CXCR4 maps to dermal immune cells in psoriasis study

CXCR4-positive staining increased in both the epidermis and dermis of imiquimod-induced psoriasis-like mouse skin, alongside higher CXCR4 mRNA, in a 2026 study that paired public human single-cell and spatial datasets with experimental validation.

Ru-nan Fang and Jian-hong Li led the team, with Y. Zhou, D. Ge and Y.-R. Pan. They published “Single-cell RNA sequencing and spatial transcriptomics characterize CXCR4 expression and associated putative cell–cell communication in psoriasis” in Archives of Dermatological Research, volume 318, issue 1, as article 529. The DOI is 10.1007/s00403-026-05009-4.

The human arm reanalyzed publicly available single-cell RNA sequencing and spatial transcriptomics datasets from psoriatic and healthy-control skin samples deposited in the Gene Expression Omnibus database. After sample-level quality control, the group integrated the datasets with Harmony, then performed cell clustering, cell-type annotation and CellChat-based ligand-receptor inference, with particular focus on the CXCL signaling pathway.

Fang and Li characterized the work as “an integrative characterization of CXCR4’s cellular distribution, spatial localization, and associated putative intercellular communication patterns in psoriasis.” The framing matters because the study identifies a plausible signaling map, not a proven functional circuit.

Human-skin findings

The single-cell analysis identified 15 major cell types in human skin. A sample-level comparison found no significant difference in the proportions of those major cell types between psoriasis and control groups after Benjamini-Hochberg correction for multiple testing.

The result matters because the CXCR4 pattern did not reflect a measurable shift in the overall cellular makeup of the analyzed samples. Instead, the investigators located the change in expression within cell populations.

CXCR4 transcripts appeared mainly in immune cells, including T cells, dendritic cells and macrophages, with limited signal in keratinocytes. Its ligand, CXCL12, showed a complementary distribution: transcript signals associated mainly with fibroblast clusters, although other populations also carried the chemokine.

CellChat-based inference identified CXCL12-CXCR4 as one candidate ligand-receptor pair with a relatively high inferred contribution to the CXCL pathway in the psoriasis group. Spatial arrangement suggested putative links between fibroblast-associated CXCL12 and multiple immune-cell populations expressing CXCR4.

Inference, not demonstration. That distinction matters.

Spatial transcriptomics added tissue-level resolution. CXCR4 transcript signals concentrated in the dermis of psoriatic skin, beneath the epidermis, where inflammatory immune infiltrates accumulate in the disease.

Experimental validation in mice

To test the computational findings, the team used an imiquimod-induced psoriasis-like mouse model. Immunohistochemistry showed increased CXCR4-positive staining areas in both epidermal and dermal compartments, while reverse transcription-quantitative polymerase chain reaction found increased CXCR4 mRNA expression.

The ligand result was more complicated. Double immunofluorescence staining for CXCL12 and Vimentin found increased tissue-level CXCL12 fluorescence in psoriasis-like skin, but the Manders overlap coefficient did not increase. The protein-level colocalization expected between CXCL12 and the stromal marker therefore went unconfirmed.

Fang and Li and their coauthors also emphasized that the precise cellular sources and functional significance of the signaling axis “require further validation.” That caution applies directly to the proposed fibroblast-immune connection: transcriptomic inference and spatial proximity cannot establish ligand binding, downstream signaling or a causal role in psoriasis.

What does this change for clinicians?

For clinicians, the change is conceptual rather than procedural. The study supplies a detailed spatial and cellular reference for CXCR4 and a testable candidate stromal-immune communication axis. It does not demonstrate functional ligand-receptor interactions or test a clinical intervention, so the paper offers no basis for changing treatment.

Several boundaries sharpen that conclusion. The human work used public datasets rather than newly collected patient samples, and the imiquimod model mimics key features of psoriasis without being human disease. The inferred ligand-receptor links come from transcriptomic computation, not direct demonstration of signaling.

Future functional studies now need to test whether the proposed CXCL12-CXCR4 stromal-immune exchange operates in human lesions and identify the dominant cellular source of CXCL12; that evidence, rather than expression maps alone, would give the axis translational weight.