These experiments were approved by the ethics committee of the University of Versailles Saint-Quentin-en-Yvelines (CEEA-47; APAFIS #38573-2022082610387745 v4) and complied with French and European laws regarding animal experimentation. The animals were housed in ventilated cages in a state-of-the-art animal care facility (2CARE animal facility, accreditation A78-322-3, France), with access to food and water ad libitum on a 12-hour light/dark cycle.

A total of 180 male Swiss mice (Janvier Labs, France, 25–37 g, 5–7 weeks old) were used in this study. A single intraperitoneal (i.p.) injection of LPS (22 mg/kg, from Escherichia coli O55:B5, L2880, Merck, Darmstadt, Germany) was used to create our preclinical endotoxemic shock model. Mice were divided into different groups: the LPS group (n = 49); the LPS + atenolol group: a selective β1-adrenergic blocker (atenolol, 0.1 mg/kg i.p, AstraZeneca, Courbevoie, France) was injected 6 hours post LPS injection (n = 33); and the LPS + sivelestat group: a selective ET inhibitor (sivelestat, ab14618, Abcam, Cambridge, UK) was injected 4 hours (10 mg/kg, i.p.), 8 hours (20 mg/kg, i.p.), and 12 hours (20 mg/kg, i.p.) post LPS injection (n = 16). Control groups were also used for each condition: the saline group (NaCl 0.9%, i.p.; n = 56), the atenolol group (atenolol, 0,1 mg/kg i.p., 6 hours post saline injection; n = 14), and the sivelestat group (NaCl 0.9%, i.p. + 3 sivelestat injections at 4, 8, and 12 hours post saline injection; n = 12) ( Supplemental Figure 1 ).

Twenty-two hours post LPS injection, the mice were used for gastrocnemius MEP recordings or harvested for tissue analysis. This delay is sufficient to induce a severe inflammatory response but does not cause the death of the animals which occurs from the twenty-fourth hour.

The mice (n = 91) were first anesthetized with 5% isoflurane in 100% O₂ and then maintained at 2.5% through a nose cone. The left carotid artery was cannulated. After euthanasia, the animals were perfused with ethylenediaminetetraacetic acid (EDTA, BP2482-1, Thermo Fisher Scientific, Waltham, MA, USA) in phosphate-buffered saline 1X (PBS, BP399-1, Thermo Fisher Scientific, Waltham, MA, USA) at 1 mL/g of body weight to remove blood from organs. The lumbar (L1–L6) spinal cord segment was then dissected out, cut into small pieces, and incubated in 2.5 mg/mL collagenase (11088858001, Sigma-Aldrich, Darmstadt, Germany) and RPMI solution (Roswell Park Memorial Institute, 1640 Gibco, 11875093, Thermo Fisher Scientific, Waltham, MA, USA) at 37°C for 20 minutes in 5% CO₂. The tissue suspension was then passed through a 100-µm cell strainer (22-363-549, Fisher Scientific, Illkirch, France), resuspended in RPMI, and distributed in a 96-well plate (650 185, Greiner bio-one, Les Ulis, France) for staining.

To identify microglia, neutrophils, and ETs (saline: n = 16; LPS: n = 13; atenolol: n = 6; LPS + atenolol: n = 10; sivelestat: n = 6; LPS + sivelestat: n = 8), the cells were resuspended in a mix composed of cell viability marker (eBioscience™ Fixable Viability Dye eFluor™ 780, Thermo Fisher Scientific, Waltham, MA, USA) and SYTOX™ Green (Invitrogen- Thermo Fisher Scientific, Waltham, MA, USA) for 1 minute in PBS 1X. Cells were then incubated with anti-TFAM antibody (20 minutes in RPMI at 4°C), followed by secondary antibody (20 minutes in RPMI at 4°C, see Table 1 for antibody details) and an antibody mix (30 minutes at 4°C, Table 1 ). H3 antibody was conjugated with an antibody conjugation kit (DyLight 594, Abcam, ab201801, Paris, France) according to the manufacturer’s instructions.

For the evaluation of β1-adrenergic receptor quantity (saline: n = 11 mice; LPS: n = 12; LPS + atenolol: n = 9), cells were incubated with the antibody mix shown in Table 2 for 30 minutes at 4°C in PBS 1X.

The data was collected with a flow cytometer BD LSRFortessa™ III (BD Life Sciences, Franklin Lakes, NJ, USA) and analyzed using FlowJo™ v10.8.1 Software (RRID : SCR_008520, BD Life Sciences, Franklin Lakes, NJ, USA).

The animals used for immunofluorescence analysis (saline: n = 6 mice; LPS: n = 6, LPS + atenolol: n=6) were anesthetized using 5% isoflurane in 100% O₂ for induction and maintained at 2.5% through a nose cone. The mice were then euthanized by intracardiac perfusion of heparinized 0.1% (Panpharma, Luitré-Dompierre, France) saline solution (NaCl 0.9%). The lumbar segments (L1–L6) of the spinal cord were collected, fixed for 24 hours at 4°C (formalin 10%, HT501128 Sigma-Aldrich, Darmstadt, Germany), and cryoprotected for 48 hours in 30% sucrose in PBS 1X (Pharmagrade, 141621, AppliChem, Darmstadt, Germany), then frozen and stored at -80°C. Transversal sections of 30 µm were obtained with a cryostat (NX70, Thermo Fisher Scientific, Waltham, MA, USA) and stored at -20°C in cryoprotectant (300 g sucrose, 300 mL ethylene glycol [BP230‐4, Fisher Scientific, Illkirch, France], and 10 g polyvinylpyrrolidone 40 [PVP40‐100G, Sigma‐Aldrich, Darmstadt, Germany]), adjusted at 1 L with PBS 1X. After washing with PBS 1X, antigen retrieval (2.94 g/L tris-sodium citrate solution at pH 6) was realized for 20 minutes at 96°C. The sections were then permeabilized (0.5% triton X-100 [X100-100ML, Sigma-Aldrich, Darmstadt, Germany]) in TBS 1X, blocked (30 minutes at room temperature) in a saturation solution (4% BSA [Bovine Serum Albumin, A0336-50ML, Sigma-Aldrich, Darmstadt, Germany], 4% NDS [Normal Donkey Serum, 566460, Sigma-Aldrich, Darmstadt, Germany], 0.05% triton X-100, 0.05% Tween 20 [BP337-100, Fisher Scientific, Illkirch, France]) in TBS 1X, and incubated overnight at room temperature with primary antibodies ( Table 3 ). After washing, secondary antibodies ( Table 3 ) were incubated for 2 hours at room temperature in saturation solution. Nuclei were labeled with DAPI (D1306, Thermo Fisher Scientific, Waltham, MA, USA) for 10 minutes at room temperature (0.0025 mg/mL in saturation solution) and then washed again. The sections were mounted on slides, dried, and cover-slipped with fluorescent mounting medium (fluoromount G, Invitrogen-Thermo Fisher Scientific, Waltham, MA, USA). Images of the different sections were captured with a Hamamatsu ORCA-R2 camera mounted on an Olympus IX83 P2ZF microscope (Tokyo, Japan) or with a confocal microscope (Leica TCS SP8 microscope, Wetzlar, Germany) for higher magnification. Pictures were analyzed using ImageJ 1.53n software (ImageJ, RRID : SCR_003070, National Institutes of Health, Bethesda, MD, USA).

The excitability of motoneurons associated with the gastrocnemius muscle was evaluated (saline: n = 15 mice; LPS: n = 10; atenolol: n = 8; LPS + atenolol: n = 8; sivelestat: n = 6; LPS + sivelestat: n = 8) by electrophysiological recordings of the gastrocnemius (MEP). Briefly, anesthesia was induced using isoflurane (5% in 100% O₂) and maintained (1%) through a nose cone. Two microneedle electrodes were inserted in the left and right gastrocnemius. A third electrode was placed subcutaneously as a ground electrode. The electrodes were left in place for the entire duration of the experiment.

Trans-spinal magnetic stimulation, applied above CNS structures (spinal cord), is known to induce recordable MEP (36). This stimulation was performed using a magnetic stimulator MAGPRO R30 (Magventure, Farum, Denmark). The device is connected to a figure-of-eight coil (Cool-B65, Magventure, Farum, Denmark) that delivers a unique biphasic pulse with a stimulus intensity expressed as a percentage of the maximum output of the machine (%MO). Each animal was placed on a non-magnetic 3D-printed custom-made stereotaxic apparatus on the figure-of-eight coil, orientated at 90° to obtain the maximum amplitude (37), with the center of the coil under the cervical level of the animal. Each animal received five magnetic stimulation (MS) pulses, from 40% to 100% MO, with 10% increases (35 pulses in total), with an interpulse duration above 15 seconds to avoid repetitive, low-frequency MS-like effects. Gastrocnemius MEPs were amplified (gain, 1 k; A-M Systems, Everett, WA, USA) and band pass-filtered (100 Hz to 10 kHz). The signals were then digitized with an eight-channel Powerlab data acquisition device (acquisition rate: 100 k/s; AD Instruments, Dunedin, New Zealand) connected to a computer and analyzed using LabChart 8 Pro software (RRID : SCR_017551, AD Instruments, Dunedin, New Zealand).

BSCB permeability was determined using Evans blue (EB) (E2129, Sigma-Aldrich, Darmstadt, Germany) extravasation. The mice (saline: n = 8; LPS: n = 8) were anesthetized using 5% isoflurane in 100% O₂ for induction and maintained at 2.5% through a nose cone. A solution of 50 mg/mL of EB in saline solution was administered intravenously by a retro-orbital route (0.1 mL/10 g body weight). After 2.5 hours, the mice were anesthetized as described previously and transcardially perfused with heparinized 0.9% NaCl after euthanasia. The spinal cord was then collected. Each spinal cord was lysed in 200 µL of a 7% perchloric acid solution (311421-50ML, Sigma-Aldrich, Darmstadt, Germany) with five Zirconium-Silicate spheres (6913500, MP Biomedicals, Illkirch, France) using a Precellys Evolution homogenizer (P000062-PEVO0-A, Bertin Technologies) and then centrifuged at 10,000 G for 30 minutes. The samples were diluted three times in ethanol, and the fluorescence intensity was measured with an Odyssey CLx Infrared Imaging System (LI-COR Biosciences). The concentration of dye/g of tissue was determined using a standard curve of EB in ethanol 100%.

To measure section fluorescence, fluorescein isothiocyanate (FITC)-albumin (A9771,Sigma-Aldrich, Darmstadt, Germany, dissolved at 20mg/mL in saline solution) was intravenously administered by a retro-orbital route (0,1 mL). The circulation period of FITC-albumin was ensured for 30 min prior to euthanasia. The mice were then anesthetized as described previously and euthanized. The spinal cord was then collected and Snap-freeze immediately in liquid nitrogen cooled isopentane. Sections of 30 µm were sliced, then mounted on slides, dried, cover-slipped with fluorescent mounting medium (fluoromount G, Invitrogen-Thermo Fisher Scientific, Waltham, MA, USA). Images were captured with a Hamamatsu ORCA-R2 camera mounted on an Olympus IX83 P2ZF microscope (Tokyo, Japan).

For gastrocnemius MEP analysis, five MEPs for each side of every animal were superimposed and averaged, and the amplitude and latency were quantified with LabChart 8 Pro software (RRID : SCR_017551, AD Instruments, Dunedin, New Zealand). Two-way ANOVA was used to compare each condition obtained with the same MO (percentage of stimulation for MEP measure from 40% to 100% MO) and animal groups.

For flow cytometry analysis, FlowJo v10.8.1 Software (RRID : SCR_008520, BD Life Sciences, Franklin Lakes, NJ, USA) was used to study cell populations and quantification of mean fluorescence intensity (MFI) for ET markers and β1-adrenergic receptor expression.

For immunohistochemistry analysis, ImageJ 1.53n software (ImageJ, RRID : SCR_003070, National Institutes of Health, Bethesda, MD, USA) was used to evaluate Iba1 labeling on the entire spinal cord section. A minimum of 5 spinal sections per animal were used to evaluate the MFI and the percentage of surface occupied by Iba1 labeling per slide for each group.

All data are presented as mean ± SD. Statistics were considered significant when p<0.05 (*p<0.05, **p<0.01, ***p<0.001). Normality was tested with the Shapiro–Wilk test. The difference between two groups was assessed with the two-tailed unpaired Student’s t-test or the non-parametric Mann–Whitney test. Between at least three groups, the difference was assessed by one-way ANOVA, followed by Fisher’s LSD post hoc or the Kruskal–Wallis test followed by Dunn’s test if normality failed. SigmaPlot 12.5 software (RRID : SCR_003210 Systat Software, San Jose, CA, USA) was used for all analyses.

We first evaluated microglia activation and the presence of neutrophils in the spinal cord 22-hours after lipopolysaccharide (LPS) injection, using immunohistochemistry and flow cytometry. Microglia were identified as CD11b^{+ and} CD45^int cells ( Supplemental Figure 2 ). No significant change was observed between all groups in the frequency of microglia among living cells in the spinal cord ( Supplemental Figure 3 ). Activation of these microglia was then identified by change morphology characterized by swelling of the microglial cell body, thickening of the proximal processes, and reduction of the distal branching (38), identifiable by immunolabeling microglia with Iba1 on spinal cord section. LPS injection induced morphological changes of microglia showing microglia activation ( Figure 1B ) compared to the saline group ( Figure 1A ), with a significant increase in surface occupied by Iba1 staining ( Figure 1D ) and in mean fluorescence intensity of Iba1 staining ( Figure 1E ) for the LPS group compared to the saline group. The administration of atenolol in animals injected with LPS (LPS + atenolol group) did not affect the morphology of activated microglia ( Figures 1C–E ).

Neutrophils were identified as CD11b⁺, CD45^high, and Ly6G⁺ cells ( Supplemental Figure 2 ). The frequency of neutrophils in the CNS among live cells was increased in the spinal cord 22 hours after LPS injection compared to the saline group (1.19 ± 0.72% vs. 0.25 ± 0.17%, respectively, p<0.001), whereas the frequency of neutrophils was similar between the LPS group and LPS + atenolol group ( Supplemental Figure 3 ). We confirmed the presence of infiltrated neutrophils in the spinal cords of LPS-treated animals by immunostaining of Ly6G on spinal cord sections ( Supplemental Figure 3 ). This effect of LPS on neutrophils was correlated with an increase of Evans blue (EB) leakage into the spinal cord tissue (saline group: 27.31 ± 7.05 vs. LPS group: 81.34 ± 27.08 µg/g of spinal cord, p<0.001), ( Supplemental Figure 4 ).

We observed the presence of β1-adrenergic receptors on the surfaces of neutrophils ( Supplemental Figures 5D, E ) and microglia ( Supplemental Figure 5F ). The administration of LPS induced an increase in the expression of these receptors in blood neutrophils (1,521 ± 1,201 vs. 499 ± 604.5 [mean fluorescence intensity, MFI] respectively, p<0.001) ( Supplemental Figures 5A, D ) and microglia (14,764 ± 2,135 vs. 1,370 ± 767.9 (MFI) respectively, p<0.001) ( Supplemental Figure 5C and 4F ) compared to the saline group. The administration of atenolol did not alter the expression of β1-adrenergic receptors in neutrophils ( Supplemental Figures 5D, E ) and microglia ( Supplemental Figure 4F ) in the LPS + atenolol group or atenolol group.

To determine whether these infiltrating neutrophils and resident microglia could form ETs in the spinal cord following LPS injection, we studied the expression of four individual markers of ETs: citrullinated histone H3 (H3 citr) involved in chromatin decondensation, myeloperoxidase (MPO, a granular enzyme), SYTOX (a DNA intercalant agent), and TFAM (a protein linked to mitochondrial DNA) (39, 40). Studying ETs with a combination of these markers allows for being more specific and avoiding false positives.

Expressions of MPO in neutrophils ( Figures 2A, B ) and microglia ( Figures 3A, B ), H3 citr in neutrophils ( Figures 2C, D ) and microglia ( Figures 3C, D ), SYTOX in neutrophils ( Figures 2E, F ) and microglia ( Figures 3E, F ), and TFAM in neutrophils ( Figures 2G, H ) and microglia ( Figures 3G, H ) were significantly increased 22 hours following the injection of LPS compared to saline group. The administration of atenolol in animals injected with LPS (LPS + atenolol group) led to a reduction in MPO ( Figures 2A, B ), H3 citr ( Figures 2C, D ), and TFAM ( Figures 2G, H ) expression in neutrophils compared to the LPS group, whereas only a tendency towards decrease could be observed for SYTOX expression ( Figures 2E, F ). In microglia, atenolol administration in animals treated with LPS (LPS + atenolol group) led to a reduction of the expression in SYTOX ( Figures 3E, F ) and TFAM ( Figures 3G, H ) and tended to reduce the expression of MPO ( Figures 3A, B ) and H3 citr ( Figures 3C, D ) compared to the LPS group. As expected, the administration of atenolol alone (atenolol group) had no significant effect on the expression of MPO in neutrophils ( Figure 2B ) and microglia ( Figure 3B ), H3 citr in neutrophils ( Figure 2D ) and microglia ( Figure 3D ), SYTOX in neutrophils ( Figure 2F ) and microglia ( Figure 3F ), and TFAM in neutrophils ( Figure 2H ) and microglia ( Figure 3H ) compared to the saline group.

We also evaluated the frequency of neutrophils and microglia co-expressing H3 citr and MPO, which is a better proxy for ET formation. An increased expression of MPO alone, for instance, could reflect cell degranulation. We found an increase in the co-expression of H3 citr and MPO following LPS injection in both neutrophils and microglia ( Figures 2J and 3J respectively) compared to the saline group (respectively Figures 2I and 3I ). The administration of atenolol in LPS-treated animals (LPS + atenolol group) reduced the proportion of neutrophils ( Figure 2K ) or microglia ( Figure 3K ) positive for MPO and H3 citr co-labeling compared to the LPS group (respectively Figures 2J and 3J ).

To deepen our analysis, we also identified ET formation in neutrophils and microglia on fixed spinal cord tissue. We visualized the intraspinal NET production using a neutrophil-specific marker (Ly6G) combined with a marker of ETs (H3 citr) in LPS-treated animals. Among the neutrophils (Ly6G⁺ cells), a H3 citr labeling was observed outside the cell contour (white arrow) ( Figure 2L ) as well as a MPO labeling ( Figure 2M ). A similar observation was performed for microglia (identified as Iba1⁺ cells) for H3 citr expression in the LPS and saline group ( Figures 3L, M ).

β1-adrenergic blockers have a negative effect on the production of ETs. We wondered whether this antagonist could also have an impact on the motoneuronal excitability during severe inflammation. To address this question, gastrocnemius motor-evoked potentials were assessed in all groups ( Figure 4 ). No significant difference was found in terms of latency (i.e., the delay between the stimulation artifact and the beginning of the motor-evoked potential, MEP), which represents the conduction velocity ( Figure 4B ). However, the administration of LPS reduced the gastrocnemius MEP amplitude 22 hours after the injection. This reduction was significant from 70% to 100% maximum output (MO) for the LPS group compared to saline-injected animals (8.15 ± 5.89 vs. 39.93 ± 15.16 mV, respectively, p<0.001). Interestingly, atenolol injection in LPS animals preserved gastrocnemius MEP amplitude (ranging from 80% to 100% MO) compared to LPS alone ( Figures 4A, C, D ). The administration of atenolol alone (atenolol group) had no significant effect on gastrocnemius MEP amplitude compared to the saline group ( Figures 4C, D ).

β1-adrenergic blockers have overall anti-inflammatory effects. We wondered whether the decrease in gastrocnemius MEP amplitude we observed could be related to the increase in ET production by neutrophils and microglia after atenolol administration. To test this hypothesis, we used sivelestat, an inhibitor of the neutrophil elastase involved in the formation of ETs (41, 42).

We first verified the effect of sivelestat on NET and MiET production in our preclinical model of severe inflammation. As expected, the injection of sivelestat in LPS animals (LPS + sivelestat group) significantly reduced MPO ( Figure 5A ), H3 citr ( Figure 5B ), SYTOX, ( Figure 5C ) and TFAM ( Figure 5D ) expression in neutrophils compared to the saline-injected LPS group. Similar results were observed in microglia with sivelestat injection in LPS animals, i.e., a significant reduction in MPO ( Figure 5E ), H3 citr ( Figure 5F ), SYTOX, ( Figure 5G ) and TFAM ( Figure 5H ) expression compared to the saline-injected LPS group. In the same way, the injection of sivelestat in LPS-treated animals (LPS + sivelestat group) reduced the proportion of neutrophils ( Figure 5K ) and microglia ( Figure 5N ) positive cells for the combined MPO and H3 citr markers compared to the LPS group (respectively Figures 5J, M ).

The administration of sivelestat alone (sivelestat group) had no significant effects on the expression of all the ET markers in neutrophils (respectively Figures 5A–D ) and microglia (respectively Figures 5E–H ) compared to the saline group.

Gastrocnemius MEPs were then recorded in animals injected with sivelestat. The sivelestat administration in the LPS and saline groups had no effect on nerve conduction evaluated by latency analysis ( Figures 6A, B ). Interestingly, the reduction in ET formation following the sivelestat injection in LPS-treated animals (LPS + sivelestat group) preserved the gastrocnemius MEP amplitude (significant from 80% to 100% MO) compared to the LPS group (26.13 ± 5.78 vs. 8.15 ± 5.89 mV, respectively, p<0.01), ( Figures 6C, D ).

The concentration of β1-blockers was determined to limit significant effects on cardiovascular parameters such as heart rate and arterial pressure in our model (heart rate: -8.7 ± 1.8% of baseline value and mean arterial pressure: -12.5 ± 4.8% of baseline value at 6 hours post LPS injection and 10 minutes post atenolol administration). Moreover, atenolol has a good β1-adrenergic receptor specificity and a sufficient half-life for our protocol (53).

Interestingly, gastrocnemius MEPs (latency and amplitude) can at least partly reflect the state of the motor neurons, the nerve conduction, the neuromuscular junction, and the state of muscular contractility. LPS model being characterized by a systemic inflammation, we cannot rule out that the observed functional effect (MEP amplitude reduction) was not also partly due to muscle inflammation through, for example, ET production by neutrophils and monocytes or monocyte-derived macrophages that infiltrate the muscles.

In conclusion, our results show that LPS increased neutrophil infiltration in the spinal cord, activated microglia cells, led to ET formation, and decreased neuromuscular function. Moreover, the results showed a beneficial effect of β1-adrenergic blocker administration on ET formation and neuromuscular function evaluated by MEP amplitude. The link between ETs and neuromuscular function was confirmed following the administration of a specific inhibitor of ETs. Thus, the modulation of ET formation during inflammatory-related diseases could have beneficial effects on patients to limit neuromuscular function disagreements. It could also be beneficial for other neuroinflammatory pathologies associated with motor function disorders, such as spinal cord injuries.