Autoimmune Encephalitis in Denmark: A translational approach towards a better understanding

Autoimmune encephalitides (AE) are a group of diseases caused antibodies targeting neural surface proteins in the brain. Patients present with subacute cognitive dysfunction, behavioural changes and a variety of neurological symptoms such as seizures, involuntary movements and speech abnormalities.

The discovery of AE at the beginning of this century has contributed to the rapid evolving field of neuroimmunology, and especially the field of antibody-mediated neurological diseases. This has resulted in effective handling of patients who previously remained undiagnosed and often untreated.

The key finding of antibodies directed against the N-methyl-D-Aspartate Receptor (NMDAR), has driven an avalanche of antibody discoveries and associated syndromes, and today more than 18 different AE causing antibodies are known. Previously, 45-50% of patients were diagnosed with “encephalitis of unknown origin”, which is reflected in the Danish ICD-10 Diagnostic Codes, which contain codes such as “encephalitis without further specification”, “viral encephalitis without further specification” and “post-infectious encephalitis not classified elsewhere”. The exact incidence of AE worldwide is unknown, but is estimated to be as frequent as infectious encephalitides. Within just 20 years, the number of patients receiving the diagnosis “encephalitis of unknown origin” has halved.

The clinical phenotypes, the diagnostic approach and the effect of immunotherapy are increasingly well-described. Thus, accurate descriptions of most of the syndromes now exist, diagnostic consensus criteria have been developed and the first evidence-based European treatment guidelines are underway. However, despite increased attention and an exponentially increasing publication rate, many of the underlying pathological mechanisms remain unknown. In addition, in majority of cases, no disease triggering cause can be identified. Furthermore, recent research indicates that long-term prognosis is not as favourable as first assumed.

Thus, many important areas have yet to be adequately covered. For instance, what is the aetiology? What is the underlying pathophysiology? Are there biomarkers that can predict prognosis and guide treatment? And what is the status of AE in Denmark?

The aim of this thesis was to use a translational approach of both basic and clinical research to: i) provide an overview of current knowledge on AE (manuscript I); ii) to evaluate information on diagnostics, medical work-up, treatment and prognosis of all NMDAR-AE patients in Denmark from 2009-2019 (manuscript II); iii) to investigate the utility of Neurofilament Light Chain (NfL) as a diagnostic or prognostic biomarker in the two most common types of AE (manuscript III); and iv) to establish an animal model of NMDAR-AE for future studies of pathological mechanisms and evaluation of novel treatments (manuscript IV).

In manuscript I, we conducted a non-systematic review to present current knowledge on sub-type specific phenotypes, disease mechanisms, diagnostics and pit-falls, treatment strategies and prognosis. We found that clinical syndromes were increasingly well-described. However, knowledge on underlying disease aetiology was scarce, treatment was still based upon expert opinion and data on long-term outcomes was limited.

Manuscript II was a retrospective observational study on all NMDAR-AE patients in Denmark from 2009-2019. We found that the incidence had been steadily increasing since 2009. In addition, the Danish NMDAR-AE cohort presented a higher median age, a more equal female:male distribution, and fewer associated tumours and ICU admissions than described in other European cohorts. Only approximately 20% of our patients were <18 years, which is significantly less than reported in other cohorts. One third of our patients had undergone brain positron-emissions-tomografi (PET) CT, and abnormal findings were noted in half of patients with a normal MRI. Additionally, older patients had a worse long-term outcome and more often paraneoplastic or post-infectious NMDAR-AE. Lastly, we observed that immunosuppressant maintenance therapy ≥3 months after discharge had no effect on relapse-rate. 

Manuscript III was a biomarker study conducted on our Danish NMDAR- and Leucine-rich Glioma Inactivated 1 (LGI1) AE patients. Here, we found that CSF-NfL levels at diagnosis were higher in LGI1- than in idiopathic NMDAR-AE. Additionally, CSF-NfL levels were lower in idiopathic/teratoma associated NMDAR-AE compared to malignant, demyelinating or post-infectious NMDAR-AE. Finally, higher CSF-NfL levels at diagnosis were associated with a worse prognosis in the combined cohort. 

In manuscript IV, we conducted a basic neurobiological study to establish a rat model of NMDARAE. We found that 14 days of continuous intraventricular infusion of patient antibodies resulted in cognitive dysfunction (memory impairment) compared to control animals. NMDAR antibody treated animals also showed a tendency towards increased anxiety levels. Lastly, we found that infusion with patient antibodies resulted in metabolic changes in the hippocampus (hypermetabolism) and the orbitofrontal cortex (hypometabolism) of rats, detected by brain PET CT.

Overall, our studies found that the NMDAR-AE syndrome may present slightly different in a real-life setting, when including entire national cohorts. In general, the incidence in Denmark has been steadily increasing, we identify and treat patients rapidly and the prognosis is favourable. One exception, however, are the paediatric patients which appear heavily underdiagnosed in Denmark when compared to the rest of the world.
In challenging cases with atypical disease presentation and a normal MRI, PET CT might be of utility in diagnosing NMDAR-AE. Additionally, CSF-NfL at the time of diagnosis can be useful to distinguish NMDAR-AE patients with concomitant diseases. Furthermore, CSF-NfL might be a prognostic biomarker in AE, as patients with higher levels at diagnosis have worse long-term outcome. Lastly, we established an animal model of NMDAR-AE and implemented PET CT in addition to behavioural and ex vivo studies. This animal model will enable further studies into pathological mechanisms and pave the way for a better disease understanding and development of novel treatments.