Prognostic and therapeutic monitoring value of plasma and urinary cytokine profile in primary membranous nephropathy: the STARMEN trial cohort

Abstract

Background: Primary membranous nephropathy (PMN) is usually caused by anti-phospholipase A2 receptor (PLA2R) autoantibodies. There are different therapeutic options according to baseline risk. Novel biomarkers are needed to optimize risk stratification and predict and monitor the response to therapy, as proteinuria responses may be delayed. We hypothesized that plasma or urinary cytokines may provide insights into the course and response to therapy in PMN.

Methods: Overall, 192 data points from 34 participants in the STARMEN trial (NCT01955187), randomized to tacrolimus-rituximab (TAC-RTX) or corticosteroids-cyclophosphamide (GC-CYC), were analysed for plasma and urine cytokines using a highly sensitive chemiluminescence immunoassay providing a high-throughput multiplex analysis.

Results: Baseline (pretreatment) urinary C-X-C motif chemokine ligand 13 (CXCL13) predicted the therapeutic response to TAC-RTX. Cytokine levels evolved over the course of therapy. The levels of nine plasma and six urinary cytokines correlated with analytical parameters of kidney damage and disease activity, such as proteinuria, estimated glomerular filtration rate and circulating anti-PLA2R levels. The correlation with these parameters was most consistent for plasma and urinary growth differentiation factor 15 (GDF15), plasma tumour necrosis factor α and urinary TNF-like weak inducer of apoptosis. Decreasing plasma GDF15 levels were associated with response to GC-CYC. Four clusters of cytokines were associated with different stages of response to therapy in the full cohort, with the less inflammatory cluster associated with remission.

Conclusion: PMN displayed characteristic plasma and urine cytokine patterns that evolved over time as patients responded to therapy. Baseline urinary CXCL13 concentration could be a prognostic marker of response to TAC-RTX.

Keywords: anti-PLA2R; clinical trial; inflammation; membranous nephropathy; treatment.

Graphical Abstract

Figure 1:

Baseline and follow-up biomarker levels and treatment outcomes. (A–F) TAC-RTX arm; (G–L) GC-CYC arm. (A, D, G and J) Baseline values for responders and non-responders for selected cytokines (urinary CXCL13, serum IL-10, serum GDF15 and serum IL-5). Results are shown as individual data points (circles) with medians (bars) and interquartile ranges (box). Whiskers extend to the minimum and maximum values excluding outliers. P-values were calculated using the Mann–Whitney U test. (B, C, E, F, H, I, K, L) Change in biomarker levels during treatment. Squares along the line indicate patient status at each point. Non-responder: non-response status at 9 months (B, E) and 6 months (H, K); responder limited, partial or complete remission status at 9 months (C, F) and 6 months (I, L). U: urine biomarker levels. Outcome defined at the end of treatment.

Figure 2:

Percentage of patients in each treatment arm with urinary CXCL13 levels above the cut-off established for the TAC-RTX arm. All participants are shown for both treatment arms, including responders and non-responders.

Figure 3:

Unsupervised hierarchical clustering analysis of inflammatory cytokine levels and association with clinical features. The cluster analysis disclosed four global cytokine clusters. (A) Unsupervised hierarchical clustering analysis of inflammatory cytokine levels. Cluster analysis was performed by a complete linkage based on city block distance. Patient proteinuria status at the time of sample collection and next visit are indicated at the bottom of the cluster. (B) Immunological remission: white rectangle indicates immunological remission (negative anti-PLA2R). (C) Proteinuria remission status. The four identified clusters are indicated at the bottom of each graph, in the horizontal axis. The statistical significances between the clusters were calculated using the Fisher's exact test. U: urine biomarker levels. Outcome defined as patient condition at the time of sample collection.

Figure 4:

Transition of cytokine clusters. (A) Cluster trajectories between two consecutive sampling points. The number represents the number of sample pairs with that trajectory. Trajectories with four or more sample pairs are represented in blue and those with only one in grey. (B) Cluster trajectories from baseline to the latest sampling point. The number of patients at baseline is indicated next to each cluster in red characters.

Figure 5:

Treatment-specific monitoring cytokine clusters for disease activity. TAC-RTX group (A, C, D) or GC-CYC group (B, E, F). (A, B) Unsupervised hierarchical clustering analysis of cytokines, which were selected based on volcano plots, yielded two clusters for disease monitoring in the TAC-RTX arm (M-TAC-RTX clusters 1 and 2: labels coloured purple in panels C and D) and three in the GC-CYC arm (M-GC-CYC clusters 1, 2 and 3: labels coloured in burgundy E and F). The cluster analysis was performed by average linkage based on Euclidean distance. Patient proteinuria status at the sample collection and next visit and status of immunological remission are indicated. (C, E) Immunological status. (D, F) Proteinuria remission status in the clusters. Statistical significances calculated using the Steel–Dwass test or Fisher's exact test. U: urine biomarker levels. Outcome is defined as the patient condition at the time of sampling.

Figure 6:

Precision approach to monitoring and treatment of membranous nephropathy: role of cytokine profiles. (A) Baseline and periodic monitoring of cytokine profiles could contribute to the early identification of patients who will respond to specific IST regimens. For responding patients, the baseline inflammatory profile of cytokines (cluster 1) will evolve to less inflammatory (clusters 2 and 3) or non-inflammatory (cluster 4) profiles. For patients predicted to not respond or not responding in periodic monitoring, alternative therapeutic approaches should be considered. This additional information may contribute to shortening the times for therapeutic decision-making. The figure is a conceptual representation of data from Fig. 3. NIST: non-immunosuppressive therapy (e.g. renin–angiotensin system blockers or other antiproteinuric therapies). (B) Context of use 1. After deciding to prescribe IST because of a KDIGO 2021 guideline high risk, urinary CXCL13 may be incorporated into decision-making about the optimal IST regimen by adding information of the chances of response to TAC-RTX. In this regard, the KDIGO 2021 guideline states that ‘Rituximab and CNIs have fewer and milder side effects than cyclophosphamide. Therefore, most physicians and patients will prefer initial treatment with rituximab or CNIs over treatment with cyclophosphamide.’ Thus, urinary CXCL13 may be useful in clinical decision-making at this point to allow gauging the probability of response to TAC-RTX. (C) Context of use 2. Conceptual representation of the potential added value of cytokine clusters when integrated with anti-PLA2R titres and proteinuria values in the follow-up of the response to IST. The precise role of cytokine clusters in follow-up is less well defined at this point than the potential role of baseline urinary CXCL13, but baseline evaluation and follow-up of cytokine clusters may provide added information that may be integrated with that provided by anti-PLA2R levels (in those patients who had positive baseline values) and proteinuria responses for a more holistic follow-up, which may signal earlier a positive response to IST than proteinuria alone. In the future, following further validation in anti-PLAR2-negative PMN, the simultaneous lack of response of proteinuria and cytokines may lead to reconsideration of the IST regimen earlier than current practice. Each column of coloured squares in panel B represents a potential follow-up combination of biomarkers and the putative interpretation. Both the contexts of use represented in panels B and C require external validation.