The lungs are constantly exposed to the external environment and a myriad of antigenic challenges within the air. Chronic exposure to allergens and other airborne antigens can result in the formation of lymphocyte aggregates in the lung, which can harbor ectopic germinal centers (GCs). After allergen exposure, GCs that form in the lung are much smaller and less densely packed with B cells than lymph node GCs. Despite this, ectopic lung GCs support somatic hypermutation and affinity-based maturation as in lymph node GCs, and export memory B cells (MBCs) directly into the lung tissue. This demonstrates that the lung can locally diversify B cell responses and supports the generation of tissue MBC populations in situ.
T follicular helper (T FH ) cells are essential for germinal center (GC) formation and long-lived humoral immunity. Here, we used an Il21 fate mapping (Il21-fm) strategy to remove T FH cells from established GCs to disentangle their function from that of other T cells in a temporal manner. We confirm their role in driving proliferation and positive selection of GC B cells but show that GCs can survive the transient absence of T FH cells. After ablation of T FH cells, both the GC response and affinity maturation recover via ingress of new T FH cells. Despite recovery of the B cell response, T FH cell numbers are never fully restored. This feeds through into the T cell memory response, resulting in diminished recall GC responses. This work shows an unappreciated resilience of primary GC responses to perturbations in T FH cells and demonstrates critical memory T cell generation during this phase.
24 February 2026 in European Journal of Immunology
We developed a 50-plex imaging protocol using the MACSima platform to characterize the microarchitecture of germinal centers in secondary lymphoid tissues. This workflow combines tissue processing, automated cyclic imaging, image preprocessing, and a new analysis method that detects morphological changes in irregularly shaped stromal cells.
Summary: Quantitative assessment of brain histology is often constrained by predefined feature sets and labor-intensive manual annotations. To overcome these limitations, we employed unsupervised deep learning to automatically extract and quantify tissue organizational patterns from myelin-stained rat brain sections without the need for prior labeling. We evaluated nonlinear convolutional autoencoders (AEs) against linear principal component analysis (PCA) for feature representation, followed by clustering with Gaussian mixture models. Compared to PCA, AEs better preserved fine axonal architecture and produced more consistent and interpretable tissue clusters across hierarchical levels. The resulting clusters revealed anatomically meaningful tissue organization, including different white matter densities and grey matter subregions. When applied to tissue from sham and mild traumatic brain injury animals, AE-derived features also captured pathology-related alterations, such as white matter loss and injury-specific microstructural changes. These findings demonstrate that unsupervised deep learning can automatically characterize tissue organization at multiple scales and detect pathological changes, offering a scalable, unbiased approach to quantitative neuropathology.