Development of a new strategy to predict early sepsis

Sepsis is defined as a life-threatening organ dysfunction caused by a
dysregulated host response to infection. It can affect people from all ages,
causing significant rates of morbidity and mortality in Canada and across the
world (Singer, 2016). Sepsis is also a leading cause of mortality in adulthood
but especially in the elderly (Starr, 2014) and the Surviving Sepsis campaign
has estimated that it causes around 5 million deaths annually.

Currently, microbiology diagnoses take at least 30 hours to obtain and
clinicians are forced to treat empirically with antibiotics while waiting for
results. In 30-50% of cases, these initial antibiotics are retrospectively
recognized as inadequate. This is partly because *last line* antibiotics are
not recommended as part of the initial therapy, since this would lead to an
increase in the prevalence of organisms that are resistant to these crucial
antibiotics. The current clinical diagnostic methods are not only variable and
subjective (Rhee, 2016), but they are also insensitive: in half of septic
patients a causative organism is never found. In our view, this approach to
sepsis diagnosis and treatment is unacceptable as the risk of death increases
by 8% each hour that inappropriate antibiotics are used.

Moreover, after the initial strong inflammatory response to primary infection,
the next phase of illness in septic patients is profound immune suppression,
that has been linked to a biological phenomenon known as endotoxin tolerance or
cellular reprogramming (CR). Beeson (Beeson, 1946) first reported CR in 1946 as
the abolition of the fever response of rabbits undergoing repeated daily
injections of the same dose of typhoid vaccine. Subsequently, it was found that
neutrophils and monocytes isolated from later-stage septic patients exhibited a
state of hypo-responsiveness, including the absence of pro-inflammatory
cytokine production and low levels of HLA-DR expression, and that these
patients are classically unable to defend against secondary infections.
Similarly, patients who survive acute septic shock have deficiencies in
monocyte cell activation that persist for two weeks (or more). CR involves a
suppressive effect induced by exposure to more than one dose of microbial
signature molecules such as endotoxin/lipopolysaccharide (or the bacteria
containing them). These molecules in single dose are potent inducers of
inflammation. It was later found that repeated doses of several different
microbial molecules that induce inflammation could lead to the same suppressive
effects and so the phenomenon is more correctly termed cellular reprogramming
(CR). Key aspects of the CR phenotype are: (i) an inability of cells to respond
to related stimuli (immune amnesia) (Biswas, 2009), (ii) cross-tolerance to
other bacterial signature molecules, and (iii) the long duration of the
reprogrammed cells consistent with the long-term nature of sepsis.

In this phase of CR during sepsis, deadly secondary infections can occur
becoming a driver of mortality rates but clinicians have had no means of
identifying when immune suppression really occurs (Zhao, 2016) and after
decades of research, clinicians have been unable to predict when a septic
patient will transition to an immunosuppressed state that signals rapid
clinical deterioration. In our effort to investigate this, we defined CR and
inflammation signatures using transcriptomics on human volunteer peripheral
blood mononuclear cells challenged either a single time (inflammatory) or twice
(CR) with endotoxin (LPS) (Pena, 2011) respectively. In a follow up study, we
demonstrated the strong presence of a gene expression signature of CR. Our
conclusions were informed by both a meta-analysis of 592 patients, including
both pediatric and adult sepsis cohorts, and by an initial single-blinded
observational clinical study in adult patients in the emergency ward (72
patients), at first clinical presentation. These observations are fundamental
and novel. With this project, our main objective is to perform a validation
phase with a larger cohort, in a multi-center, international study. The results
of this study will have enormous implications in the management of sepsis, with
respect to early diagnosis and prognosis, as well as supporting early
application of appropriate therapies including antimicrobial and
non-pharmaceutical (e.g. immunomodulatory) strategies. *

To validate and demonstrate reproducibility of the identified sepsis CR gene
signature in participants from different age groups. The validation multicenter
study will also lead to biomarker refinement of the original core set of 31
genes tailored to each age group. To determine whether the CR gene signature is
affected by variation in the genome (eg. polymorphisms) and whether changes in
gene expression affect downstream protein levels, as transcription may be
affected by epigenetic regulation. Not only will this approach allow to
identify a genetic signature associated with sepsis, but it may also provide
highly valuable information about the pathophysiology of sepsis.

This study is designed as a multicenter prospective cohort study, in which
every center is expected to include 100 subjects. Transcriptome analysis
performed on the blood sample collected upon presentation to the emergency
department (ED) will be associated with clinical parameters of sepsis, which
will be collected during ED stay and hospitalization.

Study groups:
- sepsis subjects (n = 1000, of which n = 100 will be included in the UMCG)
- surgical subjects (control group; n = 100, of which n = 10 will be included
in the UMCG)
Note: study groups will be stratified upon analysis (see *Statistical
analysis*).

Study phases:
1. signature refinement
2. signature validation
3. systems biology analysis of genetic signature

Note: each study phase has a specific objective; to answer these objectives,
data and blood samples collected from both study groups will be analyzed in two
phases. For the third phase, samples obtained from a subset of patients with
sepsis (n = 50), subsequently admitted to the ED at the UMCG, will be used.

Study participants are asked to donate blood by means of a (minimally invasive)
venapuncture during their stay at the ED at the same moment when blood is
collected for routine examinations as part of standard care to limit the
discomfort as much as possible for the study participant. Participants will not
directly benefit from the results. However, the results may be of major
relevance to increase the speed and accuracy of diagnosis in patients with
sepsis.