The research team employed imaging mass cytometry (IMC), a technology that uses metal-conjugated antibodies and mass spectrometry to detect dozens of proteins simultaneously on a single tissue section. Where conventional immunofluorescence is limited to roughly four or five markers because of spectral overlap, IMC allowed the researchers to build a panel of 32 antibodies covering 19 cell population markers, 3 activation markers and 10 ECM components. Critically, the team went beyond standard cellular analysis. Because matrix molecules sit outside cells, traditional cell-mask-based approaches fail to capture them. To solve this, the researchers developed DeepThresh, a novel deep-learning thresholding algorithm modelled on a U-Net architecture with a ResNet152 encoder. Trained on expert-annotated images, DeepThresh generated robust binary masks of ECM staining across all tissue regions, overcoming the variability that plagued classical thresholding methods such as Otsu’s algorithm.

With matrix staining segmented, the pipeline calculated Euclidean distances from every cell to each ECM component, and these distance profiles were clustered to define twelve distinct lung matrix environments. This is the study’s central methodological innovation: for the first time, every cell in the image carries both its canonical identity and a quantitative description of its immediate matrix surroundings. The approach was applied to lung sections from BALB/c and C57BL/6 mice, two widely used inbred strains, following chronic twice-weekly intranasal exposure to a cocktail of house dust mite, ragweed and Aspergillus extracts, a model known as DRA that reproduces key features of human allergic airway pathology, including collagen deposition, hyaluronan accumulation and steroid-resistant recruitment of eosinophils and neutrophils.

The resulting atlas revealed that the healthy lung is divided into two major matrix compartments. The alveolar parenchyma, the delicate gas-exchange region, is characterised by proximity to heparan sulphate, laminin gamma-1, type-IV collagen and fibrinogen. In contrast, the adventitial cuff, the connective tissue sheath surrounding airways and blood vessels, is enriched for type-I, type-III and type-VI collagens together with hyaluronan. During allergic inflammation, these spatially constrained environments changed in strikingly region-specific ways, and the two mouse strains responded differently, underscoring how genetic background shapes not just the magnitude of inflammation but its anatomical architecture.

One of the most visually dramatic findings concerned the adventitial cuff. In allergic animals, infiltrating immune cells accumulated in discrete, contiguous patches around the pulmonary artery and airway-adjacent vessels, and these patches were significantly more numerous and larger in BALB/c mice than in C57BL/6 mice. Three-dimensional precision-cut lung slices stained with fluorescent antibodies confirmed the IMC findings and revealed internal structure within the cuff: the region adjacent to the blood vessel was dominated by CD11b-positive CD64-positive monocyte-derived macrophages, while a more distal, densely nucleated zone resembled induced bronchus-associated lymphoid tissue (iBALT), rich in B cells and largely excluding macrophages. These iBALT-like structures formed only in BALB/c animals, suggesting that strain-specific type-2 immune skewing may provide the signals needed to seed organised lymphoid tissue within the inflamed lung.

The matrix itself shifted around these inflammatory patches in a coordinated fashion. Allergic BALB/c mice showed expansion of sparse ‘inflammatory zone’ matrix environments with increased distances to type-I, type-III and type-VI collagens, but reduced distances to laminin gamma-1 and hyaluronan. This represents a shift from a densely collagenous matrix to one dominated by laminin and hyaluronan, molecules known to regulate cell migration through tissue. Because hyaluronan-rich matrices can facilitate or direct immune cell infiltration, this remodelling may actively sculpt the localisation of B cell activation and myeloid accumulation within the cuff, offering a mechanistic hypothesis for why inflammation aggregates in these specific niches.

The alveolar parenchyma told a different story. Following allergen challenge, the ratio of alveolar type-I (ATI) to alveolar type-II (ATII) epithelial cells fell from roughly 3:1 in healthy controls to approximately 1:1 in allergic animals of both strains, a signature of injury and repair, since ATII cells are known to proliferate and replace damaged ATI cells. Immunofluorescent validation using RAGE as an ATI marker and surfactant protein C as an ATII marker confirmed the relative loss of the ATI compartment. Spatially, the alveolar region subdivided into resting alveoli enriched for ATI cells and basement membrane components, and activated alveolar regions enriched for ATII cells, alveolar macrophages and proximity to hyaluronan and chondroitin sulphate. These glycosaminoglycans are known regulators of immune cell migration, and chondroitinase treatment in fibrosis models reduces macrophage numbers, suggesting that this matrix shift may directly control macrophage retention in the damaged alveolus.

Perhaps the most translational insight concerns the airway subepithelial space, the region beneath the airway epithelium that undergoes extensive remodelling in human asthma. The study found that this region expanded significantly during allergic challenge in both strains and became a hub of immune-stromal interaction. Two fibroblast populations were identified there: alpha-smooth muscle actin-positive stromal cells, present at steady state, and S100a4-positive fibroblasts, which expanded dramatically during allergy. Neighbourhood analysis showed that S100a4-positive fibroblasts and CD11b-positive immune cells, including MerTK-positive macrophages, came into close contact specifically during allergic inflammation, and both stromal populations overlayed type-I and type-III collagen deposition. Because macrophages and fibroblasts are known to regulate each other reciprocally, and because type-VI collagen in this niche can promote fibroblast survival and migration, the authors propose that this immune-stromal network constitutes a regulatory circuit controlling ECM deposition during allergic airway pathology.

The study’s design choices also matter for future work. The entire pipeline was developed on formalin-fixed paraffin-embedded tissue, the predominant preservation method in human biobanks, meaning the approach can be translated directly to archived patient samples. All datasets and code, including the DeepThresh and MatrixIMC pipelines, have been made publicly available. By integrating cellular and matrix analysis within a single tissue section, the framework captures a dimension of tissue biology that single-cell RNA sequencing alone cannot, since transcriptomic profiles of matrix genes are often poor surrogates for the post-translationally modified proteins actually present in tissue. The authors caution that future work must clarify the temporal sequence of these changes and establish causality, but the atlas already generates testable hypotheses about how targeting specific matrix components or the macrophage-fibroblast interactions within remodelling niches could intervene in allergic airway disease. For a field that has long catalogued inflammation cell by cell, this study makes a compelling case that where a cell sits, and what it sits on, may matter as much as what the cell is.

Subject of Research: Spatial mapping of extracellular matrix and cellular environments during allergic airway inflammation using imaging mass cytometry in mouse models

Article Title: Extracellular matrix phenotyping by imaging mass cytometry defines distinct cellular matrix environments associated with allergic airway inflammation

Article References: Parkinson, J. E., Bryant, M., Ghafoor, M., Dodd, R. J., Tompkins, H. E., Fergie, M., Burgess, M. O., Rattray, M., & Sutherland, T. E. (2026). Extracellular matrix phenotyping by imaging mass cytometry defines distinct cellular matrix environments associated with allergic airway inflammation. Molecular Systems Biology. https://doi.org/10.1038/s44320-026-00234-5

Image Credits: AI Generated

DOI: 10.1038/s44320-026-00234-5

Keywords: imaging mass cytometry, extracellular matrix, allergic airway inflammation, asthma, airway remodelling, lung immunology, hyaluronan, collagen, fibroblasts, macrophages, spatial biology, mouse models