Improving Risk Stratification in Lupus Nephritis: The Role of Urinary Biomarkers

Lupus nephritis (LN) is one of the most severe complications of systemic lupus erythematosus, with kidney failure as a leading cause of morbidity. However, despite advances in immunosuppression, predicting which patients will experience long-term renal decline is clinically challenging.

A recent analysis presented at the 2025 American College of Rheumatology (ACR) Annual Meeting targeted this gap by exploring long-term predictors of kidney function loss. Here’s an overview of the analysis and what it found.

Limits of Proteinuria

Historically, proteinuria has served as a surrogate for treatment response in LN, with a sub-0.5 urine protein-to-creatinine ratio (UPCR) at one year broadly interpreted as a marker of disease control.

Repeat biopsy data, however, have exposed the limits of this approach. Roughly half of proteinuric responders still exhibit histologic evidence of intrarenal activity, and up to 30% go on to experience progressive renal insufficiency despite meeting clinical benchmarks.

This disconnect has reinforced the need for biomarkers that can more accurately predict long-term outcomes.

Study Design

The analysis presented at the 2025 ACR Annual Meeting followed 170 individuals with lupus nephritis from the Accelerating Medicines Partnership cohort for up to 7.8 years, with a median follow-up duration of 4.9 years.

Participants underwent urine sampling at the time of kidney biopsy and again at months 3, 6, and 12. Urine samples were analyzed using a high-throughput platform to quantify 1,200 proteins.

The primary outcome was defined as a sustained ≥40% decline in estimated glomerular filtration rate (eGFR) from baseline or the development of end-stage kidney disease.

Treatment throughout the study period remained at the discretion of managing clinicians, allowing for real-world heterogeneity in therapeutic exposure.

To identify biomarkers associated with future kidney function loss, researchers applied time-to-event models, evaluating protein levels at each sampling point in relation to long-term renal outcomes. A Random Survival Forest model was then trained to predict eGFR loss using a panel of 11 urinary proteins and validated on a 20% hold-out test set. Additionally, single-cell RNA sequencing and spatial transcriptomics were used to map the cellular origin of identified biomarkers within kidney tissue, linking urinary signals to intrarenal pathophysiology.

Tenascin C as a Key Predictor

Looking at the results, Tenascin C, a matricellular glycoprotein involved in tissue remodeling, emerged by month 3 as the most robust urinary predictor of future eGFR loss, with a hazard ratio of 4.6. Elevated levels were sustained through month 12 in patients who eventually experienced significant kidney function decline.

These findings point to Tenascin C’s reliability as a biomarker not only for early detection but also for long-term prognostication.

Spatial transcriptomics localized Tenascin C expression specifically to interstitial myofibroblasts within LN-affected kidneys, supporting its role in fibrotic remodeling rather than transient immune activity. This cellular mapping reinforced the notion that Tenascin C reflects irreversible structural changes not captured by traditional markers like proteinuria.

Proteomic Risk Score

Additionally, the 11-protein classifier, developed from longitudinal urinary data, achieved an area under the curve (AUC) of 0.91 at 48 months in predicting eGFR loss.

Its performance remained consistent across clinical subgroups, including those stratified by standard proteinuria thresholds. Notably, the model identified high-risk patients within the proteinuric responder population and low-risk patients among those not meeting UPCR goals, revealing biologic heterogeneity that standard markers fail to capture.

These findings challenge the current paradigm of response assessment in LN, where reliance on UPCR alone may result in both overtreatment and under-recognition of high-risk disease.

Inflammatory and Fibrotic Markers

Beyond Tenascin C, the study highlighted an array of biomarkers linked to fibrosis and chronic inflammation, including CD163, CD206, FABP4, IL-6, and IGFBP-6. Elevated CD163 and CD206 levels, markers of M2 macrophage polarization, were particularly notable in high-risk individuals, aligning with a pathophysiologic model of persistent intrarenal immune signaling and incomplete resolution of inflammation.

This biomarker profile underscores the importance of evaluating tubulointerstitial changes, which often remain undetected by proteinuria-centric monitoring yet drive long-term loss of renal function.

Clinical Translation and Future Stratification

The integration of urinary proteomics and machine learning represents a practical advance in lupus nephritis care. Rather than depending on glomerular outputs like proteinuria alone, clinicians may soon have access to a biologically grounded risk score that reflects the state of fibrotic and inflammatory activity within the kidney.

Such a shift would support more nuanced treatment decisions, including the intensification or de-escalation of immunosuppression based on a patient's individual risk trajectory.

These findings signal a movement toward biologically stratified care in lupus nephritis—where treatment targets are aligned with disease biology rather than proxy metrics—and lay the groundwork for trials that evaluate outcomes based on intrarenal molecular activity rather than delayed clinical endpoints.

Reference:
Lee C, Taghavi S, Zhang S, et al. Urinary tenascin C predicts kidney function loss in lupus nephritis. Arthritis Rheumatol. 2025; 77(suppl 9). Accessed October 3, 2025. https://acrabstracts.org/abstract/urinary-tenascin-c-predicts-kidney-function-loss-in-lupus-nephritis/