Received: 26 November 2021 Revised: 9 June 2022 Accepted: 16 June 2022 DOI: 10.1002/glia.24234 RESEARCH ARTICLE Microglial response promotes neurodegeneration in the Ndufs4 KO mouse model of Leigh syndrome Kevin Aguilar1,2 | Gemma Comes1,2 | Carla Canal1,2 | Albert Quintana1,2 | 1,2 1,2 Elisenda Sanz | Juan Hidalgo 1 Department of Cell Biology, Physiology, and Immunology, Animal Physiology Unit, Faculty Abstract of Biosciences, Universitat Autònoma de Leigh syndrome is a mitochondrial disease characterized by neurodegeneration, Barcelona, Barcelona, Spain 2 neuroinflammation, and early death. Mice lacking NDUFS4, a mitochondrial complex I Institut de Neurociències, Universitat Autònoma de Barcelona, Barcelona, Spain subunit (Ndufs4 KO mice), have been established as a good animal model for studying human pathology associated with Leigh syndrome. As the disease progresses, there is Correspondence Juan Hidalgo, Department of Cellular Biology, an increase in neurodegeneration and neuroinflammation, thereby leading to deteriorat- Physiology and Immunology, Animal ing neurological symptoms, including motor deficits, breathing alterations, and eventu- Physiology Unit, Faculty of Biosciences, Universitat Autònoma de Barcelona, Barcelona ally, death of the animal. However, despite the magnitude of neuroinflammation 08193, Spain. associated with brain lesions, the role of neuroinflammatory pathways and their main Email: juan.hidalgo@uab.es cellular components have not been addressed directly as relevant players in the disease Funding information pathology. Here, we investigate the role of microglial cells, the main immune cells of “la Caixa” Foundation, Grant/Award Number: ID 100010434; Agència de Gestio  d'Ajuts the CNS, in Leigh-like syndrome pathology, by pharmacologically depleting them using Universitaris i de Recerca, Grant/Award the colony-stimulating factor 1 receptor antagonist PLX3397. Microglial depletion Number: 2017SGR-323; European Regional Development Fund, Grant/Award Number: extended lifespan and delayed motor symptoms in Ndufs4 KO mice, likely by preventing RTI2018-101105-B-I00; H2020 European neuronal loss. Next, we investigated the role of the major cytokine interleukin-6 (IL-6) Research Council, Grant/Award Number: ERC- 2014-StG-638106; Ministerio de Ciencia e in the disease progression. IL-6 deficiency partially rescued breathing abnormalities and Innovacion, Grant/Award Number: modulated gliosis but did not extend the lifespan or rescue motor decline in Ndufs4 KO PID2020-114977RB-I00; Ministerio de Ciencia, Innovacion y Universidades, Grant/ mice. The present results show that microglial accumulation is pathogenic, in a process Award Number: RTI2018-101838-J-I00; n, independent of IL-6, and hints toward a contributing role of neuroinflammation in the Secretaría de Estado de Investigacio Desarrollo e Innovacion, Grant/Award disease of Ndufs4 KO mice and potentially in patients with Leigh syndrome. Numbers: SAF2014-57981P, SAF2017-88108-R KEYWORDS IL-6, Leigh syndrome, microglia, Ndufs4 KO, neuroinflammation [Correction added on August, 11, 2022, after first online publication: the grant/award number for the European Regional Development Fund has been changed from RTI2018-101105–I00 to RTI2018-101105- B-I00.] 1 | I N T RO DU CT I O N that lead to dysfunction in the energetic mitochondrial machinery by affecting oxidative phosphorylation (Gorman et al., 2016). Due to the Primary mitochondrial diseases are a group of rare genetic disorders wide range of existing mutations, these disorders are highly heteroge- characterized by mutations in nuclear or mitochondrial DNA genes neous and present a plethora of clinical symptoms and different This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. © 2022 The Authors. GLIA published by Wiley Periodicals LLC. 2032 wileyonlinelibrary.com/journal/glia Glia. 2022;70:2032–2044. AGUILAR ET AL. 2033 disease onsets, which can vary between birth and adulthood. The of cells contributing to neuronal death in Ndufs4 KO mice. In addition, most affected tissues are those that require high energy, such as mus- IL-6 was found to play a minor role, thereby necessitating further cles and the central nervous system (CNS) (Gorman et al., 2016). Leigh evaluation of the role of other cytokines. syndrome (LS) is the most common early-onset mitochondrial disease. Also known as subacute necrotizing encephalomyelopathy, LS is char- acterized by progressive focal neurodegeneration in specific brain 2 | MATERIAL AND METHODS regions caused by numerous mutations (Lake et al., 2016; Lee et al., 2020). The time of death from such complications usually occurs 2.1 | Animals around 3 years of age, and the neurological symptoms include devel- opmental delay and regression, seizures, ataxia, dystonia, or hypoto- All mice used in this study were fed ad libitum and housed in a 12-h nia. Currently, there is no effective treatment available for patients light–dark cycle under specific pathogen-free conditions at constant with LS (Lake et al., 2016; Lee et al., 2020). temperature (22 ± 2 C). All experiments were performed as per the Mutations in the nuclear gene NDUFS4 (NADH dehydrogenase approval of the Ethics Committee on Animal Experiments of the Uni- ubiquinone iron–sulfur protein 4), which encodes a subunit of complex versitat Autònoma de Barcelona (Ref. 3971 and 4155). I of the mitochondrial respiratory chain, have been described to pro- duce LS (Fassone & Rahman, 2012). In this context, the Ndufs4 KO mouse recapitulates most aspects of human pathology associated with 2.2 | Mice generation LS and has therefore been accepted as an LS mouse model (Quintana et al., 2010). Ndufs4 KO mice develop progressive encephalopathy The parental mice strains used in this study were Ndufs4+/ (Kruse beginning at postnatal day P30, which finally leads to the death of the et al., 2008) and Il6/ mice; the latter were generated in our labora- animal at around P55. During the disease, Ndufs4 KO mice manifest tory (Sanchis, Fernández-Gayol, Comes, et al., 2020). Both mouse lines growth retardation, decreased body weight, ataxia, paw paralysis, sei- were backcrossed with a C57BL/6 background for at least 10 genera- zures, and respiratory abnormalities (Johnson et al., 2013; Kruse tions. Both male and female mice were used in the present study. et al., 2008; Quintana et al., 2012). The main symptoms have been For experiments with PLX3397, male and female Ndufs4+/ mice shown to have a neurological origin since Cre-driven Ndufs4 deletion in were bred with each other to obtain WT (wild-type) and Ndufs4 KO neurons and glia (Nes-KO) essentially demonstrated the same pheno- mice. Three rounds of breeding were required to generate double- type as systemic Ndufs4 KO mice (Quintana et al., 2010). Neuroinflam- deficient mice lacking IL-6 and NDUFS4. (i) Firstly, IL-6 deficient mice mation, mainly microgliosis and astrogliosis, has been established as a (Il6/) were crossed with heterozygous Ndufs4 (Ndufs4+/) mice. hallmark of the disease in regions where neurodegeneration occurs in (ii) Il6+//Ndufs4+/ offspring were crossed with Il6/ mice. (iii) Finally, the late stages of pathology in Ndufs4 KO mice (Quintana et al., 2010). the resulting Il6//Ndufs4+/ and Il6+//Ndufs4+/ mice were bred to This is consistent with observations in human patients with LS (Arii & obtain the four genotypes in the current study: control (Il6+// Tanabe, 2000; Lake et al., 2016). Surprisingly, the role of neuroimmune Ndufs4+/), referred to as WT in the text; Il6 KO (Il6//Ndufs4+/), processes has not been studied per se but as a secondary event associ- Ndufs4 KO (Il6+//Ndufs4/), and double KO (Il6//Ndufs4/). ated with neurodegeneration. It is now well described that mitochon- drial defects or alterations can be major drivers of inflammation and disease progression in a wide range of pathologies, including neurode- 2.3 | Genotyping generative diseases (Luna-sánchez et al., 2021; West & Shadel, 2017; Zhong et al., 2019). Similarly, several immune-related processes have All mice were genotyped by PCR using DNA extracted from their tails. been found to be upregulated in both whole-body and brain-specific For DNA extraction, the tails were boiled in 100 μl of 50 mM NaOH Ndufs4 deficient mice (McElroy et al., 2020; Perry et al., 2021). solution for 7 min. The following primers were used:50 -AGCCTGTTCT- In this study, we evaluated the contributions of microglia, which CATACCTCGG-30 , 50 -GTCCTCTATGAGGGTACAGAG-30 , and 50 -GGT are specialized immune cells of the CNS (Bennett et al., 2016; Li & GCATACTTATACTACTAGTAG-30 for Ndufs4 genotyping; 50 -GAGACT Barres, 2018), to the Ndufs4 KO mouse pathology. We used a pharma- GTGAGAGAGGAGTGTG-30 , 50 -CATCTTATCTGGGCTGACCCTAG-30 , cological approach that involved chronic administration of an antago- and 50 -TCTCTGCTGGGATCTAGGGCC-30 for IL-6 deficiency (Sanchis, nist of the colony-stimulating factor 1 receptor (CSF1R), PLX3397, to Fernández-Gayol, Comes, et al., 2020). deplete mouse microglia from the CNS parenchyma (Elmore et al., 2014; Green et al., 2020). We also studied the role of the major cytokine, interleukin-6 (IL-6), which has been shown to play a signifi- 2.4 | Microglial depletion with the colony- cant role in neuroinflammation (Erta et al., 2012; Spooren et al., 2011) stimulating factor 1 receptor (CSF1R) inhibitor, and many other scenarios (Jones & Jenkins, 2018). Moreover, recent PLX3397 studies have depicted increased IL-6 expression (Balsa et al., 2020) and upregulation of the IL-6/JAK/STAT signaling pathway in the brain WT and Ndufs4 KO mice were injected with vehicle solution (Veh) or of Ndufs4 KO mice compared with controls (McElroy et al., 2020). Our 40 mg/kg of PLX3397 (MedChem, HY-16749) intraperitoneally results strongly suggest that microglial cells constitute a major subset 6 days a week beginning on postnatal days P24–P25. PLX3397 was 2034 AGUILAR ET AL. dissolved in a solution of 20% DMSO, 40% PEG-300, 5% Tween-80, previously (Prada-Dacasa et al., 2020). Briefly, the mice were allowed and 35% saline (vol/vol). to acclimatize to the plethysmograph chamber (EMMS, catalog num- ber: PLY310) for 45 min, followed by 15 min of the experimental period. Tidal volume normalized per body weight (μLg1) and respira- 2.5 | Disease staging and clinical evaluation tory frequency (breathsmin1) have been reported in this study. Disease staging of NDUFS4-deficient mice was performed according to previously described guidelines (Quintana et al., 2010). In brief, dis- 2.7 | Tissue preparation ease progression in Ndufs4 KO mice can be subdivided into three stages: early-stage (P0–P29), mid-stage (P30–41), and late-stage (over Mice were euthanized by decapitation at P47–P48 for the PLX3397 P41) of the disease in which clinical signs increase in severity (Kruse experiments and P38 and P55 for mid and late stages, respectively, in et al., 2008; Quintana et al., 2010). The mice were weighed and exam- the double deficiency experiments. Brains extracted for histology were ined daily for clinical signs (dehydration, ataxia, and gait/postural alter- fixed in 4% paraformaldehyde (PFA) solution at 4 C for 24 h and then ations). Paw clasping behavior and body twisting were examined by transferred to cryoprotective 30% sucrose in PBS (Phosphate-buffered suspending the animals by the tail for 20 s. Mice were humanely saline) (0.01 M pH 7.4) solution for 48 h. Later, the olfactory bulbs euthanized after a 20% loss of maximum body weight. (OB) were separated from the rest of the brain, creating a coronal cut in the prefrontal cortex. Brains were frozen by immersion in isobutane at 30 C, whereas OBs were frozen in dry ice and embedded in an OCT 2.6 | Behavioral tests medium. For multiplex analysis, brain areas were dissected, flash-frozen in liquid nitrogen, and kept at 80 C until further use. Mice were tested at the early, mid, and late stages at postnatal days P27–29, P38–39, and P47–49, respectively. All mice were tested after 30 min of habituation in the testing room. 2.8 | Multiplex Multiplex analysis was previously optimized and validated in our labora- 2.6.1 | Rotarod tory (Sanchis, Fernández-Gayol, Vizueta, et al., 2020). In short, both OBs and half cerebellums were mechanically homogenized using an MM-400 The rotarod test was used to monitor motor coordination. The task mixer mill (Retsch GmbH, Haan, Germany), followed by ultrasonic homog- consisted of placing the animals on a rotating bar that accelerated enization using Sartorius-LABSONIC P (Sartorius AG, Göttingen, from 4 rpm to 40 rpm (Harvard Apparatus, Holliston, MA, USA) at a Germany). Samples were homogenized in 100 μl (OB) and 200 μl (cerebel- lapse of 300 s until the animal fell to the ground. The latency time to lum) of ice-cold protein extraction buffer containing 25 mM HEPES, 10% fall from the rod was also measured. At every disease stage, mice IGEPAL, 0.1 M MgCl2, 1.3 mM EDTA (pH 8.0), and 0.1 M EGTA (pH 8.0). underwent two consecutive training trials. The test was performed Homogenization buffer was supplemented with a 1% protease inhibitor thrice, and the mean of the trials was calculated. cocktail and 0.1 M phenylmethylsulphonyl fluoride (Sigma-Aldrich P8340 and P7626, respectively). The homogenate was centrifuged at 12,000 g for 5 min. The supernatant was collected to determine the protein con- 2.6.2 | Open field centration using the Pierce™ BCA Protein Assay (Thermofisher 23.227). A protein concentration of 5–12 μg/μl of the tissue lysate was used to The mice were placed in a methacrylate white box (56 [W], 36.5 [D], detect the abundance of IL-6, TNF-α, and IL-1β using the Mouse High 31 [H] cm) and allowed to explore the apparatus freely for 10 min. Sensitivity T Cell Magnetic Bead Panel (Millipore, catalog number The measured parameters include the total ambulation distance (hori- MHSTCMAG-70 K) according to the manufacturer's instructions. Data zontal activity) and the number of rearings (vertical activity; the num- were obtained using a Luminex MAGPIX instrument system (Luminex ber of times the mouse stood on its rear limbs). Rotation behavior was Corporation) and analyzed using the xPONENT software v4.2 (Luminex assessed in the corresponding late-stage open-field test and measured Corporation). All results were normalized to the protein concentration in as the presence or absence of the behavior. The activity of the mice the lysate. Results between the blank and the first standard of the stan- was recorded through videography. The videos were analyzed using dard curve were considered as 0. IL-1β levels were unquantifiable as the EthoVision XT tracking software (Noldus Information Technology bv., levels seemed to be below the detection threshold of the technique. Both Wageningen, The Netherlands). sexes were included in this experiment and analyzed together. 2.6.3 | Whole-body plethysmography 2.9 | Histology Unrestrained whole-body plethysmography was used to assess the Coronal sections of the brain were prepared using a cryostat at 20 μm ventilatory function. The detailed protocol has been described from the bregma, 0.62 to 2.46 mm, and from 5.34 to 6.64 mm AGUILAR ET AL. 2035 following Franklin and Paxinos mouse brain atlas (Paxinos & Protease was washed away with two PBS washes (2 min each). Target Franklin, 2013). The sections were preserved in an anti-freezing solu- and negative control probes were applied directly to the sections and  tion at 20 C until staining. OBs were also cut at 20 μm, directly incubated at 40 C for 2 h in a HybEZ oven. Next, the slides were incu- mounted on Superfrost slides (Thermo Fisher Scientific), dried at room bated with preamplifier and amplifier probes (AMP1, 40 C for 30 min; temperature (RT) (20–26 C), and stored at 80 C until staining. For AMP2, 40 C for 15 min; AMP3, 40 C for 30 min). The slides were free-floating brain sections, four slices were selected for each animal then incubated with fluorescent labels (AMP4-Alt C). Finally, the brain with duplicates for each of the following regions: cortex and hippo- sections were mounted using DAPI Fluoromount. campus (between bregma 1.94 and  2.30 mm), VN, and cerebellum (between bregma 5.88 and  6.12 mm), and MCN (between bregma 6.24 and  6.48 mm). For immunostaining, sections were blocked in 2.11 | Microscopy and image quantification 1% BSA and 0.2% Triton X-100 in PBS (0.01 M pH 7.4) solution for 1 h at RT. Slides were then incubated with the chosen primary anti- Immunofluorescence images were captured using a Nikon Eclipse 90i body overnight at 4 C and diluted in blocking solution (1:1000 for microscope coupled to a Nikon Digital Camera DXM1200F using the rabbit anti-GFAP, Dako Z0334; 1:1000 for rabbit anti-IBA-1, Wako ACT-1 v2.70 capture software (Nikon, Tokyo, Japan). Whole Cornus 019-19741; 1:1000 for rabbit anti-TMEM119, ab209064; 1:500 for Ammonis 1 (CA1), VN, OB, and MCN were photographed at 20 chicken anti-NeuN, Millipore ABN91). The following day, sections magnification, while pictures of the whole cortex and cerebellum were left for 1 h at RT before incubation with specific secondary anti- slices were taken at 10 magnification. Images were captured under bodies (1:600 goat anti-rabbit Alexa Fluor 568 Thermofisher A11011; specific detector sensitivity conditions for each area. The samples 1:600 goat anti-chicken Alexa Fluor 488 Thermofisher A11039). The were quantified using ImageJ software. For GFAP, IBA-1, and slides were kept away from light for 1 h at RT. Finally, sections were TMEM119, immunofluorescence quantification of the integrated den- washed, and free-floating brain slices were mounted on Superfrost sity normalized by the measured area was performed. The total num- slides (Thermo Fisher Scientific). The slides were then dried and ber of IBA-1+ cells was manually counted. Microglial phenotypes mounted with DAPI Fluoromount (Southern Biotech). For immunohis- were visually discriminated. In situ hybridization images were cap- tochemical analysis of CD3+ cells, mounted brain sections were incu- tured using a Zeiss LSM700 confocal microscope (Carl Zeiss AG) at bated with citrate buffer (10 mM sodium citrate, 0.05% Tween, pH 6) 20 magnification. Positive gad2 and Slc17a6 neurons were counted for 20 min at 96 C for antigen retrieval and incubated with endoge- and relativized by area. nous peroxidase blocking solution (70% methanol and 3% hydrogen peroxide) for 15 min. Slices were then blocked as described previ- ously, incubated with primary antibody overnight at 4 C, and diluted 2.12 | Statistics in blocking solution (1:100 for rabbit antihuman CD3, Dako A0452). The next day, sections were incubated for 1 h at RT with secondary All graphics are shown as mean ± SEM, except for tidal volume and antibody (1:200 biotinylated goat anti-rabbit, Vector BA-1000) and respiratory frequency (box and whisker plot). The plots were created then with horseradish peroxidase-coupled streptavidin (1:600, Vector using GraphPad 8 software (GraphPad Software, Inc.). Statistical ana- SA-5004). The immunoreactivity was visualized using 0.5 mg/ml lyses were performed using the Statistical Package for Social Sciences 3,3-diaminobenzidine and 0.033% hydrogen peroxide for 10 min at (SPSS) 19 (IBM). Survival was analyzed using log-rank survival tests, RT. Finally, the slides were dehydrated and covered with DPX. A posi- with genotype as a factor. For repeated measures (weight), a general- tive control (IL-6 overexpressing mice) with CD3+ cell infiltrates was ized estimating equation (GEE) analysis was used. A generalized linear used to correctly detect CD3+ cells (Giralt et al., 2013). model (GzLM) was used to analyze the results of rotarod, open-field, tibial length, respiration parameters, immunofluorescence quantifica- tion, and in situ hybridization. A chi-square test was used for the fre- 2.10 | In situ hybridization using RNAscope quency analysis (rotation behavior). For the PLX3397 experiments, both male and female mice were included in similar proportions and In situ hybridization (ISH) using RNAscope Multiplex Fluorescent kit analyzed together (no sex effect was detected). Regarding double (Advanced Cell Diagnostics) was used to detect gad2 (GABAergic neu- deficiency experiments, male and female mice were separately ana- rons) and Slc17a6 (glutamatergic neurons) transcripts. Briefly, 48– lyzed since IL-6 has been described to exert sex-dependent effects in 49-day old mice were euthanized, and their brains were embedded in some neurodegenerative diseases (Miller et al., 2010; Sanchis, Fernán- OCT medium and frozen with dry ice. Fourteen-micrometer brain sec- dez-Gayol, Comes, et al., 2020). For tidal volume, respiratory fre- tions containing the VN region were obtained, directly mounted on quency, and multiplex results, the sexes were analyzed together slides, and kept at 80 C until use. First, sections (between bregma owing to a lower number of experimental animals and similar results 5.88 and  6.00 mm) were fixed in cold PBS containing 4% PFA for for both sexes in these variables. The major factors included in the 30 min at RT. Following two quick washes in PBS, the brain slices statistical analysis for GzLM and GEE were genotype (WT vs. KO) and were dehydrated in 50% (5 min), 70% (5 min), and 2  100% (5 min treatment (vehicle vs. PLX3397) in the case of PLX3397 experiments each) ethanol and treated with protease IV solution at RT for 30 min. and IL-6 (WT vs. KO) and Ndufs4 (WT vs. KO) in the case of double 2036 AGUILAR ET AL. F I G U R E 1 Inhibition of the CSF1R extends lifespan and delays motor symptoms in Ndufs4 KO mice. (a) Kaplan–Meier survival curve for Ndufs4 KO mice treated with vehicle (n = 12) or PLX3397 (n = 16). (b–e) Effects of PLX3397 over the motor phenotype of Ndufs4 KO. In order: Latency to fall in the accelerating rotarod, total distance traveled in the open-field test, rotation behavior in vehicle and PLX3397 treated Ndufs4 KO mice at the late stage open-field, and clasping onset in vehicle and PLX3397 treated Ndufs4 KO mice. (f) Tidal volume and respiratory frequency of WT and Ndufs4 KO mice treated with vehicle or PLX3397 at P47–P48, WT-Veh (n = 12), WT-PLX3397 (n = 12), Ndufs4 KO-Veh (n = 10), Ndufs4 KO-PLX3397 (n = 13). Long-rank test for survival, GzLM statistical tests for rotarod, open-field, tidal volume, and respiratory frequency, chi-square test for rotation behavior, and Student's t-test for clasping. Numbers within bars represent the “n” of the group. ● p ≤ .05 (WT vs. Ndufs4 KO), ★ treatment effect p ≤ .05 (Veh vs. PLX3397), ♦ p ≤ .05 interaction between both factors, * p = .05–.01, *** p ≤ .001. WT (wild-type), Veh (vehicle) deficiency experiments. In some cases, the WT/Il6 KO or Ndufs4 KO/ effects and prevent the disruption of other cellular neuroimmune Double KO groups were analyzed separately, including genotype and components, we considered it convenient to administer PLX3397 at a age as factors (as specified in the text). For comparisons between two dosage causing partial microglial depletion but still capable of exerting groups (i.e., clasping), Student's t test was used. One-way ANOVA fol- a significant effect on disease progression. In WT mice, daily injection lowed by Tukey's multiple comparison test was used for the dosage of PLX3397 caused dose-dependent microglial depletion (Figure S1A). pilot study analysis. Statistical significance was defined as p ≤ .05. A daily dose of 40 mg/kg could reduce 70% of the IBA-1+ cells across the whole brain after 14 days of treatment; thus, we used this dose for further experiments. 3 | RESULTS As expected, Ndufs4 KO mice exhibited early mortality (Figure 1a), progressive motor decline (Figure 1b–e), and altered respi- 3.1 | Microglia depletion increases lifespan and ration (Figure 1f). They also showed lower body weight (Figure S1B) improves the motor performance of Ndufs4 KO mice and growth (Figure S3B). Microglial depletion significantly rescued several abnormalities. Thus, it prominently increased the survival rate Resembling Leigh syndrome pathology (Lake et al., 2016; Lee (Figure 1a) and decreased progressive deterioration of balance and et al., 2020), Ndufs4 KO mice develop a fatal, early-onset progressive motor coordination in the rotarod (Figure 1b). While vehicle-treated encephalopathy, characterized by respiratory deficiencies, a strong Ndufs4 KO mice were mostly incapable of performing the rotarod test decline in motor function, and neurodegeneration and neuroinflam- by 2 months (late stage), PLX3397-treated mice showed improved mation in several brain areas (Johnson et al., 2013; Kruse et al., 2008; motor performance at all stages. However, this was statistically signifi- Quintana et al., 2010). However, the role of neuroinflammation is cant only at the corresponding mid (P38–P39) and late stages (P47– poorly understood. Here, we induced neuroimmune disruption by P48) of the disease (Ndufs4 KO mice treated with vehicle versus pharmacologically depleting microglial cells using the well-known PLX3397, p = .042 and p = .009, respectively). This presumably inhibitor of CSF1R, PLX3397 (Elmore et al., 2014). To avoid off-target reflects the effects on motor coordination rather than general activity AGUILAR ET AL. 2037 since microglia depletion did not rescue ambulation in the open-field reproducibility of the experiments. Ndufs4 KO mice also showed dra- test (Figure 1c). Microglia depletion also reduced the frequency of matic microglial depletion both in areas where microgliosis is known rotation behavior (mice circling on themselves, possibly due to vestib- to occur (VN, OB, and cerebellum) and in those lacking it (cortex and ular lesions) in Ndufs4 KO mice at the P47–48 stage (p = .014) hippocampus). In comparison, massive cell accumulation hindered (Figure 1d). This also delayed the onset of clasping/twisting behavior adequate cellular individualization and, thereby, a precise cell count in in NDUFS4-deficient mice (Figure 1e) at the P47–48 stage (p ≤ .001). the first three areas. This was possible in the cortex and hippocampus, PLX3397 administration showed a clear trend of a reduction in total where PLX3397 administration reduced IBA-1+ cells by 70% and body weight of male and female WT mice but did not modify the 60%, respectively. Different results depending on the brain area were decreased body weight of Ndufs4 KO mice (Figure S1B). observed regarding GFAP immunoreactivity due to PLX3397 adminis- Patients with LS normally present with respiratory abnormalities tration (Figure 2A,B). Thus, in the VN, PLX3397 did not affect GFAP that are strongly associated with mortality (Arii & Tanabe, 2000). Con- immunofluorescence in either WT or Ndufs4 KO mice, whereas in the stitutive Ndufs4 KO mice also show abnormal breathing patterns, cerebellum, a reduction was observed in both the genotypes. In con- which increase in severity as pathology progresses (Quintana trast, within the OB, cortex, and hippocampus, PLX3397 treatment et al., 2012). Respiratory rhythm generation and control include increased GFAP immunofluorescence. Taken together, the GFAP numerous CNS regions, such as the VN, cerebellum, and different immunofluorescence results suggest that astrocytes respond to the brainstem nuclei. Respiratory abnormalities in Ndufs4 KO mice are microglial reduction in a region-dependent manner. These findings are partially associated with mitochondrial dysfunction due to NDUFS4 consistent with those of previous reports (Najafi et al., 2018). loss in the VN and abnormal responses in the pre-Bötzinger complex of the brainstem (Quintana et al., 2012). In line with these findings, using whole-body plethysmography (Prada-Dacasa et al., 2020), we 3.3 | Microglial response promotes both observed an increased tidal volume (volume of air inspired by the ani- glutamatergic and GABAergic neuronal loss in Ndufs4 mal in one breath) in Ndufs4 KO mice at P47–P48 (Figure 1F, left), but KO mice no changes in respiratory frequency (Figure 1F, right). Microglial depletion significantly increased the tidal volume, while a similar trend After showing that PLX3397 treatment could prevent microglial accu- was observed for respiratory frequency (PLX3397 effect, p = .09). mulation in major neurodegenerative areas of Ndufs4 KO, we investi- gated whether a reduction in the microglial response could prevent neuronal loss in Ndufs4 KO mice in two areas with massive neuroin- 3.2 | PLX3397 treatment protects against IBA-1+ flammation, the OB and the VN. By quantifying the area occupied by cell accumulation and proliferation in brain regions NeuN immunostaining, we observed that microglial depletion partially associated with neurodegeneration in Ndufs4 KO mice protected against neuronal loss in the granular cell layer of the olfactory bulb (GrOB) (Figure 3a). In the whole VN, in contrast, microglial deple- Ndufs4 KO mice present with a characteristic fatal encephalopathy in tion did not prevent neuronal loss (Figure 3b). However, a more thor- which neurodegeneration and neuroinflammation occur in specific ough inspection indicated that although extensive microgliosis occurred brain regions. These two processes are mainly observed in the VN, in the whole VN of Ndufs4 KO mice, it was especially dramatic between OB, brainstem, cerebellum, and hippocampus (Johnson et al., 2013; the magnocellular part of the medial vestibular nucleus (MVeMC) and Miller et al., 2020; Quintana et al., 2010; Shil et al., 2021). This neuro- the lateral vestibular nucleus (LVe), where extensive bilateral symmetri- degenerative and neuroinflammatory process is the main cause of the cal lesions appeared (Figure 3d) (Quintana et al., 2010). Analysis of neu- Ndufs4 KO phenotype since brain-specific NDUFS4 deficiency (using ronal loss in this area indicated that microglial depletion tended to Nes-Cre mice) results in an almost identical phenotype to that of sys- rescue neuronal loss (Figure 3c). Since NeuN is a general neuronal temic Ndufs4 KO (Quintana et al., 2010). Furthermore, conditional marker, we sought to investigate if specific neuronal populations were deletion of Ndufs4 in different neuron subtypes was accompanied by affected by microglial depletion in the VN. To this end, we used ISH to neuroinflammation and contributed differently to the pathology assess specific transcripts as markers of glutamatergic (Slc17a6) and (Bolea et al., 2019). Here, we wanted to examine whether microglial GABAergic (Gad2) neurons (Figures 3e,f, and S2). Microglial depletion depletion could deplete or reduce microglia in Ndufs4 KO mice and rescued both glutamatergic and GABAergic neuronal loss in the VN prevent IBA-1+ cell accumulation in the known neurodegenerative (Figure S2A). When the VN was divided into non-lesioned (MVePC, areas found in late-stage mice. As expected, prominent gliosis SuVe, and the medial part of the MVeMC) (Figure 3e, left and S2B) and (as revealed by IBA-1 and GFAP immunostaining) was observed in the lesioned areas (lateral part of MVeMC and LVe) (Figure 3e, right and f), VN, OB, and cerebellum, where massive neurodegeneration was microglial depletion rescued neuronal losses in Ndufs4 KO mice, mainly described in vehicle-treated Ndufs4 KO mice (Figure 2a,b). In addition, in the lesion-associated area (Ndufs4 KO mice vehicle versus PLX3397, GFAP (but not IBA-1) immunostaining clearly increased in the cortex p = .002). This reduced neuronal loss can be attributed to both gluta- and hippocampus (Figure 2a). Treatment with PLX3397 caused a matergic (Figure 3e, middle) and GABAergic neuronal preservation robust decrease in IBA-1 fluorescence intensity in all the studied brain (Figure 3e, bottom). Collectively, these results indicate that microglial areas. WT animals showed microglial depletion like that observed in cells promote neuronal loss in Ndufs4 KO mice independent of the neu- the pilot dosage study (Figure S1A), which confirmed the ronal subtype. 2038 AGUILAR ET AL. F I G U R E 2 Inhibition of the CSF1R prevents IBA-1+ cell accumulation in neurodegeneration-related areas of Ndufs4 KO mice brains. (a) IBA-1 fluorescence intensity or total IBA-1+ positive cell and GFAP fluorescence intensity in VN, OB, cerebellum, cortex, and hippocampus of WT and Ndufs4 KO mice treated with vehicle or PLX3397 at P47–48 days of age. GzLM statistical tests. The experimental “n” is the same for all graphs and stated in VN IBA-1. (b) Representative images of IBA-1 and GFAP immunofluorescence in VN, OB, and cerebellum. ● p ≤ .05 (WT vs. Ndufs4 KO), ★ treatment effect p ≤ .05 (Veh vs. PLX3397), ♦ p ≤ .05 interaction between both factors. Scale bar 100 μm. VN (vestibular nuclei), OB (olfactory bulb) AGUILAR ET AL. 2039 F I G U R E 3 Microglia mediates neuronal loss in Ndufs4 KO mice. (a) Percentage of the area occupied by NeuN staining in the GrOB. (b) Percentage of the area occupied by NeuN staining in the whole VN and the area associated with the main lesion. (c) Image of IBA-1 staining of the VN of a Ndufs4 KO mouse showing where Ndufs4 KO mice use to develop massive lesions. (d) Number of neurons of both glutamatergic and GABAergic subtypes in the non-lesioned and the lesioned area of the VN of Ndufs4 KO mice, and the number of neurons subdivided by the neuronal subtype. (e) Representative ISH images of glutamatergic and GABAergic neurons of the different genotypes in the lesioned area of the VN. (a, b, and d) GzLM statistical tests. ● p ≤ .05 (WT vs. Ndufs4 KO), ★ treatment effect p ≤ .05 (Veh vs. PLX3397), ♦ p ≤ .05 interaction between both factors. Scale bar 100 μm. GrOB (granular cell layer of the olfactory bulb), VN (vestibular nucleus), MVePC (medial vestibular nucleus, parvicellular part), MVeMC (medial vestibular nucleus, magnocellular part), SuVe (superior vestibular nucleus), LVe (lateral vestibular nucleus) 3.4 | Lack of evidence of infiltrating CD3+ cells in presented T lymphocyte infiltrates in the most prominent neuro- motor-related neurodegenerative regions of Ndufs4 degenerative areas and whether PLX3397 treatment modified KO mice them to some extent. Interestingly, we did not find CD3 + cell infiltrates in either the cerebellum or the VN of Ndufs4 KO mice at Some reports using PLX3397 at concentrations like those used in the late stage (Figure 4a, right), but substantial CD3 infiltrates this study have described that besides depleting microglia, it may were present in the OB at this stage, which were not affected by also alter the populations of other immune cells (Han et al., 2020; microglial depletion (Figure 4b, left, right). A GFAP-IL6 mouse cer- Lei et al., 2020). Thus, we investigated whether Ndufs4 KO mice ebellum (a chronic model of neuroinflammation with CD3 + cell 2040 AGUILAR ET AL. F I G U R E 4 T-cells do not infiltrate into the VN and the cerebellum of Ndufs4 KO mice, but they do infiltrate in the OB at the late stage of the disease (a) image showing the positive control used in the CD3 immunohistochemistry (GFAP-IL6) together with representative images of the cerebellum and the VN of Ndufs4 KO mice showing no CD3+ cells infiltration in these regions at P48–P49. (b) Number of CD3+ cells in the OB of the different genotypes at P48–P49 together with representative images. Black arrows indicate CD3+ cells. GzLM statistical test. ● p ≤ .05 (WT vs. Ndufs4 KO) infiltrates) was used as a positive control (Giralt et al., 2013) areas at the late stage of the disease compared with WT mice (WT vs. (Figure 4a, left). Ndufs4 KO, p = .005 and p = .001 for OB and cerebellum, respectively) (Figure 5A, left). This observation is consistent with previous studies (Balsa et al., 2020; McElroy et al., 2020). The lack of IL-6 was confirmed 3.5 | IL-6 levels are increased in Ndufs4 KO mice in the IL-6 deficient groups. Moreover, we investigated whether other brains, but IL-6 deficiency does not protect against pro-inflammatory cytokines could also be involved and if IL-6 deficiency Ndufs4 KO clinical signs or neuropathology could modify their regulation. In line with the described neuroinflamma- tion in Ndufs4 KO mice, we also observed a significant increase in TNF- Once we demonstrated that the microglial response is noxious in α in the OB of Ndufs4 KO and Double KO mice at the late stage com- Ndufs4 KO mice, we next investigated one of the potential candidates pared with their littermates (Ndufs4 effect, p ≤ .001), and the same implicated in this process, the cytokine IL-6. IL-6 is highly induced in trend was observed in the cerebellum (Ndufs4 effect, p = .13) chronic neurodegenerative diseases, such as Alzheimer's disease, Par- (Figure 5a, right). Interestingly, despite the massive neuroinflammation kinson's disease, and multiple sclerosis, with potentially detrimental present in NDUFS4-deficient mice at late stages of pathology (Johnson roles (Erta et al., 2012; Spooren et al., 2011). To study putative IL-6 et al., 2013; Quintana et al., 2010), IL-1β levels were undetectable in contributions to the Ndufs4 KO mouse phenotype, we generated a any genotype or region (not shown). mouse line with combined IL-6 and NDUFS4 deficiencies and their Additionally, IL-6 deficiency in Ndufs4 KO male and female mice respective littermate controls. First, we investigated IL-6 protein levels did not modify survival (Figure 5b) or motor coordination (Figure 5c). In by multiplex analysis in two of the main neurodegenerative regions of addition, IL-6 deficiency did not rescue Ndufs4 KO mice in terms of Ndufs4 KO mice, the OB and cerebellum (Quintana et al., 2010). IL-6 body weight and growth, total motor activity, vertical behavior, or the levels were significantly increased in Ndufs4 KO mice in both brain onset of clasping, regardless of sex (Figure S3). In contrast, IL-6 AGUILAR ET AL. 2041 F I G U R E 5 IL-6 levels are increased in Ndufs4 KO mice brains, but IL-6 deficiency does not modify disease progression. (a) IL-6 and TNF-α levels in the olfactory bulb and cerebellum at the late stage. (b) Kaplan–Meier survival curve for WT (n = 14 males, n = 11 females), Il6 KO (n = 15 males, n = 13 females), Ndufs4 KO (n = 18 males, n = 17 females), and double KO (n = 22 males, n = 16 females) mice. (c) Latency to fall in the accelerating rotarod. (d) Whole-body plethysmography for assessing breathing alterations. For both tidal volume and respiratory frequency, WT (early, n = 9; mid, n = 9; late, n = 8), Il6 KO (early, n = 8; mid, n = 10; late, n = 9), Ndufs4 KO (early, n = 12; mid, n = 11; late, n = 7) and double KO (early, n = 12; mid, n = 17; late, n = 10). (a–d) GzLM statistical test for a, c, and d; long-rank test for survival for b. ● Ndufs4 effect p ≤ .05 (WTs vs. Ndufs4 KOs), ★ IL-6 effect p ≤ .05 (WTs vs. Il6 KOs), ♦ p ≤ .05 interaction between both factors deficiency rescued the increased respiratory tidal volume shown by lifespan or rescue motor decline in Ndufs4 KO mice, such as hypoxia Ndufs4 KO mice (Figure 5D, left), particularly at the mid-stage of the (Jain et al., 2016; Jain et al., 2019), rapamycin treatment (Johnson disease. This effect could not be explained by neuronal loss or clear et al., 2013), or doxycycline administration (Perry et al., 2021), have neuroinflammatory changes in important breathing-related regions dur- also been shown to reduce gliosis and immune-related pathways in ing the mid-stage of the disease (Figure S4). We did not observe Ndufs4 KO brains. Whether this neuroimmune process is a pathologi- changes in respiratory frequency due to NDUFS4 deficiency at any cal mechanism contributing to the Ndufs4 KO phenotype or simply is stage, whereas IL-6 deficiency tended to increase respiratory frequency, a consequence of neurodegeneration deserves adequate investigation particularly in NDUFS4 non-deficient mice (Figure 5d, right). This was in the context of complex I-associated mitochondrial disorders. Here, statistically significant when comparing IL-6 deficient and WT mice with we report one of the first pieces of evidence that depicts microglia as genotype and age as statistical factors (WT vs. Il6 KO, p = .001). a mediator of Ndufs4 KO pathology since its reduction improved the In line with the behavioral results, double KO mice did not show motor phenotype and decreased the neuropathology in these mice. substantial changes in gliosis at either the mid (Figure S5) or late-stage Importantly, these results strongly suggest that microglia may be con- of the diseases (Figures S6 and S7), in which IL-6 deficiency increased trolling neuronal survival. Interestingly, despite VN glutamatergic neu- microgliosis in Ndufs4 KO mice, particularly in male mice (Figure S7). rons that have been shown to be susceptible to NDUFS4 deficiency (Bolea et al., 2019), we found that microglia promote both glutamater- gic and GABAergic neuron loss, indicating that microglial-mediated 4 | DISCUSSION neuronal loss is not restricted to susceptible neurons. In addition, increased neuronal preservation in the GrOB can be mostly attributed As shown above, neuroinflammation is a key manifestation of the to the GABAergic neuronal subpopulation since it is eminently a Ndufs4 KO mice pathology. Interventions that either extend the GABAergic neuronal region (Nagayama et al., 2014). In contrast, the 2042 AGUILAR ET AL. VN is mainly composed of glutamatergic and GABAergic neuronal respiratory frequency in nonpathological conditions. To our knowl- subpopulations as well as a small glycinergic subpopulation (Bagnall edge, this is the first direct report showing that IL-6 can also modulate et al., 2007). The precise pathways by which microglia promote neuro- breathing. Since we found no differences in gliosis between Ndufs4 nal death remain unclear and require further investigation. KO and Double KO mice at the mid-stage, we hypothesize that IL-6 A recently published study using CSF1R inhibition with PLX3397, has an intrinsic capacity to affect breathing, as observed in the case of but at higher concentrations, which may target and eliminate a wider other cytokines (Giannakopoulou et al., 2019; Lorea-Hernández range of immune cell types than microglia and myeloid CNS- et al., 2020). This could be achieved by changing neuronal firing or associated cells, showed an increased lifespan in Ndufs4 KO mice playing a relevant role in glia–neuron communication (Peña- (Stokes et al., 2022). An interesting debate in the scientific community Ortega, 2019). exists regarding whether CSF1R inhibition targets microglial cells Regarding the strong neuroimmune response present in Ndufs4 almost exclusively or if other immune cells are also being affected, KO mice, IL-6 signaling is only one of the multiple neuroinflammatory which could potentially have relevant roles in CNS diseases. It has mediators that can be the contributing factor. Mitochondrial defects been speculated that the higher the concentration of PLX3397, the have been known to elicit different inflammatory responses, including higher are the number of targets affected (i.e., circulating monocyte TLR9, inflammasome pathway, or IFN type I responses (Luna-sánchez subpopulation, some T-cell subpopulations, or other CNS-associated et al., 2021; Zhong et al., 2019). We did not find evidence of inflam- myeloid cells) (Green & Hume, 2021; Han et al., 2020; Lei masome activation because one of the main mechanisms of action of et al., 2020). Although the concentration used in this study is within this complex is the final conversion of pro-IL-1β into its active form the standards used in the field and is supposed to act exclusively on (Yang et al., 2019), and we could not detect IL-1β in Ndufs4 KO mice. microglia, we cannot rule out the possibility that our dose and timing Interestingly, some studies have reported interferon-related response of PLX3397 also eliminate some other immune cell subpopulations. It activation in NDUFS4-deficient mice (McElroy et al., 2020; Perry seems reasonable to speculate that the dose-dependent effect of et al., 2021) and other mitochondrial disease mouse models (Y. Lei PLX3397 on the Ndufs4 KO mouse phenotype may be associated with et al., 2021). Thus, this pathway could be a potential candidate associ- the type(s) of the immune cell population being eliminated. Hence- ated with the microglial response present in Ndufs4 KO mice and forth, more efforts are needed to elucidate immune cell-type-specific deserves further attention. contributions to this type of disorder (i.e., peripheral or central ner- Overall, our results indicate that microglial depletion using vous system-associated macrophages) and to mechanistically describe PLX3397 has a protective effect in Ndufs4 KO mice, revealing the how this immune interplay affects Ndufs4 KO mouse pathology. Ear- pathogenicity of the neuroimmune response in Ndufs4 KO mice as lier, Strokes et al. claimed to obtain leukocyte depletion; however, no more than just a consequence of neurodegeneration, and serves as a other immune subpopulations apart from IBA-1+ were investigated in proof-of-concept to study neuroimmune-related processes in other the context of PLX3397 administration. To look at other leukocyte mitochondrial diseases that affect the brain. + subpopulations, apart from those of myeloid origin (IBA-1 cells), we immunohistochemically stained for CD3, a well-described pan-T-cell AUTHOR CONTRIBU TIONS marker, in the neurodegenerative areas of Ndufs4 KO mice. The Kevin Aguilar, Albert Quintana, Elisenda Sanz, and Juan Hidalgo con- absence of CD3+ cell accumulation in the main areas responsible for ceived and designed the study. Kevin Aguilar, Gemma Comes, and the major symptoms of Ndufs4 KO (VN and cerebellum) reinforces the Carla Canal performed research. Kevin Aguilar analyzed data. Kevin idea that the pathogenic neuroimmune response present in these Aguilar and Juan Hidalgo wrote the initial draft of the manuscript. All mice is mainly guided by innate immunity mechanisms, as observed in authors edited, reviewed, and revised the manuscript. RNAseq data (McElroy et al., 2020; Perry et al., 2021). IL-6 is an important mediator of neuroinflammation and is known ACKNOWLEDG MENTS to have strong detrimental effects during chronic neuroinflammation. Kevin Aguilar and Carla Canal acknowledge the support of Therefore, we hypothesized that IL-6 could potentially guide micro- FPU17/02065 and PIF UAB fellowships, respectively. This study was glial activity in Ndufs4 KO mice. Collectively, our results indicate that n y Universidades y supported by Ministerio de Ciencia, Innovacio IL-6 is not behind the major microglial effects and that other media- Fondo Europeo de Desarrollo Regional (RTI2018-101105-B-I00; Juan tors guide the response since its deficiency does not ameliorate the Hidalgo) [Correction added on August, 11, 2022, after first online Ndufs4 KO mice neuropathology. However, IL-6 deficiency partially publication: the grant/award number for the European Regional reversed the respiratory abnormalities observed in these mice. Respi- Development Fund has been changed from RTI2018-101105–I00 to ration rhythm generation and coordination require complex circuitry RTI2018-101105-B-I00.]. Albert Quintana and Elisenda Sanz received in which the VN, MCN, and brainstem are key components. Respira-  n y Cajal fellowships (RyC-2012-11873; Albert Quintana, Ramo tion is strongly influenced by inflammation (Peña-Ortega, 2019). Fur- RYC2019-028501-I; Elisenda Sanz). Elisenda Sanz received funds thermore, microglia and different inflammatory mediators can from MICIU Proyectos I+D+i “Retos  n” Investigacio modulate respiratory rhythm (Lorea-Hernández et al., 2016; Peña- (RTI2018-101838-J-I00). Albert Quintana received funds from the Ortega, 2019). For instance, IL-1β (Lorea-Hernández et al., 2020) and European Research Council (Starting grant NEUROMITO, ERC- IL-10 (Giannakopoulou et al., 2019) can modulate tidal volume and 2014-StG-638106), MINECO Proyectos I+D de Excelencia AGUILAR ET AL. 2043 (SAF2014-57981P; SAF2017-88108-R), MICINN Proyectos I + D + i Giralt, M., Ramos, R., Quintana, A., Ferrer, B., Erta, M., Castro-Freire, M., (PID2020-114977RB-I00), AGAUR (2017SGR-323) and “la Caixa” Comes, G., Sanz, E., Unzeta, M., Pifarré, P., García, A., Campbell, I. L., & Hidalgo, J. (2013). Induction of atypical EAE mediated by transgenic Foundation (ID 100010434), under the agreement LCF/PR/ production of IL-6 in astrocytes in the absence of systemic IL-6. Glia, HR20/52400018. We acknowledge Prof. Ian L. Campbell for the 61(4), 587–600. https://doi.org/10.1002/glia.22457 GFAP-IL6 mouse. We also want to thank Dr. Gemma Manich for kindly Gorman, G. S., Chinnery, P. F., DiMauro, S., Hirano, M., Koga, Y., providing anti-TMEM119. McFarland, R., Suomalainen, A., Thorburn, D. R., Zeviani, M., & Turnbull, D. M. (2016). Mitochondrial diseases. Nature Reviews. Disease Primers, 2, 16080. https://doi.org/10.1038/nrdp.2016.80 CONF LICT OF IN TE RE ST Green, K. N., Crapser, J. D., & Hohsfield, L. A. (2020). To kill a microglia: A The authors claim no competing interests. case for CSF1R inhibitors. Trends in Immunology, 41(9), 771–784. https://doi.org/10.1016/J.IT.2020.07.001 Green, K. N., & Hume, D. A. (2021). On the utility of CSF1R inhibitors. Pro- DATA AVAI LAB ILITY S TATEMENT ceedings of the National Academy of Sciences, 118(4), 1–2. https://doi. The data that support the findings of this study are available from the org/10.1073/PNAS.2019695118 corresponding author upon reasonable request. Han, J., Fan, Y., Zhou, K., Zhu, K., Blomgren, K., Lund, H., Zhang, X. M., & Harris, R. A. (2020). Underestimated peripheral effects following phar- macological and conditional genetic microglial depletion. International ORCID Journal of Molecular Sciences, 21(22), 8603. https://doi.org/10.3390/ Juan Hidalgo https://orcid.org/0000-0003-0921-1122 IJMS21228603 Jain, I. H., Zazzeron, L., Goldberger, O., Marutani, E., Wojtkiewicz, G. R., Ast, T., Wang, H., Schleifer, G., Stepanova, A., Brepoels, K., RE FE R ENC E S Schoonjans, L., Carmeliet, P., Galkin, A., Ichinose, F., Zapol, W. M., & Arii, J., & Tanabe, Y. (2000). Leigh syndrome: Serial MR imaging and clinical Mootha, V. K. (2019). Leigh syndrome mouse model can be rescued by follow-up. American Journal of Neuroradiology, 21(8), 1502–1509. interventions that normalize brain Hyperoxia, but not HIF activation. http://www.ajnr.org/content/21/8/1502 Cell Metabolism, 30(4), 824–832.e3. https://doi.org/10.1016/j.cmet. Bagnall, M. W., Stevens, R. J., & du Lac, S. (2007). Transgenic mouse lines 2019.07.006 subdivide medial vestibular nucleus neurons into discrete, Neuro- Jain, I. H., Zazzeron, L., Goli, R., Alexa, K., Schatzman-Bone, S., Dhillon, H., chemically distinct populations. Journal of Neuroscience, 27(9), 2318– Goldberger, O., Peng, J., Shalem, O., Sanjana, N. E., Zhang, F., 2330. https://doi.org/10.1523/JNEUROSCI.4322-06.2007 Goessling, W., Zapol, W. M., & Mootha, V. K. (2016). Hypoxia as a Balsa, E., Perry, E. A., Bennett, C. F., Jedrychowski, M., Gygi, S. P., therapy for mitochondrial disease. Science, 352(6281), 54–61. https:// Doench, J. G., & Puigserver, P. (2020). Defective NADPH production doi.org/10.1126/SCIENCE.AAD9642 in mitochondrial disease complex I causes inflammation and cell death. Johnson, S. C., Yanos, M. E., Kayser, E. B., Quintana, A., Sangesland, M., Nature Communications, 11(1), 1–12. https://doi.org/10.1038/ Castanza, A., Uhde, L., Hui, J., Wall, V. Z., Gagnidze, A., Oh, K., s41467-020-16423-1 Wasko, B. M., Ramos, F. J., Palmiter, R. D., Rabinovitch, P. S., Bennett, M. L., Bennett, F. C., Liddelow, S. A., Ajami, B., Zamanian, J. L., Morgan, P. G., Sedensky, M. M., & Kaeberlein, M. (2013). MTOR inhi- Fernhoff, N. B., Mulinyawe, S. B., Bohlen, C. J., Adil, A., Tucker, A., bition alleviates mitochondrial disease in a mouse model of Leigh syn- Weissman, I. L., Chang, E. F., Li, G., Grant, G. A., Hayden drome. Science, 342(6165), 1524–1528. https://doi.org/10.1126/ Gephart, M. G., & Barres, B. A. (2016). New tools for studying micro- SCIENCE.1244360 glia in the mouse and human CNS. Proceedings of the National Academy Jones, S. A., & Jenkins, B. J. (2018). Recent insights into targeting the IL-6 of Sciences, 113(12), 1738–1746. https://doi.org/10.1073/pnas. cytokine family in inflammatory diseases and cancer. Nature Reviews 1525528113 Immunology, 18, 773–789. https://doi.org/10.1038/s41577-018- Bolea, I., Gella, A., Sanz, E., Prada-Dacasa, P., Menardy, F., Bard, A. M., 0066-7 Machuca-Márquez, P., Eraso-Pichot, A., Mòdol-Caballero, G., Kruse, S. E., Watt, W. C., Marcinek, D. J., Kapur, R. P., Schenkman, K. A., & Navarro, X., Kalume, F., & Quintana, A. (2019). Defined neuronal popu- Palmiter, R. D. (2008). Mice with mitochondrial complex I deficiency lations drive fatal phenotype in a mouse model of Leigh syndrome. develop a fatal Encephalomyopathy. Cell Metabolism, 7(4), 312–320. eLife, 8, e47163. https://doi.org/10.7554/ELIFE.47163 https://doi.org/10.1016/J.CMET.2008.02.004 Elmore, M. R. P., Najafi, A. R., Koike, M. A., Dagher, N. N., Lake, N. J., Compton, A. G., Rahman, S., & Thorburn, D. R. (2016). Leigh Spangenberg, E. E., Rice, R. A., Kitazawa, M., Matusow, B., Nguyen, H., syndrome: One disorder, more than 75 monogenic causes. Annals of West, B. L., & Green, K. N. (2014). Colony-stimulating factor 1 receptor Neurology, 79, 190–203. https://doi.org/10.1002/ana.24551 signaling is necessary for microglia viability, unmasking a microglia pro- Lee, J. S., Yoo, T., Lee, M., Lee, Y., Jeon, E., Kim, S. Y., Lim, B. C., Kim, K. J., genitor cell in the adult brain. Neuron, 82(2), 380–397. https://doi.org/ Choi, M., & Chae, J. H. (2020). Genetic heterogeneity in Leigh syn- 10.1016/j.neuron.2014.02.040 drome: Highlighting treatable and novel genetic causes. Clinical Genet- Erta, M., Quintana, A., & Hidalgo, J. (2012). Interleukin-6, a major cytokine ics, 97(4), 586–594. https://doi.org/10.1111/CGE.13713 in the central nervous system. International Journal of Biological Sci- Lei, F., Cui, N., Zhou, C., Chodosh, J., Vavvas, D. G., & Paschalis, E. I. ences, 8(9), 1254–1266. https://doi.org/10.7150/ijbs.4679 (2020). CSF1R inhibition by a small-molecule inhibitor is not microglia Fassone, E., & Rahman, S. (2012). Complex I deficiency: Clinical features, specific; affecting hematopoiesis and the function of macrophages. biochemistry and molecular genetics. Journal of Medical Genetics, Proceedings of the National Academy of Sciences, 117(38), 23336– 49(9), 578–590. https://doi.org/10.1136/JMEDGENET-2012-101159 23338. https://doi.org/10.1073/PNAS.1922788117 Giannakopoulou, C. E., Sotiriou, A., Dettoraki, M., Yang, M., Perlikos, X. F., Lei, Y., Guerra Martinez, C., Torres-Odio, S., Bell, S. L., Birdwell, C. E., Toumpanakis, D., Prezerakos, G., Koutsourelakis, I., Kastis, G. A., Bryant, J. D., Tong, C. W., Watson, R. O., West, L. C., & West, A. P. Vassilakopoulou, V., Mizi, E., Papalois, A., Greer, J. J., & (2021). Elevated type I interferon responses potentiate metabolic dys- Vassilakopoulos, T. (2019). Regulation of breathing pattern by IL-10. function, inflammation, and accelerated aging in mtDNA mutator mice. American Journal of Physiology. Regulatory, Integrative and Comparative Science Advances, 7(22), 1–16. https://doi.org/10.1126/SCIADV. Physiology, 317, 190–202. https://doi.org/10.1152/ajpregu.00065.2019 ABE7548 2044 AGUILAR ET AL. Li, Q., & Barres, B. (2018). Microglia and macrophages in brain homeostasis the National Academy of Sciences, 107(24), 10996–11001. https://doi. and diseases. Nature Reviews Immunology, 18(4), 225–242. https://doi. org/10.1073/PNAS.1006214107 org/10.1038/nri.2017.125 Quintana, A., Zanella, S., Koch, H., Kruse, S. E., Lee, D., Ramirez, J. M., & Lorea-Hernández, J. J., Camacho-Hernández, N. P., & Peña-Ortega, F. Palmiter, R. D. (2012). Fatal breathing dysfunction in a mouse model (2020). Interleukin 1-beta but not the interleukin-1 receptor antago- of Leigh syndrome. The Journal of Clinical Investigation, 122(7), 2359– nist modulates inspiratory rhythm generation in vitro. Neuroscience 2368. https://doi.org/10.1172/JCI62923 Letters, 734, 134934. https://doi.org/10.1016/J.NEULET.2020. Sanchis, P., Fernández-Gayol, O., Comes, G., Aguilar, K., Escrig, A., 134934 Giralt, M., Palmiter, R. D., & Hidalgo, J. (2020). A new mouse model to Lorea-Hernández, J. J., Morales, T., Rivera-Angulo, A. J., Alcantara- study restoration of interleukin-6 (IL-6) expression in a Cre-dependent Gonzalez, D., & Peña-Ortega, F. (2016). Microglia modulate respiratory manner: Microglial IL-6 regulation of experimental autoimmune rhythm generation and autoresuscitation. Glia, 64(4), 603–619. encephalomyelitis. Journal of Neuroinflammation, 17(1), 1–17. https:// https://doi.org/10.1002/GLIA.22951 doi.org/10.1186/S12974-020-01969-0 Luna-sánchez, M., Bianchi, P., & Quintana, A. (2021). Mitochondria- Sanchis, P., Fernández-Gayol, O., Vizueta, J., Comes, G., Canal, C., induced immune response as a trigger for neurodegeneration: A path- Escrig, A., Molinero, A., Giralt, M., & Hidalgo, J. (2020). Microglial cell- ogen from within. International Journal of Molecular Sciences, 22(16), derived interleukin-6 influences behavior and inflammatory response 8523. https://doi.org/10.3390/ijms22168523 in the brain following traumatic brain injury. Glia, 68(5), 999–1016. McElroy, G. S., Reczek, C. R., Reyfman, P. A., Mithal, D. S., https://doi.org/10.1002/GLIA.23758 Horbinski, C. M., & Chandel, N. S. (2020). NAD+ regeneration rescues Shil, S. K., Kagawa, Y., Umaru, B. A., Nanto-Hara, F., Miyazaki, H., lifespan, but not ataxia, in a mouse model of brain mitochondrial com- Yamamoto, Y., Kobayashi, S., Suzuki, C., Abe, T., & Owada, Y. (2021). plex I dysfunction. Cell Metabolism, 32(2), 301–308.e6. https://doi. Ndufs4 ablation decreases synaptophysin expression in hippocampus. org/10.1016/J.CMET.2020.06.003 Scientific Reports, 11(1), 10969. https://doi.org/10.1038/s41598-021- Miller, H. C., Louw, R., Mereis, M., Venter, G., Boshoff, J.-D., Mienie, L., 90127-4 Reenen, M., Venter, M., Lindeque, J. Z., Domínguez-Martínez, A., Spooren, A., Kolmus, K., Laureys, G., Clinckers, R., de Keyser, J., Quintana, A., & Westhuizen, F. H. (2021). Metallothionein 1 overex- Haegeman, G., & Gerlo, S. (2011). Interleukin-6, a mental cytokine. pression does not protect against mitochondrial disease pathology in Brain Research Reviews, 67, 157–183. https://doi.org/10.1016/j. Ndufs4 knockout mice. Molecular Neurobiology, 58(1), 243–262. brainresrev.2011.01.002 https://doi.org/10.1007/S12035-020-02121-Y Stokes, J. C., Bornstein, R. L., James, K., Park, K. Y., Spencer, K. A., Vo, K., Miller, V. M., Lawrence, D. A., Coccaro, G. A., Mondal, T. K., Andrews, K., Snell, J. C., Johnson, B. M., Morgan, P. G., Sedensky, M. M., Dreiem, A., & Seegal, R. F. (2010). Sex effects of Interleukin-6 defi- Baertsch, N. A., & Johnson, S. C. (2022). Leukocytes mediate disease ciency on neuroinflammation in aged C57Bl/6 mice. Brain Research, pathogenesis in the Ndufs4(KO) mouse model of Leigh syndrome. JCI 1318, 11–22. https://doi.org/10.1016/J.BRAINRES.2009.12.091 Insight, 7(5), e156522. https://doi.org/10.1172/jci.insight.156522 Nagayama, S., Homma, R., & Imamura, F. (2014). Neuronal organization of West, A. P., & Shadel, G. S. (2017). Mitochondrial DNA in innate immune olfactory bulb circuits. Frontiers in Neural Circuits, 98(8), 1–19. https:// responses and inflammatory pathology. Nature Reviews Immunology, doi.org/10.3389/fncir.2014.00098 17, 363–375. https://doi.org/10.1038/nri.2017.21 Najafi, A. R., Crapser, J., Jiang, S., Ng, W., Mortazavi, A., West, B. L., & Yang, Y., Wang, H., Kouadir, M., Song, H., & Shi, F. (2019). Recent Green, K. N. (2018). A limited capacity for microglial repopulation in advances in the mechanisms of NLRP3 inflammasome activation and the adult brain. Glia, 66(11), 2385–2396. https://doi.org/10.1002/glia. its inhibitors. Cell Death & Disease, 10(2), 1–11. https://doi.org/10. 23477 1038/s41419-019-1413-8 Paxinos, G., & Franklin, K. B. J. (2013). Paxinos and Franklin's the mouse Zhong, F., Liang, S., & Zhong, Z. (2019). Emerging role of mitochondrial DNA brain in stereotaxic coordinates. Academic Press. as a major driver of inflammation and disease progression. Trends in Peña-Ortega, F. (2019). Clinical and experimental aspects of breathing Immunology, 40, 1120–1133. https://doi.org/10.1016/j.it.2019.10.008 modulation by inflammation. Autonomic Neuroscience, 216, 72–86. https://doi.org/10.1016/J.AUTNEU.2018.11.002 Perry, E. A., Bennett, C. F., Luo, C., Balsa, E., Jedrychowski, M., SUPPORTING INF ORMATION O'Malley, K. E., Latorre-Muro, P., Ladley, R. P., Reda, K., Wright, P. M., Additional supporting information can be found online in the Support- Gygi, S. P., Myers, A. G., & Puigserver, P. (2021). Tetracyclines pro- ing Information section at the end of this article. mote survival and fitness in mitochondrial disease models. Nature Metabolism, 3(1), 33–42. https://doi.org/10.1038/s42255-020- 00334-y Prada-Dacasa, P., Urpi, A., Sánchez-Benito, L., Bianchi, P., & Quintana, A. How to cite this article: Aguilar, K., Comes, G., Canal, C., (2020). Measuring breathing patterns in mice using whole-body pleth- Quintana, A., Sanz, E., & Hidalgo, J. (2022). Microglial response ysmography. Bio-Protocol, 10(17), 3741. https://doi.org/10.21769/ promotes neurodegeneration in the Ndufs4 KO mouse model BIOPROTOC.3741 of Leigh syndrome. Glia, 70(11), 2032–2044. https://doi.org/ Quintana, A., Kruse, S. E., Kapur, R. P., Sanz, E., & Palmiter, R. D. (2010). Complex I deficiency due to loss of Ndufs4 in the brain results in pro- 10.1002/glia.24234 gressive encephalopathy resembling Leigh syndrome. Proceedings of