Background: Axial spondyloarthritis (axSpA) and psoriatic arthritis (PsA) are chronic inflammatory diseases within the spondyloarthritis (SpA) spectrum that share genetic risk factors and inflammatory pathways, yet differ substantially in clinical manifestations, disease course, and therapeutic responses [1]. Although immune dysregulation is central to both diseases, direct high-dimensional comparisons of immune cell composition, activation states, and soluble inflammatory mediators between axSpA and PsA remain scarce. Improved understanding of disease-specific immunophenotypes may provide mechanistic insights into disease heterogeneity and support biomarker-driven patient stratification.

Objectives: To systematically compare immune cell subsets, activation and exhaustion markers, and circulating cytokine and inflammatory mediator profiles in axSpA and PsA in order to identify immune signatures that discriminate between these diseases.

Methods: Peripheral blood mononuclear cells (PBMCs) were obtained from 179 patients (88 axSpA, 91 PsA) and 49 healthy donors. High-dimensional immune profiling was performed using spectral flow cytometry, encompassing 230 million acquired events and 55 million high-quality immune cells. Comprehensive analyses included major innate and adaptive immune subsets, B- and T-cell differentiation states, activation and exhaustion markers, and antigen-presenting cell populations. In parallel, serum cytokines, chemokines, and soluble inflammatory mediators were quantified. Group comparisons were performed to define disease-specific immune features. Multivariable regularized regression models were applied to integrate immune parameters and identify discriminative immune signatures, with performance assessed by cross-validation

Results: Deep immunophenotyping revealed overlapping yet distinct immune profiles in axSpA and PsA. Shared features included a shift toward pro-inflammatory classical monocytes (axSpA p=0.0007, PsA p=0.0229) and increased frequencies of Th1 and T follicular helper cells (all p<0.0001) compared with healthy donors. In addition, Th1-like regulatory T cells were expanded in both disease entities (axSpA p=0.0004, PsA p=0.0091).

AxSpA was characterized by a pronounced expansion of innate-like lymphocyte populations, including double-negative T cells and γδ T cells (p=0.0051, p=0.0017), alongside increased plasmablasts (p=0.0017) and CD21 ^- /low B cells (p=0.0276). AxSpA patients further exhibited elevated Th17 cells (p=0.0091) with a concomitant reduction of Th2-like cells (p=0.0019). Across multiple immune subsets, axSpA showed broadly increased expression of activation markers, particularly CD80, CD86, CD95, and HLA-DR (p=0.0242-<0.001), accompanied by features consistent with immune senescence/exhaustion in T cells.

In contrast, PsA displayed increased frequencies of dendritic cells (p=0.0269), expansion of IgM ⁺ IgD ⁻ memory B cells (p=0.0450), and higher proportions of central memory CD4 T cells (p=0.0371), indicative of stronger adaptive immune engagement.

Serum cytokine profiling further discriminated disease entities. PsA was associated with elevated levels of sIL-2R (p=0.0011), IL-17A (p=0.0380), sTNFR1 (p=0.0016), and mediators linked to vascular inflammation. Conversely, axSpA exhibited higher concentrations of inflammasome-related cytokines IL-1β (p=0.0002) and IL-18 (p=0.0265), innate cytokines such as GM-CSF (p=0.0017) and IFN-α2 (p=0.0002), fibroblast-associated IL-11 (p=0.0005), and increased IL-15 (p<0.0001) and IL-12p40 (p=0.0018) compared to PsA patients.

Integration of cellular and soluble immune features using elastic-net regularized logistic regression robustly distinguished PsA from axSpA (cross-validated AUC ≈0.90). A stable immune signature dominated by activation-associated markers, particularly CD80, achieved strong discrimination at the individual patient level (AUC≈0.94), highlighting distinct immunopathological programs underlying axSpA and PsA.

Conclusions: Despite shared inflammatory pathways, axSpA and PsA exhibit distinct immune architectures. AxSpA is characterized by broader innate immune activation and heightened T-cell activation/exhaustion, whereas PsA displays a more adaptive, memory-oriented immune profile. These findings provide mechanistic insight into disease heterogeneity within the SpA spectrum and support the potential of immune-based biomarkers for disease classification and patient stratification.

REFERENCES: [1] Dougados M, Baeten D. Spondyloarthritis. Lancet. 2011;377(9783):2127-37.

Acknowledgments: NIL.

Disclosure of Interests: Natalie Frede Lilly, Nils Craig-Müller: None declared, Silke Kohrt: None declared, Marei-Theresa Bott: None declared, Ulrich Salzer: None declared, Ana Venhoff: None declared, Jens Thiel: None declared, Julia Frede: None declared, Reinhard Voll: None declared, Marta Rizzi: None declared, Raquel Lorenzetti: None declared, Nils Venhoff UCB, Janssen, Novartis, AbbVie, AstraZeneca, Boehringer Ingelheim, Bristol Myers Squibb, Chugai, CSL Vifor, GSK, Novartis, Pfizer, Sanofi, Janssen., AbbVie, AstraZeneca, Bristol Myers Squibb, Chugai, CSL Vifor, GSK, Novartis, Pfizer, Sanofi, Janssen, AbbVie, Novartis, Pfizer.