Human IgLON5 Antibodies Reduce Neuronal IgLON5 Clusters and Cause Motor Dysfunction in Mice

Abstract

Background and Objectives

Anti-IgLON5 disease is a progressive neurologic disorder characterized by sleep disturbances, gait instability, involuntary movements, and bulbar dysfunction. In long-standing cases, autopsy studies reveal a brainstem-predominant neuronal tauopathy. The disease is defined by antibodies against the neuronal adhesion molecule IgLON5 (IgLON5-abs), which reduce IgLON5 membrane clusters and disrupt the cytoskeleton in vitro. Our aim was to investigate whether these pathogenic effects occur in vivo through passive antibody transfer.

Methods

A passive transfer model was established by infusing CSF from patients with anti-IgLON5 disease or controls into the lateral ventricles of adult mice for 14 days via osmotic pumps. Motor and behavioral performance was evaluated using tests of coordination, sociability, anxiety-like behavior, and spatial memory. Mice were sacrificed at days 7, 18, and 30 for analysis of brain-bound human antibodies and quantification of total and synaptic IgLON5 clusters by confocal microscopy. Additional analyses included immunohistochemistry for phosphorylated tau, gliosis, and microglial activation.

Results

Mice receiving anti-IgLON5 CSF exhibited impaired motor coordination in the beam-walking and rotarod performance. Behavioral alterations included reduced social interaction, increased anxiety-like behavior, weight loss, and increased liquid intake. Human IgG deposition was predominantly localized in the hippocampus and periventricular regions, coinciding with a reduction in total and synaptic IgLON5 clusters whereas levels of the postsynaptic marker PSD95 remained unchanged. The reduction in IgLON5 clusters persisted through day 30. Microglial activation was consistently observed in affected regions, but tau pathology and gliosis were absent.

Discussion

This passive transfer model demonstrates that IgLON5 antibodies reduce neuronal membrane IgLON5 clusters, accompanied by microglial activation and motor and behavioral alterations. These results support a pathogenic role of IgLON5 antibodies in anti-IgLON5 disease.

Introduction

Anti-IgLON5 disease is a complex neurologic disorder at the intersection of autoimmunity and neurodegeneration, defined by the presence of antibodies targeting IgLON5, a neuronal cell adhesion molecule, whose function remains largely unknown.^1,2 Patients typically present with a heterogeneous constellation of symptoms, including a distinctive sleep disorder with non-REM and REM parasomnias and sleep apnea, gait instability, abnormal movements, and bulbar dysfunction.³ Unlike other autoimmune encephalitides with antibodies against surface proteins, anti-IgLON5 disease follows a predominantly chronic course and shows limited response to immunotherapy,⁴ highlighting the need to elucidate the distinct mechanisms of IgLON5 antibodies (IgLON5-abs). Moreover, neuropathologic studies have shown hyperphosphorylated tau accumulation in neurons of the brainstem tegmentum and hypothalamus.^1,5

Although the pathogenesis of anti-IgLON5 disease remains unclear, several lines of evidence support an autoimmune basis: (1) a strong genetic association with the HLA-DRB1*10:01-DQB1*05:01 haplotype^6-8; (2) postmortem studies showing a time-dependent progression of hypothalamus/brainstem tau pathology, indicating a secondary process that may contribute to, but not initiate, the disease^5,9; and (3) in vitro studies demonstrating that prolonged exposure of neurons to IgLON5-abs irreversibly reduces membrane IgLON5 clusters and disrupts the cytoskeleton.^10,11

Despite consistent in vitro evidence supporting the pathogenic effects of IgLON5-abs,^10-12 existing animal models have not thoroughly investigated whether the in vitro alterations in IgLON5 clustering occur in vivo or correlate with clinical manifestations of the disease.^13-15

In this study, we used a passive transfer model using patients' CSF antibodies to assess their effects on neuronal membrane IgLON5 clusters and their association with motor and behavioral changes.

Methods

CSF Samples From Patients and Controls

CSF from 6 patients with IgLON5-abs was pooled and infused to the cerebroventricular system of mice via osmotic pumps for 14 days (Figure 1A). The clinical features of these patients are listed in eTable 1. As controls, we used pooled CSF from 6 patients with schizophrenia spectrum disorders without signs of inflammation and negative for neuronal autoantibodies. The pooled CSF samples from patients and controls were dialyzed (Slide-A-Lyzer 20K, Thermo-Fisher, Rockford, IL) against sterile phosphate-buffered saline (PBS) overnight at 4°C. Neuronal reactivity of the pooled CSFs was assessed using (1) rat brain immunohistochemistry; (2) a cell-based assay with HEK293 cells expressing the full-length IgLON5 human protein, which indicated an IgLON5-abs titer of 1:320; and (3) live immunostaining of cultured rat hippocampal neurons, as previously described (eFigure 1). Incubation of rat hippocampal neuronal cultures with pooled CSF (1:20 dilution), as well as subsequent quantification of IgLON5 clusters, as previously described, was used to determine the effects of the pooled samples on IgLON5^10,11 (eFigure 1).

Figure 1. Behavioral Alterations in Mice After Cerebroventricular Infusion of Anti-IgLON5 Antibodies.

(A) Schematic time line of the behavioral test battery performed in animals infused with IgLON5-abs or control CSF. (B) Body weight of IgLON5 mice was significantly lower compared with controls at both days 18 and 30 but not at baseline. (C) Beam-walking performance at day 17 indicates increased crossing time. (D) Impaired motor coordination and balance in IgLON5 mice, shown by reduced time on the accelerating rotarod and lower maximal speed endured before falling at day 25 after infusion. (E) The novel object location (NOL) test at day 7 shows reduced exploratory behavior in IgLON5 mice. (F) Reduced sociability at day 18, reflected by a lower social interaction index in the three-chamber test (3ChT). (G) Depressive-like behavior in IgLON5-treated mice evidenced by reduced sucrose preference (anhedonia) in the sucrose consumption test. Locomotor activity (LA) measured during 24 hours did not show statistical differences (in eFigure 2). Data are presented in box plots and whiskers, indicating Tukey analysis. Statistical significance was assessed using 2-way ANOVA with Tukey correction for multiple comparisons (*p = 0.02; **p = 0.003, ***p = 0.0001). Number of animals per group and time point: 7 days (4 IgLON5 mice/3 control mice); 18 days (19 IgLON5 mice/18 control mice); 30 days (19 IgLON5 mice/14 control mice).

Animals

Male C57BL6/J mice (Charles River), 8–10 weeks old, were housed in cages of 5 until 1 week before surgery, when they were housed individually. The room was maintained at a controlled temperature (21 ± 1°C) and humidity (55 ± 10%), with a 12-hour light/dark cycle for illumination; food and water were available ad libitum. All experiments were performed during the light phase, and animals were habituated to the experimental room before starting the tests.

Surgery and Placement of Ventricular Catheters and Osmotic Pumps

Cerebroventricular infusion of CSF was performed as previously described.^16-18 In brief, 2 osmotic pumps per animal were loaded with 100 µL of pooled patient CSF (mice with IgLON5) or control CSF (control mice) prepared 24 hours before surgery. The next day, mice were deeply anesthetized with inhaled isoflurane and received subcutaneous buprenorphine for analgesia. Animals were then placed in a stereotaxic frame and underwent bilateral catheter (PlasticsOne, model 3280PD-2.0/SP) insertion into the ventricles (0.2 mm anterior and 1.00 mm lateral from bregma, depth 2.2 mm), secured with dental cement. Each arm of the catheter was then connected to one osmotic pump, which was subcutaneously implanted in the animal's back.

Behavioral Testing

A panel of standardized behavioral tests was conducted by researchers blinded to the experimental conditions, according to the schedule shown in Figure 1A. These included the novel object location (NOL) test to assess spatial memory, the sucrose preference test to evaluate depressive-like behavior, and the 3-chamber test to assess social preference. Motor coordination and balance were evaluated using the rotarod and beam balance tests alongside assessment of the footstep patterns (gait analysis). Spontaneous locomotor activity was automatically monitored in activity boxes equipped with 2 parallel rows of photocell detectors (9 × 20 × 11 cm; Imetronic, Pessac, France). The number of animals per group and time point was as follows: 7 days (4 IgLON5 mice/3 control mice); 18 days (19 IgLON5 mice/18 control mice); 30 days (19 IgLON5 mice/14 control mice). A detailed description of each test is provided in eMethods.

Brain Tissue Processing

To determine the effects of patients' antibodies in mouse brains, animals were deeply anesthetized with isoflurane and perfused with PBS at the indicated time points (days 7, 18, and 30). Two distinct tissue processing protocols were followed: (1) To analyze human IgG deposits, IgLON5 clusters, and neuropathologic abnormalities, brains were harvested and transferred to ice-cold PBS. Half of the brain was fixed by immersion in 4% paraformaldehyde (PFA) for 1 hour at 4°C, cryoprotected with 40% sucrose for 48 hours at 4°C, embedded with freezing media, and snap-frozen with prechilled isopentane. The contralateral half-brain was fresh-frozen and the hippocampus dissected for extraction of brain-bound human IgG. (2) For neuropathologic studies, a subset of 10 brains (5 IgLON5 and 5 control mice) were fixed in 4% PFA for 24 hours, sliced in 2-mm-thick coronal sections, and dehydrated using a sequential battery of alcohol concentrations (70% for 3 hours twice, 96% overnight, and 100% for 3 hours) followed by 3 hours in xylene. Tissue was embedded in paraffin and stored at room temperature (RT).

Quantification of Human IgG in Mice Brain

To assess the presence of brain-bound human IgG, tissue sections were incubated with biotinylated goat anti-human IgG (H+L) (1:2000, #BA-3000, Vector Laboratories, Inc. Newark, CA) overnight at 4°C, and the reactivity was developed using a conventional avidin-biotin-immunoperoxidase technique. Quantification of human IgG deposits in different regions of the hippocampus was performed as previously described.¹⁸ In brief, 5-µm-thick sagittal sections were blocked with 1% bovine serum albumin for 60 minutes at RT and incubated overnight at 4°C with Alexa Fluor 488 goat anti-human IgG (#A11013, Thermo-Fisher Scientific, MA, diluted 1:500). Slides were then mounted with ProlongTM Gold antifade reagent with DAPI (P36935, Thermo-Fisher Scientific). Results were scanned under a LSM710 Zeiss confocal microscope (Carl Zeiss, Jena, Germany) with EC-Plan NEOFLUAR CS ×100/1.3 NA oil objective. Standardized z-stacks, including 25 optical images, were acquired from 5 different, equally spaced hippocampal areas of CA1 and CA3 and dentate gyrus (DG) using sequential scanning, 1,024 × 1,024 lateral resolution, and Nyquist-optimized z-sampling frequency. Images were deconvolved with 20 iterations using theoretical point spread functions and maximum likelihood estimation algorithms of Huygens Essential software 25.10 (Scientific Volume Imaging, Hilversum, The Netherlands).

Analysis of Neuronal IgLON5 Clusters

To determine the effects of IgLON5-abs on total and synaptic IgLON5 clusters, we followed protocols previously described.^10,18 In brief, tissue sections from IgLON5 mice or control mice were incubated with IgLON5-abs (serum 1:200 diluted) for 2 hours at RT, permeabilized with Triton X-100, and sequentially incubated with rabbit polyclonal antibody against PSD95 (diluted 1:500, rabbit, Clone 18258 Abcam) and the corresponding secondary antibodies, both diluted 1:1,000, Alexa Fluor 594 goat anti-human IgG and Alexa Fluor 488 goat anti-rabbit IgG (A-11014, A-11008, Molecular Probes). Slides were mounted as above and results scanned with a confocal microscope (Zeiss LSM710).

A spot detection algorithm from Imaris suite 10.1 (Oxford Instruments, Abingdon, United Kingdom) was used for IgG deposit quantification and cluster density analysis based on automatic segmentation of the images to spots. The density of clusters was expressed as spots/µm³. Three-dimensional co-localization of clusters (e.g., IgLON5 and PSD95) was obtained using a spot co-localization algorithm implemented in Imaris suite 7.6.4. Synaptic localization was defined as co-localization of IgLON5 with postsynaptic PSD95. Synaptic cluster density was expressed as co-localized spots/µm³.

Confirmation of Brain-Bound IgLON5-Abs by Acid Extraction and Immunoprecipitation

To ensure that IgLON5-abs are specifically bound to the tissue, we followed 2 methods: (1) Acid extraction of human antibodies from fresh-frozen hippocampal tissue was performed, as previously reported.¹⁷ In brief, hippocampi from 3 IgLON5-abs–infused mice and 3 control CSF-infused mice were dissected, pooled separately, weighted, and homogenized in 0.5 mL of PBS containing protease and phosphatase inhibitors (Sigma-Aldrich). After extensive washing with PBS, the last wash was kept to confirm the absence of unbound antibodies (pre-extraction fraction). Brain-bound antibodies were then eluted with acid (0.1M Na-Citrate buffer, pH 2.7) and neutralized, as reported.¹⁷ Both the pre-extraction fraction and the acid-extracted fraction were tested undiluted for the presence of antibodies by immunofluorescence on HEK293 cells expressing human IgLON5. (2) Hippocampal lysates were prepared from 100 mg of fresh-frozen tissue obtained from IgLON5 and control mice. After centrifugation, the supernatant was incubated with protein A/G beads to capture IgLON5-abs–protein complexes. The beads were washed with PBS and resuspended and boiled in 25 µL of roti-load buffer (K929.1, Carl Roth, Karlsruhe, Germany). The supernatant was separated by electrophoresis, transferred to a nitrocellulose membrane, and incubated with commercial IgLON5 antibodies (#ab122763; Abcam, Cambridge, United Kingdom) and the reactivity visualized with chemiluminescence.

Neuropathologic Studies

Neuropathologic studies were performed using immunohistochemistry on coronal 4-µm paraffin sections and incubation overnight at 4°C with the following antibodies: biotinylated phospho-tau (AT8, Ser202, Thr205 mouse monoclonal antibody #MN1020B, Invitrogen, Waltham, MA) to detect phosphorylated tau (dilution: 1/1,000), glial fibrillary acidic protein (GFAP) (1/1,000, rabbit polyclonal, #G9269, Sigma Aldrich, St. Louis, MI) to assess gliosis (1/1,000), and Iba-1 (1/1,000, #019-19741, made in rabbit, Fujifilm, Waco, Richmond, VI) to evaluate microglia (1/1,000). Anti-rabbit IgG (H+L) secondary antibody (# 31460, Thermo-Fisher Scientific) was applied at a dilution of 1/1,000 for GFAP and Iba-1 and incubated for 1 hour at RT, and the reactivity was demonstrated by the avidin-biotin peroxidase method for all primary antibodies. Microglial density was quantified using the ImageJ Cell Counter plugin. Five images per animal from either perivascular or hippocampal regions were randomly selected at identical magnification and region-of-interest dimensions. Only clearly identifiable microglial somata were counted. Cell counts were normalized to the imaged area and expressed as cells per mm². Analyses were performed on tissue from 3 IgLON5 mice and 3 control animals. Results were photographed under a brightfield microscope (Zeiss Axio imager.M2, Carl Zeiss, Jena).

Statistical Analyses

Statistical analyses were performed using unpaired or 2-way ANOVA, followed by appropriate post hoc tests. When comparing all possible group means across treatments and time points, the Tukey multiple comparisons test was applied. When only specific planned comparisons were performed (e.g., control mice vs IgLON5 mice at each time point), Šidák correction was used to adjust for multiple testing. Data are presented as mean ± SEM, visualized using boxplots and whiskers for group comparisons, and statistical significance was set at p < 0.05. Statistical tests and graphs were performed using Prism (v7; GraphPad Software, San Diego, CA).

Standard Protocol Approvals, Registrations, and Patient Consents

Written informed consent for the storage and use of these samples was obtained for all patients. The study was approved by the ethics committee of Hospital Clínic de Barcelona, Spain. The procedures in animals have been approved by the ethics committee of Universitat de Barcelona (reference number 616/24), and the directive 2010/63EU has been followed.

Data Availability

All data supporting the conclusions of this article are included within the article and additional files.

Results

Cerebroventricular Infusion of Patients' CSF Alters Gait and Behavior in Mice

IgLON5 mice exhibited a significant reduction in body weight measured at day 18, coinciding with the peak of IgG deposition, and this weight loss persisted through day 30 (Figure 1B). At day 17, they showed impaired fine motor coordination and balance in the beam balance test, spending more time crossing the beam compared with both their baseline performance and that of control mice (Figure 1C). During this task, a fine body tremor was observed in 6 of 14 animals, accompanied by an atypical bounding-like gait (Video 1), although the number of steps did not differ significantly from controls (eFigure 1A). Motor and balance impairments were further confirmed on day 25 in the rotarod test, where IgLON5 mice remained on the rod for a significantly shorter duration and at a lower speed before falling (Figure 1D). The presence of compensatory behaviors during the test (e.g., wall touching) further supported the presence of fine motor coordination deficits (Video 2). By day 24, IgLON5 mice displayed subtle gait abnormalities, characterized by uneven, zigzag locomotion suggestive of balance and coordination deficits (eFigure 1B). Nonetheless, differences in stride length or width did not reach statistical significance in the evaluable traces (eFigure 2C-D-C).

Video 1

Beam balance test. Representative videos of a control mouse performing the beam balance test and IgLON5 mouse illustrating the observed atypical bounding-like gait observed in 6 of 14 animals..Download Supplementary Video 1 ^{(3.2MB, mp4)} via http://dx.doi.org/10.1212/200584_Video_1

Video 1

Rotarod test. Representative videos of a control mouse performing the rotarod test showing a normal behavior and IgLON5 mouse trying to gain balance by touching the walls of the rod.Download Supplementary Video 2 ^{(5.7MB, wmv)} via http://dx.doi.org/10.1212/200584_Video_2

IgLON5 mice did not exhibit memory deficits in the NOL test (eFigure 2E). However, they exhibited a marked decrease in overall exploratory behavior, as reflected by a significantly shorter total exploration time compared with control mice at 7 days after surgery (Figure 1F). In the three-chamber test, IgLON5 mice also demonstrated lower social engagement (Figure 1F). Moreover, they displayed depressive-like behavior, as indicated by a lack of preference for sucrose-containing water and an increased total water intake relative to controls (Figure 1G). No other significant behavioral differences were observed between IgLON5 and control mice (eFigure 2 F).

In Vivo Effects of IgLON5 Antibodies Replicate In Vitro Findings

IgLON5 mice, but not controls, showed hippocampal deposition of human IgG as early as day 7, which peaked at day 18 and declined by day 30 (eFigure 3). Region-specific analysis revealed that, at day 18, human IgG accumulated preferentially in the CA1 and DG of the hippocampus, whereas only background staining was detected in control mice (Figure 2, A–C). By day 30, a partial clearance of antibodies was observed, with IgG deposits remaining significant only within the CA1 region (Figure 2D). Apart from the hippocampus, weak IgG deposition was inconsistently found in other brain regions of IgLON5 mice, such as the hypothalamus, but was absent in the brainstem and in all brain regions of control mice (data not shown). To confirm that the observed IgG deposits corresponded to brain-bound IgLON5-specific antibodies, IgG was acid-extracted from hippocampal tissue and tested in HEK cells expressing IgLON5. Pre-extraction fractions showed no reactivity (eFigure 4, A–C), whereas the extracted IgG specifically recognized IgLON5 in a cell-based assay (eFigure 4, D–F), confirming the selective binding of patient-derived antibodies to IgLON5 within brain parenchyma. Moreover, immunoprecipitation of hippocampal IgG pulled down the IgLON5 protein, further validating the specificity of brain-bound antibodies for IgLON5 (eFigure 4G). Patient CSF containing IgLON5-abs induced an in vivo reduction in both total (Figure 3, A and C) and synaptic (Figure 3, B and D) IgLON5 cluster density within the hippocampus at day 18, coinciding with the peak concentration of IgLON5-bound antibodies. By contrast, the postsynaptic marker PSD95 remained unchanged (Figure 3, B and C). The most pronounced loss of IgLON5 clusters was observed in the CA1 and DG (Figure 3, B–D). By day 30, IgLON5 cluster levels had recovered in the DG (Figure 4, A and B), while remaining reduced in the CA1 and CA3 regions.

Figure 2. Distribution of Human IgG Deposits in the Hippocampus of Mice Infused With Anti-IgLON5 Antibodies.

(A) Representative immunofluorescence image of the hippocampus from an IgLON5-infused mouse at 18 days after surgery, showing human IgG deposits (green) using anti-human IgG antibody. (B) Control mouse infused with pooled control CSF, showing minimal background staining. (C) Quantification of the intensity of fluorescence corresponding to human IgG immunoreactivity in different hippocampal subregions (CA1, CA3) and dentate gyrus (DG) at 18 days after infusion. (D) Quantification of intensity of human IgG immunolabeling in the same brain areas at 30 days after infusion, showing regional differences and overall reduction in signal over time. Number of animals per group and time point: 18 days (19 IgLON5 mice/18 control mice); 30 days (19 IgLON5 mice/14 control mice). Scale bar = [200 µm] (****p < 0.0001,***p = 0.0005).

Figure 3. Decrease of Total and Synaptic IgLON5 Clusters In Vivo in Hippocampal Subregions at Day 18 After Surgery.

(A) Representative images showing total IgLON5 clusters in the hippocampus. Upper panel: total IgLON5 clusters in control mice in CA1, CA3, and dentate gyrus (DG); lower panel: total IgLON5 clusters in IgLON5 mice in the same hippocampal areas. (B) Quantification of synaptic and total IgLON5 clusters using the spot algorithm in Imaris software. Panels (a–f): total IgLON5 clusters (green spots) in (a, b) CA1, (c, d) CA3, and (e, f) DG, comparing control (a, c, e) and IgLON5 (b, d, f) mice. Panels (g–l): PSD95 puncta (red) representing excitatory postsynaptic sites in (g, h) CA1, (i, j) CA3, and (k, l) DG; control mice shown in g, i, and k and IgLON5 mice in h, j, and l. Panels (m–r): synaptic IgLON5 clusters, defined as co-localized IgLON5 and PSD95 signals, in (m, n) CA1, (o, p) CA3, and (q, r) DG. Control mice in m, o, and q; IgLON5 mice in n, p, and r. Scale bars = [5 µm]. (C) Quantification of total IgLON5 and PSD95 clusters using the spot algorithm of Imaris software. (D) Quantification of synaptic IgLON5 clusters with Imaris software (N = 7 animals per group). Data shown as mean ± SEM. Statistical comparisons performed using 2-way ANOVA; ****p < 0.0001.

Figure 4. Reduction of Total and Synaptic IgLON5 Clusters in Hippocampal Subregions at 30 Days After Infusion.

(A) Quantification of total IgLON5 clusters in hippocampal subregions (CA1, CA3, and dentate gyrus) at 30 days after infusion. IgLON5 mice show a significant reduction in cluster density in CA1 and CA3 compared with controls, while dentate gyrus levels are restored. (B) Quantification of synaptic IgLON5 clusters, defined by co-localization with PSD95, shows a similar reduction in CA1 and CA3 of IgLON5 mice. PSD95 density remained unchanged, indicating selective loss of IgLON5 at the synapse. Data are expressed as mean ± SEM. Statistical analysis performed using 2-way ANOVA with the Tukey multiple comparison test; p < 0.05 considered significant; ****p < 0.0001, **p = 0.0013 (N = 7 animals per group).

Neuropathologic Findings

Mild microglial activation was detected in the hippocampus, where Iba1-positive microglia were distributed throughout the CA1–CA3 regions and the dentate gyrus. In IgLON5 mice, microglia displayed a moderately increased density (mean = 120 cells/mm² ±33 in IgLON5 mice vs mean = 94 cells/mm² ±29 in control mice; p = 0.01) (eFigure 5) and occasional signs of activation, characterized by enlarged somata and thicker, retracted processes, particularly within the pyramidal and subgranular layers. By contrast, control mice exhibited a uniform distribution of Iba1-positive cells, consistent with a resting microglial phenotype (Figure 5). No evidence of tau pathology or gliosis was observed in the brains of IgLON5 mice at day 30.

Figure 5. Microglial Activation in IgLON5 Mice.

(A) Representative image showing microglial activation in the hippocampus of an IgLON5 mouse, labeled with Iba1 antibody. (B) Higher magnification of the boxed area in (A), illustrating amoeboid, activated microglia (arrows). (C) Iba1 staining in a control mouse demonstrates a uniform distribution of microglia with ramified morphology, characteristic of a resting phenotype. (D) Magnified view of the boxed area in (C) highlighting ramified, resting microglia with small cytoplasm (arrows). (N = 5 mice per group). Scale bars = [200 µm and 20 µm]. Quantitative analysis of Iba-1–positive microglia revealed an increased number of detected microglial particles in IgLON5 mice compared with controls (quantification in eFigure 5).

Discussion

We demonstrate that IgLON5-abs are pathogenic in vivo, linking antibody-mediated neuronal IgLON5 loss to motor and behavioral deficits. IgLON5 mice showed consistent motor deficits across multiple behavioral paradigms, peaking with maximal human IgG brain deposition and persisting beyond the infusion period, indicating a lasting functional effect of antibody exposure. These alterations mirror the predominant motor features of the CSF donor patients, characterized by gait ataxia, imbalance, and limb dysmetria.³ Moreover, IgLON5 mice showed progressive weight loss. Although the underlying mechanism remains uncertain and may reflect the locomotor deficits described above, it could also indicate early bulbar or hypothalamic dysfunction.^3,19

Beyond motor deficits, IgLON5 mice exhibited impaired social interaction and exploratory behavior, indicating reduced sociability and increased anxiety-like traits. They also exhibited higher liquid intake, suggesting alterations in homeostatic or reward-related mechanisms, potentially linked to hypothalamic changes observed in the autopsy of patients with anti-IgLON5 disease.¹⁹ Although these features have not been extensively reported in patients, they may represent underrecognized aspects of the disease, warranting closer attention in clinical assessments. A limitation inherent to intraventricular passive-transfer approaches is the heterogeneous diffusion of antibodies across brain regions. Accordingly, apart from the hippocampus and perivascular areas, only weak and inconsistent immunostaining was observed in other regions, such as the hypothalamus. Therefore, we cannot exclude the possibility that antibodies reached these regions at levels below the detection threshold and contributed to the observed behavioral alterations.

Notably, IgLON5 mice did not exhibit memory deficits, in contrast to other reported animal models of autoimmune encephalitis, such as those involving the passive transfer of NMDAR, LGI1, or CASPR2 antibodies.^16-18 This observation is consistent with the clinical differences between these encephalitides and anti-IgLON5 disease, in which memory impairment is not a typical feature.

Our model reproduces the antibody effects previously observed in cultured neurons, where IgLON5 clusters were cross-linked by the antibodies and internalized,^10,11 a point not fully addressed before. This antibody-mediated reduction of neuronal IgLON5 clusters was maximal at the peak of IgG deposition and persisted until day 30, well beyond the end of infusion, indicating long-lasting effects. These findings are different from those reported for the passive transfer of NMDAR-abs, in which total and synaptic NMDAR clusters were already restored at this time point.¹⁸ The reduction and internalization of IgLON5 clusters observed in our study are likely mediated by the IgG1 fraction of IgLON5-abs, as demonstrated in our in vitro experiments. However, the specific role of IgLON5 IgG4-abs was not addressed in this study, representing a limitation and an important avenue for future investigations, particularly in light of autopsy findings showing co-localization of IgG4 deposition with IgLON5 by immunohistochemistry.

The spatial overlap of microglial activation with sites of human IgG deposition and IgLON5 cluster loss suggests that microglia respond to local antigen-antibody interactions. This activation may initially be protective by clearing immune complexes and synaptic debris, but persistent activation could drive chronic neuroinflammation, secondary synaptic loss, and neurodegeneration. Therefore, the mild microgliosis observed in our model, also seen in patient autopsies,⁵ may represent an early stage of anti-IgLON5–associated neurodegeneration, requiring prolonged antibody exposure or additional cofactors to develop tau pathology.

The absence of tau pathology in IgLON5 mice may reflect several factors: (1) The difficulty of modeling the slowly progressive nature of anti-IgLON5 disease, which in humans involves chronic antibody exposure over months to years. Indeed, neuropathologic studies have shown greater tau accumulation in long-duration cases, while shorter courses often have minimal or no tau.⁹ The 14-day intraventricular infusion in our model may be insufficient to induce the cumulative effects needed for tau aggregation. (2) Species-specific differences, as murine tau differs in isoform composition and may be less susceptible to misfolding or post-translational modifications²⁰ triggered by human IgLON5 antibodies.

Another limitation, shared by all passive antibody transfer models, is the short study duration imposed by brain surgery, preventing assessment of whether IgLON5 cluster loss is irreversible, as seen in vitro, or slowly recoverable. This could be addressed in future studies using a mouse model generated by active immunization with IgLON5, as successfully performed for anti-NMDA receptor encephalitis.²¹

However, compared with previous models,^13-15 this study advances the field by using patient CSF rather than purified IgG from a single serum sample and by correlating antibody presence with molecular changes in IgLON5 clusters and behavioral outcomes. Earlier models using hippocampal or ventricular injections reported memory deficits, anxiety-like behavior, microglial activation, and neuronal loss but did not establish a direct link between these findings and antibody co-localization and specific reduction of surface and synaptic IgLON5 clusters. Other antibody-transfer models targeting the substantia nigra reported dopaminergic neuronal loss and increased tau phosphorylation; however, dysfunction of the dopaminergic system alone does not fully account for the clinical manifestations observed in patients, and antibody specificity for IgLON5 was not demonstrated.¹³ In a previous study aimed to analize the sleep disorder,¹⁵ sleep disturbances were systematically assessed, yet no clear abnormalities were observed. Moreover, under blinded assessment, phospho-tau immunoreactivity was occasionally detected in control mice, particularly in males, suggesting that potential sex-related differences may influence these findings and should be carefully addressed in future studies.

Our findings closely mirror the motor dysfunction observed in patients. In our animal model, we infused intraventricularly patient CSF, providing in vivo evidence that human IgLON5 antibodies induce behavioral changes, reduce IgLON5 clusters, and trigger microglial activation. Demonstrating in vivo pathogenicity underscores a direct antibody-driven component in disease progression and supports targeting humoral immunity as a therapeutic strategy.

Glossary

DG

dentate gyrus

NOL

novel object location

PBS

phosphate-buffered saline

PFA

paraformaldehyde

RT

room temperature

Author Contributions

A.B. Serafim: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data. E. Aguilar: major role in the acquisition of data; analysis or interpretation of data. M. Alba: major role in the acquisition of data. E. Maudes: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data. L. Marmolejo: major role in the acquisition of data. J. Planagumà: major role in the acquisition of data. M. Guasp: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. C. Papi: major role in the acquisition of data. L. Molina-Porcel: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data. C. Gaig: major role in the acquisition of data. J.O. Dalmau: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. F. Graus: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. L. Sabater: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data.

Study Funding

This study has been funded by Instituto de Salud Carlos III through the project FIS21/00165 LS, CG (co-funded by the European Regional Development Fund “Investing in your future”). The authors thank the European Joint Programme on Rare Diseases (EJPRD) for funding a networking event (contract number 463001015) on IgLON5 disease LS, CG, FG, JD from the European Union's Horizon 2020 research and innovation programme under the EJP RD COFUND-EJP N°825575. IDIBAPS belongs to CERCA (Centres de recerca de Calalunya). A.B. Serafim thanks Fundação para a Ciência e Tecnologia (FCT) of Portugal for the doctoral fellowship grant number 2022.13121.BD. EM is a recipient of the ECTRIMS Research Fellowship. L. Marmolejo is supported by ISCIII-PFIS contract FI24/00021. This work was supported by FCT - Fundação para a Ciência e Tecnologia, I.P., by project reference and DOI identifier doi.org/10.54499/2022.13121.BD.

Disclosure

J.O. Dalmau receives royalties from licensing fees to Athena diagnostics for a patent for the use of Ma2 and autoantibody test, licensing fees to Euroimmun for patents for the use of NMDAR and GABAbR as autoantibody tests, and licensing fees for the use of DPPX, GABAaR, and IgLON5 antibodies as diagnostic tests. F. Graus receives royalties from licensing fees to Euroimmun for the use of IgLON5 as a diagnostic test. The remaining authors declare that they have no competing interests for this publication. Go to Neurology.org/NN for full disclosures.