peripheral cytokine levels are integrated in the brainstem medulla oblongata with efferent (motor) vagus neurons, which control cytokine levels and inflammation (1, 9). Electrical vagus nerve stimulation of the cervical portion of the vagus nerve has been instrumental in studying the inflammatory reflex. This approach has been shown to significantly suppress serum TNF and other pro-inflammatory cytokine levels in murine endotoxemia and other inflammatory conditions (7, 10). Electrical vagus nerve stimulation has also been successfully explored in treating human inflammatory disorders, including rheumatoid arthritis (11, 12), and inflammatory bowel disease (13–15), and this research has been instrumental for establishing the field of Bioelectronic Medicine (10, 16).

In contrast to this progress, our understanding of the brain mechanisms that govern the inflammatory reflex and vagus nerve anti-inflammatory cholinergic signalling remains limited (17–19). During inflammation, increased levels of peripheral pro-inflammatory cytokine levels trigger activation of neural signaling along afferent vagus neurons, which terminate in the brainstem nucleus tractus solitarius (1, 20). This information is communicated to the adjacent dorsal motor nucleus of the vagus (DMN) and other brain regions within a central autonomic network (21–24). DMN is a major source of efferent vagus cholinergic neurons with anti-inflammatory output as we recently identified using optogenetic stimulation (25). While the future clinical translation of optogenetics, including optogenetic DMN stimulation, remains challenging because of the need for expressing light sensitive opsin molecules in the target area, electrical neuronal stimulation has been widely used in brain modulation therapeutic approaches (9). However, whether electrical stimulation can be applied to activate DMN efferent anti-inflammatory signalling remained unknown. The DMN is a major source of efferent preganglionic vagus nerve fibers controlling visceral functions, including cardiac function (26–30). While the functional role of DMN vagal input to the heart is predominantly associated with modulation of the ventricular excitability and contractility (31–33), there is some evidence that activation of this input may also cause bradycardia (28–30). Therefore, it is an imperative to assess whether anti-inflammatory effects can be achieved using electrical DMN stimulation (eDMNS) without affecting the heart rate (HR) and avoiding undesirable bradycardia.

Here, using eDMNS, we show a current-dependent decrease in the HR of mice. Importantly, we demonstrate for the first time that eDMNS, at intensities that do not affect the HR, suppresses serum pro-inflammatory cytokine levels in mice with endotoxemia. We also reveal the broader scope of anti-inflammatory effects of this approach and its therapeutic efficacy in a mouse model of cecal ligation and puncture induced bacterial infection and deleterious inflammation - a clinically relevant scenario of polymicrobial sepsis (34). Sepsis is a complex disease defined as life-threatening organ dysfunction caused by a dysregulated host response to infection (35). Sepsis affects approximately 1.7 million adults in the United States each year and is linked to more than 250 000 deaths, which makes it a leading cause of death in the US hospitals and a major public health problem (36–38). Currently, there are no clinically approved specific treatments for sepsis, which dictates the need for evaluating new therapeutic approaches.

Overall, our findings indicate new possibilities for selective brain immune regulation and inflammatory control using electrical/bioelectronic neuromodulation targeting the DMN and other brain regions.

2. Materials and methods

2.1. Animals

The Institutional Animal Care and Use Committee (IACUC) and the Institutional Biosafety Committee of the Feinstein Institute for Medical Research, Northwell Health, Manhasset, NY approved all procedures described, in accordance with NIH guidelines. Male, C57BL/6 mice were purchased from Jackson Laboratories. All mice were given access to food and water ad libitum and maintained on a 12h light and dark schedule at 25°C and fed a standard Purina rodent chow diet (Rodents - Feed and nutrition products | Purina (multipurina.ca)). Mice were allowed to acclimate to the environment for at least 10 days before being used in experiments. Experiments were conducted on mice aged 8–10 weeks old. Sex-specific differences in rodent models of sepsis have been previously documented and attributed to heterogeneous and even opposing effects of male and female sex steroid hormone immunoregulatory effects (39). Therefore, for consistency we used only male mice in all experiments.

2.2. DMN localization and stimulation

Mice were subjected to sham stimulation or eDMNS as follows. Animals were anesthetized by inhalation of isoflurane gas mixed with air at 2.5%. Once anesthetized, mice were transferred to a stereotaxic frame equipped with a nozzle for continuous isoflurane administration at 1–1.5%. A midline skin incision was performed, and the underlying muscles retracted exposing the dura matter. The head was tilted forward to ensure adequate viewing of the fourth ventricle through the dura matter. A slit was made in the dura matter with a 23G needle, and the cerebrospinal fluid was cleared and the obex point exposed. A concentric bipolar electrode was fixed to the stereotaxic frame and then placed into the obex. The electrode was stereotaxically guided to the coordinates of the left DMN (0.25 mm lateral to the obex and 0.48 mm deep to the brainstem surface)(25). In specified instances the right DMN was localized at the same coordinates; however, the electrode was moved in the opposite lateral direction from the obex point. Stimulation was performed, or not (in sham mice), using a multi-channel systems STG4000 series stimulator at 500 μA, 250 μA, or 50 μA, 30 Hz, 260 μsec for five minutes (in LPS and CLP sepsis experiments). Subsequently, the electrode was withdrawn, and the animal was sutured, removed from the stereotaxic frame, and processed further according to the experimental design. The following groups of experimental animals were used in endotoxemia experiments. Mice (n = 8–9) were subjected to sham stimulation or left side 250 μA eDMNS. Mice (n = 9 prior to vehicle (saline) administration and n = 13–16 prior to endotoxin administration) were subjected to sham stimulation or left side 50 μA eDMNS. Groups of mice (n = 11–15) were sham operated or subjected to unilateral vagotomy and then sham stimulation or left side 50 μA eDMNS was performed. Corticosteroids were determined in mice (n = 9) subjected to sham stimulation or left side 50 μA eDMNS. Mice (n = 8–9) were subjected to sham stimulation or right side 50 μA eDMNS. The following groups of experimental animals were used in CLP experiments. Cytokines were examined in groups of mice (n = 10–15) subjected to sham stimulation or left side 50 μA eDMNS. Survival was studied in groups of mice (n = 13–15) subjected to sham stimulation or left side 50 μA eDMNS.

2.3. Heart rate recording

Heart rate recording was performed in groups of mice (n = 8–10) and analyzed as previously published (40, 41). Heart rate was measured prior, during and after eDMNS using the Harvard apparatus - ‘small animal physiological monitoring system’ and external needle electrodes. Recording electrodes were placed sub-dermally in the upper limbs of the mice contralaterally, while a grounding electrode was attached to the lower left limb and the recording was started. The stimulating electrode was guided to the left or right DMN as described. Baseline heart rate was obtained for five minutes, and next electrical stimulation was applied with varying amplitudes (500, 250 and 50 μA) for one min.

2.4. Vagus nerve transection (vagotomy)

Mice were subjected to unilateral left vagus nerve transection or sham operation prior to performing sham stimulation or eDMNS. Mice were anesthetized with 2.5% isoflurane gas mixed with air and maintained at 1.5%. A cervical midline incision was performed, and the underlying muscle was carefully retracted. The left vagus nerve was identified, and the carotid sheath removed. The left vagus nerve was isolated and transected solely to avoid damage to surrounding tissue and mechanical activation. In sham animals the carotid sheath was removed, and the left vagus nerve identified. Animals were sutured and allowed to recover for one week prior to sham or eDMNS as previously described (42, 43).

2.5. LPS administration

Directly prior to placing the mice in the recovery cage, lipopolysaccharide (LPS, endotoxin, Sigma-Aldrich, Serotype O111:B4, Lot# L4310–100mg) (0.5 mg/kg) was administered intraperitoneally and mice were allowed to recover in a heated cage for 90 minutes prior to euthanasia via CO2 asphyxiation. Blood, withdrawn through cardiac puncture, and the spleen were collected and processed for further analyses.

2.6. Cecal ligation and puncture (CLP)

CLP was performed as we have previously reported (44, 45) in a cohort of C57BL/6, male mice, aged 10–12 weeks. Briefly, mice were anesthetized using isoflurane gas mixed with air at 2.5%; once anesthetized mice were maintained at 1.5%. An abdominal midline incision was made, followed by a laparotomy. The cecum was identified and partially removed through the opening, a suture was tied around the length of the cecum and pierced with a 22-gauge needle twice, the cecum was subsequently gently squeezed to expel a measure of stool from the piercings. The cecum was returned to the abdominal cavity and the animal was sutured. A single dose of buprenorphine (s.c., 0.05 mg/kg in 100 μl) was given as an analgesic according to the IACUC’s recommendation after the end of the CLP procedure. As previously reported, administration of this single dose of buprenorphine did not alter serum cytokine levels following CLP in mice (46). Subsequently the animal was flipped over and eDMNS or sham stimulation was performed as described above. Animals were recovered in a heated cage and then returned to their home cage for 24 hours until euthanasia by CO2 asphyxiation. Blood was withdrawn through cardiac puncture and centrifuged for serum analysis. Additionally, the spleen was removed and processed in tissue protein extraction reagent for cytokine analysis and protein quantification. Survival was routinely monitored for 14 days as mice were evaluated twice daily and their condition was recorded.

2.7. Analyte measurement

Blood was withdrawn and centrifuged for serum analysis. The spleen was removed and processed in tissue protein extraction reagent for cytokine analysis and protein quantification using a standard Bradford protein assay kit (Biorad). Subsequently tissues and serum were analyzed for cytokine levels using standard ELISA kits: TNF (88–7324: Invitrogen); IL-6 (R&D Systems DY406); and IL-10 (R&D Systems DY417). Serum cytokines were expressed as picogram per milliliter of serum collected and splenic cytokines were expressed as picograms per mg splenic protein.

2.8. Statistics

Data are presented as means ± standard error of the mean. Statistical analyses of experimental data were conducted using GraphPad Prism 9.5.0 (GraphPad Software, Inc., La Jolla, CA). After evaluating the data for normality using a Shapiro-Wilk test, differences between two groups were assessed using an unpaired two-tailed Student’s t test (for normally distributed data) or the Mann-Whitney test (when the data did not meet the assumptions of normality). These tests were used in analyzing cytokine and corticosteroid data from experiments with mice subjected to sham stimulation or eDMNS during endotoxemia or CLP. Typically, data was pooled from experiments with multiple small cohorts of animals. For more than 2 group comparisons one-way ANOVA (for HR analysis) or two-way ANOVA (for analyzing cytokine data from experiments in which mice were administered with vehicle or endotoxin or mice with sham operation or vagotomy were subjected to sham stimulation or left side eDMNS as well as endotoxin administration) were used. Survival was analyzed using the Mantel-Cox Test. HR was expressed as averaged traces ± standard error of the mean. To analyze HR the mean beats per minute (BPM) was averaged over the 1 min recording times, standardized to each distinct pre-stimulation average HR and then the change in HR was compared for each individual trace (pre-, during-, and post-stimulation) – a paired one-way ANOVA was subsequently performed. All individual points refer to one animal. P values less than 0.05 were considered significant. All specific statistical tests are listed in Results and in the figure legends.

3. Results

3.1. Electrical left DMN stimulation affects heart rate in a current-dependent manner

The DMN efferent vagal output regulates cardiac function and in addition to the established control on the ventricular excitability and contractility, there is evidence that DMN stimulation using optogenetic stimulation and other approaches may also decrease the HR (28–31). Therefore, to provide insight into the effects of eDMNS on the HR, we examined three different regimens of eDMNS in which 30 Hz and 260 μsec were kept constant and combined with 500 μA, 250 μA, or 50 μA current on the HR. Briefly, anesthetised mice were placed on the stereotaxic frame, and surgical intervention was performed to slowly insert a concentric bipolar electrode in the left DMN using previously established coordinates (25). The HR was recorded for 1 min prior, during, and post-stimulation. As shown in Figure 1A and B, left side eDMNS significantly inhibited the HR when using a current of 500 μA (*P = 0.0145; one-way ANOVA: Dunn’s multiple comparison test) or 250 μA (*P = 0.0182; one-way ANOVA: Dunn’s multiple comparison test). However, no significant effect of eDMNS was observed when 50 μA were used. These results indicate that eDMNS can cause current-dependent bradycardia.

Figure 1. Electrical left side DMN stimulation regulates the heart rate.

Electrical DMN stimulation causes a current dependent decrease in the heart rate as evident by (A) HR recordings and (B) Standardized change in heart rate with stimulation (500 μA: *P = 0.0159 & 250 μA: *P = 0.0429). Data are represented as individual mouse data points with mean ± SEM. One-way ANOVA was used with Dunn’s multiple comparisons. See Materials and Methods for details.

3.2. Electrical left DMN stimulation suppresses inflammation without affecting the heart rate

As bradycardia is an undesirable side effect, we next studied whether eDMNS can be applied to selectively alter inflammation in a manner which does not affect the HR. We used eDMNS in mice with endotoxemia, induced by peripheral LPS (endotoxin) administration, a standardized model that has been widely utilized in investigating cytokine responses and inflammation in both animals and humans (34, 47). Recently, in endotoxemic ChAT-ChR2-eYFP mice, we showed that optogenetic stimulation of the left DMN, which contains a substantial cluster of cholinergic neurons projecting within the efferent vagus nerve, suppresses serum TNF levels (25). In the current study, sham stimulation (electrode inserted, but no electricity) or left eDMNS was carried out prior to administering LPS (i.p.) and the animals were euthanized 90 mins later as schematically depicted in Figure 2A. We first examined whether a regimen of eDMNS that altered the HR (i.e., 30 Hz, 260 μsec, 250 μA), for 5 mins would affect serum TNF levels. As shown in Figure 2B, eDMNS (that decreased the HR) compared with sham stimulation resulted in a significant decrease in serum TNF (**P = 0.037; Student’s t-test). Then, in another cohort of mice we applied sham stimulation or a regimen of eDMNS, which did not cause any alterations in the HR – i.e., 30 Hz, 260 μsec, 50 μA, for 5 mins prior to subjecting mice to endotoxemia. The lack of effect of this regimen of eDMNS on the HR was confirmed for the entire duration (5 min) of stimulation compared with pre- and post-stimulation (Supplementary Figure 1). eDMNS, compared with sham stimulation significantly decreased serum TNF levels (**P = 0.001; Student’s t-test) (Figure 2C). This result was indicative for the possibility to beneficially alter cytokine responses without affecting the HR. Therefore, to broaden the insight, we also examined the effect of eDMNS on the serum levels of the anti-inflammatory IL-10 (48). As shown in Figure 2C, serum levels of the anti-inflammatory cytokine IL-10 were significantly increased in the eDMNS group compared with sham stimulated controls (*P = 0.021; Student’s t-test). As the spleen is a major target organ of the efferent arm of the inflammatory reflex (49, 50), we also studied the effects of left side eDMNS on splenic cytokine levels. As shown in Figure 2D, eDMNS significantly decreased splenic TNF levels (***P = 0.0001; Student’s t-test) and had no significant effect on splenic IL-10 levels (P > 0.05; Student’s t-test). In addition, we examined the effect of left eDMNS in control mice, not subjected to an inflammatory insult. Serum TNF and IL-10 levels in groups of control mice, i.e., subjected to sham stimulation or eDMNS and injected with saline instead of LPS were very low and below the stipulated assay sensitivity. Splenic TNF levels in control mice (subjected to sham stimulation and injected with vehicle) were significantly lower compared with levels in sham stimulated mice injected with endotoxin (Supplementary Figure 2). In contrast to its effect in mice injected with LPS, eDMNS (compared with sham stimulation) did not alter significantly splenic TNF in control mice (2-Way ANOVA: Stim – F(1, 29) = 9.46, **P = 0.0045; LPS – F(1, 13) = 148.2, ****P <0.0001; Stim X LPS interaction – F(1, 13) = 10.26, **P = 0.0069; Sidak’s multiple comparisons – Sham-operated Saline injection vs Sham-operated LPS injection ****P <0.0001; Sham-operated Saline injection vs eDMNS Saline injection P = 0.9973; Sham-operated LPS injection vs eDMNS LPS injection ****P < 0.0001). Splenic IL-10 levels in sham stimulated control mice were also significantly lower than IL-10 levels in sham stimulated mice injected to LPS and eDMNS compared with sham stimulation had no significant effect in either of the groups (2-Way ANOVA: LPS F(1, 42) = 56.16, ****P <0.0001; Sidak’s multiple comaprisons – Sham-operated saline injection vs sham-operated LPS injection ****P < 0.0001; Sham-operated saline injection vs eDMNS saline injection P = 0.9994; Sham-operated LPS injection vs eDMNS LPS injection P = 0.9964) (Supplementary Figure 2). These results indicate that an eDMNS regimen that does not alter the HR can significantly reduce inflammation in mice with endotoxemia.

Figure 2. Electrical left side DMN stimulation (eDMNS) alters serum and splenic cytokines in endotoxemic mice.

(A) Schematic depiction of the experimental design using eDMNS in LPS administered mice. Anesthetized mice were placed on a stereotactic frame and following a surgical intervention to insert an electrode in the left DMN, sham stimulation or eDMNS was performed for 5 mins and LPS (0.5 mg/kg, i.p.) was administered. Mice were euthanized 90 mins later and blood and spleen were collected and processed for analysis. (B) eDMNS (30 Hz, 260 μsec, 250 μA) compared with sham stimulation significantly decreases serum TNF (**P = 0.037; Student’s t-test), (C) eDMNS (30 Hz, 260 μsec, 50 μA) significantly decreases serum TNF (**P = 0.001; Student’s t-test) and increases IL-10 levels. (*P = 0.021; Student’s t-test) (D) eDMNS (30 Hz, 260 μsec, 50 μA) significantly suppresses splenic TNF (***P = 0.0001; Student’s t-test) but does not significantly alter splenic IL-10. Data are represented as individual mouse data points with mean ± SEM. See Materials and Methods for details.

3.3. Electrical DMN stimulation does not alter cytokine levels in vagotomised mice during endotoxemia

We next assessed the role of the vagus nerve in mediating the anti-inflammatory effect of eDMNS. Mice were subjected to surgical transection of the left vagus nerve (unilateral vagotomy) or sham operation one week prior to applying sham stimulation or left side eDMNS in the two groups of mice. As shown in Figure 3A, eDMNS compared with sham stimulation (i.e., the electrode was inserted into the target area, followed by no stimulation) significantly reduced serum TNF levels in sham operated control mice (in which the cervical vagus nerve was surgically localized, but not transected). However, this anti-inflammatory effect was abrogated in mice with ipsilateral vagotomy (2-way ANOVA: Stim x Surgery interaction F(1, 54) = 4.279, *P = 0.0434; Stim F(1, 54) = 7.979, **P = 0.0066; Tukey’s multiple comparisons – sham operated control vs stim mice **P = 0.0036; – vagotomized control vs stim mice P = 0.9554). Similarly, while eDMNS significantly reduced splenic TNF levels in sham operated controls, the effect was abolished in mice with vagotomy (Figure 3B) (2-way ANOVA: Stim x Surgery interaction F(1, 50) = 6.843, *P = 0.0117; Tukey’s multiple comparisons multiple comparisons – sham operated control vs stim mice *P = 0.0128; – vagotomized control vs stim mice P = 0.9429). In addition, while serum IL-10 levels were significantly increased in sham operated mice subjected to eDMNS, the levels of this anti-inflammatory cytokine in mice with vagotomy and subjected to eDMNS were not significantly altered (Figure 3C) (2-way ANOVA: Stim F(1, 53) = 6.592, *P = 0.0131; Tukey’s multiple comparisons multiple comparisons – sham operated control vs stim mice *P = 0.0119; – vagotomized control vs stim mice P = 0.9470). These results confirm a critical role of the vagus nerve in mediating the anti-inflammatory effect of eDMNS.

Figure 3. The anti-inflammatory effect of electrical left side DMN stimulation (eDMNS) is abrogated in mice with ipsilateral cervical vagotomy and is not related to corticosteroid level alterations during endotoxemia.

eDMNS, compared with sham stimulation significantly suppresses (A) serum TNF (**P = 0.0036) and (B) splenic TNF (*P = 0.0128), and (C) increases serum IL-10 (P= 0.012) in mice with sham operation and these effects are abolished in mice with unilateral cervical vagotomy (performed one week earlier) during endotoxemia. Two-way ANOVA was used with Tukey’s multiple comparisons. (D) eDMNS, compared with sham stimulation does not significantly change serum corticosterone levels during endotoxemia (*P = 0.5481; Students t-test) Data are represented as individual mouse data points with mean ± SEM. See Materials and Methods for details.

3.4. Electrical DMN stimulation does not alter serum corticosterone levels during endotoxemia

Corticosteroids, as a result of activation of the hypothalamic pituitary adrenal (HPA) axis by exogenous neuronal stimuli, can be involved in the regulation of inflammation (10, 51). Therefore, to examine whether eDMNS might activate this axis, we also assessed the effects of left side eDMNS on serum corticosterone levels. As shown on Figure 3D, there was no significant difference in these levels between sham stimulated mice and mice subjected to eDMNS (P = 0.5481; Student’s t-test). These results suggest that no modulation of a neuroendocrine mechanism with altered corticosteroid release is associated with eDMNS in our experimental procedures.

3.5. Right side eDMN stimulation also activates an anti-inflammatory response without changing the heart rate

While the left cervical vagus nerve stimulation has been predominantly utilized in treating inflammation, there is also evidence for anti-inflammatory effects of right VNS (10, 52). Therefore, we also examined the possibility to alter cytokine responses via applying right side eDMNS. First, we investigated the effects of right side eDMNS on the HR utilizing the same regimens we applied in studying the effects of the left side eDMNS. As shown in Figure 4A and B, right side eDMNS resulted in regimen-dependent HR decrease; while 500 μA or 250 μA caused significant HR reduction, 50 μA did not alter the HR (One-way ANOVA: Dunn’s multiple comparisons; 500 μA (pre-stim vs stim), *P = 0.0112; 500 μA (stim vs post stim), ***P = 0.0007: and 250 μA, *P = 0.0192). Accordingly, we next examined the effect of right side eDMNS using 30 Hz, 260 μsec, 50 μA, for 5 mins on cytokine responses in LPS treated mice. Applying this eDMNS resulted in significantly decreased serum TNF levels (**P = 0.0086; Student’s t-test) and increased serum IL-10 levels (**P = 0.0055; Student’s t-test) (Figure 5A). Moreover, no significant differences in splenic cytokines between the stimulated and the sham stimulated groups of mice were observed (Figure 5B) (TNF: P = 0.7195, Student’s t-test & IL-10: P = 0.9087, Student’s t-test). These results indicate that applying right side eDMNS can cause bradycardia, but an eDMNS regimen which does not affect the HR can significantly reduce serum TNF levels and enhance IL-10 without significantly affecting the splenic cytokine levels.

Figure 4. Electrical right side DMN stimulation (eDMNS) regulates the heart rate.

eDMNS causes a current dependent decrease in the heart rate: (A) HR recordings; (B) Standardized change in heart rate with stimulation (500 μA (pre-stim v. stim); *P = 0.0112: 500 μA (stim v. post stim); ***P = 0.0007 and 250 μA: *P = 0.0192). Data in B are represented as individual mouse data points with mean ± SEM. One-way ANOVA was used with Dunn’s multiple comparisons. See Materials and Methods for details.

Figure 5. Electrical right side DMN stimulation (eDMNS) alters serum cytokines in endotoxemic mice.

(A) eDMNS (30 Hz, 260 μsec, 50 μA) significantly decreases serum TNF (**P = 0.0086; Students t-test) and increases serum IL-10 levels (**P = 0.0055; Student’s t-test). (B) eDMNS (30 Hz, 260 μsec, 50 μA) does not significantly alter splenic TNF and IL-10. Data are represented as individual mouse data points with mean ± SEM. See Materials and Methods for details.

3.6. Electrical DMN stimulation alters serum and splenic cytokine levels and improves survival in mice with CLP-induced polymicrobial sepsis

While LPS induced endotoxemia, associated with robust systemic cytokine release and inflammation is considered by some as a model (albeit with limited clinical relevance) of gram negative sepsis (34), cecal ligation and puncture (CLP) is a widely used clinically relevant model of polymicrobial sepsis (34). Therefore, to broaden the insight on the therapeutic utility of eDMNS we studied the effects of this approach in mice subjected to CLP. Based on our results in the LPS model and considering the much broader utilization of left vs right cervical VNS in preclinical and clinical inflammatory studies, we investigated the anti-inflammatory efficacy of left eDMNS during CLP. As depicted in Figure 6A, we examined the therapeutic effects of eDMNS on cytokine responses, implementing an experimental design in which animals were first subjected to CLP. Briefly, C57BL/6 male mice were anesthetised, the CLP surgery was performed, and the animals were sutured and then placed into a stereotaxic frame for a surgical intervention to apply sham stimulation or eDMNS (50 μA, 30 Hz, 260 μsec, for 5 mins) as described in detail in Material and Methods. Twenty-four hours after the CLP, mice were euthanized, and blood and spleen were collected and processed for cytokine analysis. As shown in Supplementary Figure 3, eDMNS (compared with sham stimulation) significantly supressed serum TNF (*P = 0.0291; Student’s t-test) and decreased (albeit not significantly) splenic TNF levels (P = 0.0739; Student’s t-test) in mice with CLP. While in endotoxemia, TNF is a validated indicator of systemic inflammation (53, 54), the role of TNF in CLP sepsis pathogenesis has not been definitively established (34, 55, 56). However, the pro-inflammatory cytokine IL-6 has proven useful as a marker of sepsis severity in preclinical settings and in humans (57, 58) and a therapeutic target in experimental sepsis (59). Therefore, we also analysed the effect of eDMNS on serum and splenic IL-6 and IL-10 levels in mice with CLP sepsis. As shown on Figure 6B, eDMNS (compared with sham stimulation) significantly reduces serum IL-6 levels (*P = 0.0384; Mann-Whitney test) and did not significantly change serum IL-10 levels (P = 0.5980; Student’s t-test). eDMNS (compared with sham stimulation) also significantly reduced splenic IL-6 levels (*P = 0.0472; Student’s t-test). and increased splenic IL-10 levels (*P = 0.0353; Student’s t-test) (Figure 6C). In addition, eDMNS (compared with sham stimulation) significantly improved the 14-day survival rate of mice subjected to CLP (**P = 0.004; Mantel-Cox Test) (Figure 6D). These results demonstrate the differential, cytokine specific anti-inflammatory effects of eDMNS in mice with CLP-induced polymicrobial sepsis and the efficacy of this approach in improving CLP survival.

Figure 6. Electrical left side DMN stimulation (eDMNS) suppresses inflammation and improves survival in mice with CLP sepsis.

(A) Schematic depiction of the experimental approach for eDMNS in mice with CLP sepsis. eDMNS or sham stimulation was performed in anesthetized mice after CLP surgery and cytokines at 24h, or 14-day survival were analyzed. (B) eDMNS (30 Hz, 260 μsec, 50 μA) significantly decreases serum IL-6 (*P = 0.0384; Mann-Whitney test) and has no effect on IL-10 levels. (C) eDMNS significantly decreases splenic IL-6 (*P = 0.047; Student’s t-test) and increases IL-10 (*P = 0.035; Mann-Whitney test) levels. Data are represented as individual mouse data points with mean ± SEM. Unpaired Student’s t test was used. (D) eDMNS compared with sham stimulation (30 Hz, 260 μsec, 50 μA) significantly improves survival (**P = 0.004, Mantel-Cox Test) See Materials and Methods for details.

4. Discussion

Here, we show for the first time that electrical stimulation can be applied to the brainstem DMN to beneficially alter inflammatory indices in male mice with endotoxemia and CLP-induced sepsis without potentially deleterious effects on the HR.

Inflammation is an essential protective response to infection, injury, and other conditions and stimuli (2, 60). This response is localized, balanced, and timely resolved. However, disruption of this beneficial scenario leads to excessive, unresolved, and chronic inflammation that is deleterious to the host (2, 3) and drives the pathogenesis of many inflammatory and autoimmune diseases (61, 62). The vagus nerve plays a major role in the regulation of cytokine levels and inflammation (1, 7, 8, 63–66). Electrical stimulation of the cervical vagus nerve, which suppresses aberrant inflammation in murine models of endotoxemia and other inflammatory conditions (7, 19, 50, 67–69) has been successfully explored in the clinical treatment of rheumatoid arthritis, inflammatory bowel disease and other disorders (7, 10–15).

The vagus nerve contains afferent (about 80%) and efferent (about 20%) fibers and cervical VNS stimulates both types, which decreases selectivity and theoretically increases the possibility of side effects. The brainstem DMN is a major source of efferent vagus fibers and provides a locus for targeted stimulation of these neurons in anti-inflammatory exploration. Using optogenetic stimulation we recently revealed the role of DMN cholinergic neurons in suppressing serum TNF levels during murine endotoxemia (25). This study and other reports employing optogenetic stimulation emphasize the importance of studying targeted brain neuromodulation in new strategies for controlling inflammation (17, 70–72). While optogenetic stimulation provides a cell specific means of neuromodulation, the clinical applicability of this approach remains challenging. In contrast, there is great progress in using electrical stimulation to modulate specific brain areas for therapeutic benefit (73, 74). For instance, deep brain stimulation is a standard of care for Parkinson’s disease, essential tremor and dystonia (73, 74) and this approach of targeting various brain regions has been explored in the treatments of epilepsy, Alzheimer’s disease and other disorders (75). However, the possibility of applying electrical stimulation on the DMN to control cytokine responses and inflammation remained unexplored.

Among other organs, the left and the right DMN efferent vagus nerve fibers innervate the heart and regulate cardiac function (26–30). The physiological role of DMN vagal cardiac innervation is manly linked to modulation of the ventricular excitability and contractility (31, 32, 76). However, chronotropic effects as a result of activation of DMN cardiac projections have also been reported (28–30). Therefore, prior to investigating its effects on inflammation we examined the effects of the left and the right eDMNS on the HR, because desirable anti-inflammatory effects of eDMNS should not be associated with undesirable bradycardia. Altering the stimulation parameters, we show a current-dependent effect of left eDMNS on the HR and the lack of effect when a current of 50 μA is used. Then we show that in LPS administered mice this eDMNS significantly decreases serum TNF and increases the levels of the anti-inflammatory cytokine IL-10 compared with sham stimulation. In addition, eDMNS significantly decreases TNF in the spleen – a major target organ of the efferent arm of the inflammatory reflex, in which the efferent vagus nerve interacts with the splenic nerve in the celiac-superior mesenteric ganglion complex (25, 77–80).

Our results obtained in experiments with mice subjected to vagotomy or sham surgery prior to eDMNS clearly indicate that the anti-inflammatory effect of eDMNS is mediated through the vagus nerve, because ipsilateral vagotomy abrogated the effect. In addition, eDMNS does not change serum corticosteroid levels, the peripheral end product of the HPA axis with anti-inflammatory effects. This observation rules out activation of the HPA axis in the brain as another (neuroendocrine) mechanism mediating the eDMNS anti-inflammatory effects. Activation of the HPA axis and increased serum corticosteroid levels in endotoxemia and other inflammatory conditions have been previously reported (81–83). Increased serum levels of TNF and other pro-inflammatory cytokines have been directly associated with triggering HPA axis activation (83, 84). Accordingly, one would expect that lower TNF levels in animals with eDMNS compared with sham stimulation would be paralleled with lower corticosteroid levels, which we did not observe. A possible explanation could be that eDMNS also causes an increase in serum levels of IL-10, which has also been linked to HPA axis regulation (85, 86). Thus, the net effect of eDMNS on serum cytokines may reflect the observed lack of significant alterations in corticosteroid levels.

We also demonstrate the current-dependent effects of right side eDMNS on the HR and the lack of effect at 50 μA. This eDMNS regimen, which does not affect the HR also results in anti-inflammatory effects. However, in contrast to left eDMNS, these effects do not involve changes in the splenic cytokines analyzed. These results reveal a differential, side dependent DMN anti-inflammatory regulation using the same parameters of eDMNS. They also indicate that reducing serum TNF using left eDMNS occurs in parallel with and may be related to decreasing the splenic levels of these cytokines. However, the reduction of serum TNF using right eDMNS is not associated with decreases in the splenic levels of this cytokine - an observation that suggests that suppression of cytokine production and release in other organs targeted through right eDMNS may play a role. These observations are in line with previous reports indicating differences in anti-inflammatory regulation applying right cervical VNS in other murine models of inflammatory disorders (52). Similarly, there was a lack of significant effect of left or right eDMNS on the splenic levels of the prototypical anti-inflammatory cytokine IL-10 (48).

Our results demonstrating the effects of eDMNS in settings of experimental sepsis substantially broaden the insight about the therapeutic efficacy of this approach. We have shown for the first time that eDMNS alters serum cytokines and improves survival in mice with CLP. Left side eDMNS applied after CLP suppresses serum IL-6 and serum TNF - an effect also observed in mice with endotoxemia. However, in contrast to endotoxemia, eDMNS does not significantly affect serum IL-10 levels in CLP mice. Interestingly, while the reduction in serum IL-6 is associated with lower splenic IL-6 following eDMNS, there is also a significant increase in splenic IL-10 levels in the stimulated group, which does not reflect changes in the serum levels of this anti-inflammatory cytokine. Importantly, eDMNS does not merely influence serum and splenic cytokine levels, but also results in significant survival improvement in this lethal CLP model. These effects indicate the therapeutic utility of a single eDMNS in this widely used clinically relevant model of polymicrobial sepsis. They also rationalize performing future studies utilizing other timeframes and paradigms of stimulation that will provide additional relevant insight.

The brain in sepsis has been mostly evaluated in terms of possibilities to treat neuropsychiatric manifestations, including cognitive deterioration, delirium, and comma, which have been documented within the scope of sepsis-associated encephalopathy (24, 87). Our results indicate that the brain also provides a therapeutic target for controlling cytokine responses in sepsis and improving the disease survival. Future technological developments may allow the use of surgically implanted (under anaesthesia) devices/interfaces for brain neuromodulation in non-anesthetised patients with inflammatory disorders. This is a validated approach in the use of deep brain stimulation in patients with Parkinson’s disease and dystonia (74). It is important to note that future experiments with female mice may provide additional insight of interest for developing further gender specific therapeutic strategies. While our results indicate the lack of effect of eDMNS on splenic TNF in control, not subjected to an inflammatory insult mice, future studies should further evaluate the safety profile of electrical brain stimulation in clinical settings.

5. Conclusion

Our results reveal that during inflammation, eDMNS using parameters which do not cause bradycardia, none-the-less beneficially alters cytokine responses through signaling that requires an intact vagus nerve and does not affect corticosteroid levels. The results also indicate the therapeutic anti-inflammatory efficacy of eDMNS in a preclinical scenario of a complex disease such as sepsis. These findings are of interest for enabling future bioelectronic approaches targeting DMN and other brain regions for therapeutic benefit in human conditions characterised by immune dysregulation and excessive inflammation.

Highlights.

• Electrical stimulation of the dorsal motor nucleus of the vagus inhibits inflammation with no heart rate effects.

• This approach beneficially alters cytokine levels in mice with experimental sepsis.

• This approach also improves survival in experimental sepsis.

Data availability statement

Data is available at the authors’ discretion upon direct request to the corresponding authors.