The Clinical Spectrum of Autoimmune-Mediated Neurological Diseases in Paediatric Population

Neurological autoimmune diseases have various origins and pathogeneses. Specific antibodies are associated with paraneoplastic syndromes, other infectious agents, or inherited disorders. We aim to evaluate the relation between the autoantibodies, the chosen symptoms, demographic characteristics, and infection history. We retrospectively analysed 508 children during neurological diagnostics. We investigated serum antineuronal, IgG, IgM anti-ganglioside, and anti-aquaporin-4 in both the serum and cerebrospinal fluid (CSF) anti-cell surface and anti-synaptic protein antibodies in 463, 99, 44, 343, and 119 patients, respectively. The CSF polymerase chain reaction detection of Herpesviridae, enterovirus, B19 parvovirus, adenovirus, and parechovirus involved 261 patients. We included available clinical information and electroencephalographic, radiologic, and microbiological results. The IgM anti-ganglioside antibodies increased the risk of tics and positive symptoms (p = 0.0345, p = 0.0263, respectively), the anti-glutamic acid decarboxylase particle of paresis (p = 0.0074), and anti-neuroendothelium of mutism (p = 0.0361). Anti-neuroendothelium, IgM anti-ganglioside, and CSF anti-N-methyl-D-aspartate antibodies were more often associated with consciousness loss (p = 0.0496, p = 0.0044, p = 0.0463, respectively). Anti-myelin antibodies co-occured with Herpes simplex virus (HSV)-2 IgG (p = 0.0415), anti-CV2 with HSV-1 IgM (p = 0.0394), whereas anti-glial fibrillary acidic protein was linked with past Epstein-Barr virus infection. The anti-ganglioside IgM and anti-myelin particles were bilaterally correlated (p = 0.0472). The clinical pictures may overlap, requiring specialistic diagnostics. We noticed the links between the infection aetiology and the specific autoantibody’s positivity.


Introduction
Autoimmune diseases are intensively studied. The diversified aetiopathological mechanisms remain under continuous research. However, several hypotheses include, e.g., molecular bacterial and viral mimicry, individual susceptibility to autoagression, and other genetic predispositions [1,2].
Specific antibodies appear as a part of paraneoplastic neurological syndromes (PNS) triggered by malignancies [3]. Mentioned rare diseases do not directly occur because of local tumour infiltration or distant metastasising but due to immunisation to antigens common for healthy and cancerous tissues [3,4]. Amongst autoantibodies, the strongest correlation with neoplastic disorders was described for anti-Ma/Ta, antineuronal nuclear antibody (ANNA) type 1 (anti-Hu) and type 2 (anti-Ri), against Purkinje cell cytoplasmic antigen type 1 (anti-PCA1) (anti-Yo) and type 2 (anti-PCA2), amphiphysin, collapsin response mediator protein 5 (CRMP5) (anti-CV2), and anti-Tr antibodies, together named anti-onconeuronal antibodies. Although the clinical course may overleap, specific antibodies also co-occur with certain syndromes [4].
Non-paraneoplastic autoimmune neurological syndromes are a large group of diseases in which the immune system abnormally targets the proper central nervous system's (CNS) antigens, but the underlying trigger remains outside the neoplastic transformation [5]. The particles causing the neurological impairment, aimed against neural antigens, involve anti-myelin, anti-myelin-associated glycoprotein (MAG), anti-neuroendothelium (NET), anti-glutamic acid decarboxylase (GAD), anti-non-myelinated fibres, and anti-glial fibrillary acidic protein (GFAP) antibodies.
The study compares current or previous infection indicators, the analysed autoantibodies, and the systemic inflammatory markers in a paediatric population. We planned to observe the autoantibodies' associations with age and the chosen symptoms incidence in the disease course.

Materials and Methods
We retrospectively analysed 508 paediatric patients between 2.5 months and 18 years of age suspected of autoimmune disease due to acute or persistent neurologic symptoms. The included participants were hospitalised in a tertiary centre between 3 January 2017 and 2 December 2019. Each child was diagnosed and treated with an individually suited plan.
We provided the statistical assessment with Statistica 13.3 (TIBCO Software Inc., 2017, Palo Alto, CA, USA). Depending on the data distribution and the number of groups, we used the χ 2 , the Mann-Whitney U (MW) test. A p-value of less than 0.05 we considered statistically significant.
In routine CSF examination, a significantly higher protein concentration occurred in a group with CSF anti-NMDAr antibodies than in children without them (p1 = 0.0499). What is more, both in whole and the autoantibody-positive populations, patients with antimyelin antibodies had significantly lower glucose concentrations than children without the particles (p1 = 0.0054; p2 = 0.0094, respectively). The median and IQR of analysed values appear in Table 2.
In the antineuronal autoantibody-positive group, patients suffered most frequently from consciousness alternation (25.6%), involuntary movements (24%), and headaches (22.5%). A large share presented focal signs (20.2%) and sight deterioration (17.1%). The symptoms' frequency did not statistically differ from autoantibody-negative patients. Separately, patients with serum anti-NET, anti-ganglioside IgM, and CSF anti-NMDAr antibodies had a significantly higher consciousness loss incidence than children without mentioned immunoglobulins, both in the total research population (p1 = 0.0496, p1 = 0.0044, p1 = 0.0463, respectively) and in the autoantibody-positive patients' group (p2 = 0.0185, p2 = 0.0258, p2 = 0.0367, respectively). Moreover, children with anti-MAG antibodies had a lower cognisance alternation rate (p1 = 0.0159; p2 = 0.0229). In addition, antiganglioside IgM positive patients statistically more often experienced involuntary movements (p1 = 0.0345), obsessive-compulsive disorder (OCD) (p1 = 0.0263), or positive symptoms (p1 = 0.0263). Similar observations did not apply to other autoantibodies. Our report suggests that the children with anti-GAD are more often diagnosed with varying severity paresis (p1 = 0.0074; p2 = 0.0078). Another result indicated that defecation problems might be significantly more frequent in the anti-GAD positive population (p2 = 0.0375); however, in our study, only one out of six anti-GAD positive children suffered from the mentioned ailment. We noticed mutism in 6.5% of patients with anti-NET antibodies. Compared to patients without such antibodies, it appeared to be a significant symptom (p1 = 0.0361). The CSF anti-NMDAr immunoglobulins' presence was associated with headache in our population (p1 = 0.0271; p2 = 0.0294). The frequency of focal and meningeal signs, balance, gait, sight disturbances, seizures, and impaired urination between patients with and without specific autoantibody did not differ significantly. We placed the complete data in Table 3.

Anti-MAG Ab
Antimyelin Ab

Anti-NET Antibodies
Anti-NET antibodies appear in rheumatic diseases [8], demyelinating autoimmune disorders including multiple sclerosis (MS), peripheral polyneuropathies [9], and paraneoplastic syndromes [10]. The targets spread in cerebral and nerve-associated vessels of the peripheral (PNS) and CNS. Chronic diseases may induce autoantibodies, which later, due to injury to the blood-brain barrier (BBB), may penetrate the intrathecal compartment [9]. The neuroendothelial cells participate in haemostasis and inflammatory response [11,12]; in addition, the cognitive impairment correlates with antibodies titre [13]. Our results partially fit the data in other research, in which seizures were observed in 33% of patients, upper or lower motor neuron involvement in 27% and 22% of patients, respectively, cranial nerve impairment bothered half of all patients, and cerebellar syndrome occurred in 65% of them [14]. The positive psychiatric symptoms' incidence was comparable to other autoantibodies and autoantibody-negative patients. In contrast, OCD exceeded the share amongst autoantibody-negative (6.5% vs. 3.8%) or participants with any autoantibodies detected (6.5% vs. 4.2%). However, the statistical significance was not found. In contrast to previous studies, we did not observe an increased cardiovascular incident rate in autoantibody-positive patients-ischaemic stroke appeared in one patient [13].
The pairwise tests revealed significant anti-NET and anti-ganglioside IgM coexistence. The publications describe the in vitro direct anti-monosialotetrahexosylganglioside antibodies' cross-reaction on microvascular endothelial cells in neuropathy, leading to disruption of the barrier between vessels and nerves, causing the exposure of other antigens [15]. However, the hypothesis concerns the peripheral nervous system and does not reflect the clinical reports.

Anti-GFAP Antibodies
The antibodies to GFAP cause inflammations involving the brain, cerebellum, meninges and brain stem, resulting in encephalopathy [16]. The clinical picture contains severe headaches, consciousness deterioration, and involuntary movements and less frequently partial paresis, ataxia, and other walking difficulties. Patients show memory loss, psychotic episodes, or autonomic involvement such as urine and stool incontinence or retention [17]. Fang et al. estimate that every fifth anti-GFAP-positive patient requires prolonged immunosuppression to sustain the remission. In our population, the most significant differences concern consciousness deterioration (18.5% vs. 68.8%), headaches (22.2% vs. 81.3%), and involuntary movements (18.5% vs. 33.3%), whereas paresis frequency was similar in both reports (18.5% vs. 18.8%) [18]. The cited authors selectively analysed an adult population with autoimmune disorders, whereas we based our research on a heterogeneous paediatric cohort, including children with infective or unclear aetiology. In the case series, the authors described two children who presented gait impairment and nystagmus. What is more, the first diagnosed with meningoencephalitis showed dysautonomia, bladder dysfunction, and cerebellar syndrome, whereas the second suffered from headaches and consciousness alternation, followed by extrapyramidal syndrome [19]. Our four anti-GFAP-positive participants had other autoantibodies detectable. We did not observe children with positive psychiatric symptoms in our population, and only one patient experienced autonomic urination impairment. Our data revealed a significantly higher anti-GFAP incidence in children with detected anti-VCA EBV IgG or anti-EBNA IgG. However, the literature lacks data on the correlation between anti-GFAP and infectious agents.

Anti-Myelin Antibodies
Autoantibodies against myelin compounds participate in various clinical syndromes such as cerebellar dysfunction, extrapyramidal symptoms, motor neurons, and vision impairment. The involvement of cranial nerves differs between reports [14,20]. The antimyelin particles appear in physiological protective reactions to CNS trauma but also occur in patients with a past oncological history or accompany primary autoimmune disorders and rheumatic diseases [14,21,22]. Our anti-myelin-positive patients suffered most frequently from involuntary movements, paresis, or sight deterioration and consciousness alternation.
Myelin oligodendrocyte glycoprotein (MOG) occurs exclusively in the CNS [23]. The antibodies against MOG predominantly affect an optic nerve and a spinal cord, accompanying difficulties with speech, swallowing, vision, and oculomotor function. The research in the paediatric population correlated the anti-MOG antibodies with neuromyelitis optica (NMO) and transverse myelitis (TM) in monofocal and polyfocal forms. The antibody frequently appears in acquired demyelinating diseases in childhood-especially in the younger population presenting acute disseminated encephalomyelitis (ADEM). The authors also concluded that anti-MOG-positivity during the acute phase in almost 50% is transient, as well as with non-MS disease course [24]. The positivity concerns 20% of patients with NMO and TM but twice more with ADEM [25,26]. Moreover, anti-MOG persistent positivity brings relapse risk 3-5 times greater [27]. The HSV may cause peripheral neuropathies, including facial nerve palsy [28]. We found a correlation between anti-myelin positivity and anti-HSV-2 IgG particles, but previous studies suggest anti-MOG antibodies' peripheral presence preceding such infection. The viral disease impairs the BBB, enabling the circulating autoantibodies to conjugate with CNS and trigger symptoms [29]. Literature also confirms that influenza and EBV participate in the pathogenesis of anti-MOG ADEM [30], having a predictive value in the relapsing multiphasic course risk [27]. The viral infections preceding the NMO confirm that infectious diseases facilitate autoantibodies penetration to the intrathecal compartment [6]. The animal model independently proved the HSV-1 potential to trigger encephalitis and multifocal demyelinated CNS lesions [6]. Zheng et al. noticed the cross-reactivity between MOG and CMV in the rat experimental autoimmune encephalomyelitis (AIE) model [31]. However, postinfectious anti-MOG positive NMO appears after bacterial infections (e.g., B. burgdorferi and Mycobacterium tuberculosis (M. tuberculosis)) [7,32]. We found the association with M. pneumoniae IgM that fits previous reports hypothesising the molecular mimicry triggering anti-myelin antibodies production and immunocomplexes deposition causing perivenular inflammation [1,26]. Christie et al. confirmed the cross-reaction with galactocerebroside C in 38% of M. pneumoniae encephalitic patients [33].

Anti-MAG Antibodies
The anti-MAG antibodies are responsible for various neuropathies. Several kinds of research successfully confirmed the pathological role of the particles [34], e.g., the passive transfer of human-derived anti-MAG antibodies to chickens. The investigation revealed the development of a similar pathophysiological pattern of deposits in myelin sheets and structural change of the layers [35]. The most typical course begins with numbness and upcoming paraesthesia, leading to distal, symmetrical sensory impairment, mainly due to the injury of myelinated fibres. Motor involvement with muscle weakness and ataxia rarely happen [36]. In our population, patients most frequently experienced involuntary movements, sight deterioration, and gait impairment in 40%, 26.7%, and 20% of cases, respectively. Compared to the available report in which balance impairment involved 17.7% of adult patients, the incidence reached 13.3% in our population. We did not observe cranial nerve involvement or epileptic seizures, which bothered 50% and 66% of cited research patients. The same authors detected motor neurons involvement in over 34% and myopathy in 60% of participants [14]. Findings associated with anti-MAG antibodies more typical for the paediatric population include ataxia after chicken pox [37] and their presence in 70% of patients diagnosed with autism [38].
The detected link between anti-MAG positivity and antecedent CMV infection was described in a previous study in which a CMV DNA occurred in 88% of anti-MAG positive patients with chronic neuropathy. The mechanism remains unknown, but molecular mimicry was suspected [39]. However, several works undermine the correlation and suggest the latent CMV reactivation, but the acquirable data cannot justify anti-MAG production [40]. In addition, Lunn at. al. failed to confirm the hypothesis in patients with peripheral neuropathy and paraproteinemia [41].

Anti-PCA2 Antibodies
The anti-PCA2 is a partially characterised onconeuronal antibody associated with small-cell lung carcinoma (SCLC). In our research, the tics dominated amongst children's symptoms; only one suffered from the walking impairment or experienced a consciousness decrease. However, the previous publications describe a wide range of symptoms potentially associated with the mentioned antibody due to encephalitis, myelitis, or neuropathy. Cognitive impairment appears as a short-term memory decrease, disorientation, and consciousness clouding. The paraneoplastic syndrome may manifest with cerebellar ataxia, regional paraesthesia, spastic paresis, or myasthenic disorder [42,43].

Anti-GAD Antibodies
The anti-GAD antibodies participate in autoimmune DM pathogenesis [44]. The neurological syndromes are rare, have marked female predominance, and appear in patients with anti-GAD detectable in CSF [45]. They involve cerebellar ataxia, stiff-person syndrome, AIE, and epileptic episodes resulting from impaired GABAergic transmission [46]. Paediatric data suggest that high anti-GAD titres provoke temporal-based focal seizures, psychiatric symptoms, and cognitive decrease, manifesting in remembrance difficulties and developmental slowdown [5]. On the other hand, up to 1.7% of healthy and 5% of neurological patients with different aetiology are anti-GAD positive [47]. We observed pareses and impaired defecation in our population in 66.7% and 16.7% of participants, respectively. The 16.7% of patients also experienced balance difficulties. Honnorat et al. described gait impairment in all patients; however, the study focused on patients with cerebellar involvement. The seizure and visual symptoms rates were lower, reaching 33.3% vs. 53% and 16.7% vs. up to 86%. We noticed a similar consciousness alternation frequency than previously described for anti-GAD-associated cerebellitis. We did not note memory deterioration, which appeared in 67% of patients in the cited research [44]. In research concerning children with AIE, authors described a change of behaviour and alternated consciousness in all three anti-GAD positive patients; two presented hallucinations and sleepiness, and one patient had an autonomic disorder or tremor [48].

Anti-Yo Antibodies
Anti-Yo belongs to well-described autoantibodies associated with ovarian and breast cancer, but 2-8.5% of non-oncologic patients occur positive [4,49]. The most typical clinical manifestation bothering 90% of cases includes cerebellar degeneration; however, brainstem encephalitis or peripheral neuropathy happens rarely [49]. In our non-oncologic population, we diagnosed paresis in two patients; in addition, two experienced consciousness deterioration and one patient had tics. In literature, varied ataxia and nystagmus concerned all patients; peripheral neuropathy appeared in 47% and cognitive impairment in 18% of paraneoplastic syndrome cases [50]. In the paediatric population, anti-Yo-related paraneoplastic syndrome with typical PCD, including ataxia, speech impairment, and vertigo, coappear with Hodgkin's lymphoma; in children without malignancy, the anti-Yo positivity rate increases in patients diagnosed with attention deficit hyperactivity disorder [51].
Interestingly, significant anti-CV2 positivity appeared amongst children with anti-HSV-1 IgM. A similar association was not previously described, nor with another infectious agent. The analysis requires further research because of the low positive patient quantity and potential statistical bias.

Anti-Tr Antibodies
Antibodies against Tr protein are partially characterised paraneoplastic molecules predominantly observed in Hodgkin lymphoma, but 11% of positive patients had no detectable neoplasm [4]. Amongst analysed children, we detected the anti-Tr antibodies in a single patient with autism and epilepsy hospitalised due to transient consciousness impairment. What is more, the boy also produced anti-GFAP antibodies. Typically, anti-Tr antibodies cause PCD or limbic encephalitis [49], but in children, paraneoplastic syndromes are casuistic, including vision impairment, ataxia, tremor, and muscular hypotonia [53]. Characteristic ataxia may appear as an isolated syndrome or combine with encephalopathy and peripheral neuropathy [42].

Anti-Ma/Ta Antibodies
This well-described onconeural antibody appears in germinal testicular tumours or non-SCLC and less frequently in other solid neoplasms. The main clinical presentations include limbic or brainstem encephalitis and PCD, with their frequencies assessed for 58%, 21%, and 16%, respectively [49]. Patients with underlying tumours constitute 58-96% of positive participants [4]. They experience consciousness alternation, cataplexy, sleep disorders and diencephalic disorders [54]; children are more prone to focal seizures, behaviour change, or speech and muscle tone impairment [5]. In paediatric non-paraneoplastic AIE typical symptoms include behavioural change and, speech and consciousness impairments, but sometimes positive psychiatric symptoms appear [48]. In our research, the boy with anti-Ma/Ta antibody suffered from blurred vision and headache associated with optic nerve inflammation, foretelling MS.

Anti-NMDAr Antibodies
Autoantibodies against NMDAr most frequently cause AIE. In the general population, 58% of patients are diagnosed with ovarian teratoma [55], but antecedent infection appears typically amongst children. The best-described infective anti-NMDAr encephalitis triggering factor is HSV-1; the other viruses' roles are ambiguous [2,6,7,56]. Our research suggests the link between CSF anti-NMDAr and CMV IgM positivity. The clinical spectrum includes dyskinesis, choreoathetosis, dystonic posturing, and rigidity. Psychiatric symptoms such as insomnia, paranoia, anxiety disorders, and cognitive deterioration appear less frequently [3,42]. Our patients with CSF anti-NMDAr antibodies presented consciousness loss and severe headaches. Imbalance and motion problems probably resulted from generalised weakness. Amongst patients with those antibodies in serum, 13% experienced deterioration in cognitive functions. However, the most common complaint was vision impairment (50%), and 21% suffered from involuntary movements; psychiatric disorders appeared in 13% and paresis in 16% of cases. The results partially cover the previously observed clinical pictures of paediatric AIE; however, decreased consciousness occurred in less than 50% of cases. The report describes behavioural alternation in all patients, speech disorder in 73%, facial dyskinesias in 64%, and hallucinations in 18% of children [48].

Antiganglioside Antibodies
Gangliosides, crucial for signal transition, spread on neurons' surfaces. The antibodies against them participate in chronic and acute neuropathies' pathogeneses; the role in paediatric idiopathic epilepsy needs further research [57]. Antiganglioside antibodies may act as primary or secondary aetiological factors. Although specific targets characterise particular neuropathies, we analysed them collectively.
The children who presented IgG antibodies suffered more frequently from muscle weakness and gait impairment. Nevertheless, less often exhibited symptoms included consciousness alternation or positive psychiatric symptoms. In the literature, IgG antiganglioside antibodies link with specific symptomatic disorders or diseases with other known aetiology, e.g., Guillain-Barre syndrome (GBS), MS, and Alzheimer's or Parkinson's disease [57]. Research has widely described the association of GBS with previous CMV, EBV, and M. pneumoniae infections [58], but our analyses did not bring sufficient evidence to confirm the findings in our population.

Anti-Aquaporin-4 Antibodies
In our research, we detected one patient diagnosed with non-specific stress-associated vision impairment. She reported headache, sight deterioration, presented involuntary movements, and balance impairment. The NMO spectrum disorder most frequently appears as myelitis (84%) and optic neuritis (63%). A brain, brainstem, and area postrema are involved in around 15% of cases. The disease may occur idiopathically or after viral infection with a tendency to relapse [59,60]. Anti-AQP4 antibodies typically manifest with transient vision loss, peripheral sensory and motor impairment, and paroxysmal movements episodes in 20% of patients [42].

Conclusions
The paper presented the diseases' manifestations and sought pathogenesis. The clinical pictures partially differed from courses described in adult patients. The main reasons may involve involve a lower incidence of malignancies and a higher prevalence of infectious diseases in children. The protocol involving autoantibodies' assessment in non-specific neurologic diagnostics might contribute to the detection of pathogenic particles in the pre-clinical phase, enabling the close patient's observation and prompt treatment in the early stage.
The main drawback resulting from a retrospective study is the difficulty in comparing symptoms between patients and low quantity in subgroups decreasing the tests' strength. The other source of bias remains the inclusion of patients in different stages of diseases. This aspect concerns both infective and non-infective aetiologies; however, in the first case, the serological testing would be helpful, especially after the acute phase. The report's objectivity could be improved by the standardised protocol used for patient anamneses and examinations.  Institutional Review Board Statement: The Institutional Ethical Committee stated that formal approval and informed consent is not required due to the retrospective character of the study, including anonymised data analysis from clinically justified procedures.