New‑onset refractory status epilepticus (NORSE) denotes the clinical presentation of refractory status epilepticus occurring in individuals without prior epilepsy or relevant neurological disease and without an identifiable acute structural, toxic, or metabolic cause. NORSE is associated with high mortality and often leads to long‑term neurologic sequelae including chronic, drug‑resistant epilepsy and cognitive impairment. Among identifiable causes, autoimmune encephalitis (AIE) is a major contributor; however, up to half of cases remain cryptogenic (c‑NORSE).
Traditionally, investigations into NORSE have emphasised innate immune mechanisms such as cytokine‑driven neuroinflammation and microglial activation. This review synthesises evidence that adaptive immune components—particularly B cells, autoantibodies, and T cells—also play critical roles in AIE‑associated NORSE and in sustaining neuroinflammation that may promote seizure generation and epileptogenesis.
This narrative review used a non‑systematic literature search of PubMed, Scopus, and Web of Science for studies published up to 2026. The search targeted literature on NORSE, B cell and T cell‑mediated immune mechanisms, and cytokine profiling. Original research (cohort studies, case series, case reports) were prioritised according to methodological strength; well‑documented case reports were included when they yielded mechanistic insights. References were also identified via manual screening. The review critically evaluated study design, sample size, methodological rigour, consistency of findings, and relevance to NORSE pathogenesis.
Patients with NORSE frequently exhibit markedly elevated serum pro‑inflammatory cytokines. Reported mediators include IL‑6, IL‑12p70, TNF‑α, and chemokines such as CXCL8 (IL‑8), CCL2 and CCL3. Higher concentrations of these cytokines have been associated with poorer clinical outcomes in the literature. In many cases, serum pro‑inflammatory cytokines decline after status epilepticus resolves; however, c‑NORSE patients may show persistent or evolving immune activation during a chronic phase lasting at least three months after SE resolution. Persistent elevations reported in this phase include IL‑12p70, IL‑17A, and TNF‑α, which coincide temporally with development of post‑NORSE epilepsy in observational studies.
Dysregulated innate immunity—manifest as monocyte and microglial activation and a pro‑inflammatory cytokine milieu—is hypothesised to drive blood–brain barrier (BBB) dysfunction. BBB disruption facilitates leukocyte extravasation into CNS parenchyma and exposure of peripheral immune cells to neuronal antigens. Mechanistically, IL‑6 promotes neuroinflammation and can impair BBB integrity, while TNF‑α amplifies inflammatory cascades. Chemokines such as CXCL8 recruit neutrophils and CCL2/CCL3 recruit monocytes, promoting microglial activation via receptors such as CCR2 and CCR5.
Imaging and neuropathological studies report increased BBB permeability in NORSE, especially in the thalamus and basal ganglia, compared with encephalitis without SE. The proposed sequence is that an acute cytokine surge enables non‑specific trafficking of peripheral lymphocytes into the CNS, where B cells may encounter neuronal surface antigens (for example, NMDAR). Within a pro‑inflammatory CNS microenvironment, sustained IL‑6 signalling can promote B‑cell survival, plasmablast differentiation, clonal expansion, and intrathecal antibody production. As innate cytokine levels subside, a transition toward adaptive immune‑dominated activity may occur with IL‑12p70, IL‑17A and TNF‑α supporting effector T‑cell differentiation and maintenance. These adaptive responses can further drive B‑cell activation and sustain neuroimmune activation, potentially contributing to epileptogenesis.
It must be emphasised that the temporal ordering and causal relationships between innate cytokine storms, BBB disruption, and subsequent adaptive immune activation in AIE‑associated NORSE or c‑NORSE are currently hypothetical; definitive, direct evidence delineating these steps is not established in the cited literature.
Evidence supports increased T‑cell presence in the CNS of NORSE patients. Cerebrospinal fluid (CSF) often shows pleocytosis with lymphocytic predominance (median CSF white cell count reported as 18/µl in referenced studies). Single‑cell transcriptomic analyses indicate T‑cell signatures in NORSE CNS tissue exceeding baseline immune surveillance. Neuropathological series document perivascular leukocyte infiltration in a subset of cases (reported in the literature as approximately 14% of examined cases), with infiltrates predominantly composed of CD8+ T cells.
BBB disruption and a pro‑inflammatory milieu likely facilitate T‑cell trafficking into the CNS. From a therapeutic viewpoint, modulating T‑cell mediated inflammation has been suggested as a potential strategy, although definitive interventional trial data specific to NORSE are not detailed in the source.
CD8+ cytotoxic T cells are implicated in direct neuronal injury in NORSE, particularly when intracellular neuronal antigens are involved. Gene expression profiles show greater functional activity of T cells in hippocampal tissue compared with cortex in NORSE patients, with signatures related to cytotoxic activity, apoptosis, and immune infiltration.
In conditions where intracellular antigens are the target (for example, GAD65, ANNA‑1/Hu, Ma2/PNMA2), antigen processing and presentation on MHC class I can trigger autoreactive CD8+ responses. Activated CD8+ cells release perforin and granzymes, inducing membrane pore formation and caspase‑dependent apoptosis in neurons. Perforin has also been linked experimentally to increased vascular permeability and BBB disruption, which may further perpetuate neuroinflammation and seizure activity. Animal epilepsy models (pilocarpine‑induced) report increased peripheral CD3+ and CD8+ T‑cell populations, supporting a contributory role for cytotoxic lymphocytes in seizure development.
Notably, cytotoxic T‑cell mediated neuronal injury can occur independently of detectable autoantibodies, highlighting that absence of serum or CSF autoantibodies does not exclude T‑cell driven pathology.
The reviewed literature presents a mechanistic framework in which an early innate cytokine surge promotes BBB disruption and non‑specific lymphocyte entry into the CNS. Subsequent antigen encounter and local pro‑inflammatory signalling support clonal expansion of B cells and T cells, with intrathecal autoantibody synthesis and cytotoxic T‑cell activity sustaining neuroinflammation and promoting seizure generation and perpetuation.
Therapeutically, the review mentions existing immunomodulatory approaches that target adaptive immunity, including monoclonal antibody therapies and immunosuppressive agents that inhibit T‑cell activation or cytokine signalling. Innovative therapeutic concepts inspired by chimeric antigen receptor (CAR) T‑cell technology are discussed as emerging strategies. Specific efficacy data, protocols, or comparative outcomes for these therapies in NORSE are not reported in detail in the source and therefore cannot be summarised here.
Overall, a deeper understanding of the interplay between innate and adaptive immunity, BBB dynamics, and antigen‑specific responses is essential to guide mechanism‑based therapies for NORSE. The source emphasises that many mechanistic links remain hypothetical and that further mechanistic and clinical studies are required to establish causality and to evaluate targeted immunotherapies.