Lupus Patients Genomically Stratified to Explain Treatment Responses Save

Systemic lupus is a clinically heterogeneous disorder, unified by requisite clinical features and exuberant humoral response to unknown triggers.  While the diagnosis is easy, the disease course and management can be complicated and challenging.  

There has been significant frustration with newer targeted therapies (e.g., monoclonal antibodies against B cells or IFNα, etc.) that have either failed or yielded modest benefits in a disorder that demands more.  One approach to these disappointing responses would be to establish if lupus is one or several disorders that can be distinguished on clinical, immune or molecular grounds.

In the current online edition of Cell, Pascual and coworkers at the Baylor Research Institute have published the results of their detailed clinical and genomic analyses of 158 pediatric SLE patients, wherein they show that transcriptome analyses enables the stratification of 7 “types” of lupus, with each having an identifiable pathology that may be selectively targeted with personalized immunosuppression.

These researchers and others have previously shown the importance of plasmacytoid dendritic cells, IFNα, neutrophil activity and netosis on the pathogenesis of lupus.  In these investigations, they sought to identify immune correlates to lupus activity by transcriptionally profiling 924 serial blood samples from 158 pediatric lupus patients, while at the same time collecting demographics, treatments, disease activity measures, SLEDAI components, and renal biopsy/nephritis classes.

At each time point they identified a whole blood signature or finger print by transcriptome analyses. Using unsupervised hierarchical clustering of the 15,386 transcripts, they found their previously reported IFNα signature in 85% of 924 SLE samples. They then looked for differentially expressed transcripts between lupus patients and controls, and found a highly reproducible SLE “fingerprint” with gene overexpression indicating IFN response, neutrophil, inflammation, cell cycle, erythropoiesis, and histone gene activity. Conversely, there was underexpression of genes linked to NK cell/cytotoxicity, lymphoid lineage, B cells, T cells, and protein synthesis.  

Other findings detailed in this paper include:

• By examining correlates with disease activity they showed the plasmablast signature was most correlated with disease activity and this was confirmed by subset quantification by cell sorting.
• African-Americans showed enrichment of plasmablast, cell cycle, and erythropoiesis modules that were associated with higher SLEDAI, ESR and C3 levels. (Conversely, Hispanics and Caucasians showed enrichment of neutrophil, myeloid lineage, and inflammation-related modules)
• Plasmablast signatures were decreased most by mycophenolate and intravenous cyclophosphamide; the same therapies were also the most effective in decreasing the IFN response.
• While IFN, plasmablast, and B cell gene expression was correlated with SLEDAI activity;  neutrophil, myeloid, and inflammation modules were strongly expressed in patients with renal involvement and nephritis.  The authors believe that an initial IFN/plasmablast/B cell disease initiation is magnified and damaging (especially to the kidneys) with systemic inflammation represented by myeloid and neutrophil activity.
• Individual patients clinical traits were followed over time and correlated with weighted gene co-expression network analyses to identify modules and transcripts that best correlated with disease status.  Clusters of lupus patients were identified according to their SLEDAI and clinical status and gene expression. (Stronger gene expression indicated by greener dots in the figure).  
• They identified seven molecularly unique SLE patient groups (PG1–7), each displaying a specific combination of five immune signatures correlating with the SLEDAI. Hence these 7 groups (P1-P7) are base on gene expression for erythropoiesis (ER), IFN response, myeloid lineage/neutrophils (ML), plasmablasts (PB), and lymphoid lineage (LL). For instance, in PG3 only plasmablast modules correlated with the SLEDAI and in PG5 only IFN and myeloid lineage modules correlated with SLEDAI.

Overall, these investigators were able to identify immune signatures correlating with SLEDAI across different genotypes and phenotypes and that such an approach may be useful in the development of customized treatment strategies. The authors believe that the molecular heterogeneity identified in their pediatric SLE patients provides an explanation for the past failures in several lupus clinical trials.