Patient Recruitment to GAinS (now closed)
Patients admitted to an intensive care unit (ICU) with sepsis due to community acquired pneumonia or faecal peritonitis were eligible for enrolment in the original GAinS study. For each patient, we obtained a blood sample for DNA extraction upon recruitment. On the first, third, and fifth day in ICU, we also took plasma and urine samples, and stored peripheral blood leukocytes for RNA extraction.
Genome wide association study for sepsis survival
Our genome wide association study for survival from sepsis was published in the Lancet Respiratory Medicine. This was the first GWAS for outcome performed in sepsis patients, carried out as part of the European GenOSept (Genetics Of sepsis and Septic shock) study. In a meta-analysis of 4 cohorts of pneumonia patients, we identified common variants in the FER gene associated with a reduced risk of death from sepsis.
Sepsis endotypes from transcriptomics
We generated genome wide gene expression data for peripheral blood leukocytes from 552 sepsis patients. Using this data, we identified and validated two distinct sepsis response signatures (SRS1 and SRS2) with pronounced differences in gene expression patterns and mortality rates.
We also demonstrated that these disease endotypes may be resolved using a small gene set. These findings, which are supported by comparable results in other studies (e.g. MARS), highlight an opportunity for patient stratification and a precision medicine approach to the management of sepsis.
Differential treatment responses
In the context of the VANISH clinical trial, we found that the SRS endotypes showed different responses to hydrocortisone.
Expression quantitative trait loci
By analysing gene expression as a quantitative trait, it is possible to identify putative regulatory variants and their target genes. As the regulatory landscape is highly context specific, and known to be dynamic in the immune response, we mapped expression quantitative trait loci (eQTL) in sepsis patients and identified associations for a number of disease-relevant genes.
For more than half of patients recruited to GAinS with pneumonia, the causative pathogen was not identified using standard clinical tests. We are using next generation sequencing methods to detect and classify all genetic material present in the blood of sepsis patients, and thus improve knowledge of infectious aetiology in the GAinS cohort.
We are currently generating mass spectrometry proteomic data for a subset of patients enrolled to GAinS in order to further investigate heterogeneity in the sepsis response. Early results indicate that the SRS endotypes we describe in transcriptomic data are also reflected in the plasma proteome.
We are now recruiting all cause sepsis patients admitted to a general ward or to a high dependency/intensive care unit in Oxford, to this integrated, multi-omic study of the host response to sepsis. Samples are also being obtained from convalescent survivors.
Since the start of the pandemic GAinS2 has been repurposed to recruit patients with COVID of varying severity. We have adopted an integrated, multi-omic approach to identify features of the immune response associated with COVID severity. We also compare the response to COVID with that seen in sepsis or influenza.