CSF. Even though our blog focuses on neurogenetics, there is a much wider range of rare diseases in child neurology, including many conditions that may not even have names yet. One emerging category is pediatric neuroimmune disorders. In a recent publication generating CSF immune profiles in these conditions, we contributed by providing the framework for pediatric sample acquisition. Here is what we learned and why ongoing biobanking is critically important in child neurology.
Figure 1. Pediatric CSF immune profiling highlights a distinct intrathecal signature in MS. This figure shows that while B cells are increased across acquired demyelinating syndromes (ADS), pediatric multiple sclerosis (MS) is uniquely characterized by marked expansion of antibody-secreting cells (ASCs) and relative reduction of CD14+ myeloid cells in the cerebrospinal fluid (CSF). The upper panels display cell frequencies and the lower panels absolute counts (cells/mL) across diagnostic groups: non-inflammatory neurological disease (NIND), peripheral inflammatory neurological disease (PIND), autoimmune encephalitides (AIE), inherited disorders of white matter (IDWM), other ADS, myelin oligodendrocyte glycoprotein antibody–associated disease (MOGAD), and MS. Together, the enrichment of ASCs and relative depletion of CD14+ myeloid cells in MS support the concept that MS establishes compartmentalized humoral immune activity within the CNS early in disease, distinguishing it from MOGAD and other ADS. Adapted from Espinoza et al., 2025 (CC BY-NC-ND 4.0).
Rare disease. In the United States, a rare disease is defined as a condition affecting fewer than 200,000 people. Virtually all genetic epilepsies and neurodevelopmental disorders fall into this category. Accordingly, it may be tempting to equate rare disease with genetic causes. However, this perception exists largely because genomic sequencing has made extraordinary progress over the last 20 years. Without high-throughput sequencing, few of the more than 150 genetic epilepsies recognized today would have a name in 2026. This raises a broader question: what else is out there? What other rare diseases are we currently missing, and what will it take to find them?
Landscapes. The honest answer is that we do not know. Neurology in general, and child neurology in particular, remains a field of “mystery diagnoses,” where a significant number of individuals have conditions for which we currently have no explanation. One way to build a framework for identifying and treating future rare diseases is the systematic availability of biosamples and reference values for biomarkers. While neuroimmune disorders are increasingly recognized in adults, their pediatric presentations often differ. In the recent publication by Espinoza and collaborators, we aimed to provide a reference framework by profiling CSF immune landscapes across childhood in both inflammatory and non-inflammatory conditions.
Profiles. In total, Espinoza and collaborators assessed CSF immune profiles in 85 individuals between 3 and 18 years of age, including children with neuroinflammatory conditions and non-inflammatory brain disorders as controls. Pediatric multiple sclerosis (MS, n = 15) and myelin oligodendrocyte glycoprotein antibody–associated disease (MOGAD, n = 10) were the two largest neuroimmune groups in the study. A standardized 16-parameter flow cytometry panel was used to assess CSF immunophenotypes. The panel included markers for T cells, B cells, CD14+ myeloid cells, and natural killer cells, as well as markers of activation and differentiation. This allowed detailed quantification of immune subsets such as antibody-secreting cells and activated memory T cells, even from small CSF volumes typical of pediatric practice.
Dynamic CSF. Several findings were particularly insightful, even for someone like me who does not primarily work in neuroimmunology. First, non-inflammatory pediatric CSF is not simply a scaled-down version of adult CSF. While the overall immune architecture resembles that of adults, the CD8+ compartment undergoes an age-associated shift from naive to memory phenotypes. This indicates that CNS immune surveillance matures across childhood, which most likely reflects antigen exposure over time and the establishment of resident memory populations within the CNS.
MS versus MOGAD. A second important finding was the clear biological distinction between the two major neuroinflammatory groups. MS and MOGAD can look strikingly similar at first, but they are different at the level of CSF immunophenotypes. MS shows compartmentalized, intrathecal immune activity with increased antibody-secreting cells and relative depletion of CD14+ myeloid cells. MOGAD, in contrast, lacks a strong intrathecal antibody-secretion signature and often shows normal or elevated myeloid populations. In other words, MS behaves like a CNS-resident immune disorder from early on, whereas MOGAD is more antibody-defined and less compartmentalized within the CNS. Both pediatric MS and MOGAD are rare diseases, and the early divergence in their underlying biology is striking. In MS, the immune response takes up residence within the CNS. In MOGAD, the disease-defining antibodies are generated in the periphery and carried into the CNS, where they trigger inflammation.
Pediatric biobanking. What does this study tell us about pediatric biobanking more broadly? For me, the conclusion is straightforward: we need to bank where we are not yet looking. Our field currently has a strong focus on genetics, which likely explains a substantial proportion of rare diseases. However, there will be additional mechanisms. Future discoveries will require reference datasets, baseline immune landscapes, and the ability to detect biomarker deviations across large cohorts to identify small subsets of individuals with rare conditions. This cannot occur solely during routine clinical care, but it requires systematic biobanking. Therefore, pediatric biobanks, especially repositories that store CSF under standardized conditions, will be invaluable for future discoveries. We can only treat rare conditions if we can first identify them, and that identification depends on infrastructure. For example, the next rare disease may not be found in a genome, but it may be found in a frozen CSF vial.
