Integrating large language models in care, research, and education in multiple sclerosis management.

Use of techniques derived from generative artificial intelligence (AI), specifically large language models (LLMs), offer a transformative potential on the management of multiple sclerosis (MS). Recent LLMs have exhibited remarkable skills in producing and understanding human-like texts. The integration of AI in imaging applications and the deployment of foundation models for the classification and prognosis of disease course, including disability progression and even therapy response, have received considerable attention. However, the use of LLMs within the context of MS remains relatively underexplored. LLMs have the potential to support several activities related to MS management. Clinical decision support systems could help selecting proper disease-modifying therapies; AI-based tools could leverage unstructured real-world data for research or virtual tutors may provide adaptive education materials for neurologists and people with MS in the foreseeable future. In this focused review, we explore practical applications of LLMs across the continuum of MS management as an initial scope for future analyses, reflecting on regulatory hurdles and the indispensable role of human supervision.

Effects of obstructive sleep apnea treatment on neurodegenerative biomarker neurofilament light chain and cognitive performance.

Obstructive sleep apnea is associated with cognitive impairment and increased risk for neurodegenerative diseases. Obstructive sleep apnea treatment with positive airway pressure therapy helps to improve cognitive symptoms and reduces long-term dementia risk. To test whether these treatment effects are due to a reduction in neuronal damage, we examined longitudinal changes in the neurodegenerative serum neurofilament light chain and cognitive performance of patients with obstructive sleep apnea. In this study, 17 patients with obstructive sleep apnea completed baseline and follow-up (9 month after starting PAP treatment) investigation of sleep, daytime symptoms, cognitive testing and serum neurofilament light chain measurements. Depending on treatment adherence and efficacy, participants were assigned either to the effective treatment (n = 10) or non-effective treatment group (n = 7). As results at baseline lower mean oxygen saturation during sleep was associated with higher serum neurofilament light chain. Patients in the non-effective treatment group showed a significant increase of age-adjusted percentile of serum neurofilament light chain levels at follow-up, whereas serum neurofilament light chain values remained constant in the effective treatment group. At a functional level, effective treatment leads to an improvement in processing speed, which was not the case in the non-effective treatment group. Longitudinal changes of age-adjusted serum neurofilament light chain levels were associated with changes in cognitive performance. To conclude, this longitudinal observational study showed that effective obstructive sleep apnea treatment positively affects the amount of neuronal damage as well as working memory performance. As cognitive symptoms might not only be attributed to obstructive sleep apnea-related sleep deficiency, but also neurodegeneration, our results underline the importance of treatment adherence and efficacy for the prevention of neuronal damage and cognitive consequences.

Development and multi-center validation of a fully automated digital immunoassay for neurofilament light chain: toward a clinical blood test for neuronal injury.

OBJECTIVES: Neurofilament light chain (NfL) has emerged as a promising biomarker for detecting and monitoring axonal injury. Until recently, NfL could only be reliably measured in cerebrospinal fluid, but digital single molecule array (Simoa) technology has enabled its precise measurement in blood samples where it is typically 50-100 times less abundant. We report development and multi-center validation of a novel fully automated digital immunoassay for NfL in serum for informing axonal injury status. METHODS: A 45-min immunoassay for serum NfL was developed for use on an automated digital analyzer based on Simoa technology. The analytical performance (sensitivity, precision, reproducibility, linearity, sample type) was characterized and then cross validated across 17 laboratories in 10 countries. Analytical performance for clinical NfL measurement was examined in individual patients with relapsing remitting multiple sclerosis (RRMS) after 3 months of disease modifying treatment (DMT) with fingolimod. RESULTS: The assay exhibited a lower limit of detection (LLoD) of 0.05 ng/L, a lower limit of quantification (LLoQ) of 0.8 ng/L, and between-laboratory imprecision <10 % across 17 validation sites. All tested samples had measurable NfL concentrations well above the LLoQ. In matched pre-post treatment samples, decreases in NfL were observed in 26/29 RRMS patients three months after DMT start, with significant decreases detected in a majority of patients. CONCLUSIONS: The sensitivity characteristics and reproducible performance across laboratories combined with full automation make this assay suitable for clinical use for NfL assessment, monitoring in individual patients, and cross-comparisons of results across multiple sites.

Raman Spectroscopy of Optically Trapped Living Human T Cell Subsets and Monocytes.

In recent years, Raman spectroscopy has garnered growing interest in the field of biomedical research. It offers a non-invasive and label-free approach to defining the molecular fingerprint of immune cells. We utilized Raman spectroscopy on optically trapped immune cells to investigate their molecular compositions. While numerous immune cell types have been studied in the past, the characterization of living human CD3/CD28-stimulated T cell subsets remains incomplete. In this study, we demonstrate the capability of Raman spectroscopy to readily distinguish between naïve and stimulated CD4 and CD8 cells. Additionally, we compared these cells with monocytes and discovered remarkable similarities between stimulated T cells and monocytes. This paper contributes to expanding our knowledge of Raman spectroscopy of immune cells and serves as a launching point for future clinical applications.

Serum neurofilament light concentrations are associated with cortical thinning in anorexia nervosa.

BACKGROUND: Anorexia nervosa (AN) is characterized by severe emaciation and drastic reductions of brain mass, but the underlying mechanisms remain unclear. The present study investigated the putative association between the serum-based protein markers of brain damage neurofilament light (NF-L), tau protein, and glial fibrillary acidic protein (GFAP) and cortical thinning in acute AN. METHODS: Blood samples and magnetic resonance imaging scans were obtained from 52 predominantly adolescent, female patients with AN before and after partial weight restoration (increase in body mass index >14%). The effect of marker levels before weight gain and change in marker levels on cortical thickness (CT) was modeled at each vertex of the cortical surface using linear mixed-effect models. To test whether the observed effects were specific to AN, follow-up analyses exploring a potential general association of marker levels with CT were conducted in a female healthy control (HC) sample (n = 147). RESULTS: In AN, higher baseline levels of NF-L, an established marker of axonal damage, were associated with lower CT in several regions, with the most prominent clusters located in bilateral temporal lobes. Tau protein and GFAP were not associated with CT. In HC, no associations between damage marker levels and CT were detected. CONCLUSIONS: A speculative interpretation would be that cortical thinning in acute AN might be at least partially a result of axonal damage processes. Further studies should thus test the potential of serum NF-L to become a reliable, low-cost and minimally invasive marker of structural brain alterations in AN.

Differential effects of selective versus unselective sphingosine 1-phosphate receptor modulators on T- and B-cell response to SARS-CoV-2 vaccination.

BACKGROUND: Sphingosine 1-phosphat receptor modulators (S1PRMs) have been linked to attenuated immune response to SARS-CoV-2 vaccines. OBJECTIVE: To characterize differences in the immune response to SARS-CoV-2 vaccines in patients on selective versus unselective S1PRMs. METHODS: Monocentric, longitudinal study on people with multiple sclerosis (pwMS) on fingolimod (FTY), siponimod (SIP), ozanimod (OZA), or without disease-modifying therapy (DMT) following primary and booster SARS-CoV-2 vaccination. Anti-SARS-CoV-2 antibodies and T-cell response was measured with electro-chemiluminescent immunoassay and interferon-γ release assay. RESULTS: Primary vaccination induced a significant antibody response in pwMS without DMT while S1PRM patients exhibited reduced antibody titers. The lowest antibodies were found in patients on FTY, whereas patients on OZA and SIP presented significantly higher levels. Booster vaccinations induced increased antibody levels in untreated patients and comparable titers in patients on OZA and SIP, but no increase in FTY-treated patients. While untreated pwMS developed a T-cell response, patients on S1PRMs presented a diminished/absent response. Patients undergoing SARS-CoV-2 vaccination before onset of S1PRMs presented a preserved, although attenuated humoral response, while T-cellular response was blunted. CONCLUSION: Our data confirm differential effects of selective versus unselective S1PRMs on T- and B-cell response to SARS-CoV-2 vaccination and suggest association with S1PRM selectivity rather than lymphocyte redistribution.

Influence of Pre-Analytic Conditions on Quantity of Lymphocytes.

Lymphocytes are key players in the pathogenesis of multiple sclerosis and a distinct target of several immunomodulatory treatment strategies. In this study, we aim to evaluate the effect of various pre-analytic conditions on immune cell counts to conclude the relevance for clinical implications. Twenty healthy donors were assessed for the effects of distinct storage temperatures and times after blood draws, different durations of tourniquet application, body positions and varying aspiration forces during blood draws. Immune cell frequencies were analyzed using multicolor flowcytometry. While storage for 24 h at 37 °C after blood draws was associated with significantly lower cell counts, different durations of tourniquet application, body positions and varying aspirations speeds did not have significant impacts on the immune cell counts. Our data suggest that immune cell counts are differently affected by pre-analytic conditions being more sensitive to storage temperature. Pre-analytic conditions should be carefully considered when interpreting the laboratory values of immune cell subpopulations.

Brain Atrophy Is Substantially Accelerated in Neurological Wilson's Disease: A Longitudinal Study.

BACKGROUND: Although brain atrophy is common in neurological Wilson's disease, longitudinal studies are lacking. OBJECTIVE: The objective of this study was to measure longitudinal brain atrophy rate and to relate it to the change in neurological impairment in Wilson's disease. METHODS: We included patients with brain imaging done at diagnosis and at least 12 months later. The atrophy rate was measured as percentage change in ventricular volume, whereas the change in neurological impairment was scored on the Unified Wilson's Disease Rating Scale. RESULTS: Of 57 patients, 36 had neurological presentation, 17 had hepatic presentation, and 4 were presymptomatic. The annualized atrophy rate was significantly greater in patients with the neurological presentation than in other patients (P = 0.001). In the neurological presentation, the atrophy rate correlated with the change in impairment (rho = 0.39, P = 0.018) and was significantly greater in those with worsening after diagnosis than in those without worsening (P < 0.001). CONCLUSIONS: Brain atrophy rate appears as a promising marker of neurodegeneration in Wilson's disease. © 2022 International Parkinson and Movement Disorder Society.

Serum Neurofilament Light Chain as a Biomarker of Brain Injury in Wilson's Disease: Clinical and Neuroradiological Correlations.

BACKGROUND: Clinical scales and neuroimaging are used to monitor nervous system injury in Wilson's disease, while data on serum markers are scarce. OBJECTIVE: To investigate whether serum concentrations of neurofilament light chain (sNfL) correlate with brain injury in Wilson's disease patients. METHODS: In 61 treatment-naïve patients, the Unified Wilson's Disease Rating Scale and a validated semiquantitative brain magnetic resonance imaging scale were compared with concentrations of sNfL. RESULTS: Concentrations of sNfL were significantly higher in patients with neurological disease compared with patients presenting with other forms (39.7 ± 73.4 pg/mL vs. 13.3 ± 9.2 pg/mL; P < 0.01). Moreover, the sNfL concentration positively correlated with neurological severity scores and with acute and chronic brain damage based on the neuroimaging scale. CONCLUSIONS: Neurofilament light chain concentrations may be used as a marker of brain injury in Wilson's disease, in addition to the clinical and neuroimaging disease severity scales. © 2022 International Parkinson and Movement Disorder Society.

Brain structure and neurocognitive function in two professional mountaineers during 35 days of severe normobaric hypoxia.

BACKGROUND AND PURPOSE: Animal studies suggest that exposure to severe ambient hypoxia for several days may have beneficial long-term effects on neurodegenerative diseases. Because, the acute risks of exposing human beings to prolonged severe hypoxia on brain structure and function are uncertain, we conducted a pilot study in healthy persons. METHODS: We included two professional mountaineers (participants A and B) in a 35-day study comprising an acclimatization period and 14 consecutive days with oxygen concentrations between 8% and 8.8%. They underwent cerebral magnetic resonance imaging at seven time points and a cognitive test battery covering a spectrum of cognitive domains at 27 time points. We analysed blood neuron specific enolase and neurofilament light chain levels before, during, and after hypoxia. RESULTS: In hypoxia, white matter volumes increased (maximum: A, 4.3% ± 0.9%; B, 4.5% ± 1.9%) whilst gray matter volumes (A, -1.5% ± 0.8%; B, -2.5% ± 0.9%) and cerebrospinal fluid volumes (A, -2.7% ± 2.4%; B, -5.9% ± 8.2%) decreased. Furthermore, the number (A, 11-17; B, 26-126) and volumes (A, 140%; B, 285%) of white matter hyperintensities increased in hypoxia but had returned to baseline after a 3.5-month recovery phase. Diffusion weighted imaging of the white matter indicated cytotoxic edema formation. We did not observe changes in cognitive performance or biochemical brain injury markers. DISCUSSION: In highly selected healthy individuals, severe sustained normobaric hypoxia over 2 weeks elicited reversible changes in brain morphology without clinically relevant changes in cognitive function or brain injury markers. The finding may pave the way for future translational studies assessing the therapeutic potential of hypoxia in neurodegenerative diseases.

Acute Effects of Ocrelizumab Infusion in Multiple Sclerosis Patients.

B cell-depleting therapies such as ocrelizumab (OCR) are highly effective in people with multiple sclerosis (MS). Especially at treatment start and initial infusion, infusion-related reactions (IRR) are a common adverse event. The relevance of acute changes of cell-depleting therapies on peripheral immune compartments and routine lab testing is important for clinical practice. We systematically analyzed routine blood parameters, detailed blood immunophenotyping and serum cytokine profiles in 45 MS patients starting on OCR. Blood samples were collected before and after corticosteroid premedication and directly after each OCR infusion of the first three ocrelizumab infusions. Blood B cells were rapidly depleted and accompanied only by a mild cytokine release at the first OCR infusion. Cytokine release was not significantly detectable from a third application in line with decreasing IRRs. B cell depletion was accompanied by short-lived changes in other immune cell populations in number, activation and cytokine secretion after each OCR infusion. Standard lab parameters did not show any clinically relevant changes. Our data demonstrate only mild changes during the first OCR infusion, which are not present any more during long-term treatment.

Ella versus Simoa Serum Neurofilament Assessment to Monitor Treatment Response in Highly Active Multiple Sclerosis Patients.

The measurement of serum neurofilament light chain (sNfL) is of growing importance in the field of neurology. In the management of multiple sclerosis, it can serve as a useful marker to assess disease activity and treatment response. This paper compares two available methods, namely the Single Molecule Array (Simoa) and the Ella microfluid platform, to measure longitudinal sNfL levels of 42 highly active multiple sclerosis patients treated with alemtuzumab over a period of 36 months. In order to assess the methods agreement, Bland-Altman plots and Passing-Bablok regression were analyzed. Here, we show that despite the fact that Ella measures around 24% higher values than Simoa, both are equally suitable for longitudinal sNfL monitoring.

Impedimetric Microfluidic Sensor-in-a-Tube for Label-Free Immune Cell Analysis.

Analytical platforms based on impedance spectroscopy are promising for non-invasive and label-free analysis of single cells as well as of their extracellular matrix, being essential to understand cell function in the presence of certain diseases. Here, an innovative rolled-up impedimetric microfulidic sensor, called sensor-in-a-tube, is introduced for the simultaneous analysis of single human monocytes CD14+ and their extracellular medium upon liposaccharides (LPS)-mediated activation. In particular, rolled-up platinum microelectrodes are integrated within for the static and dynamic (in-flow) detection of cells and their surrounding medium (containing expressed cytokines) over an excitation frequency range from 102 to 5 × 106  Hz. The correspondence between cell activation stages and the electrical properties of the cell surrounding medium have been detected by electrical impedance spectroscopy in dynamic mode without employing electrode surface functionalization or labeling. The designed sensor-in-a-tube platform is shown as a sensitive and reliable tool for precise single cell analysis toward immune-deficient diseases diagnosis.

Interleukin-17 and Th17 Lymphocytes Directly Impair Motoneuron Survival of Wildtype and FUS-ALS Mutant Human iPSCs.

Amyotrophic lateral sclerosis (ALS) is a progressive disease leading to the degeneration of motor neurons (MNs). Neuroinflammation is involved in the pathogenesis of ALS; however, interactions of specific immune cell types and MNs are not well studied. We recently found a shift toward T helper (Th)1/Th17 cell-mediated, pro-inflammatory immune responses in the peripheral immune system of ALS patients, which positively correlated with disease severity and progression. Whether Th17 cells or their central mediator, Interleukin-17 (IL-17), directly affects human motor neuron survival is currently unknown. Here, we evaluated the contribution of Th17 cells and IL-17 on MN degeneration using the co-culture of iPSC-derived MNs of fused in sarcoma (FUS)-ALS patients and isogenic controls with Th17 lymphocytes derived from ALS patients, healthy controls, and multiple sclerosis (MS) patients (positive control). Only Th17 cells from MS patients induced severe MN degeneration in FUS-ALS as well as in wildtype MNs. Their main effector, IL-17A, yielded in a dose-dependent decline of the viability and neurite length of MNs. Surprisingly, IL-17F did not influence MNs. Importantly, neutralizing IL-17A and anti-IL-17 receptor A treatment reverted all effects of IL-17A. Our results offer compelling evidence that Th17 cells and IL-17A do directly contribute to MN degeneration.

[MSProDiscuss - Development of a Digital Anamnesis Tool to Identify Disease Progression in Multiple Sclerosis]. / MSProDiscuss ­ Entwicklung eines digitalen Tools zur standardisierten Patientenanamnese, um Progredienz bei Multipler Sklerose zu identifizieren.

During the course of Multiple Sclerosis (MS), most patients with relapsing remitting MS (RRMS) convert to secondary progressive MS (SPMS), an MS-phenotype associated with a steady deterioration of functional ability independent from relapses and worsened prognosis. Due to the heterogeneity of this conversion, SPMS-diagnosis is often challenging and made retrospectively with a delay of several years. In this review, we first discuss advantages and limitations of screening tools for early SPMS-detection such as the SPMS nomogram, the MS prediction score, and the best SPMS definition approach. These screening tools might help to shorten the phase of diagnostic uncertainty. We then focus on the development of MSProDiscuss, a novel web-based tool that helps the treating neurologist to systematically assesses parameters highly relevant for SPMS-conversion during routine anamnesis. These parameters involve disease activity, symptoms, and impacts of the patient's overall symptoms. In a recent validation study, MSProDiscuss demonstrated high sensitivity, specificity, and interrater reliability. MSProDiscuss does not impose an additional time burden on the treating neurologist and its results are easy to interpret by a simple traffic light system. In first usability tests, it was therefore assessed as a helpful tool for the clinical routine. The early detection of clinically significant progression by diagnostic tools such as MSProDiscuss could open a time-window for therapeutic interventions.

How minimal variations in neuronal cytoskeletal integrity modulate cognitive control.

Until recently, investigating microscopic changes in the integrity of human brain matter has not been possible in vivo. It has hence remained unknown whether and how small non-pathological variations in cytoskeletal neuronal integrity affect human cognitive functioning. We investigated the role of neuronal cytoskeleton integrity for complex multicomponent behavior, which is relevant to real-life situations, as complex goals are often achieved by assembling a series of sub-tasks. For this, we quantified scaffolding proteins (i.e. neurofilament light; NF-L) using a single-molecule array (SIMOA), a new and uniquely ultra-sensitive method, and integrated this with behavioral and neurophysiological (EEG) data. For the first time, we showcase that slightest non-pathological variations in cytoskeletal integrity strongly modulate the efficiency of cognitive control processes. We show that the architecture and efficiency of theta-oscillations networks during cognitive control processes reflects a mechanism that establishes the relationship between neuronal cytoskeleton integrity and multicomponent behavior. Attentional selection processes do however not seem to play a role. The efficiency and network architecture of theta oscillations provides an important missing neural link that helps to explain how diffuse and seemingly miniscule variations in neuronal integrity may lead to reduced or even impaired cognitive functioning that is important for everyday activities.

Molecular biomarkers in multiple sclerosis.

Multiple sclerosis (MS) is an inflammatory-neurodegenerative disease of the central nervous system presenting with significant inter- and intraindividual heterogeneity. However, the application of clinical and imaging biomarkers is currently not able to allow individual characterization and prediction. Complementary, molecular biomarkers which are easily quantifiable come from the areas of immunology and neurobiology due to the causal pathomechanisms and can excellently complement other disease characteristics. Only a few molecular biomarkers have so far been routinely used in clinical practice as their validation and transfer take a long time. This review describes the characteristics that an ideal MS biomarker should have and the challenges of establishing new biomarkers. In addition, clinically relevant and promising biomarkers from the blood and cerebrospinal fluid are presented which are useful for MS diagnosis and prognosis as well as for the assessment of therapy response and side effects.

Engrafting human regulatory T cells with a flexible modular chimeric antigen receptor technology.

As regulatory T cells (Tregs) play a fundamental role in immune homeostasis their adoptive transfer emerged as a promising treatment strategy for inflammation-related diseases. Preclinical animal models underline the superiority of antigen-specific Tregs compared to polyclonal cells. Here, we applied a modular chimeric antigen receptor (CAR) technology called UniCAR for generation of antigen-specific human Tregs. In contrast to conventional CARs, UniCAR-endowed Tregs are indirectly linked to their target cells via a separate targeting module (TM). Thus, transduced Tregs can be applied universally as their antigen-specificity is easily adjusted by TM exchange. Activation of UniCAR-engrafted Tregs occurred in strict dependence on the TM, facilitating a precise control over Treg activity. In order to augment efficacy and safety, different intracellular signaling domains were tested. Both 4-1BB (CD137) and CD28 costimulation induced strong suppressive function of genetically modified Tregs. However, in light of safety issues, UniCARs comprising a CD137-CD3ζ signaling domain emerged as constructs of choice for a clinical application of redirected Tregs. In that regard, Tregs isolated from patients suffering from autoimmune or inflammatory diseases were, for the first time, successfully engineered with UniCAR 137/ζ and efficiently suppressed patient-derived effector cells. Overall, the UniCAR platform represents a promising approach to improve Treg-based immunotherapies for tolerance induction.