Glutamate is the most abundant, potent, and important excitatory neurotransmitter in the mammalian central nervous system (CNS). Glutamate is released primarily by nerve cells in the brain, activates glutamate receptors of multiple types (Reiner and Levitz, 2018; Hansen et al., 2021), and induces multiple effects critical for intact neural activities and essential brain functions, including cognition, memory, and learning. Glutamate also plays major roles in the development of the CNS, including synapse induction and elimination and cell migration, differentiation, and death (Hansen et al., 2021; Gill and Pulido, 2001; Nedergaard et al., 2002; Hinoi et al., 2004; Niswender and Conn, 2010; Colombo and Francolini, 2019). In addition, glutamate plays a signaling role in peripheral organs and tissues, among them the heart, kidney, intestine, lungs, liver, ovary, testis, bone, pancreas and the adrenal, pituitary and pineal glands, and several other organs and tissues (Gill and Pulido, 2001; Nedergaard et al., 2002; Hinoi et al., 2004).

Glutamate induces its numerous effects via binding and activating a very large number of glutamate receptors (GluRs) (Reiner and Levitz, 2018).

The two main families of GluRs are the ionotropic glutamate receptors (iGluRs) (Hansen et al., 2021) and the metabotropic glutamate receptors (mGluRs) (Niswender and Conn, 2010), each consisting of several types, subtypes, and subunits.

Colombo et al. report that glutamate and acetylcholine are the most common neurotransmitters used in the central excitatory synapses and at the neuromuscular junction (NMJ) (Colombo and Francolini, 2019). Glutamate receptors localize postsynaptically at neuromuscular junctions in mice, and glutamatergic transmission occurs at these junctions (Berger et al., 1995; Grozdanovic and Gossrau, 1998; Mays et al., 2009; Personius et al., 2022; Personius et al., 2016). Ettorre et al. (2012) found that glutamatergic neurons induce the expression of functional glutamatergic synapses in primary myotubes and review the past and recent experimental evidences in support of the role of glutamate as a mediating neurotransmitter at the synapse between the motor nerve ending and the skeletal muscle fiber at the vertebrate neuromuscular junction.

Personius et al. studied the role and the neuromuscular glutamate receptors, mainly of the NMDA type, and found that NMDA receptors modulate developmental synapse elimination in mice (Personius et al., 2016). Based on all their findings in that study, the researchers concluded that neuromuscular NMDA receptors play important roles in neuromuscular activity and in elimination of excess synaptic input during development (Personius et al., 2016). Personius et al. found that blockage of neuromuscular iGluRs, either NMDA receptors or AMPA receptors, impairs reinnervation following nerve crush in adult mice (Personius et al., 2022). The researchers also showed the presence of NMDA receptors at the endplate after a nerve crush, documented muscle responses to NMDA after the crush, and found that blocking iGluR of the NMDA or AMPA type, during regeneration, slows polyneuronal innervation and behavioral recovery after a nerve crush. Based on these findings, and additional ones discovered in that study, the researchers concluded that iGluRs play a significant role in promoting recovery after a sciatic nerve crush (Personius et al., 2022).

Despite the multiple studies and findings with regards to glutamate in the NMJ, to the best of our knowledge, no study has shown thus far neither the expression of functional iGluRs on the cell surface of human skeletal muscle cells nor that glutamate by itself can induce direct activating effects on human muscle cells, via their iGluRs.

Glutamate is the primary amino acid absorbed by resting and active muscles. It participates in various metabolic pathways in the skeletal muscle, both at rest and during contraction (Personius et al., 2016; Ettorre et al., 2012; Rutten et al., 2005).

Glutamate is known to be delivered to the skeletal muscle via different pathways: it is actively absorbed from the circulation, released by intracellular protein degradation, and synthesized by transamination of the branched-chain amino acids (leucine, isoleucine and valine), which are oxidized in the skeletal muscle (Rutten et al., 2005). Glutamate may also be delivered to the muscles by neurons (Rutten et al., 2005). Glutamate seems to play a determining role in the synthesis of glutathione, the most abundant intracellular antioxidant. During exercise, glutamate plays a central role in energy provision, via different metabolic pathways (Rutten et al., 2005).

Importantly, reduced glutamate levels in the skeletal muscle is a consistent finding in several diseases, and several disease-related factors such as hypoxia and oxidative stress were found to be involved in the disturbed glutamate metabolism (Rutten et al., 2005). In a review entitled “Skeletal muscle glutamate metabolism in health and disease: state of the art,” Rutten et al. presented the knowledge updated to 2005 regarding the metabolic function and regulation of glutamate in skeletal muscle under physiological and pathophysiological circumstances (Rutten et al., 2005).

Engelen et al. found that glutamate levels are reduced in patients with chronic obstructive pulmonary disease (COPD) and reported that altered glutamate metabolism is associated with reduced muscle glutathione levels in patients with emphysema (Rutten et al., 2005; Engelen and Schols, 2003; Engelen et al., 2000).

Recently, Caredio et al. reported that prion diseases disrupt glutamate/glutamine metabolism in the skeletal muscle (Caredio et al., 2024). The researchers found that the GLUL gene is upregulated in the skeletal muscle of prion-infected mice and humans with Creutzfeldt–Jakob disease, leading to disruptions in glutamate and glutamine metabolism and a reduction in muscle glutamate levels (Caredio et al., 2024).

The current study focuses only on GluR3, the ionotropic glutamate receptor of the AMPA GluR3 type subunit 3, encoded by the GRIA3 gene. Figure 1A shows a schematic drawing of AMPA GluR3. AMPA receptors that contain the GluR3 subunit can be either homomeric GluR3(o) receptors, heteromeric GluR3(o)/GluR2(o), or GluR3(o)/GluR2(i) receptors (Hansen et al., 2021).

GluR3 is best known for its expression and function in cells of the central nervous system, mainly in neurons (excitatory neurons and interneurons, and to a lesser degree in glia/astrocytes in the CNS. In addition, a number of studies have reported GluR3 expression in other cell types, primarily in T cells of various types and origins (Ganor et al., 2003; Ganor et al., 2009; Ganor and Levite, 2014; Ganor et al., 2007; Levite et al., 2020) (see more on this topic in a separate subsection below) and podocytes (kidney glomerulus (Li et al., 2019).

In the nervous system, GluR3 is involved in the regulation of neural networks. GluR3 is expressed in the spinal cord, brainstem, thalamus and cortex (Engelen et al., 2000; Caredio et al., 2024; Ganor et al., 2003; Ganor et al., 2009). According to Steenland et al. (2008), these GluR3 expression sites suggest that GluR3 regulates functions spanning the neuraxis. Furthermore, Steeland et al. predicted that the localization of GluR3 subunits at respiratory control regions (Sato et al., 1993) and cardiac control regions (Corbett et al., 2003) implies that GluR3 may have an impact on cardiorespiratory function (Steenland et al., 2008). It is also predicted that GluR3 subunit knockout will affect motor tone as these receptors/subunits are located on motor neurons of the ventral spinal cord (Sato et al., 1993). Strikingly, Steenland et al. found that GluR3^−/− knockout mice virtually lack electroencephalographic and signatures of NREM sleep and that three of nine GluR3 (−/−) mice expressed seizure activity during wakefulness and sleep (Steenland et al., 2008).

These findings suggest that deletion of the GluR3 gene by itself may predispose to seizures. Moreover, GluR3 gene knockout also produced state-dependent respiratory modulation, with a selective reduction in the breathing rate during behavioral inactivity (Steenland et al., 2008).

Within the brain, Moga et al. found GluR3 immunoreactivity in all pyramidal neurons and astrocytes and in most interneurons (Moga et al., 2003). Selective GluR3 expression is significantly elevated in the somata of parvalbumin-containing interneurons, which are potent inhibitors of cortical pyramidal neurons, and are vulnerable in brains of epilepsy patients (Moga et al., 2003). Taken together, these findings indicate that GluR3 has diverse neurophysiological impacts and is involved in sleep, wakefulness regulation, breathing, and generation of cortical seizures (Steenland et al., 2008).

We previously found that normal human as well as human T cell lymphoma and autoimmune T cells express functional GluR3 (Ganor et al., 2003; Ganor et al., 2009; Ganor and Levite, 2014; Ganor et al., 2007; Levite et al., 2020).

We also discovered that glutamate activates directly its GluRs expressed in T cells and induces/increases all these T cell functions and features: integrin-mediated adhesion to laminin and fibronectin, chemotactic migration, elevation of CD147/EMMPRIN (a cancer-associated matrix metalloproteinase (MMP) inducer, MMP-9 secretion, engraftment of cells in vivo, CD3 zeta expression, and most importantly T cell killing of cancer cells (Ganor et al., 2003; Ganor et al., 2009; Ganor and Levite, 2014; Levite et al., 2021; Saussez et al., 2014). We also found that glutamate reduces PD-1 expression in T cells (Levite et al., 2021). All these findings indicate that glutamate and GluR3 seem to have very important roles in T cell functions and in the entire function of the immune system.

Apart from being an important physiological glutamate receptor/subunit, GluR3 is an antigen of very detrimental autoimmune GluR3 antibodies (Levite and Goldberg, 2021). Autoimmune GluR3 antibodies play a key pathological role in “Autoimmune Epilepsy” (Levite et al., 2020; Levite and Goldberg, 2021; Levite, 2002). In our own multiple studies on “Autoimmune Epilepsy,” and in studies of other groups, autoimmune GluR3 antibodies, especially GluR3 B antibodies directed against the extracellular junctional and antigenic peptide “B” peptide of GluR3 (shown schematically in Figure 1A), were found to be pathological antibodies that cause multiple detrimental effects in vitro and in vivo. In particular, the GluR3 B antibodies can kill neural cells by three mechanisms: induction of excitotoxicity, reactive-oxygen-species (ROS), and/or complement-fixation. The GluR3 B antibodies were also found to induce or facilitate in animal models all the following effects: brain damage, seizures, and behavioral impairments (for the most recent review, see Levite and Goldberg (2021)), and for some of the original research papers, see Levite et al. (2020), Cohen-Kashi Malina et al. (2006), Levite et al. (1999), Levite and Hermelin (1999), Ganor et al. (2004), Ganor et al. (2014), Ganor et al. (2005a), Ganor et al. (2005b), Goldberg-Stern et al. (2014).

The autoimmune GluR3B peptide antibodies were found so far in the serum of ∼27% of >300 persons with severe, intractable, and enigmatic epilepsy of various types (for the most updated review and original paper that summarize and cite all the relevant studies see Levite and Goldberg (2021); Taiwo et al. (2025). In addition, in a recent study, we found that epileptic Nodding Syndrome (NS) patients have autoimmune GluR3 B antibodies and that the NS patient’s GluR3 B peptide autoimmune antibodies bind both neural cells and immune T cells (that express GluR3), induce ROS in both neural cells and T cells, and kill both of these cell types (Levite et al., 2020).

NS is a form of epilepsy affecting children in Sub-Saharan Africa, marked by head-nodding seizures, progressive cognitive decline, growth stunting, and developmental delays (Abd-Elfarag et al., 2021; Samia et al., 2022). NS is now recognized as an idiopathic epileptic encephalopathy with tauopathy. NS is strongly associated with muscle wasting (loss of muscle mass) and low muscle mass and can involve a temporary loss of neck muscle tone (Abd-Elfarag et al., 2021; Samia et al., 2022).

The characteristic head nodding results from a temporary loss of neck extensor tone. While muscle strength is generally maintained, the disorder can progress to include peripheral muscle wasting and generalized wasting. The head nodding, which can occur multiple times per minute, is thought to be caused by a loss of muscle tone in the neck and potentially other muscle groups (Abd-Elfarag et al., 2021; Samia et al., 2022).

Recent studies show that autoimmunity, the parasite Onchocerca volvulus, possible environmental triggers, and vitamin B6 deficiency contribute to NS. While treatments for symptoms exist, an effective cure is lacking. Current clinical trials test the effectiveness of doxycycline as a treatment for NS in children and adolescents (Idro et al., 2024).

In the present study, we asked three main questions: do human skeletal muscle cells express functional GluR3, are these cells activated by glutamate, and do autoimmune GluR3 B antibodies of either NS patients or other intractable epilepsy patients bind and affect human skeletal muscle cells.

The novel findings of this study provide a positive answer to all of these questions. However, further studies in various directions are of course needed to confirm, expand, and deepen these findings and to understand and examine their implications for muscle cells and motor function of healthy or diseased individuals.

The primary culture of human muscle cells (myoblasts) was derived from the gastrocnemius of a 32-year-old female following an isolation procedure as described (Amsili et al., 2007). The culture was obtained from the Muscle Tissue Culture Collection (MD-NET, service structure S1, 01GM0302, BMBF, Eurobiobank) at the Friedrich-Baur Institute. The biopsy material was obtained from the Ludwig-Maximilians Universitat Munchen (Germany. IRB approval no 45-14-2014). The primary muscle cells were cultured in Promocell Human muscle growth medium (PromoCell, C-23160). The medium was changed every 3 days. The cells were used in experiments as myoblasts. In IncuCyte experiments, however, the medium was not changed during the entire incubation time, as indicated for each experiment, and from days 3 to 4 on culture, their morphology pointed to initiation of differentiation in spite of the fact that a differentiation medium was not used.

The Ministry of Health of South-Sudan had provided us an allowance to conduct a clinical diagnostic study, and withdrawal of small blood samples for diagnostic research, from South-Sudanese Nodding Syndrome patients and healthy subjects, based on an IRB approval provided 25 May 2012. This investigation (both cohorts) was performed at the town of Mundri, Western Equatoria, South-Sudan. All South Sudanese NS patients and control healthy signed an informed consent.

The relevant clinical information on the NS patients included in the present study is depicted in our previous study on the presence and pathogenic effects of GluR3B antibodies in these NS patients (Levite et al., 2020).

All the NS patients in the study group were diagnosed by a pediatric neurologist. For some of them, a deeper diagnosis was performed, and most of these clonic–tonic seizures and additional clinical symptoms were documented. Healthy South Sudanese subjects, at similar age range, ratio between males and females, and geographical locations to that of the NS patients, were recruited for the study. A small volume of blood was withdrawn from all the NS patients and healthy subjects by the Sudanese clinicians, serum was separated in place, and all samples were shipped to Israel for all the subsequent in vitro and in vivo studies described in this paper and in Levite et al. (2020).

The present study on the antibodies of intractable epilepsy patients received an IRB approval No. 0339-09 from the ethic committee of Rabin Medical Center, Israel, which the Schneider Medical Center in Israel is affiliated to. The epilepsy patients signed informed consent forms. All the studied epilepsy patients are/were treated by Prof. Hadassa Stern Goldberg (author), Head of the Epilepsy Unit in Schneider Medical Center. Table 1, showing the epilepsy patient’s clinical information, was prepared by Prof. Hadassa Stern Goldberg. The epilepsy patients are addressed throughout the study in coded names, and the manuscript does not disclose any confidential information about the patients, which can help identify them.

RNA was extracted from muscle cells by the Tri Reagent isolation solution (Cat# TR 118, Molecular Research Center, United States). Brain1 RNA was purchased (Thermofisher Cat #AM7962). Brain2 RNA was a kind gift from Dr Eisenberg. cDNA synthesis was performed by reverse transcription of the RNAs using the M-MLV Reverse Transcriptase kit (Cat#M1701, Promega, WI) according to the manufacturer’s protocol and screened for GluR3 by PCR and sequencing. Two sets of primers were used, at a concentration of 0.4 mM/reaction, as described in Ganor et al. (2003); the first, spanning exons 3 to 6: upstream primer (GluR3 E3), GACGCAGATGTGCAGTTTGTCATC; downstream primer (GluR3 E6), TAGTGGTGCATTCTTGGCTTCAGG, resulting in a 516-bp product; the second, spanning exons 4 to 9: upstream primer (GluR3 E4), CGATACTTGATTGACTGCGA; downstream primer (GluR3 E9), TACTATGGTCCGATTCTCTG, resulting in a 632-bp product As an internal control for the integrity of the samples, a ribosomal RNA S14 region was amplified using the following primers: upstream primer, GTCCATGTCACTGATCTTTCTGGC; downstream primer, GTTTGATGTGTAGGGCGGTGATAC, resulting in a 166-bp product.

Conditions for PCR were as follows: 94 °C for 1 min, 60 °C for 40 s, and 72 °C for 40 s (29 cycles for S14 PCR and 38 cycles for GluR3 PCR). The cDNA sequencing was performed with an automated sequencer at the sequencing unit of the Hebrew University of Jerusalem.

The mouse anti-human/rat/pig GluR3 B peptide monoclonal antibody (mAb) was originally produced owing to the request of the research group of M.L. (first author herein), at the Life Science Core Facilities, Faculty of Biochemistry, Weizmann Institute of Science, Israel. Since then, the GluR3B mAb has been used successfully for several published studies, among them Levite et al. (2020), and commercialized by the Weizmann Institute of Science to Medimab, Canada (Medimab, Cat no. #GLU149.29.61).

The GluR3 B peptide (also called the ‘B’ region of GluR3) is a 24-amino acid peptide, whose amino acid (aa) sequence is NEYERFVPFSDQQISNDSSSSENR), corresponding to aa372-395 of the glutamate/AMPA receptor/subunit GluR3. The GluR3 B peptide used in ELISA in Levite et al. (2020) was synthesized by Thermo Fisher Scientific. The biotinylated GluR3 B peptide Biotin - GSGSNEYERFVPFSDQQISNDSSSSENR-OH, used herein for affinity-purification of human GluR3 B antibodies from the era of NS patients was synthesized by Pepscan, Netherlands.

Sera of the epilepsy patients and healthy individuals were tested for the presence of GluR3B antibodies by ELISA, as done previously in several of our studies (see, for example, Goldberg-Stern et al. (2014)). In the first (coating) step, microtiter wells of a Maxisorp microtiter immunoplate (Nunc, Roskilde, Denmark) were covered with 50 µL per well of 10^-7 M of GluR3B peptide suspended in a coating buffer of 0.1 M NaHCO3 pH 8.2 for detecting the respective GluR3B antibodies. In parallel, other microtiter wells of another control microtiter plate were covered with 50 μL per well of phosphate-buffered saline (PBS) with 1% bovine serum albumin (BSA) only (i.e., without any antigenic peptide), to detect nonspecific control binding to BSA.

The two microtiter plates were incubated overnight at 4 °C. The following day, the microtiter wells of each plate were washed twice with 100 µL of PBS with 1% BSA per well and then once with 100 µL per well of double-distilled water (DDW). The next step was the blocking of the nonspecific binding, performed by adding 50 µL per well of PBS+1%BSA to all the microtiter wells, and subsequent 2-h incubation at room temperature (RT). At the end of the blocking stage, the microtiter wells were washed, first with PBS with 1% BSA and then with DDW, as in the previous washing step. Thereafter, the serum of each patient or control individual underwent serial dilutions in PBS with 1% BSA, to final dilutions of 1:10, 1:100, and 1:1,000. Then, 100 µL of each serum at each of the three dilutions was added to two adjacent duplicate microtiter wells, in each of the two plates: the GluR3 B -coated and the PBS+1%BSA-coated plates. The two microtiter plates were then incubated overnight at 4 °C. The next day, the wells were washed once more, first with PBS+1%BSA and then with DDW, as in the previous washing steps. Thereafter, horseradish peroxidase (HRP)-conjugated goat anti-human IgG (Jackson Immunoresearch, West Grove, PA) was diluted in PBS+1%BSA to a final dilution of 1:1,000, and 50 µL of this dilution was added to each well in the microtiter plates. Then, the plates were incubated for 2 h at RT. Following an additional washing step, as described above, ABTS peroxidase substrate (Kirkegaard and Perry Laboratories, Gaithersburg, MD) was added to each well, and the optical density (OD) at 405 nm was measured with an ELISA reader. Repeated OD measurements of the plates were performed 3–4 times, at approximately 20-min intervals, from the time the substrate was added. The first OD reading was done when the green color in some wells became clear.

To obtain the level of the specific GluR3 B peptide antibodies in each serum, a two-step calculation was performed as follows: 1. Calculation of the average value of every set of adjacent duplicate wells, 2. Subtraction of the average OD of the nonspecific binding to PBS+1%BSA of each serum sample, in each dilution and each OD reading, from the average OD of the specific binding of that specific serum to the GluR3 B peptide, of the very same serum sample, at the same dilution, and the same OD reading.

Thus, the final values of specific GluR3 B peptide antibodies were determined for each serum dilution using the equation: [average OD of duplicate wells in the GluR3 B -coated plate] - [average OD of duplicate wells in the PBS+1%BSA -coated plate]. In most cases, the results of the second OD reading (∼30 min after adding the substrate) were the clearest.

Findings were considered positive if the final value of specific GluR3 B peptide antibodies was equal or higher than the experiment’ cut off. The cutoff was the calculated value of the average OD + 2 X SD of the GluR3 B antibodies detected in all the healthy control individuals, tested in the same experiment. In most experiments, the calculated cutoff was approximately 0.25–0.4 OD.

The protocol consisted of the following steps: 1. Equilibration of 150-µL packed beads streptavidin–agarose (Thermo Scientific Pierce, Cat no. 20353) with PBS on 2-mL open columns. 2. Loading on columns and reloading four times 1 mg of biotinylated peptide resuspended in 2 mL PBS. 3. Washing peptide affinity column with 5 × 1 mL PBS before peptide specific antibody purification. 4. Diluting 0.5 mL serum in 2 mL PBS, spinning 10 min, 13,000 rpm and discharging pellets. 5. Loading on the peptide affinity column and reloading unbound another four times. 6. Washing with 1 mL PBS several times up to reaching OD 280 nm of washing <0.05 OD280 (blank PBS). 7. Adding 250-µL elution buffer (0.1 M acetic acid pH 3.0) and then fitting a stopper to the column, followed by waiting for approximately 2 min at 4 °C and spinning. 8. Collecting the 250 µL on a tube containing 100 µL 1M TrisHCl pH 8.5 (in order to neutralize the solution immediately). Repeating five times. 9. Checking the protein concentration by OD 280 nm, and pooling IgG-rich fractions. 10. Dialyzing vs. 100 mL PBS, changing buffer after 2 h, and final overnight dialysis vs. fresh PBS. 11. Aliquoting and freezing purified antibodies.

Sera of South Sudanese NS patients and of intractable epilepsy patients (see Table 1) and respective healthy control subjects were centrifuged at 1,600 RPM for 10 min 4 °C. Clean supernatants were diluted to final 4 mL with PBS (10 mM NaPO₄, 150 mM NaCl pH7.4). The samples were loaded onto a PBS equilibrated 1 mL Protein G Sepharose FF (GE Healthcare) affinity chromatography column using AKTA AVANT (GE Healthcare). Then, the column was washed with PBS until reaching low OD 280 nm; IgG was eluted with 0.1 M glycine pH 3.0 buffer and 1 mL eluted fractions were collected in tubes containing 0.08 mL 1.5 M TriHCl pH 8.5 in order to get immediate pH neutralization. IgG fractions were pooled according to the OD 280 nm profile and dialyzed 3 h vs. 100 mL PBS and ON 4 °C using fresh PBS using 13-kDa cutoff dialyze tubes. Finally, they were concentrated by ultrafiltration devices (cut-off 30 kDa of Amicon), and purified IgG samples were analyzed for contents by A280 (E.C 1.35) and by SDS-PAGE (Novex 4%–12%, Thermo), using reduced and non-reduced sample buffer. PageRuler (Thermo) used as MW markers. Purified IgG was aliquoted and frozen.

Immunostaining of GluR3 was done in two steps. First, fixed human skeletal muscle cells (growing on slides) were incubated with a mouse GluR3 B mAb (which we previously produced specifically against the isolated GluR3 B peptide, corresponding to GluR3 aa 372 to 395, and showed later that it binds the GluR3 B peptide in its natural configuration in human neural cells and T cells (Ganor et al., 2007; Levite et al., 2020).

Second, the human skeletal muscle cells were immunostained with anti-desmin antibody to confirm they were indeed muscle cells and with DAPI to detect intact cells.

Human primary skeletal muscle cells (growing on slides) were treated along the following steps (and three washes were performed in between key steps): 1. Fixation with 4% paraformaldehyde; 2. Incubation with affinity-purified GluR3 B antibodies of NS patients or healthy subjects or with purified IgGs of intractable epilepsy patients (Table 1) or respective control healthy subjects (either 1 h in RT, or overnight (ON) in 4C); 3. Blocking with PBS 2% BSA (30 min RT); 4. Incubation with a secondary Cy3- Goat and human F (ab)₂ fragment (30 min RT).

Primary human skeletal muscle cells (myoblasts), growing on slides, were treated along the following steps (and three washes were performed in between key steps): 1. Incubation with total human IgG (∼0.2 mg/mL) of either intractable epilepsy patients or a healthy control subject (either 1 h in RT or ON in 4C); 2. Blocking with PBS 2% BSA (30 min RT); 3. For staining dead cells: Incubation with Sytox green (ThermoFisher Scientific’ 1 μM, 30 min, RT); 4. Fixation with 4% paraformaldehyde; 5. Blocking with PBS 2% BSA (30 min RT); 6. Incubation with a secondary Cy3- Goat and human F (ab)2 fragment (Jackson ImmunoResearch code: 109-166-088) (30 min RT); 7. Permeabilization and further blocking (0.2% Triton, 2% BSA in PBS, 30 min, RT); 8. Finally, to confirm they were indeed muscle cells, the human skeletal muscle cells were immunostained with mouse anti-human desmin mAb and then with goat anti-mouse IgG-Alexa 647. At the last stage, the cells were stained with DAPI to detect intact cells.

Human primary skeletal cells (myoblasts), growing in tissue culture plates, were loaded with CoroNa™ Green, AM, cell permeant dye (Molecular Probes and ThermoFisher Scientific Cat. No 36676) according to the manufacturer instructions. The CoroNa Green dye is a Sodium ion (Na+) indicator that exhibits an increase in green fluorescence emission intensity upon binding sodium (Na+ ions). After loading the muscle cells with CoroNa™ Green, the cells were distributed evenly in wells of microtiter plates, and the plates were placed inside the IncuCyte incubator, for few hours of rest, before starting the experiment/s.

Then, we set up the IncuCyte system to photograph the skeletal muscle cells, (both phase contrast and corona green fluorescence) in all the microtiter wells in the plate (according to the early planning) few times. Few hours later, we started the experiment: we quickly and gently removed the microtiter plate from the IncuCyte incubator, placed it in an adjacent sterile hood, and added either glutamate, AMPA, NMDA, or IgG of a given epilepsy patient or healthy subject, at the desired concentration (all prepared freshly from frozen stocks of much higher concentrations just before each experiment) to replicate microtiter wells in the right predetermined position in the plates. Immediately afterward, we returned the microtiter plate very gently to the IncuCyte and photographed the very same wells again (both Phase contrast and CoroNa™ Green Fluorescence). Thereafter, in the long-term experiments, the IncuCyte was designed to measure, analyze, quantify both the confluence and the Corona Green Fluorescence, proportional to the levels of intracellular Sodium ions, and to photograph the cells every 3 h. In the short-term (responses to glutamate within 2 min), experiment whose representative photos are shown in Figure 9, the time that elapsed between the first photography (before addition of any molecule), and the second photograph was exactly 2 min after addition of either glutamate, AMPA, NMDA or human IgGs of a given epilepsy patient or of healthy control subject.

Statistical analysis was performed using the Student’s t-test. The p values are written inside the figures and/or specified in the “Results” in all experiments in which the statistical analysis could be performed.

RNAs of primary human skeletal myoblasts growing in tissue culture were tested for their expression of GluR3, glutamate receptor of the AMPA-type subunit 3, by GluR3-specific RT-PCR, using two sets of GluR3-specific primers, spanning exons 3–9 of the gene (Ganor et al., 2003). As a positive control, human brain RNA was tested in parallel since GluR3 is highly expressed in several areas of the brain (Ganor and Levite, 2014; Levite et al., 2020). As seen in Figure 1B, we found that the human primary skeletal muscle cells express GluR3 RNA.

Next, we found that primary human skeletal muscle cells (myoblasts) express GluR3 and its extracellular junctional GluR3 B peptide (shown schematically in Figure 1A) on their cell surface. This is demonstrated by the specific binding and immunostaining of the skeletal muscle cells by a mouse GluR3 B monoclonal antibody (mAb) and subsequent confocal microscopy images (Figures 2A–D).

To find out if GluR3 is an active functional receptor in human skeletal muscle cells, we had to target it selectively, without affecting other GluRs. However, to the best of our knowledge, there is no selective GluR3 receptor agonist or antagonist, and therefore none could serve for targeting only GluR3. There are only glutamate receptor agonists and antagonists that bind to a broad family of GluRs, including among them AMPA and NMDA agonists and antagonists.

In the absence of the selective GluR3 receptor agonist, we decided to use the mouse GluR3 B mAb and test if it affects the skeletal muscle cells. Our hypothesis was that the GluR3 B mAb may activate GluR3 since it was previously found that the GluR3 B peptide defines a novel GluR subunit-specific agonist binding and activation site within GluR3, remote and independent of the “classical” glutamate binding site, and that GluR3 B antibodies can activate GluR3 by binding this “B” region (Cohen-Kashi Malina et al., 2006; Levite et al., 1999; Carlson et al., 1997; McDonald et al., 1999; Twyman et al., 1995).

We found that a single addition of GluR3 B mAb increased, in a dose-dependent manner, the number of human skeletal muscle cells growing in the tissue culture for 1 week (Figure 3). The GluR3 B mAb was added to the skeletal muscle cells only once at different concentrations: 1 μg/mL, 10 μg/mL, 25 μg/mL, 100 μg/mL, 400 μg/mL, or 1 mg/mL and found to increase the cell number by: 1.3-, 1.8-, 1.4-, 1.5-, 1.8-, and 2.3-fold, respectively (Figure 3). We speculate (but did not prove directly) that the GluR3 B mAb induced this effect by activating and inducing de novo proliferation of the skeletal muscle cells.

Next, we tested if the GluR3 B mAbs increase the confluence of the human skeletal cells growing in the tissue culture and used the IncuCyte technology for studying that. The positive results are shown in Figure 4. As a reminder: in general, in cell culture biology, confluence refers to the percentage of the surface of a culture dish that is covered by adherent cells, while the cell number refers to, trivially, the number of cells in a given region. Cell confluency can affect cell behavior and growth, so it is important to accurately and reliably measure cell confluency. For further facts on the importance of cell confluence, please refer to https://bitesizebio.com/63887/cell-confluency/.

We found that a single addition of GluR3 B mAb, at either 1 mg/mL, 400 μg/mL or 10 μg/mL, increased significantly the confluence of the skeletal muscle cells during few days. This finding is observed in the following figures: Figure 4A, that presents muscle cells’ confluence curves of all the experimental groups during 4 days from addition of the GluR3B mAb; Figure 4B, that presents histograms of the confluence level at day 3 and also the averages, SD, and statistical values of the experimental groups at this day; Figures 4C–E, that show representative photographs of the skeletal muscle cells in this experiment, taken by the IncuCyte microscope and camera at day 4. The representative photographs of the untreated muscle cells are seen in Figure 4C, the photographs of the muscle cells added with 400 μg/mL GluR3 B mAb are seen in Figure 4D, and the photographs of cells added with 10 μg/mL GluR3 B mAb are seen in Figure 4E. These images visually illustrate the confluence-increasing effect of the GluR3 B antibody.

In the next step, we asked if glutamate, the natural neurotransmitter that binds GluR3 and all other types of GluRs, activates directly primary human skeletal muscle cells (myoblasts). We found that a single addition of glutamate, at either 10⁻⁴M, 10⁻⁵M, or 10⁻⁶M, increased by ∼1.5 fold, in a statistically significant manner, the number of human skeletal muscle cells growing in culture for 1 week (Figure 5). To the best of our knowledge, this is the first evidence that glutamate can independently activate directly human primary skeletal muscle cells and increase their number.

Next, we performed different functional experiments, testing if glutamate can independently increase intracellular sodium (Na⁺) levels in human myoblasts. We next briefly summarize the background and reason for testing intracellular sodium levels.

Regulation of intracellular sodium (Na⁺) levels and mitochondrial sodium plays an important role in multiple critical cellular processes in health and disease. Sodium concentration in the extracellular fluid is very carefully maintained in a range of 135–145 mmol/L and in the intracellular fluid at a much lower concentration of ∼10 mmol/L. Sodium entry into cells from the extracellular milieu takes place primarily via either of the following: 1. Sodium channels, which can be either voltage-gated or ligand-gated (in many cases neurotransmitter-gated), 2. Sodium transporters, or 3. Sodium pumps. Once sodium ions enter cells, they can act as a critical intracellular second messenger that regulates many cellular functions (Murphy and Eisner, 2009). Furthermore and importantly, the intracellular sodium concentration (10–15 mM) is regarded as a marker for cellular viability.

For testing if glutamate increases intracellular sodium levels in human skeletal muscle cells, we designed a multistep experimental protocol. Human primary skeletal muscle cells growing in tissue culture were loaded with CoroNa Green. a fluorescent dye that is a highly specific sodium (Na⁺) ion indicator, and that exhibits an increase in green fluorescence emission intensity upon binding sodium ions. Subsequently, glutamate was added at either of four concentrations, 10⁻⁵M to 10⁻⁸M to predetermined six replicate wells containing CoroNa Green-labeled skeletal muscle cells. For control, we added only buffer (PBS) to other wells. From that point onward, the skeletal muscle cells were repeatedly measured and analyzed for both their confluence and their CoroNa Green fluorescence, proportional to the level of intracellular sodium ions, and photographed. Finally, the IncuCyte technology yielded graphs of the CoroNa Green fluorescence before and after the addition of glutamate. Figures 6A–D, presenting the CoroNa Green fluorescence curves during 5 days after adding glutamate, and Figure 6E–presenting histograms of the CoroNa Green fluorescence only at day 5 only, as well as the averages, SD, and statistical values of all the experimental groups, show that adding glutamate at either 10⁻⁵M (Figure 6A; Figure 6E), 10⁻⁶M (Figures 6B,E), or 10⁻⁷M (Figures 6C,E) increased significantly the levels of intracellular sodium ions in the human skeletal muscle cells, evident by the elevated levels of CoroNa Green fluorescence. The effect was dose-dependent: the strongest effect was induced by glutamate at 10⁻⁵M, lower at 10⁻⁶M, even lower at 10⁻⁷M, and the lowest at 10⁻⁸M (Figures 6A–E).

Next, we tested if treatment of the cells with two iGluR agonists, AMPA and NMDA, also increase the intracellular sodium levels in human skeletal muscle cells. In this experiment, we also tested again the effects of glutamate.

The results are shown in two types of figures: Figures 7A–C, presenting the CoroNa Green fluorescence curves during 3 days, and in Figures 7D–F, presenting histograms of the CoroNa Green fluorescence only at day 3.

In this experiment, we found that glutamate at a concentration of 10⁻⁵M - 10⁻⁷M (Figures 7A,D), as well as AMPA at 10⁻⁵M - 10⁻⁷M (Figures 7B,E), and NMDA - at either 5 × 10⁻⁴M or 5 × 10⁻⁵M (Figures 7C,F), increased the levels of intracellular sodium ions in the human primary skeletal muscle cells. It should be noted that all the tested concentrations of both AMPA and NMDA are the well-known active concentrations of these agonists.

Next, we studied if glutamate induces very fast responses in skeletal muscle cells and found that glutamate increased the levels of the intracellular sodium ions within 2 min only (Figure 8), and maybe even sooner in shorter times we could not test.

This finding is shown in the eight representative images of CoroNa Green fluorescence, proportional to the levels of the intracellular sodium ions, inside each of four different human skeletal muscle cells, before and 2 min after adding glutamate (compare in Figure 8: A to B, C to D, E to F, and G to H). The fact that glutamate elevated sodium levels within few minutes (and maybe even less) suggests that glutamate induced these effects via direct activation of GluRs, and that glutamate may contribute to muscle excitation. However, this asssumption requires further investigation and proofs, primarily by electrophysiological recordings.

Next, we tested if affinity-purified human autoimmune GluR3 B antibodies of few epileptic NS patients bind human skeletal muscle cells, alike the mouse GluR3 B mAbs does. The method we used for purification of NS patient’s GluR3 B antibodies is drawn schematically in Figure 9A, and described in the “Methods” and in Levite et al. (2020).

We found that the NS patient’s GluR3 B antibodies indeed bind the human primary skeletal muscle cells (Figures 9B–E). In contrast, control healthy eluents do not bind the muscle cells (Figures 9F,G).

Next, we isolated all the IgG antibodies from the NS patient’s sera rich in autoimmune GluR3 B antibodies, but that also contain many other antibodies, and tested if they bind human skeletal muscle cells. The reason for testing this was that in vivo, in the patient’s blood, the GluR3 B antibodies are surely present in a heterogeneous milieu of antibodies, not in an isolated form, so if they bind in vivo GluR3 expressed on the cell surface of cells, they ought to do so when they are surrounded by many other types of antibodies. We found that the IgGs of the NS patients indeed bind the human skeletal muscle cells (Figures 9H–L). In contrast, control IgGs of healthy subjects do not bind these cells (Figures 9M–O).

In multiple previous studies on “Autoimmune Epilepsy,” we and others found that some enigmatic intractable epilepsy patients who suffer from very severe and prolonged epilepsy (and very different from NS) have elevated levels of autoimmune GluR3 B antibodies in their serum and that these GluR3 B antibodies induce multiple pathological effects on the neural cells in vitro and in vivo and damage the brain (for our most updated comprehensive review on “Autoimmune Epilepsy” GluR3B antibodies, which summarizes and cites most of the original studies done by various groups, and for just two recent original studies on the pathological effects in vitro and in vivo of GluR3B antibodies of intractable epilepsy patients and NS patients, the readers are referred to Levite et al. (2020), Levite and Goldberg (2021), Taiwo et al. (2025). These findings indicate that these patients suffer from “Autoimmune Epilepsy.” In some epilepsy patients, it may be their main disease, and in others, it may accompany and contribute substantially to their seizures and other neurological impairments. Some of the intractable epilepsy patients having autoimmune GluR3 B antibodies suffer also from motor problems in addition to the seizures, but others do not. So far, the possible binding and harmful effects of epilepsy patients’ GluR3 B antibodies or their IgG antibodies that contain high levels of GluR3 B antibodies to muscle cells were not tested.

In this specific study focusing primarily of GluR3 in muscle cells and the effects of glutamate on muscles cells, we performed only a very small pilot investigation on antibodies of few intractable epilepsy patients, which differ in many epilepsy characteristics but that all have elevated GluR3 B antibodies in serum (as found in our previous studies). The clinical information about the seven epilepsy patients we chose for this small pilot study is shown in Table 1.

Figure 10A presents the results obtained in only one of the ELISAs we performed for detecting GluR3 B antibodies in the serum of intractable epilepsy patients and shows that in comparison to nine healthy controls, four epilepsy patients: IE-3, IE-4, IE-6, and IE-7 (Table 1), have elevated levels of GluR3 B antibodies in serum. Elevated levels of autoimmune GluR3 B antibodies in the other three epilepsy patients: IE-2, IE-13, and IE-14 (data not shown herein) were observed in similar ELISA tests that we performed at other time points and included in previous reports (among them Ganor et al. (2005a), Goldberg-Stern et al. (2014).

Next, since we did not have enough serum from these patients to affinity-purify their specific autoimmune GluR3 B antibodies (as we did for the NS patients), we limited ourselves to isolating all their IgGs and testing them. We found that purified IgGs of few of the intractable epilepsy that have high levels of GluR3 B antibodies in serum indeed bind human skeletal cells. This is seen in the representative photos of the positive binding of the IgGs of four of these epilepsy patients (Figures 10B–G).

In contrast, the IgGs of four healthy control subjects did not bind the muscle cells (Figures 10H–M).

In the next step of the study, we found out that the purified IgGs of just one intractable epilepsy patient: IE-3* (others not tested), that contain elevated levels of GluR3B antibodies, decrease intracellular Na⁺ levels in human skeletal muscle cells, within few minutes (Figures 10N,O). In contrast, the IgGs of healthy subjects do not induce this effect (Figures 10P,Q). *Of note, IE-3 was the only epilepsy patient whose antibodies we tested in this study for their effect on intracellular sodium levels in muscle cells. We chose this specific patient due to two main facts: 1. This patient is in an extremely bad condition (bedridden for the last several years; the patient’s clinical information is summarized in Table 1. In addition, a detailed description of the patient can be found in our recently published article on the multiple in vitro and in vivo pathogenic effects of the autoimmune antibodies of this patient (Taiwo et al., 2025), 2. We already found and published that IE-3 has elevated IgG antibodies, and three types of glutamate receptor antibodies, directed against AMPA-GluR3 B , NMDA-NR1 and NMDA-NR2 peptides, and that the IgG antibodies of IE-3 bind and kill human neural cells in vitro, bind neural cells in the hippocampus and cortex and cause neural loss in these brain regions in vivo in naïve rats, and most important: induce recurrent generalized tonic–clonic seizures in these naïve rats (Taiwo et al., 2025).

In the last set of experiments in this study, we found that IgGs of few intractable epilepsy patients which have elevated GluR3 B antibodies not only bind human skeletal muscle cells, but also kill some of these cells within 1 h only (Figures 10U–W). In contrast, IgG antibodies of healthy subjects do not induce cell death (Figures 10S,T).

In this experiment, cell death was studied by staining the muscle cells with Sytox green–a fluorescent dye that enters only dead/dying necrotic cells. The killing of the muscles cells by the epilepsy patients’ IgG is clearly seen when comparing the representative confocal microscopy images of the Sytox Green⁺ dead cells aside the muscle-specific desmin⁺ cells in the: A. untreated muscle cells (Figure 10R), B. muscle cells added with IgG of 2 control healthy subjects (Figures 10S,T), C. muscle cells added with IgG of 5 intractable epilepsy patients: IE-2 (Figure 10U), IE-3 (Figure 10V), IE-6 (Figure 10W), IE-14 (Figure 10X), and IE-15 (Figure 10Y).

The quantitative graph of the percentage of dead cells (Figure 10Z) shows that the untreated muscle cells had 19.5% ± 16.2 dead cells; the muscle cells added with the IgGs of the three healthy subjects had 2.1% ± 3.6, 4.2% ± 8.4% and 12.3% ± 8 dead cells, respectively; and the muscle cells added with the IgGs of the five epilepsy patients had 45.3% ± 10.6 (*p), 55.7% ± 16.3 (*p), 56.7% ± 12.1 (**p), 55.8% ± 23.5 (*p) and 51.6% + 26 (*p) dead cells, respectively.

These findings suggest that the autoimmune GluR3 B antibodies or IgGs of intractable epilepsy patients that contain elevated levels of GluR3 B antibodies can damage muscle cells and motor function and that their effect is different from that of glutamate itself, and of GluR3 B mAb.