Neuronal and Neuroaxonal Damage Cerebrospinal Fluid Biomarkers in Autoimmune Encephalitis Associated or Not with the Presence of Tumor

Abstract

The aim of this study was to evaluate the association of neuronal damage biomarkers (neurofilament light chain (NFL) and total tau protein (T-tau)) in the CSF of patients with autoimmune encephalitis (AE) with the presence of an underlying malignancy and to determine correlations with patient characteristics. The study comprised 21 patients with encephalitis associated with antibodies against intracellular (n = 11) and surface/synaptic antigens (extracellular, n = 10) and non-inflammatory disease controls (n = 10). Patients with AE associated with intracellular antigens had increased CSF-NFL (p = 0.003) but not T-tau levels compared to controls. When adjusted for age, CSF-NFL but not CSF-T-tau was higher in patients with encephalitis associated with intracellular antigens as compared to those with encephalitis associated with extracellular antigens (p = 0.032). Total tau and NFL levels were not significantly altered in patients with encephalitis associated with extracellular antigens compared to controls. NFL in the total cohort correlated with neurological signs of cerebellar dysfunction, peripheral neuropathy, presence of CV2 positivity, presence of an underlying tumor and a more detrimental clinical outcome. AE patients with abnormal MRI findings displayed higher NFL levels compared to those without, albeit with no statistical significance (p = 0.07). Using receiver operating characteristic curve analysis, CSF-NFL levels with a cut-off value of 969 pg/mL had a sensitivity and specificity of 100% and 76.19%, respectively, regarding the detection of underlying malignancies. Our findings suggest that neuronal integrity is preserved in autoimmune encephalitis associated with extracellular antigens and without the presence of tumor. However, highly increased NFL is observed in AE associated with intracellular antigens and presence of an underlying tumor. CSF-NFL could potentially be used as a diagnostic biomarker of underlying malignancies in the clinical setting of AE.

Keywords: autoimmune encephalitis; autoimmune neurological syndrome; neurofilament light chain; non-paraneoplastic; paraneoplastic neurological syndrome; total tau protein.

Figure 1

Biomarkers of neuronal and neuroaxonal injury at disease exacerbation. (A,B) Cerebrospinal fluid (CSF) NFL levels and (C,D) total tau levels were compared between the control group (n = 10) and patients with autoimmune encephalitis (AE). In panels A and C, each group of patients was subdivided into those associated with the presence of autoantibodies against extracellular antigens and those associated with the presence of autoantibodies against intracellular antigens. In panels B and D, each group of patients was subdivided into the those associated with the presence of an underlying tumor and those not associated with the presence of tumor. (E,F) Receiver operating characteristic curve (ROC) analysis displaying the trade-off between sensitivity (true-positive rate) and (100% specificity) (false-positive rate). ROC curves tested the ability of each biomarker to discriminate the reference group (the control group and patients with AE without tumors) from the group of interest (patients with AE and tumors). For NFL: area under the curve = 0.9286, standard error = 0.04694 and 95% confidence interval: 0.8366 to 1.000; p-value = 0.0001. This analysis indicated that the best cut-off for CSF-NFL was >969 pg/mL, with the corresponding sensitivity and specificity to predict the presence of tumor set at 100% and 76.19%, respectively (E). ROC curve statistics for total tau were not statistically significant: area under the curve = 0.6950, p = 0.0862 (F). The horizontal lines in the dots represent the median. The vertical lines indicate the 95% CI, confidence interval. The p-values were obtained via unpaired t-test analysis using the Mann–Whitney test. p-values are indicated in the figure, and statistically significant comparisons were considered as those with p < 0.05. CSF = cerebrospinal fluid; AE = autoimmune encephalitis; NFL = neurofilament light chain; ROC = receiver operating characteristic curve.

Figure 2

Representative MRI findings of our study cohort. (A,B) Brain MRI from a 72-year-old male patient with a diagnosis of anti-CV2 paraneoplastic encephalitis. Axial (D) and coronal (E) T2-weighted FLAIR images (at 1.5T). T2-weighted FLAIR signal hyperintensities can be seen in the hippocampi and amygdalas bilaterally. (C) Example of brain MRI from an 82-year-old female patient with a diagnosis of anti-CV2 paraneoplastic encephalitis. Sagittal FLAIR sequence (at 3T) showing signs of cerebellar vermal atrophy. The atrophy of the anterior lobe of cerebellar vermis (with dilatation of the primary fissure) is disproportionate to posterior lobe atrophy of cerebellar vermis and relative sparing of other fissures. (D,F) Example of brain MRI from a 58-year-old female patient with a diagnosis of anti-GAD autoimmune encephalitis demonstrating dilatation of the ventricular system and cortical and hippocampal atrophy over the course of the disease. (D) Left: Axial T1-weighted 3D spoiled gradient echo sequence (at 1.5T); image acquired at 2 years after disease onset. Right: axial T1-weighted Turbo field echo sequence (at 3T); image acquired at 4 years after disease onset. (E) Left: Coronal T2-weighted fast spin echo sequence (at 1.5T); image acquired at 2 years after disease onset. Right: Coronal T2-weighted 3D sequence (at 3T); image acquired at 4 years after disease onset. (F) Left: Axial T2-weighted FLAIR sequence (at 1.5T); image acquired at 1 year after disease onset. Right: Axial T2-weighted FLAIR sequence (at 3T); image acquired at 4 years after disease onset. Note the enlargement of the temporal horns of the lateral ventricles (Scheltens 2 bilaterally) in panels A and B, and signs of cortical atrophy (GCA 1) four years after disease onset, especially in panels B and C. The rate of ventricular enlargement and cortical atrophy (in particular, as depicted by temporal horn enlargement) is particularly high for a 3-year period. MRI = magnetic resonance imaging; FLAIR = fluid-attenuated inversion recovery; GAD = glutamic acid decarboxylase; CV2/CRMP5 = collapsin response mediator protein; GCA= cortical atrophy.

Figure 3

General principles of immune-pathogenesis of paraneoplastic and non-paraneoplastic autoimmune encephalitic syndromes and the role of neuronal and neuroaxonal damage biomarkers. The presence of tumor is more commonly found in autoimmune encephalitis associated with intracellular antigens (lower panel). Autoantigens expressed or released (either after apoptotic or necrotic cell death) by tumor cells are thought to initiate an autoimmune response especially after being captured by antigen-presenting cells (APCs) (e.g., dendritic cells) in peripheral lymphoid tissue (lower panel). In the absence of a tumor, the underlying pathology is obscure and the most common triggering factors are thought to be a viral infection or genetic/environmental factors affecting tissue homeostasis (upper panel). Extracellular antigens are processed and presented by APCs to CD4+ helper T cells in the context of MHC II molecules and fully activate B cells. B cells differentiate in antibody-producing plasma cells. B cells, plasma cells and IgG autoantibodies enter the central nervous system through a leaky blood–brain barrier. IgG autoantibodies have the potential to cause synaptic autoimmune disorders directly through antigen internalization, complement activation, antibody-dependent cell cytotoxicity (ADCC) and receptor blocking (upper panel). Intracellular antigens processed by APCs, in the context of MHC class I molecules and in the presence of a pro-inflammatory cytokine milieu, favor the activation of cytotoxic CD8+ effector T cells. Cytotoxic CD8 T cells invade the CNS to exert cellular neuron-directed autoimmunity via the release of perforin and granzyme B, leading to the apoptosis of the target cells (lower panel). CSF-NFL levels are elevated in cases of encephalitis associated with the presence of an underlying tumor, where cytotoxic CD8 T causes more destructive effects in the affected tissue (lower panel). CSF-NFL and tau are not elevated in patients with encephalitis associated with extracellular antigens, possibly reflecting less destructive effects of autoantibody-mediated neuroaxonal injury (e.g., dysfunction of neuronal synapsis, indicative of distinct antibody-mediated effects) (upper panel). APC = antigen-presenting cells; HLA = human leukocyte antigen; MHC = major histocompatibility complex; ADCC = antibody-dependent cell cytotoxicity; NFL = neurofilament, HSV = herpes simplex virus; BCR = B cell receptor, TCR = T cell receptor.