Hypothalamic-Bulbar MRI Hyperintensity in Anti-IgLON5 Disease with Serum-Restricted Antibodies: A Case Report and Systematic Review of Literature

INTRODUCTION

Anti-IgLON5 disease is a rare autoimmune neurologic disorder clinically characterized by four-main disease-phenotypes that may often overlap: 1) a distinctive sleep disorder (abnormal non-rapid eye movement sleep (NREM) sleep initiation and motor activation associated with frequent vocalizations and movements, undifferentiated NREM sleep, REM sleep Behavior Disorder; 2) bulbar dysfunction; 3) progressive supranuclear palsy (PSP)-like syndrome; and 4) dementia with or without chorea or parkinsonism [1–4]. Among the most characteristic symptoms in anti-IgLON5 disease, laryngeal stridor is both a cause of significant morbidity and a relevant clinical clue, observed only in multiple system atrophy (MSA) among other parkinsonian and dementing disorders. Still, current knowledge on anti-IgLON5 disease pathology did not explore the mechanisms and structures underlying this presentation.

Brain MRI is generally normal or nonspecific. Although the exact pathophysiology of the disease is still incompletely understood, a strong association with specific HLA Class II molecules has been reported, including HLA-DRB1^*10:01 –HLA-DQB1^*05:01 or, less frequently, HLA-DQA1^*01 - HLA-DQB1^*05:01, which supports a primary autoimmune etiology [5].

We report a case of bulbar onset anti-IgLON5 disease with laryngeal stridor, distinctive brain MRI features, and serum-restricted antibodies in association with the alternative HLA haplotype, its clinico-radiological correlations, and a review of the current literature focused on the main clinical and instrumental findings.

METHODS

We reviewed the English literature for full papers on anti-IgLON5 disease published until July 31, 2020, searching the PubMed database of the National Center for Biotechnology Information and Scopus database for the term “IgLON5” (n = 82). We sorted through the titles to identify papers reporting on anti-IgLON5 encephalopathy (n = 59). Full texts were scrutinized to only include clinical studies presenting: a) patients with detected anti-IgLON5 antibodies either in serum or cerebrospinal fluid (CSF) specimens; b) either single case reports or case series (n = 34). Duplicated cases were carefully identified wherever possible, and in instances of unresolvable overlapping the less informative study was presented separately. Two readers (a neurologist, M.T., and a neurology resident, D.C.) assessed the full texts separately to abstract the following data with the support of a pre-defined form: sex, age at onset, follow-up length, outcome, treatment institution and type, response to treatment (null, incomplete/relapsing or complete), CSF investigations including neurodegenerative markers, HLA analysis, MRI findings detection of anti-IgLON5 antibodies in serum and CSF as confirmed by cell-based assay using HEK 293 cells transfected with the IgLON5 complementary DNA.

A set of clinical signs and symptoms was collected from case reports and case series with sufficient data identification at a single patient level, based on previous reports [1], Movement Disorder Society clinical diagnostic criteria for PSP [6] and MSA Second Consensus criteria [7]. Criteria timeframes were not considered. For each patient, each condition was considered present only if specifically stated in the full text, and discordance in data extraction between the two readers was resolved by consensus.

We identified patients that could be labelled as probable or possible MSA or PSP according to the aforementioned criteria. Descriptive analysis of clinical, laboratory, and radiological variables are reported.

RESULTS

Case report

A 65-year-old woman was evaluated at another hospital because of an 8-month history of unremitting cough and progressively increasing dysphagia to solids and intermittently to liquids, eventually precipitating an aspiration pneumonia. The patient also reported a change in her voice quality, fatigue, constipation, new onset of frontal headache radiating to the neck, right hemifacial spasm, and trismus. Hissing sounds during sleep were disclosed by her partner, highly suggestive of stridor.

She had a past history of restless legs syndrome and bruxism since she was in her twenties, in addition to long standing diabetes mellitus type 2 and bilateral carpal tunnel syndrome. Family history was unremarkable.

On admission, the patient spoke with a hoarse voice, had involuntary contractions of facial and masticatory muscles, and presented bilateral gaze-evoked nystagmus in the horizontal plane.

Direct fiberoptic laryngoscopy revealed bilateral vocal cord paresis with paradoxical inspiratory adductor spasm, and a barium swallow examination showed esophageal dysmotility with defective relaxation of the lower esophageal sphincter.

Electromyographic investigation of the trigeminal nerve reflexes revealed normal R1–R2 latencies and responses of the blink reflex, as opposed to the absence of the masseter inhibitory reflex and loss of both silent periods (SP1 and SP2) after electrical stimulation of mentalis nerve while the patient was clenching the yaws. Standardized brain imaging at 1.5T MRI scanner did not show significant abnormalities, and total-body CT scanning and positron-emission tomography excluded underlying malignancies.

A lumbar puncture was performed and CSF analysis revealed a total protein level of 65 mg per deciliter with no cells and no oligoclonal bands.

An extensive panel of serologic and CSF tests for autoimmune disorders, including antibody-mediated and paraneoplastic neurologic syndromes (anti-Hu/ANNA1, Ri/ANNA2, Yo, Ma2/Ta, CV2, SOX1, Zic4, Tr, amphiphysin, recoverin, titin, LGI1, CASPR2, NMDAR, GABAR, AMPAR, DPPX, GAD65) was negative.

A working diagnosis of seronegative autoimmune syndrome was considered and intravenous immune globulin (IVIG) at a dose of 2 g/kg of body weight over 5 days was administered and repeated after 1 month. Resolution of the symptoms, except for the persistence of dysphonia, was observed, and a maintenance therapy with oral prednisone at 0.6 mg/kg/die was started, at 11 months from the onset. Six months later, soon after prednisone tapering to 0.4 mg/kg/die the patient presented dysphagia and severe stridor. Also on this occasion, rescue therapy with a single course of IVIG followed by an increase of oral prednisone (0.8 mg/kg/die) and initiation of azathioprine (1.2 mg/kg/die) resulted in a good response over the next few weeks. Another tapering trial at two years from onset induced a further relapse necessitating a rescue therapy. In particular, respiratory insufficiency due to severe laryngospasm led to the placement of a permanent tracheostomy tube. The patient was referred to our Institute for a re-evaluation still being steroid-dependent at three years from onset, obliged to tracheostomy tube and in indolent clinical worsening with mild dysphagia.

CSF studies confirmed a mild protein elevation with elevated CSF:serum albumin ratio (10.13, reference value <7.00). Normal IgG:albumin ratio, absence of B lymphocytes on CSF flow cytometry and negative oligoclonal bands excluded intrathecal synthesis. CSF and serum testing for anti-IgLON5 (anti-IgLON5 IFA kit, EUROIMMUNE Lubeck) detected autoantibodies restricted to serum. A similar pattern was found on stocked frozen samples from the first assessment. Similar to previous reports [2, 3], staining of the molecular layer of the cerebellum was observed on immunohistochemistry but only with serum samples (Fig. 1).

OPEN IN VIEWER

Fig. 1

A sagittal section of rat cerebellum immunolabeled with the patient’s serum shows staining of the molecular and granular layers. Nonspecific staining of Purkinje cells is also observed.

Neurodegenerative markers were within reference values (Aβ_1 - 42 550 pg/mL, Aβ_1 - 40 5,574 pg/ml, p-tau 20 pg/mL, t-tau 174 pg/mL, Aβ_1 - 42/p-tau ratio 27.5, Aβ_1 - 42/Aβ_1 - 40 ratio 0.1).

DNA sequencing detected DQB1^*05 in homozygosis, DQA1^*01:01/01:02 and DRB1^*01/15 (Seq-uence Specific Primers method) and MAPT H1 in homozygosis (Sanger sequencing).

3.0T MRI of the brain showed an area of T2-hyper-intense, non-enhancing signal alteration spanning the brainstem tegmentum from the open medulla oblon-gata to the caudal pons, a region containing the bulbar parasympathetic nuclei (dorsal motor nucleus of the vagus, nucleus ambiguus), sensory afferents (nucleus of the solitary tract, spinal trigeminal nucleus) and passing fibers of the glossopharyngeal and vagus nerve (Fig. 2a–h). Similar findings were observed in the hypothalamus (Fig. 2i, j). Auto-nomic studies, including Ewing’s battery tests, sympathetic skin response, and ultrasound evaluation of bladder voiding, were significant for impairment of cardiovagal parasympathetic function with sparing of the orthosympathetic and sacral parasympathetic systems. Polysomnography (PSG) was recorded with synchronized audiovisual recording and included electroencephalogram (EEG) and submental electromyogram. PSG studies showed sleep fragmentation and bouts of central apnea-hypopnea without severe desaturation episodes or overt parasomnias. A normal representation of NREM stages, including K-complexes and sleep spindles was observed, in the absence of abnormal movements and erratic activity.

OPEN IN VIEWER

Fig. 2

Axial T2-weighted (a, b, c, d) and FLAIR images (e, f, g, h) demonstrate bilateral high signal intensity in the dorsolateral medulla (arrows in a, b, e, f) extending cranially to the floor of the fourth ventricle (arrows in c, d, g, h) at the site of spinal trigeminal nucleus, nucleus ambiguus, nucleus of the solitary tract, dorsal motor nucleus of the vagus, fibers of the glossopharyngeal nerve and fibers of the vagus nerve. Coronal T2-weighted (i) and FLAIR (j) show subtle hypothalamic hyperintensity (arrows in a and b). No diffusion restriction on DWI or pathological enhancement were observed (not shown).

Formal cognitive assessment revealed a decreased performance in attentive/executive domains on Addenbrooke’s Cognitive Examination-Revised and a score of 27/30 on Mini-Mental State Examination.

A protocol of monthly IVIG infusions with reinstitution of azathioprine 1.2 mg/kg were proposed as a steroid-sparing strategy, resulting in a stabilization of dysphagia.

Twelve months later the patient developed mild upper limbs weakness with muscular wasting but diffusely brisk reflexes. Persistent laryngeal adductor spasm still prevented the removal of the tracheos-tomy tube. Neurophysiological studies revealed a diffuse lower motor neuron syndrome with fasciculations, fibrillations, and positive sharp waves in all the explored muscles groups (bilateral genioglossus, biceps brachii, trapezius, and tibialis anterior muscles). Motor evoked potentials study showed increased latencies on peripheral stimulation with preserved central conduction time but reduced cortical amplitudes. At last visit, six months later, the patient did not show further progression.

DISCUSSION

We described distinctive brainstem findings on MRI in a case of anti-IgLON5 disease with predominant bulbar symptoms, in particular laryngeal stridor, an atypical immunological background, and insufficient response to immunotherapy. Characteristic changes on brain MRI are rarely reported in the published literature on anti-IgLON5 syndrome and are mostly limited to cases with acute or subacute presentation, specifically as widespread tumefactive and enhancing lesions in acute encephalitis [11], restricted diffusion in the setting of refractory seizures [13], and hypothalamic T2-hyperintensity in a case with subacute encephalopathy presenting with insomnia, confusion, and gait imbalance [8].

In our case, MRI abnormalities were located not only in the hypothalamus but also on the bilateral dorsolateral aspects of the mid and rostral me-dulla oblongata, a pattern that mirrors the diffuse involvement of tegmental brainstem comprising dorsal medullary nuclei observed on pathological studies either as lymphocytic infiltrates [19] or p-tau deposition [20], thus suggesting an in vivo depiction of local neuronal damage. The detection of hypothalamic-bulbar hyperintensity at 3.0T brain MRI, but not at a field strength of 1.5T, suggests that similar abnormalities could be more common than currently appreciated; in alternative, these discrepancies could simply reflect differences between early and late disease stages.

Bulbar syndrome with prominent respiratory and deglutitory symptoms is a frequent manifestation in anti-IgLON5 disease and often present since onset [21]. The inspiratory paradoxical vocal cord motion causing laryngeal stridor is reported in almost a third of patients with anti-IgLON5 encephalopathy, although figures may be as high as 83% if polysomnography is implemented [1], yet its pathophysiology is unexplored.

Besides anti-IgLON5 disease, the observation of acquired neurogenic non-traumatic stridor is mostly limited to MSA patients. The peculiar pathologic mechanism which underlies this uncommon entity has been previously described in this synucleinopathy: a differential spatial involvement of motoneurons within the nucleus ambiguus is postulated to result in a relative weakening of the vocal cord abductors (posterior cricoarytenoid muscle) leading to a narrowed glottis. In this setting, inspiratory efforts against the narrowed glottis cause a drop in intratracheal pressure that elicits reflex activation of adductor muscles, ultimately leading to stridor [22]. The involvement of tegmental brainstem observed on neuroimaging in our patient and on pathology in autoptic case series possibly advocates for a similar pathomechanism in anti-IgLON5 disease [19, 20]. Accordantly, brainstem pathology was not observed in the only pathology case that did not report laryngeal motion disorders [17]. Proximity is a likely explanation for the frequent occurrence of stridor in dysphagia, as neurons to pharyngo-esophageal muscles have been localized cephalad to laryngeal abductors in animal models [23, 24]. At a single patient level, we report the parallel fluctuation in severity of these symptoms, although this could also be related only to a similar susceptibility of the two systems.

Sudden death due to respiratory insufficiency is commonly reported, often in the setting of laryngospasm [3] or soon after tracheostomy tube removal despite apparent overall clinical regression [2, 21]. Ultimately, laryngeal stridor seems to be an important factor of both morbidity and mortality in anti-IgLON5 disease, although the available data prevent an accurate description of the phenomenon.

Elucidating the mechanism underlying laryngeal stridor could pave the road for more specific approaches, i.e., in the assumption that stridor necessitates a laryngeal closure reflex, interventions aimed at normalizing intratracheal pressure during inspiration as continuous positive airway pressure should damp the afferent arch and abate the symptoms, as has been documented in selected cases [2, 4]. However, there is insufficient evidence to date on the safety and the limitations of this approach. Signs of an impending relapse and the high-risk profile thus convinced us in maintaining a tracheostomy tube indefinitely.

Apart from supportive interventions for specific symptoms, current treatment of anti-IgLON5 disease is derived from previous experience on autoimmune encephalitides and relies on empirical administration of immunomodulant and immunosuppressive drugs, resulting in high variability of clinical response between patients. There is currently insufficient data to guide the clinician in the decision of a specific regimen, and further studies are needed to identify predictors of therapeutic response and the effectiveness of each option.

Anti-IgLON5 disease represents an infrequent disorder but amenable to treatment, with a protean presentation that in selected cases fully overlap with neurodegenerative diseases as PSP or MSA. Besides our analysis, Honorat et al. retrospectively found anti-IgLON5 antibodies in 1/50 PSP patients and 1/40 MSA patients, although a referral bias could not be excluded [3]. The peculiar sleep disorder could differentiate these entities, although it is likely to be misinterpreted as REM behavior sleep disorder in the absence of a formal polysomnographic evaluation. Clinical progression in months, altered CSF analysis, the presence of stridor, chorea, orofacial dyskinetic movement, or a clinical profile showing both PSP and MSA features could be possible red flags for antibodies testing in selected candidates.

We observed anti-IgLON5 reactivity restricted to serum, a finding described only in three other well-characterized patients presenting in one case with corticobasal syndrome [25] and the other with a PSP-like phenotype, all in the absence of sleep parasomnias [1, 26]. Three further patients are mentioned in a recent case series but possibly in overlap with previous reports [5]. Moreover, a serum restricted positivity is also reported in a PSP patient used as an historical control [2]. HLA genotyping of these four well-described patients revealed that only one harbored the classic HLA DRB1^*10:01-DQB1^*05:01 haplotype [26], instead the remaining three patients, ours included, harbored an isolate HLA-DQB1^*05:01 in the absence of the highly characteristic HLA-DRB1^*10:01. The alternative HLA-DQA1^*01 allele was observed in our patient, but not tested in the remaining two cases [1, 25]. There is still insufficient data to ascertain an association between this alternative haplotype and specific clinical features.

Autoantibodies of a confirmed pathogenic role but often restricted to serum have been described in other autoimmune encephalitis, i.e., in anti-LGI1 disease [27]. Previous studies hypothesized a potential pathogenic role of anti-IgLON5 [28], a view supported by the high specificity of anti-IgLON5 antibodies and the clinical response to immunotherapy in at least some patients [3, 4]. Under the hypothesis that the clinical course could be influenced by the intrathecal presence of the antibody and possibly correlating with CSF concentration as has been described in anti-LGI1 encephalitis [29], testing both CSF and serum should be recommended.

To conclude, alterations at high-field MRI in anti-IgLON5 disease demonstrate a good correlation with clinical symptoms and could represent a depiction of the process as seen by pathology. Selective vulnerability of brainstem structures carries a high-risk prognosis and warrants an accurate assessment and thoughtful therapeutic options.

DISCLOSURE STATEMENT

Authors’ disclosures available online (https://www.j-alz.com/manuscript-disclosures/20-1105r2).