Improved risk stratification and prognosis prediction in sepsis is a critical unmet need. Clinical severity scores and available assays such as blood lactate reflect global illness severity with suboptimal performance, and do not specifically reveal the underlying dysregulation of sepsis. Here, we present prognostic models for 30-day mortality generated independently by three scientific groups by using 12 discovery cohorts containing transcriptomic data collected from primarily community-onset sepsis patients. Predictive performance is validated in five cohorts of community-onset sepsis patients in which the models show summary AUROCs ranging from 0.765–0.89. Similar performance is observed in four cohorts of hospital-acquired sepsis. Combining the new gene-expression-based prognostic models with prior clinical severity scores leads to significant improvement in prediction of 30-day mortality as measured via AUROC and net reclassification improvement index These models provide an opportunity to develop molecular bedside tests that may improve risk stratification and mortality prediction in patients with sepsis.

Rationale: The plasma lipidome has the potential to reflect many facets of the host status during severe infection. Previous work is limited to specific lipid groups or was focused on lipids as prognosticators. Objectives: To map the plasma lipidome during sepsis due to community-acquired pneumonia (CAP) and determine the disease specificity and associations with clinical features. Methods: We analyzed 1,833 lipid species across 33 classes in 169 patients admitted to the ICU with sepsis due to CAP, 51 noninfected ICU patients, and 48 outpatient controls. In a paired analysis, we reanalyzed patients still in the ICU 4 days after admission (n = 82). Measurements and Main Results: A total of 58% of plasma lipids were significantly lower in patients with CAP-attributable sepsis compared with outpatient controls (6% higher, 36% not different). We found strong lipid class–specific associations with disease severity, validated across two external cohorts, and inflammatory biomarkers, in which triacylglycerols, cholesterol esters, and lysophospholipids exhibited the strongest associations. A total of 36% of lipids increased over time, and stratification by survival revealed diverging lipid recovery, which was confirmed in an external cohort; specifically, a 10% increase in cholesterol ester levels was related to a lower odds ratio (0.84; P = 0.006) for 30-day mortality (absolute mortality, 18 of 82). Comparison with noninfected ICU patients delineated a substantial common illness response (57.5%) and a distinct lipidomic signal for patients with CAP-attributable sepsis (37%). Conclusions: Patients with sepsis due to CAP exhibit a time-dependent and partially disease-specific shift in their plasma lipidome that correlates with disease severity and systemic inflammation and is associated with higher mortality.

Sepsis is a deleterious inflammatory response to infection with high mortality. Reliable sepsis biomarkers could improve diagnosis, prognosis, and treatment. Integration of human genetics, patient metabolite and cytokine measurements, and testing in a mouse model demonstrate that the methionine salvage pathway is a regulator of sepsis that can accurately predict prognosis in patients. Pathway-based genome-wide association analysis of nontyphoidal Salmonella bacteremia showed a strong enrichment for single-nucleotide polymorphisms near the components of the methionine salvage pathway. Measurement of the pathway’s substrate, methylthioadenosine (MTA), in two cohorts of sepsis patients demonstrated increased plasma MTA in nonsurvivors. Plasma MTA was correlated with levels of inflammatory cytokines, indicating that elevated MTA marks a subset of patients with excessive inflammation. A machine-learning model combining MTA and other variables yielded approximately 80% accuracy (area under the curve) in predicting death. Furthermore, mice infected with Salmonella had prolonged survival when MTA was administered before infection, suggesting that manipulating MTA levels could regulate the severity of the inflammatory response. Our results demonstrate how combining genetic data, biomolecule measurements, and animal models can shape our understanding of disease and lead to new biomarkers for patient stratification and potential therapeutic targeting.

Sepsis biomarker panels that provide diagnostic and prognostic discrimination in sepsis patients would be transformative to patient care. We assessed the mortality prediction and diagnostic discriminatory accuracy of two biomarkers reflective of cell death (apoptosis), circulating cell-free DNA (cfDNA), and nucleosomes. The cfDNA and nucleosome levels were assayed in plasma samples acquired in patients admitted from four emergency departments with suspected sepsis. Subjects with non-infectious systemic inflammatory response syndrome (SIRS) served as controls. Samples were acquired at enrollment (T0) and 24 h later (T24). We assessed diagnostic (differentiating SIRS from sepsis) and prognostic (28-day mortality) predictive power. Models incorporating procalcitonin (diagnostic prediction) and APACHE II scores (mortality prediction) were generated. Two hundred three subjects were included (107 provided procalcitonin measurements). Four subjects exhibited uncomplicated sepsis, 127 severe sepsis, 35 septic shock, and 24 had non-infectious SIRS. There were 190-survivors and 13 non-survivors. Mortality prediction models using cfDNA, nucleosomes, or APACHEII yielded AUC values of 0.61, 0.75, and 0.81, respectively. A model combining nucleosomes with the APACHE II score improved the AUC to 0.84. Diagnostic models distinguishing sepsis from SIRS using procalcitonin, cfDNA(T0), or nucleosomes(T0) yielded AUC values of 0.64, 0.65, and 0.63, respectively. The three parameter model yielded an AUC of 0.74. To our knowledge, this is the first head-to-head comparison of cfDNA and nucleosomes in diagnosing sepsis and predicting sepsis-related mortality. Both cfDNA and nucleosome concentrations demonstrated a modest ability to distinguish sepsis survivors and non-survivors and provided additive diagnostic predictive accuracy in differentiating sepsis from non-infectious SIRS when integrated into a diagnostic prediction model including PCT and APACHE II. A sepsis biomarker strategy incorporating measures of the apoptotic pathway may serve as an important component of a sepsis diagnostic and mortality prediction tool.

RNA editing is a widespread epigenetic process that can alter the amino acid sequence of proteins, termed ‘recoding’. In cephalopods, recoding occurs in most proteins and is hypothesized to be an adaptive strategy to generate phenotypic plasticity. However, how animals use RNA recoding dynamically is largely unexplored. Using microtubule motors as a model, we found that squid rapidly employ RNA recoding to enhance kinesin function in response to cold ocean temperature. We also identified tissue-specific recoded squid kinesin variants that displayed distinct motile properties. Finally, we showed that cephalopod recoding sites can guide the discovery of functional substitutions in non-cephalopod dynein and kinesin. Thus, RNA recoding is a dynamic mechanism that generates phenotypic plasticity in cephalopods and informs the functional characterization of conserved non-cephalopod proteins.