Electrical stimulation of the vagus nerve ameliorates inflammation and disease activity in a rat EAE model of multiple sclerosis

Significance

Multiple sclerosis (MS), the most common disabling neurologic disease in young adults, results in demyelination within the central nervous system, decrements in functional ability, and progressive accumulation of disability. Electrical stimulation of the vagus nerve (VNS) has been shown to provide clinical benefits for disorders including rheumatoid arthritis, epilepsy, depression, and Crohn’s disease. In this report, we demonstrate that VNS decreases both clinical symptoms and molecular pathology in a standard rat model of MS, experimental autoimmune encephalomyelitis. We further show that VNS modulates gene expression, including those encoding inflammatory mediators, inflammatory reflex components, and oligodendrocyte differentiation and myelin synthesis. These data indicate that VNS may be a promising approach to treat MS and possibly impact remyelination.

Abstract

Multiple sclerosis (MS) is a demyelinating central nervous system (CNS) disorder that is associated with functional impairment and accruing disability. There are multiple U.S. Food and Drug Administration (FDA)-approved drugs that effectively dampen inflammation and slow disability progression. However, these agents do not work well for all patients and are associated with side effects that may limit their use. The vagus nerve (VN) provides a direct communication conduit between the CNS and the periphery, and modulation of the inflammatory reflex via electrical stimulation of the VN (VNS) shows efficacy in ameliorating pathology in several CNS and autoimmune disorders. We therefore investigated the impact of VNS in a rat experimental autoimmune encephalomyelitis (EAE) model of MS. In this study, VNS-mediated neuroimmune modulation is demonstrated to effectively decrease EAE disease severity and duration, infiltration of neutrophils and pathogenic lymphocytes, myelin damage, blood–brain barrier disruption, fibrinogen deposition, and proinflammatory microglial activation. VNS modulates expression of genes that are implicated in MS pathogenesis, as well as those encoding myelin proteins and transcription factors regulating new myelin synthesis. Together, these data indicate that neuroimmune modulation via VNS may be a promising approach to treat MS, that not only ameliorates symptoms but potentially also promotes myelin repair (remyelination).

Multiple sclerosis (MS) is an inflammatory, demyelinating disorder of the central nervous system (CNS) that afflicts approximately 900,000 adults in the United States alone and is associated with axonal damage, impairment in neurologic function, and progressive disability (1). The etiology of MS is incompletely defined, although an interplay of genetic and environmental factors is thought to underlie the disorder (1). From a pathophysiological perspective, astrocyte activation and CNS influx of autoreactive lymphocytes (eg, CD4⁺ Th1 and Th17 cells that secrete IFN-γ, IL-12, and IL-17) and other inflammatory mediators (e.g., the plasma protein fibrinogen) across a disrupted blood–brain barrier (BBB) are associated with initiation and perpetuation of demyelination (2–11). Drug therapies approved for the treatment of MS have been shown to inhibit CNS inflammation and slow progression of disability, however, these agents neither work well for, nor are they tolerated by, every MS patient; serious adverse events can preclude or limit their use (1, 12–14). There is a significant need for new treatments that are efficacious and well tolerated. Recent advances in bioelectronic medicine indicate that nonpharmacologic approaches may be less immunosuppressive and have fewer off-target effects.

One such approach may be the electrical stimulation of the vagus nerve (VNS). The VN is the longest cranial nerve, extending bidirectionally between the brainstem and the viscera (15). Due to its accessibility in the cervical region and connections with brain centers of clinical interest, the VN has long been a target for neuromodulation in CNS indications (16). The safety of active implanted devices that electrically stimulate the VN has been demonstrated in >125,000 patients (17). VNS using specific stimulation parameters has been demonstrated to reduce inflammation, symptoms of disease, and tissue damage in multiple animal models and clinical trials of peripheral and central inflammation-driven disorders (neuro-immunomodulation) (18–27). Termed the “Inflammatory Reflex,” the VN provides a neural mechanism to sense peripheral inflammation and reflexively regulate innate immunity by reducing proinflammatory cytokines. Information regarding presence of peripheral cytokines is sent via afferent vagus axons to the brain, which responsively sends signals through efferent VN fibers. Neuronally derived norepinephrine, which binds to β2 adrenergic receptors on tissue-resident lymphocytes (e.g., in the spleen), stimulates ChAT⁺ lymphocytes to release ACh. Neuron- and lymphocyte-derived ACh binding specifically to α7 nicotinic acetylcholine receptors (α7nAChR) on tissue-resident and circulating immunocytes leads to altered immunocytic phenotype, typified by a decrease in proinflammatory cytokine release, downregulation of inflammatory cell trafficking and extravasation into tissues, and protection against inflammation-mediated damage (28, 29). Intracellular signaling cascades downstream of α7nAChR activation include NF-κB, inflammasome, and the JAK/STAT pathways and lead to downregulation of inflammation (30).

Here, we hypothesized that VNS may be an effective treatment for MS. In this study, we explored the effects of subacute VNS in the standard rodent model, experimental autoimmune encephalomyelitis (EAE). We observed that VNS decreased the severity and duration of EAE, reduced inflammatory lesion formation and demyelination, decreased BBB disruption and CNS entry of neutrophils and pathogenic lymphocytes (Th1/Th17), reduced fibrinogen deposition, shifted microglia/macrophage phenotype toward repair, and modulated gene expression of IL-12, IFN-γ, IL-17, inflammatory reflex components, as well as myelin synthesis.

Results

VNS Reduces Disease Severity and Duration in EAE.

To investigate whether VNS is effective in reducing EAE severity and duration, device (SI Appendix, Fig. S1)-implanted rats were stimulated daily (1 mA, 60 s, TID), beginning 7 d post-EAE induction (DPI), the approximate day of EAE symptom onset, through day 21 DPI (Fig. 1A). For analysis of grouped data, individual animals were rebaselined to the day post symptom onset (DPSO). VNS significantly reduced EAE disease severity and duration compared with Sham (device implanted, no stimulation) and unimplanted disease control rats (Fig. 1B and SI Appendix, Table S1; by two-way mixed model ANOVA: treatment P = 0.005, DPSO P < 0.0001). VNS significantly attenuated EAE disease severity and duration, as demonstrated by data representing area under the curve (AUC = clinical score X number of symptomatic days; Fig. 1C; mean ± SEM; disease control = 14.88 ± 0.80, Sham = 13.50 ± 0.76, and VNS = 9.45 ± 1.01, P = 0.0013 vs. Sham), maximum clinical score (Fig. 1D; mean ± SEM: disease control = 3.72 ± 0.17, Sham = 3.36 ± 0.13, and VNS = 2.4 ± 0.31, P = 0.0005 vs. Sham), and number of symptomatic days (Fig. 1E; mean ± SEM: disease control = 7.13 ± 0.35, Sham = 7.95 ± 0.19, and VNS = 6.30 ± 0.26, P = 0.0001 vs. Sham). Consistent with reduction in disease severity and duration, VNS ameliorated the weight loss that was observed in Sham and disease control rats (SI Appendix, Table S2; by two-way ANOVA: treatment P = 0.003, DPSO P < 0.0001). Moreover, the magnitude of VNS-mediated suppression of EAE disease severity and duration was not significantly different from the effect achieved with 3 mg/kg/day of the FDA-approved oral MS drug, teriflunomide (SI Appendix, Fig. S2; AUC mean ± SEM: Vehicle = 14.38 ± 1.43, 1 mg/kg/day teriflunomide = 12.75 ± 1.64, 3 mg/kg/day teriflunomide = 7.00 ± 2.10, VNS = 9.45 ± 1.01, P = 0.594 vs. 3 mg/kg/day teriflunomide).

Fig. 1.

VNS reduces symptoms and duration of EAE. (A) Schematic representation of implantation, immunization, and VNS treatment. VNS device was implanted on day −7 and rats were immunized with MBP on day 0. Daily VNS treatment began on day 7 and continued through day 21. EAE symptom onset occurred on approximately days 7 and 8. (B) Clinical score over time (days post symptom onset; DPSO). The data are presented as median ± IQR and daily group medians were compared with Kruskal–Wallis tests. *P < 0.05, **P < 0.01. The effect of VNS treatment vs. Sham and Disease Control are quantified by (C) area under the curve, (D) maximum clinical score, and (E) the total number of symptomatic days. (B) Group means were compared by two-way mixed model ANOVA (independent variables: treatment, DPSO). (C–E) Group means were compared by ANOVA followed by Tukey’s multiple comparison test. **P < 0.01, ***P < 0.001, ****P < 0.0001.

VNS Reduces Inflammatory Lesion Formation and Demyelination in the EAE Spinal Cord (SC) at Peak Disease.

To explore mechanisms underlying VNS-mediated reduction in disease severity and duration, histochemistry analyses were utilized to assess inflammatory lesions and demyelination of VNS, Sham, and naïve rat lumbar SC at 3 to 4 d DPSO (Peak EAE) (Fig. 2 and SI Appendix, Fig. S3). As expected, there were no abnormal findings in naïve rat SC by hematoxylin and eosin (H&E) or Luxol fast blue (LFB) with nuclear fast red (NFR) staining (Fig. 2 A and B). H&E staining at 3 to 4 DPSO revealed that VNS reduced total inflammatory lesion area by 49% compared with elevated levels observed in both white matter (WM) and gray matter (GM) SC regions (arrows) of Sham rats (Fig. 2 A and C; mean % Sham ± SEM: Sham = 100.00 ± 28.36, VNS = 51.33 ± 19.20, Naïve = 0.00 ± 0.00; n = 3 to 4 per group; Sham vs. VNS: P = 0.33, Sham vs. Naive: P = 0.04, VNS vs. Naive: P = 0.34). These numerous WM and GM lesion areas colocalized with cellular infiltrates, providing evidence that VNS prevented infiltration of lesion-associating immunocytes (Fig. 2A). LFB staining at symptom peak revealed that VNS reduced the extent of demyelination in WM by 44% compared with Sham, with several observed VNS SC segments completely free of perceptible demyelination (Fig. 2 B and D; mean % Sham ± SEM: Sham = 100.00 ± 20.35, VNS = 55.72 ± 15.55, Naïve = 0.00 ± 0.00; n = 3 to 4 per group; Sham vs. VNS: P = 0.20, Sham vs. Naive: P = 0.0085, VNS vs. Naive: P = 0.13). Higher levels of both cellular infiltration-associated lesions and demyelination were also observed in Sham vs. VNS and naïve rats at 0 to 2 DPSO (Worsening EAE), and 5 to 7 DPSO (Remitting EAE) timepoints (SI Appendix, Fig. S4 A–D).

Fig. 2.

VNS reduces inflammatory lesion formation and demyelination in the EAE SC at peak disease. 3 µm lumbar SC sections harvested at peak disease were stained with H&E or LFB with NFR (LFB & NFR) to visualize inflammatory lesions containing cellular infiltrates and WM demyelination, respectively. Representative sections are shown. (A) H&E revealed that VNS prevents the formation of inflammatory lesions (arrows) in both WM and GM regions. (B) Reduced loss of myelin was observed in the WM region of VNS vs. Sham rats (arrows). (A and B) No perceptible inflammatory lesions or loss of myelin was observed in naïve rats. (C and D), Quantification of inflammatory lesion and demyelinated area. Data presented as mean ± SEM. Group means (n = 3 to 4) were compared by ANOVA corrected for multiplicity with Tukey’s test. *P < 0.05, **P < 0.01. Mean Sham lesion area = 73,030 µm². Mean Sham demyelinated area = 51,800 µm². WM, White matter; GM, gray matter.

VNS Suppresses Activation of Astrocytes, Preserves BBB Integrity, and Restricts Fibrinogen Deposition at Peak EAE.

Immunofluorescence studies were conducted to investigate the presence of astrocyte activation, BBB disruption, and parenchymal deposition of fibrinogen. The extent of BBB-destabilizing astrocyte activation during EAE was interrogated with glial fibrillary acidic protein (GFAP) staining. The Sham group displayed significantly higher GFAP expression relative to VNS throughout the GM and WM regions (Fig. 3 A and C). The reduction in GFAP staining indicated that VNS reduced astrocyte activation (GFAP mean fluorescence intensity (MFI) normalized to Sham: Sham = 1.00 ± 0.05, VNS = 0.407 ± 0.1, Naïve = 0.388 ± 0.15; Sham vs. VNS: P = 0.0006, Sham vs. Naïve: P = 0.0005, VNS vs. Naïve: P = 0.98). VNS maintained distribution of CNS endothelial tight junction protein claudin-5 in a defined mesh-like appearance, in contrast to the dysregulated expression observed in Sham (Fig. 3B). CNS parenchymal deposition of the plasma protein fibrinogen was highly evident in Sham SC segments (Fig. 3 B and D), which was elevated in both WM and GM and occasionally colocalized with areas of endothelial cell markers claudin-5 and CD31 disruption (Fig. 3C and SI Appendix, Fig. S6). In contrast, evident deposition of fibrinogen was scarce in VNS SC (Fig. 3 B and D), with the density of fibrinogen staining significantly reduced in VNS versus Sham SC (mean normalized to Sham ± SEM: Sham = 1.00 ± 0.22, VNS = 0.17 ± 0.04, Naïve = 0.04 ± 0.03; n = 3; Sham vs. VNS: P = 0.010, Sham vs. Naive: P = 0.005, VNS vs. Naive: P = 0.79). Together, these data suggest that VNS treatment suppresses astrocyte activation, preserves the integrity of the BBB, and limits deposition of fibrinogen into the CNS parenchyma.

Fig. 3.

VNS suppresses activation of astrocytes, preserves BBB integrity and restricts fibrinogen deposition at peak EAE. (A) Representative 40X confocal images of naive and EAE rat lumbar SC sections immunolabeled for astrocyte marker GFAP as an index of inflammation. (B) Representative 63X confocal images of lumbar SC sections stained for the inflammatory plasma protein fibrinogen and tight junction protein claudin-5. (C) Quantification of GFAP MFI reveals significant downregulation of GFAP expression with VNS treatment, to comparable levels with naïve animals. (D) VNS significantly reduced fibrinogen deposition and maintained claudin-5 structural integrity in the CNS. Data presented as the integrated density, normalized to Sham, of individual focal point integrations, as well as mean ± SEM. Group means (n = 3) were compared by ANOVA corrected for multiplicity with Tukey’s test, *P < 0.05, **P < 0.01, ***P < 0.001.

VNS Suppresses Activation of Macrophage/Microglia, and Shifts Their Phenotype Toward Resolution at Peak EAE.

Immunofluorescence studies were conducted to investigate the activation state of both lesion-adjacent and lesion-distal microglia/macrophage by colabeling microglia/macrophage marker Iba1 with the proinflammatory “M1” phenotype marker iNOS and the anti-inflammatory/proresolving “M2” phenotype marker CD206. Iba1⁺ area was significantly reduced at the peak of EAE disease in VNS compared with Sham rats (mean Iba1⁺ area fraction (%): Sham = 6.462 ± 0.898, VNS = 1.562 ± 0.411; n = 3, P = 0.008), suggesting a suppression of microglial activation and proliferation (Fig. 4 A and B). Costaining of Iba1 with the M1 and M2 markers indicated that VNS significantly skewed microglia from predominantly iNOS-expressing toward CD206-expressing phenotype (mean % iNOS-expressing microglia: Sham = 25.45 ± 4.14, VNS = 3.53 ± 1.66; n = 3, P = 0.008, mean % CD206-expressing microglia: Sham = 4.33 ± 0.39, VNS = 10.47 ± 2.22; n = 3, P = 0.053, mean ratio between Iba1⁺iNOS⁺ and Iba1⁺CD206⁺: Sham = 5.79 ± 0.47, VNS = 0.35 ± 0.18; n = 3, P = 0.0004) (Fig. 4B).

Fig. 4.

VNS reduces microglial/macrophage activation and proliferation, shifting from proinflammatory to proresolving state. (A) Representative 40X confocal images of EAE rat lumbar SC sections immunolabeled with microglia/macrophage marker Iba1 colocalized with either the M1 phenotype marker iNOS (proinflammation) or M2 phenotype marker CD206 (proresolution). (B) Quantification of Iba1, iNOS, and CD206 reveals significant reduction in both Iba1⁺ area fraction and percentage of iNOS⁺ Iba1⁺ population with concomitant increase in percentage of CD206⁺ Iba1⁺ population with VNS as compared to sham. Data presented as mean ± SEM. Group means (n = 3) were compared by the unpaired t test. **P < 0.01, ***P < 0.001.

VNS Reduced Entry of Pathogenic Immunocytes into the CNS.

Neutrophil infiltration into the CNS during the initial days of symptomatic disease was determined by immunofluorescent staining of CD66b positive cells on 0 to 2 DPSO. The number of CD66b⁺ neutrophils was significantly reduced in the VNS group as compared to the Sham group (# CD66b⁺ cells/focal area ± SEM: Sham = 34.94 ± 6.81, VNS = 5.16 ± 2.34, Naïve = 0.70 ± 0.10; n = 2 to 4; Sham vs. VNS: P = 0.005, Sham vs. Naive: P = 0.006, VNS vs. Naive: P = 0.78) (SI Appendix, Fig. S5 A and B). Immunofluorescence and flow cytometry analyses were employed to determine whether the cellular infiltrates observed in SC at peak EAE (3 to 4 DPSO) were composed of lymphocytes known to be pathogenic in MS and EAE (31). Immunofluorescent staining of SC sections with antibodies recognizing T lymphocyte marker CD4 and cytokines IL-17 and IFN-γ showed greater numbers of IL-17⁺ and IFN-γ⁺ CD4⁺ T-cells in Sham than in VNS (Fig. 5 A and B). Flow cytometry analysis of isolated SC cells stained for leukocyte marker CD45, as well as CD4, IL-17, and IFN-γ, also showed a 64.9 ± 5.6 % reduction in numbers of CD45⁺/CD4⁺/IL-17⁺ (% CD45⁺ population and normalized to Sham; mean ± SEM, P = 0.0003) and 63.0 ± 9.0 % reduction in numbers of CD45⁺/CD4⁺/IFN-γ⁺ (% CD45⁺ population and normalized to Sham, mean ± SEM, P = 0.095) lymphocytes in VNS rats compared to Sham rats (Fig. 5 C and D). Together, these findings suggest that VNS inhibits lesion formation and demyelination by reducing pathogenic cellular infiltration into the CNS.

Fig. 5.

VNS reduces entry of pathogenic lymphocytes into the CNS during peak EAE. (A and B) Representative 20X confocal images showing reduced presence of CD4, IL-17, and IFN-γ in VNS SC. 8 µm lumbar SC sections harvested at peak disease were probed with antibodies recognizing CD4, IL-17, IFN-γ, and with DAPI. (C and D) Lumbar SC cells were digested, isolated, probed with antibodies (against CD45, CD4, IL-17, and IFN-γ), and counted by flow cytometry. VNS reduced CD45⁺/CD4⁺/IL17⁺ and CD45⁺/CD4⁺/IFN-γ⁺ cells. (C) The gating strategy with representative results. (D) CD45⁺/CD4⁺/IL17⁺ and CD45⁺/CD4⁺/IFN-γ⁺ cells reported as a fraction of CD45⁺ cells and normalized to Sham. Data presented as mean ± SEM, for Sham CD45⁺/CD4⁺/IL17⁺ (%CD45⁺) = 33.3 ± 2.7, mean ± SEM for Sham CD45⁺/CD4⁺/IFN-γ⁺(%CD45⁺) = 18.5 ± 5.6. n = 3 to 6. ns, not significant; ***P < 0.001 by the unpaired t test with Welch’s correction.

VNS Modulates Expression of Genes Involved in Th1 and Th17 Inflammatory Pathways, Key Inflammatory Reflex Components, and Myelin Synthesis.

The EAE disease course of the VNS group diverged from that of disease control and Sham within 24 to 72 h of VNS initiation (0 to 2 DPSO; Fig. 1B). To explore how VNS impacts key pathways underlying EAE pathology, qPCR was performed on signature genes of the inflammatory reflex, genes involved in Th1 and Th17 inflammatory pathways, and genes involved in myelin synthesis and oligodendrocyte differentiation in SC at 0 to 2 DPSO. Gene expression levels of Ifng, Il12, and Il17, inducers of pathogenic Th1 and Th17 activity, were all significantly down-regulated in VNS rats (mean fold change normalized to Sham ± SEM: Ifng: 0.05 ± 0.04, n = 4, P = 0.0002; Il12: 0.29 ± 0.12, n = 4, P = 0.01; Il17: 0.08 ± 0.06, n = 4, P = 0.001) (Fig. 6A). Cytokine gene expression levels of Ifng, Il12, and Il17 were consistent with a protective effect of VNS in EAE. Expression of α7 nicotinic acetylcholine receptors (Charna7) gene was significantly increased (mean fold change normalized to Sham ± SEM: 2.59 ± 0.35, n = 4, P = 0.02) and gene expression levels of β2 adrenergic receptors (Adrb2) and choline acetyltransferase (Chat) trended upward (mean fold change normalized to Sham ± SEM: Adrb2: 3.180 ± 0.75, n = 4, P = 0.062, Chat: 2.425 ± 0.59, n = 4, P = 0.0966) (Fig. 6B). Gene expression levels of myelin basic protein (Mbp), myelin oligodendrocyte glycoprotein (Mog), and proteolipid protein (Plp), which encode several major protein components of myelin and are associated with oligodendrocyte differentiation and new myelin synthesis, were all significantly up-regulated in VNS rats (mean fold change normalized to Sham ± SEM: Mbp: 2.35 ± 0.24, n = 4, P = 0.0105; Mog: 3.47 ± 0.39, n = 4, P = 0.0078, Plp: 3.23 ± 0.22, n = 4, P = 0.0021) (Fig. 6C). The transcription factors specificity protein 1 (Sp1), SRY-box transcription factor 10 (Sox10) and Pur-α (Pur alpha), which bind at the Mbp promoter region were all significantly up-regulated in VNS rats (mean fold change normalized to Sham ± SEM: Sp1: 1.81 ± 0.25, n = 4, P = 0.049; Sox10: 2.02 ± 0.14, n = 4, P = 0.006, Pur alpha: 2.88 ± 0.38, n = 4, P = 0.0158) (Fig. 6D). Gene expression data suggest that VNS may enhance oligodendrocyte differentiation, myelin production, and inflammatory reflex signal transduction.

Fig. 6.

VNS modulates gene expression pathways at 0 to 2 DPSO. RNA was purified from lumbar SC at 0 to 2 DPSO and quantitative PCR to genes of interest was performed. (A) VNS modulates Th1 and Th17-related inflammatory gene expression, Ifng, Il12, and Il17. (B) VNS modulates gene expression of key inflammatory reflex components Adrb2, Chat, and Charna7. (C) VNS up-regulates expression of Mbp, Mog, Plp, genes involved in myelin synthesis and oligodendrocyte differentiation and (D) Mbp gene transcription factors Sp1, Sox10, and Pur alpha. (A–D) VNS-mediated gene expression levels were calculated as relative fold change over the Actin beta control gene and normalized to Sham (dotted line). Data presented as mean ± SEM. Change from Sham values was analyzed by the two-tailed unpaired t test with Welch’s correction. *P < 0.05, **P < 0.01, ***P < 0.001.

These data collectively indicate that VNS treatment decreased rat EAE disease severity and duration by preventing pathological neutrophil and lymphocyte infiltration, myelin damage, BBB disruption, and fibrinogen deposition, skewed microglia/macrophage toward repair phenotype, and beneficially modified expression of genes encoding mediators of inflammation, the inflammatory reflex, and myelin synthesis.

Discussion

We report herein that treatment with VNS decreased the severity and duration of disease in a standard rat EAE model of MS. VNS reduced astroglial activation, skewed microglia/macrophage toward resolution phenotype, maintained BBB integrity, decreased parenchymal fibrinogen deposition, and modulated the expression of genes encoding components of secreted inflammatory mediators, inflammatory reflex signaling, and myelin synthesis. VNS also decreased CNS infiltration of neutrophils and pathological T lymphocytes and reduced the extent of inflammatory lesions and demyelination. Furthermore, in a direct comparison with the FDA-approved oral disease-modifying therapy (DMT) teriflunomide, the magnitude of reduction in EAE disease severity and duration achieved by VNS mirrored that attained by teriflunomide, consistent with a previous publication that studied teriflunomide in the same EAE Lewis rat model (32). The EAE model has also provided predictive validity in the development of many drugs approved for the relapsing remitting form of MS, which similarly have demonstrated efficacy in EAE (33, 34).

VNS has been established as a safe treatment for drug-resistant epilepsy, depression, and ischemic stroke-associated motor deficits, with VNS devices implanted in >125,000 patients (17, 18, 35). We and others have demonstrated the potential for VNS to access the inflammatory reflex to nonpharmacologically reduce symptoms and disease pathology, and even in some cases recover function, in a wide variety of animal models and clinical trials of inflammation-driven diseases (18, 20, 23, 25, 36–38). Several small prospective clinical studies of VNS as a treatment for the autoimmune disorders rheumatoid arthritis and Crohn’s disease have reported positive results, and a large double-blind pivotal study of VNS in subjects with drug-resistant rheumatoid arthritis is currently underway (RESET-RA; NCT04539964) (21, 24, 26, 27, 39).

While no animal model perfectly replicates human disease, EAE recapitulates many of the pathophysiologic and clinical properties of MS. In this study, we observed therapeutic benefit of VNS on several key features of the EAE model (40). In both MS and EAE, loss of myelin is associated with formation of WM and GM lesions, impairment in neurologic and cognitive function, and mood disturbances such as depression (40–44). In our study, VNS reduced focal areas of WM and GM lesions and areas of WM demyelination by 49% and 44%, respectively. While these area reductions assessed at a single timepoint (3 to 4 DPSO) and with relatively small sample size did not reach statistical significance, evaluation of clinical symptoms across the entire observation period revealed that VNS treatment significantly attenuated both severity and duration of EAE, which is also reflected by a corresponding significant attenuated loss of body weight.

In MS and EAE, autoreactive lymphocytes and secreted mediators damage components of the myelin sheath (e.g., MBP, PLP, and MOG) and the oligodendroglia that generate CNS myelin (4, 7). Astrocytes in active lesions become hypertrophic and express high levels of activation marker GFAP as well as proinflammatory cytokines and chemokines that lead to and exacerbate BBB disruption (5). BBB integrity also depends on support from pericytes, endothelial cells, and the end-foot processes of astrocytes; loss of astrocytic end-foot contact with pericytes and endothelial cells (2, 3) and disruption of glial-blood vessel interactions via multimeric gap and tight junction proteins (e.g., claudin-5) lead to enhanced vascular permeability (8, 11, 45). Enhanced BBB permeability facilitates CNS infiltration of normally excluded myelin-reactive immunocytes (in particular, autoreactive CD4⁺ Th1 and Th17 cells that produce IFN-γ, IL-12, and IL-17), neutrophils, activated macrophages, dendritic cells, and plasma proteins such as fibrinogen (6, 9–11). Extravascular deposition of fibrinogen triggers chemokine release and initiates inflammatory demyelination in the CNS, and fibrinogen has also been identified in CNS lesions as both biomarker and driver of disease in MS and EAE (6, 46–48).

Data from the current study indicate that VNS decreased BBB disruption and myelin damage when assessed at 3 to 4 DPSO (Peak EAE). VNS suppressed activation of astrocytes and showed patterns of claudin-5 immunostaining suggestive of maintenance of normal BBB integrity. This was substantiated by a significant reduction in parenchymal fibrinogen deposition in VNS-treated rat SCs. Both qualitative (immunofluorescence staining) and quantitative (flow cytometry) analyses showed reductions in numbers of pathogenic T-lymphocytes that colocalized with GM and WM lesions (statistically significant for CD45⁺/CD4⁺/IL-17⁺, numerically for CD45⁺/CD4⁺/IFN-γ⁺). It is well established that pathogenesis of both MS and EAE is associated with both Th-1 and Th-17 subsets of T cells, through their secretion of cytokines such as IFN-γ, IL-12, and IL-17. IL-12 is also implicated in the propagation of additional IFN-γ producing Th-1 cells (31). Expression of the proinflammatory cytokine genes Ifng, Il12, and Il17 are respectively linked with these Th-1 (IL-12, IFN-γ) and Th-17 (IL-17, IL-23) cytokine pathways. The observed down-regulated expression of Ifng, Il12, and Il17 in EAE therefore supports the therapeutic potential of VNS in MS.

Inflammatory activation of innate immune cells in the CNS is implicated in MS pathogenesis (5, 49, 50). Release of proinflammatory cytokines and chemokines from Iba1⁺/iNOS⁺ resident microglia and infiltrated macrophage lineage cells recruit pathogenic lymphocytes to the CNS, while production of high levels of reactive oxygen species (ROS) damage local tissues, including the myelin sheaths that surround and protect neuronal axons (49, 50). Yet, just as these cells mediate CNS damage, microglia/macrophages are also required for normal resolution of the damage caused by inflammation, including remyelination. Following a course of damaging inflammation, healthy resolution is mediated by a phenotypic shift in these cells from M1 Iba1⁺/iNOS⁺ proinflammatory cytokine- and ROS- producing to M2 iNOS⁻/CD206⁺ anti-inflammatory and debris clearing (51). Effective clearance of myelin and cellular debris is a prerequisite for terminal differentiation of oligodendrocyte progenitor oligodendrocyte progenitor cells (OPC) into myelin-producing mature oligodendrocytes, and a lack of microglia prevents normal remyelination (52–54). The present study reveals that both Iba1⁺ area and the fraction of iNOS-expressing Iba1⁺ cells decreased following VNS, indicating that VNS reduces the proinflammatory state of innate immune cells in the CNS. In addition, the shift in these cells to a CD206⁺ resolution phenotype following VNS indicates the induction of an accelerated repair response, which correlates to the observed shortened duration of EAE symptomatology.

Neutrophil infiltration has been implicated in MS and EAE pathogenesis and reduction in infiltration may ameliorate disease (55–59). Quantitative neutrophil data from the most involved immunofluorescence fields of view (FOVs) demonstrated that VNS significantly decreased the number of infiltrating neutrophils when assessed at 0 to 2 DPSO. This finding is consistent with VNS- and a7nAChR- induced restriction of neutrophils (60).

Dysfunction in the acetylcholine neurotransmitter system has been reported and strategies that enhance cholinergic transmission have been demonstrated to ameliorate both MS and EAE (61, 62). Specific or nonspecific α7nAChR agonism (e.g., with nicotine, galantamine) resulted in suppression of neuroinflammation and symptoms of disease in numerous models of EAE (63–65) and MS (66–71). In addition, treatment with the DMT dimethyl fumarate, but not IFN-β, enhanced cholinergic transmission in MS patients (72). Here, we reported that VNS increased SC α7nAChR gene expression. In addition, genes encoding β2AR and ChAT, other key molecular conductors of the inflammatory reflex, were also acutely increased by VNS. Intriguingly, preliminary data generated from a small pilot chronic study investigating the durability of VNS in EAE demonstrated that splenocytes and isolated splenic T-cells harvested from VNS-treated rats showed a 1.6 to 2.4× increase in gene expression of Adrb2, Chat, and Charna7 after 5 mo of daily stimulation compared with Sham and naïve rats. These data support the hypothesis that chronic VNS can increase expression of key inflammatory reflex genes in early EAE, as well as over a longer-term course of treatment, without indication of reflex tachyphylaxis.

Increased expression of Mbp, Mog, and Plp is required for OPC differentiation, with OPC maturation in turn required for remyelination (73–75). It is of interest that VNS also differentially regulates expression of genes encoding myelin sheath components and transcription factors that are involved in myelin synthesis and oligodendrocyte differentiation. VNS significantly up-regulated expression of Mbp, Mog, Plp, as well as transcriptional regulators Sp1, Sox10, and Pur alpha at 0 to 2 DPSO. Intriguingly, activation of nAChRs on OPCs has been shown to up-regulate these genes and induce rapid differentiation into mature oligodendrocytes (76). Independently, a group has reported that VNS increased the rate of remyelination in the corpus callosum of rats injected with lysolecithin, while a second group has reported that upon cessation of a cuprizone demyelination diet, VNS not only accelerated remyelination in the motor cortex of mice, but also significantly increased the extent to which remyelination occurred (77, 78). Taken together, these data suggest not simply an immunoregulatory role for VNS, but perhaps a reparative role as well; a potential for VNS treatment to enhance remyelination in MS.

In addition to the inflammatory reflex, another mechanism through which VNS may positively impact inflammation, leukocyte infiltration, BBB integrity, and macrophage/microglia phenotype is by increasing specialized proresolving mediators (SPMs), lipid classes that orchestrate inflammation and its resolution. Vagotomy reduces SPMs, inhibits normal resolution, and results in exacerbation of disease, while VNS enhances resolution of inflammation and subsequent tissue damage via the production of endogenous SPMs (79–86). The VN itself stores SPMs and releases them upon stimulation to reduce both proinflammatory prostaglandins and leukotrienes (80, 83). Recent data show a disruption in typical homeostatic patterns of SPMs in patients with MS (87, 88) and treatment of isolated human brain endothelial cells and monocytes from MS patients with relevant SPMs enhances BBB integrity and reduces both monocytic inflammatory responses and transendothelial migration (87). Furthermore, treatment with either resolvin D1 or maresin-1 ameliorated disease in mouse models of EAE (89, 90).

There are several limitations to this study. While it is highly advantageous to use a fully implanted system to autonomously stimulate awake animals without the stresses of daily human interaction with immobilization and infection-prone percutaneous leads, the complete system size restricted our ability to study EAE in mice. The widely used Lewis rat model produces a brief monophasic disease course which limited the window for therapeutic intervention. Therefore, the experimental design was neither purely therapeutic nor preventative, rather a hybrid of the two (a “semiestablished” disease model with treatment initiation during concurrent subclinical CNS inflammation). In addition, this rat model does not exhibit robust CNS demyelination (34), so the ability to fully assess the extent of myelin protection by the therapy was limited. Additionally, only female animals were studied and independent animal sample sizes for secondary endpoints were relatively small.

These results of VNS in rat EAE reveal therapeutic effects comparable to the MS drug teriflunomide at allosteric equivalents 2.1 to 4.1 fold greater than those approved for human use (91), as well as amelioration of pathophysiologic drivers and correlates of disease. The high lifetime incidence (50%) of depression in MS, together with demonstrated VNS efficacy in depression, supports that VNS may represent a safe and effective long-term therapy that may address not only MS but also a mood disorder comorbidity (42). Perhaps the most provocative findings are that VNS can shift microglial phenotype to promote clearance and repair and up-regulate the expression of genes involved in myelin synthesis, indicating that VNS may not only ameliorate inflammation and clinical symptoms of disease but also enhance myelin formation and repair. Additional animal model studies are currently underway to further investigate the capacity of VNS to independently promote remyelination.

Materials and Methods

Animals.

Five- to six-week-old female Lewis rats (Charles River Laboratories, Raleigh, NC) were housed in the Center for Comparative Physiology (CCP) facility, Feinstein Institute for Medical Research in temperature- and humidity-controlled rooms on a 12 h dark/12 h light cycle, with rat chow and water provided ad libitum. All experimental procedures involving animals were preapproved by the Institutional Animal Care and Use Committee (IACUC) of the Feinstein Institutes for Medical Research, Northwell Health, Manhasset, NY in accordance with NIH guidelines for the Care and Use of Laboratory animals (92).

Stimulation Device Implantation.

The subacute to chronic implantation methodology was adapted from well-developed principles and techniques (36, 93, 94). Rats were anesthetized with isoflurane (3% induction, 2 to 2.5% maintenance) and incision sites prepared by hair removal and disinfection (povidone iodine solution and 70% isopropyl alcohol). Rats were placed supine and the left VNS isolated through a 1 cm medial incision. Rats were then placed prone and a 1.5 to 2 cm incision made on the dorsal skin. A pocket to receive the stimulation device was created using blunt scissors by separating skin from underlying tissue. A subcutaneous tunnel for the leads was formed between the pocket and the incision in the neck. The device (SI Appendix, Fig. S1) was positioned under the skin on the dorsal side and secured with nonabsorbable nylon suture. The electrode cuff was gently pulled through the tunnel and secured around the left VNS. Sham-stimulated rats were implanted in the same way, but with the sham device lacking a terminal electrode cuff. Incisions were closed with surgical staples (Stoelting^TM EZ Clip wound closure kit, Wood Dale, IL). Rats were given local anesthetic (Bupivacaine, 0.5%, s.c., Hospira Inc., Lake Forest, IL), analgesia (Buprenorphine 0.03 mg/kg, sc., Par Pharmaceutical, Chestnut Ridge, NY), antibiotics (Baytril 10 mg/kg, sc., Bayer, Berlin), and resuscitation fluids (4 to 5 mL of warm normal saline, s.c, SAI Infusion Technologies, Lake Villa, IL) and then placed on a heating pad until sternal recumbency was regained. Analgesia (Meloxicam infused diet gel, 1.5 mg/mL, Covetrus, Portland, ME) was also provided for 3 d postsurgery.

EAE Induction and VNS.

Six to seven days after nerve stimulator device implantation and while acutely under anesthesia (isoflurane, 3%), EAE was induced by injection (s.c) of emulsified peptide from guinea pig myelin basic protein (gpMBP_69-88, 100 μg/rat + complete Freund adjuvant; Hooke laboratories, Lawrence, MA) into both dorsal hind leg flanks (100 μL/side). All surgical staples from device implantation were also removed at this time. Seven days after immunization (DPI; the approximate day of EAE symptom onset), daily electrical VNS stimulation (1 mA, 10 Hz, 60 s, 0.25 ms pulse width, TID) was initiated and continued through 21 DPI or day of euthanasia (92). Rats in the disease control group were naïve to manipulation prior to EAE induction.

Treatment of EAE with Teriflunomide.

EAE was induced as above in rats that did not undergo implantation procedures. Teriflunomide stock; 50 mg of teriflunomide (Cat #5069/50, Tocris Bioscience, Minneapolis, MN) dissolved in 2 mL of 0.5% DMSO (Cat #D2650, Sigma, St. Louis, MO) and mixed with 2 mL of carboxymethyl cellulose (Across Organics, Cat #AC332601000, Thermo Fisher Scientific, Fair Lawn, NJ) dissolved in 25 mM Tris-HCl pH 7.5 (Cat #15567-027, Invitrogen, Carlsbad, CA) was prepared and diluted into fresh working solution with 25 mM Tris-HCl after the rats were weighed that day. Each rat was administered 0.2 mL teriflunomide (1 or 3 mg/kg/day) or control (25 mM Tris-HCl) by oral gavage, from 7 to 21 DPI.

Body Weight and Clinical Score Recording.

Clinical scores based on observable disease symptoms (from 0—no symptoms to 5—moribund or death) were recorded in a blinded manner, according to a standard scoring guideline (Hooke Laboratories, Lawrence, MA). Body weights were recorded from 0 to 21 DPI.

SC Samples Collection.

EAE rat SC samples were collected at 0 to 2 (worsening), 3 to 4 (peak), and 5 to 7 (remitting) DPSO (SI Appendix, Fig. S3). Rats were perfused with approximately 150 to 200 mL of ice-cold PBS via cardiac puncture under deep anesthesia, and SCs were collected for flow cytometry analysis. A small SC segment close to the lumbar region was collected and stored in RNA later solution (Invitrogen, Cat # AM7020, Carlsbad, CA) for qPCR analysis. A section of the lumbar SC was dissected and stored in 4% paraformaldehyde (PFA) (Electron Microscopy Sciences, Hatfield, PA) solution for IHC analysis.

Histochemistry.

Paraffin-embedded lumbar SC blocks were stained with LFB and NFR or H&E to visualize demyelination and cellular infiltrates, respectively. Briefly, decalcified lumbar SC sections were serially dehydrated in ethanol. Then the samples were cleared in xylene and paraffinized at 60 °C for 1 h. Paraffin sections (3 and 8 μm) were prepared and utilized for histochemical and immunofluorescence staining, respectively. H&E staining: Sections were deparaffinized and hydrated following standard procedure then immersed in hematoxylin solution for 4 min and washed in running tap water for 15 min. Slides were then incubated with eosin solution for 2 min and differentiated in 70% ethanol. Following differentiation, slides were dehydrated in ethanol, cleared in xylene, and mounted with Eukitt quick-hardening mounting medium (Cat #03989, Sigma-Aldrich, St. Louis, MO).

LFB and NFR Staining.

Lumbar SC sections (3 μm) were deparaffinized and hydrated following standard procedure and incubated with LFB solution (0.1 g LFB, 100 mL 95% ethanol, and 0.5 mL 10% acetic acid) overnight at 58 °C. Slides were washed in running tap water for 10 min. Slides were differentiated with 0.05% lithium carbonate solution for 10 to 20 s followed by differentiation with 70% ethanol until gray and WM were distinguished. Slides were then washed in distilled water and immersed in NFR solution (Cat # IW-1401, IHC World LLC., Ellicott City, MD) for 5 min, washed immediately in distilled water to remove excess staining. Slides were then dehydrated in absolute ethanol, cleared with xylene, and mounted with Eukitt quick-hardening mounting medium (Sigma-Aldrich, St. Louis, MO).

Immunofluorescence Staining.

8 μm SC sections were deparaffinized and antigen retrieval was performed using antigen unmasking solution (H-3300; Vector labs, Newark, CA) in a steamer for 10 min. Sections were cooled at room temperature and washed 3× 5 min in PBS. Sections were permeabilized with PBS containing 0.3% Triton-X for 30 min at room temperature. Sections were washed in PBS for 3 × 5 min, then blocked with 4% normal goat serum and 4% normal donkey serum for 45 to 60 min at room temperature. Sections were incubated with respective primary antibodies such as Alexa-488 conjugated mouse anti-rat CD4 (domain 1, Cat #MCA55A488, Biorad, Hercules, CA); eFluor660 conjugated mouse anti-rat IFN-γ (clone DB1, Cat #50-7310-80, Invitrogen, Carlsbad, CA); goat anti-IL-17 (Cat #SC-6078, Dallas, TX); sheep anti-fibrinogen (Cat #F4203-02F, US Biologicals, Salem, MA); Alexa-488 conjugated anti-GFAP (Cat #53-9892-82, Invitrogen, Carlsbad, CA), Alexa-488 conjugated claudin-5 (Cat #352588, Thermo Fisher Scientific, Waltham, MA), rabbit anti-CD31(Cat #PA5-32321, Thermo Fisher Scientific, Waltham, MA); mouse anti-Iba1 antibody (Cat #ab283319, Abcam, Waltham, MA); Alexa-568 conjugated rabbit anti-iNOS (Cat #Ab20995, Abcam, Waltham, MA); Alexa-647 conjugated rabbit anti-CD206 (Cat #Ab195192, Abcam, Waltham, MA); Alexa-647 conjugated rabbit anti-CD31 (Cat #Ab218582, Abcam, Waltham, MA); and mouse anti-CEACAM8/CD66b (Cat #NB100-77808, Novus Biological LLC, Centennial, CO) at 1:100 concentration in 1% normal goat serum at 4 °C overnight. Sections incubated with unconjugated antibodies were fluorescently labeled with respective secondary antibodies and incubated for 2 h at room temperature. Sections were then washed and mounted with DAPI containing mounting medium (Fluoroshield, Cat #F0657, Sigma, St. Louis, MO). Images were taken either using a KEYENCE BZ-X800 or Zeiss LSM 900 confocal microscope.

SC Sample Preparation for Flow Cytometry.

SC samples were minced separately using surgical blades in the presence of 6 mL digestion mix (1 mg/mL collagenase D, 50 μg/mL of DNase-I) and incubated at 37 °C for 30 to 40 min in a rotating apparatus. After digestion, the tissue samples were gently filtered through a 70 μm mesh strainer. 10 mL of PBS was added, the tube centrifuged at 2,000 rpm for 5 min at 4 °C, and the cell pellet reserved for flow analysis. To eliminate myelin debris from SC samples, the digested cell suspension was gently layered onto 6 mL of 15% BSA in PBS in a 15 mL conical tube and spun for 15 min at 1,900 rpm at 4 °C, without applying the brake in the centrifuge. Separated myelin debris was carefully aspirated from above the cell pellet of unmyelinated glia and lymphocytes that had settled at the bottom of the tube. The pellet was washed with ice-cold PBS and used for fluorochrome staining.

Flow Cytometry Analysis.

Isolated SC cells were divided equally into two tubes (approximately 2 × 10⁶ cells per tube), one serving as unstained control and the second for antibody staining. Cells were incubated with 100 μL of FACS buffer (2% FBS in PBS and 1 mM EDTA) in a round-bottom 96-well plate on ice. Staining cell samples were incubated with mouse anti-rat CD32 antibody (Cat #550270; BD Biosciences, Franklin Lakes, NJ) (1 μL/well/100 μL FACS buffer) on ice for 30 min to block nonspecific binding, then incubated with Aqua live/dead fluorochrome (0.25 μL/well) for 30 min on ice and washed with FACS buffer. Cells were then resuspended in FACS buffer and incubated with 0.5 μL/well of the respective fluorochrome-conjugated antibodies such as pacific blue conjugated mouse anti-rat CD45 (Cat #202226, San Diego, CA); Alexa-488 conjugated mouse anti-rat CD4 (domain 1, Cat #MCA55A488, Biorad, Hercules, CA) and incubated for 30 additional minutes on ice. Cells were washed with ice-cold FACS buffer (200 μL/well) and centrifuged at 2,000 rpm for 3 min at 4 °C. Cells were resuspended in 25 μL of FACS buffer before being mixed with 100 μL of Cytofix/Cytoperm (Franklin Lakes, NJ) solution and incubated for 20 min at 4 °C. Cells were then washed two times with Permwash buffer (Franklin Lakes, NJ) (hereafter, we keep the cells for the rest of the procedures in Permwash buffer) and incubated with cytosolic antibodies such as PE-eFluor610 conjugated mouse anti-rat IL-17(clone eBio1787, Cat # 61-7177-82, Invitrogen, Carlsbad, CA); eFluor660 conjugated mouse anti-rat IFN-γ antibody (clone DB1, Cat #50-7310-80, Invitrogen, Carlsbad, CA) for 30 min on ice. After incubation, cells were again washed with Permwash buffer twice and filtered through Falcon filtertop flow cytometry tubes (Cat #352235) to remove any debris prior to flow cytometry. Compensation control: compensation controls for each color were set using both negative and antibody-binding positive beads (BD Biosciences, Franklin Lakes, NJ) following manufacturer instructions. Briefly, in each tube, one drop of each kind of bead was added and incubated with and without the conjugated antibodies (one color per tube) and incubated at room temperature for 15 to 20 min. Beads were then washed with FACS buffer and used as compensation controls for each fluorochrome. Fluorescein minus one (FMO) control: to validate the spillover effect of each fluorochrome into other spectra, we performed an FMO control. Blood was collected from EAE rat, and the red-blood cells were lysed with ACK lysis buffer (Cat #A1049201, Fisher Scientific, Biochemia, NY). 4 × 10⁴ blood cells were incubated with the cocktail of all fluorochromes in the same tube, except one. We followed the same strategy for each fluorochrome in their respective tubes to measure spillover, if any, and 2 × 10⁴ cell samples were gated in the BD FACSymphony A3 cell analyzer (Franklin Lakes, NJ).

qPCR Assay.

Following the Trizol (Life technologist, Grand Island, NY) nucleic acid purification method, total RNA was isolated from SC samples. cDNA was synthesized using the Super Script VILO cDNA kit (Invitrogen, Carlsbad, CA), following manufacturer instructions. Quantitative real-time PCR was performed using Power-up SYBR green master mix solution in the Roche480 thermal cycler, for Th1 (Ifng, Il12) and Th17 (Il17) associated genes, key components of the inflammatory reflex (Adrb2, Chat, and Charna7), and myelin synthesis-related genes (Mbp, Mog, Plp), and known transcription factors that bind at the promoter region of the Mbp gene (Sp1, Sox10, and Pur alpha) (95). The full list of primer sequences is provided in SI Appendix, Table S3. Each sample was analyzed in duplicate, with each experiment repeated at least twice. Gene expression was presented as relative fold change over control gene (Actin beta) calculated based on the ∆∆^CT values. The relative expression of analyte to the Actin beta gene was normalized to the mean expression of the Sham group.

Quantification of Inflammation and Demyelination.

The inflammatory lesion area of the rat lumber spinal (at least two sections from each rat; approximately 0.5 cm apart) were calculated in both WM and GM regions using ImageJ (NIH). To restrict total area to definite lesions, only lesions with areas larger than the 25th percentile (3,000 µm²) were tabulated and normalized to mean Sham lesion area. Demyelination was quantified in at least two sections from each rat; approximately 0.5 cm apart. The loss of LFB staining in the WM region was tabulated by manually drawing outlines around each demyelinated region using ImageJ. The total areas of inflammatory lesions and demyelination were normalized to Sham and calculated from at least three rats per group.

Immunofluorescence Analysis.

For measurement of GFAP intensity, ImageJ was used to outline the entire SC, and the MFI within the outlined structure was recorded for statistical analysis. At least two SC sections were analyzed per animal, and all animals were normalized to Sham.

Fibrinogen deposition into the WM and GM regions of the SCs was quantified using ImageJ, following a previously published method measuring myelinated nerve fibers (96). Briefly, the images (most involved FOV; 0.010 mm²) were converted into black and white at a constant intensity threshold, and the image dimensions were calibrated to a known scale. Total pixel intensity was measured and the integrated pixel intensity (arbitrary units) over the surface area of the image was calculated and normalized to the mean Sham. To compare the number of infiltrating neutrophils among naïve, Sham, and VSN rats, CD66b⁺ cells were manually counted from at least two sections per animal and 4 to 7 FOVs (0.10 mm²) per section.

For microglia/macrophage analysis in both lesion and nonlesion areas of the Sham and VNS rat SCs, 4 to 11 sections were analyzed per animal. To measure the area fraction occupied by microglia/macrophage, images were binarized using an automated intensity threshold and the percentage thresholded pixels was reported as Iba1⁺ area fraction. To assess microglial/macrophage states, microglial/macrophage expression of iNOS vs. CD206 was quantified as putative M1 vs. M2 state, respectively (49–51). iNOS and CD206 images were thresholded and binarized. The overlap between microglia/macrophage and either CD206 or iNOS was measured by multiplying the binarized Iba1 with either CD206 or iNOS images. The percentage of CD206- or iNOS-expressing microglia was calculated by dividing the number of colocalized signal pixels (Iba1⁺ CD206⁺ or Iba1⁺ iNOS⁺) by the total Iba1⁺ pixels. The ratio between iNOS-expressing and CD206-expressing microglia was subsequently calculated.

Statistical Analyses.

Clinical scores are plotted as median ± IQR, and the Kruskal–Wallis test was used to compare medians on individual days. The complete curves were analyzed by two-way mixed model ANOVA with treatment and time (DPSO) as independent variables, with multiple comparisons with Dunnett’s post hoc test to compare means of Sham and Disease Control to mean of VNS on individual days. The AUC, Maximum clinical score, and number of symptomatic days were first analyzed for statistical outliers (ROUT; Q = 1%) and compared to the VNS group by ANOVA followed by Tukey’s multiple comparison test.

All other data are presented as the mean ± SEM. Quantification of inflammation, demyelination, and quantitative immunofluorescence targets between VNS and Sham groups were analyzed by two-tailed unpaired Student’s t test or by ANOVA followed by Tukey’s multiple comparison test, as appropriate. Infiltration of CD45⁺/CD4⁺/IL17⁺ and CD45⁺/CD4⁺/IFN-γ⁺ cells by flow cytometry are reported as a fraction of CD45⁺ cells and normalized to Sham. Differences in mean between VNS and Sham groups were calculated by the unpaired two-tailed t-test with Welch’s correction.

Analyses of qPCR data between Sham and VNS groups (0 to 2 DPSO) were calculated by the two-tailed unpaired t test with Welch’s correction. All statistical analyses were performed using GraphPad Prism v.10 and the P-values calculated < 0.05 were considered statistically significant.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.