Introduction

The COVID-19 pandemic caused by the SARS-CoV-2 virus has now affected more than 2% of the world population with over 185 million cases and 4 million deaths [1]. In addition to effects on health and mortality, the spread of the virus across the globe as well as governmental responses to the pandemic have had dire effects on human contacts and global economies and cases of anxiety and depression have increased in parallel, even in the non-infected population [2]. More directly, one study showed that approximately 30% of those infected by the virus and who experienced a severe course of the disease also developed psychological complaints, such as post-traumatic stress disorder (PTSD) [3]. Another study showed that of 125 severe cases registered as part of the CoroNerve study with neurological and psychiatric presentations of COVID-19 infection, 39 (31%) presented with altered mental status and 23 (18%) of these fulfilled the clinical case definitions for psychiatric disorders including new-onset psychosis, neurocognitive syndrome and affective disorder [4].

Yapici-Eser and colleagues recently described a potential pathomechanism based on molecular mimicry that may contribute to development of COVID-19-associated neuropsychiatric symptoms [5]. Structural similarities between the N-methyl-d-aspartic acid receptor (NMDAR) GluN1 (synonym NR1) and GluN2a (synonym NR2a) subunits with the SARS-CoV-2 nonstructural protein 8 (NSP8) and 9 (NSP9), respectively, may induce immune-mediated cross-reactivity to the NMDAR. These proteins are essential for replication of the virus and can interact directly with glutamate receptors of the NMDA and metabotropic families, leading to changes in membrane resting-state and action potentials [5]. Molecular mimicry may lead to generation of immunoglobulin G (IgG) antibodies against the NMDAR after SARS-CoV-2 infections. Anti-NMDAR encephalitis, mediated by IgG antibodies to the GluN1 subunit, is a common form of autoimmune encephalitis characterized by presentation of neurological and psychosis-like symptoms [6]. In this disease, antibodies bind to the NMDAR, induce crosslinking and receptors are subsequently internalized and thus are no longer available for excitatory glutamatergic transmission [6]. Viral diseases have been identified as potential triggers. For instance, anti-NMDAR-encephalitides can occur as a secondary disease after infection with viruses such as herpes simplex 1 or varicella zoster [7]. Moreover, past influenza A or B infections were identified as predisposing factors for NMDAR autoantibody seropositivity [8]. Accordingly, we hypothesized that SARS-CoV-2 might similarly induce anti-NMDAR encephalitis as a direct consequence of infection or secondarily through subsequent activation of autoimmune processes. Viruses such as SARS-CoV-2 hijack the cellular machinery of the host cell in order to reproduce themselves. In the process, their mimicry of key motifs of host proteins can lead to disruption of vital cellular functions, activate inflammation pathways and alter the immune response [9].

Here, we aimed to identify published cases of anti-NMDAR encephalitis with concurrent neuropsychiatric symptoms temporally associated with SARS-CoV-2 infections.

Methods

We searched the PubMed and Google Scholar databases using the search terms “NMDA encephalitis” or “NMDAR encephalitis” or “NMDA receptor encephalitis” and “SARS-CoV-2” or “COVID-19” to identify relevant cases. The last search was performed on September 20th 2021. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA, http://prisma-statement.org) were applied. The process for selecting studies (identification, screening, eligibility and inclusion in the systematic review) is reported in a flow diagram in Fig. 1. Studies were checked for eligibility and selected by two persons (VV and PCG). Initially, many more papers were identified in Google Scholar than in PubMed. However, most of these articles did not meet the search criteria as, on close inspection, they were comments on already published articles, meta-analyses, etc. For the final evaluation, only case reports or case series in original reports published in peer-reviewed international journals were considered. Articles in languages other than English or German were translated with the help of DeepL Translator (https://www.deepl.com). Finally, only 7 English [10,11,12,13,14,15,16] and one Spanish article [11] met the search criteria. The quality of the identified case reports was checked according to the CARE Case Report Guidelines (2013 CARE Checklist: www.care-statement.org) as summarized in Additional file 1: Table S1. Cases were included if they were positive for SARS-CoV-2 in nasopharyngeal swab or blood or cerebrospinal fluid (CSF) tests and suffered from anti-NMDAR encephalitis as defined by Graus et al. [17]. We extracted information regarding patient age and medical history, reason for hospitalization, respiratory symptoms, neuropsychiatric symptoms, blood-based biomarkers, magnetic resonance imaging (MRI) of the brain, CSF biomarkers and electroencephalography (EEG) readings, as well as therapies attempted and clinical outcomes, as available.

Fig. 1
figure 1

Selection procedure of the case reports. After reading of titles and abstracts, 469 publications could be excluded out of 477 potentially relevant articles, resulting in eight relevant cases finally. One unpublished congress poster matching search criteria was excluded due to the impossibility to perform a qualitative check (https://www.charcot-ms.org/files/Annual-Meetings/28/Abstracts-2020/42_ECF2020_Abstract_CL_Ramos_E.pdf)

Full size image

Results

We identified eight case reports in which anti-NMDAR encephalitis (characteristic disease symptoms and detection of NMDAR antibodies in CSF) occurred with a time delay of 3 days to 3 weeks after the manifestation of COVID-19 disease (Table 1). Four of the identified 8 patients developed symptoms with a time delay of less than one week. The age of the patients ranged from 23 months to 53 years and consisted of four males and four females. All patients had positive GluN1 antibodies in the CSF. Only patients 1–4 showed pronounced respiratory symptoms including low O2 saturation, bilateral milk glass opacities on chest x-ray and patchy bibasilar consolidation on chest computed tomography (CT) imaging [10,11,12], and severe hypoxemia with bilateral infiltrations [13]. Patients 2, 3 and 5–8 developed epileptic seizures during the course [11, 12, 14,15,16, 18]. All patients had psychiatric or neurological symptoms, manifested as a disturbance of consciousness, delirium, psychosis or catatonia. Patients 1 and 5–8 exhibited motor failures or dysarthria [10, 14,15,16, 18]. Cases 4 and 8 had increased levels of C-reactive protein (CRP) and patients 3, 4 and 8 had decreased lymphocytes [12, 13, 16]. Increased interleukin-6 (IL-6) levels were found in blood and CSF in one subject (patient 5) [14] and only in CSF in another (patient 1) [10]. Patients 1–6 had elevated lymphocytes in CSF [10,11,12,13,14,15]. Cases 3 and 8 were SARS-CoV-2 positive in the CSF [12, 16]. Patient 2 had ovarian teratoma as an associated malignancy and potential alternative trigger of disease [11]. EEG recordings from patients 1, 4, 5 and 8 showed a slowdown, and one of these (patient 5) showed an extreme delta brush-like pattern, typical for the diagnosis [10, 13, 14, 16]. In five patients, imaging examinations of the brain appeared normal but patient 4 showed hyperintensities on fluid-attenuated inversion recovery (FLAIR) images in left amygdala, left anterior putamen with negligible effects in the right amygdala [13]. It should be noted that in this latter case, an initial SARS-CoV-2 swab PCR test was negative, raising the possibility that the infection may have been hospital-acquired. Each patient received guideline-compliant steroid therapy with intravenous immunoglobulins, patients 5 and 8 underwent plasmapheresis [14, 16], and conditions improved in all cases.

Table 1 Summary of COVID-19 cases identified involving anti-NMDAR encephalitis with neuropsychiatric symptoms
Full size table

Discussion

This is the first systematic review to demonstrate a potential link between SARS-CoV-2 infection and secondary occurrence of anti-NMDAR encephalitis presenting with characteristic neuropsychiatric disorders. It is possible that mimicry of non-structural proteins of the SARS-CoV-2 virus with NMDA receptor subunit epitopes may have been the underlying cause of the autoimmune response against brain NMDA receptors (Fig. 2). However, further studies are required to pinpoint the precise mechanism of how such effects of the virus lead to central nervous system (CNS) autoimmunity.

Fig. 2
figure 2

Annotation: This schematic illustration does not correspond to the natural size proportions

Possible pathophysiology of anti-NMDAR encephalitis induced by SARS-CoV-2. During acute COVID-19 infection, viral particles, including NSP8 and NSP9, are released. The released proteins are recognized by T cells, leading to activation of B cells, which become plasma cells and produce IgM and later IgG antibodies against NSP8 and NSP9. I SARS-CoV-2-associated endothelitis and IL-17 produced by activated T cells disrupt the blood–brain barrier, allowing NMDAR antibodies to enter the CNS. IL-6 alters glial activity and initiates neutrophil granulocyte migration leading to further blood-brain barrier destruction and inflammation. II Due to molecular mimicry, antibodies produced by plasma cells in the CNS can cross-react with the NMDAR subunits GluN1, leading to receptor internalization with subsequent degradation.

Full size image

IL-6 and Th17 cells appear to be involved in the pathophysiology of COVID-19-associated neuropsychiatric symptoms [19, 20] (Fig. 2). IL-6 is a key proinflammatory cytokine that can alter neuronal and glial activity and induce axonal degeneration [21, 22]. Moreover, IL-6 levels correlate with the severity of disease course in COVID-19 [23]. IL-6-dependent Th17 activation and differentiation appears to be essential for neutrophil granulocyte migration [23]. Consequently, the proinflammatory cytokine IL-17 produced by Th17 CD4 + T cells may disrupt the blood–brain barrier (BBB) and contribute to the strong association of Sars-Cov-2 with a variety of different secondary autoimmune diseases (e.g. systemic lupus erythematosus, Guillain-Barré syndrome, antiphospholipid syndrome, large vessel vasculitis and thrombosis, psoriasis, inflammatory arthritis, autoimmune thyroiditis, type-I-diabetes) and circulating autoantibodies (e.g., anti-nuclear antibodies/ANA, anti-cardiolipin/aCL antibodies, cytoplasmic and perinuclear anti-neutrophil cytoplasmic antibodies/cANCA and pANCA, antiprothrombin IgM) [24]. Accordingly, Sars-Cov-2 has been referred to as an "autoimmune virus" [24].

Alternatively, microvascular damage caused during the systemic inflammatory response to SARS-CoV-2 infection may promote development of encephalopathy in severe cases [21] (Fig. 2). Inflammatory mediators produced by the alveolar epithelium, macrophages, and leukocytes may contribute to endothelial inflammation, increased vascular permeability, edema, and increased turnover of coagulation factors. The increased vascular permeability can also cause disturbances of the microcirculation including BBB impairment [21]. This disruption of the BBB gives rise to another possible mechanism involving facilitated transfer of NMDAR-directed antibodies, into the CNS [25].

This study was limited by the small number of patients identified with SARS-CoV-2-related autoimmune anti-NMDAR encephalitis. In addition, an aetiological role of SARS-CoV-2 in the generation of anti-NMDAR encephalitis is not definitive. In one of the cases an initial SARS-CoV-2 swab was negative and the patient only tested positive after the onset of anti-NMDAR symptoms [13]. In another case, an ovarian teratoma was present which is a potential alternative trigger of anti-NMDAR encephalitis [11]. However, more carefully-characterised studies should become available as the pandemic progresses and eventually recedes. This is important as COVID-19 patients with severe infections might have a higher risk for developing neurological manifestations such as autoimmune encephalitis. This highlights the importance of appropriate screening of patients using serological and CSF assessment via cell-based-assays for antineuronal antibodies, as well as brain imaging and EEG recordings [26]. Potentially, one might also include biomarkers for glial or neuronal damage such as S100B, myelin basic protein, neurofilament light chain (NfL), or neuron-specific enolase to screen for brain alterations [27,28,29]. Furthermore, it is important to monitor patients for neurological warning signs of autoimmune encephalitis including seizures, personality and memory changes, psychotic symptoms, delusional thinking, headache, dizziness, catatonia and dyskinesias [17, 30, 31].

The eight cases of anti-NMDAR encephalitis detailed in this paper also calls attention to the increased risk associated with the sudden onset or intensification of neuropsychiatric symptoms. All of the autoimmune encephalitis patients described in this study improved following high-dose steroids and immunoglobulin therapy, thus highlighting the importance of early immunotherapy once the diagnosis of autoimmune encephalitis has been made.