Chin, J. L. | Search bioRxiv

bioRxiv2025

Tensor-Based Integration of Time-Series Measurements Reveals Relationships Between Underlying Disease, Early Injury, and CD4+ Polarization in Liver Transplantation

For patients with end-stage liver failure, liver transplantation (LT) remains the standard-of-care, though five-year allograft survival rates remain below 80%. Liver ischemia reperfusion injuries (LIRI) arise during transplant and contribute to allograft dysfunction. While many drivers of LIRI have been well-characterized, the relationships between LIRI and later immunological signatures remain poorly understood, possibly because of limited integrated studies examining immunological signatures across the LT process. Furthermore, the role of underlying disease in LT remains poorly understood as it is unknown if later immunological signatures are universal or etiology dependent. To better examine etiology- and LIRI-driven immunological mechanisms across the LT lifetime, we collected cytokine and liver function test (LFT) measurements from LT donors and recipients at multiple pre- and post-operative timepoints. By developing a tensor-based factorization method to integrate these longitudinal measurements, we reduced the datasets to four immunological signatures that manifested across cytokines, LFTs, and time. Correlative analyses revealed associations between these integrated immunological signatures with etiology and LIRI. Later integration of allograft survival into this factorization method reduced these four signatures to two that were predictive of five-year allograft survival. Examination of these signatures prognostic of allograft survival found that CD4+ hyperpolarization towards Th1 or Th2 subtypes impaired growth factor expression and led to increased allograft loss risk. Clinical correlates revealed that donor and recipient age alongside etiology and donor liver health modulate this balance of CD4+ polarization in LT. Collectively, these results link immunological mechanisms to etiology and LIRI offering immediately actionable clinical insights alongside etiology- and donor-specific therapeutic targets for improving LT outcomes.

immunology↗

bioRxiv2022

Cytokine-expression patterns reveal coordinated immunological programs associated with persistent MRSA bacteremia

Methicillin-resistant Staphylococcus aureus (MRSA) bacteremia is a common, life-threatening infection that imposes up to 30% mortality even when appropriate therapy is used. Despite in vitro efficacy, antibiotics often fail to resolve the infection in vivo, resulting in persistent MRSA bacteremia. Recently, several genetic, epigenetic, and proteomic correlates of persistent outcomes have been identified. However, the extent to which single variables or composite patterns operate as independent predictors of outcome or reflect shared underlying mechanisms of persistence is unknown. To explore this question, we employed a tensor-based integration of host transcriptional and proteomic data across a well-characterized cohort of patients with persistent and resolving MRSA bacteremia outcomes. Tensor-based data integration yielded high correlative accuracy with persistence and revealed immunologic signatures shared across both the transcriptomic and proteomic datasets. We find that elevated proliferation of mature granulocytes associates with resolving bacteremia outcomes. In contrast, patients with persistent bacteremia heterogeneously exhibit correlates of granulocyte dysfunction or immature granulocyte proliferation. Collectively, these results suggest that transcriptional and proteomic correlates of persistent versus resolving bacteremia outcomes are complex and may not be disclosed by conventional modeling. However, a tensor-based integration approach can help to reveal consensus molecular mechanisms in an interpretable manner. Significance StatementWhile antibacterial therapies effectively resolve MRSA in vitro, these treatments often fail to clear MRSA bacteremia in vivo, suggesting that host-pathogen interactions are essential to persistent MRSA bacteremia. Recent studies have identified genetic, transcriptomic, and proteomic determinants of MRSA persistence. These determinants independently, however, provide insufficient mechanistic insight and it is unclear if they indicate unique or overlapping persistence mechanisms. Here, we use tensor-based decomposition to jointly analyze cytokine and transcriptomic measurements from patients with MRSA bacteremia. Results indicate that persistence mechanisms integrated across biological modalities reflect diverging mechanisms of persistent bacteremia. Ultimately, these results may help to identify future therapeutic targets for treating persistent MRSA bacteremia.

immunology↗