Final Report Summary - SHOCKOMICS (MULTISCALE APPROACH TO THE IDENTIFICATION OF MOLECULAR BIOMARKERS IN ACUTE HEART FAILURE INDUCED BY SHOCK)
Executive Summary:
The management of hemodynamic stability in shock patients is of paramount importance in the intensive care unit (ICU). New, cutting edge knowledge is necessary to overcome the shortcomings of available therapies, in order to cope with the challenges that anesthesiologists, intensive care specialists and emergency physicians face when dealing with shock patients.
Current therapies are targeted to reduce symptoms of shock and multiple organ failure, but they are unable to target the root cause or to act at the “beginning of the cascade”, because of the lack of a model explaining the molecular basis of shock-induced tissue injury and ensuing multiple organ failure. Hence, the effectiveness of anti-hypotensive interventions such as fluid resuscitation is limited. Fluid may restore blood pressure within minutes, but complications such as pulmonary edema may arise on a longer time scale in patients who are unresponsive to fluid infusion, and cause more severe stress to the patient’s hemodynamic stability, respiratory efficiency, and immune system.
The ShockOmics project focused on the molecular triggers of acute heart failure in association to shock, in the presence of uncontrolled proteolytic activity, in order to identify inflammatory mediators and markers which are activated in shock, through a systematic analysis of expression levels of genes and their protein products, and novel targets for the delivery of new therapies.
In this framework, several biosample and data sources, integrated by clinical records and hemodynamic monitoring: i) plasma samples from existing biobanks; ii) clinical studies in septic shock (SS) and cardiogenic shock (CS) patients in ICU sampled at T1<16h, T2=48h, T3=7d; iii) animal shock models include hemorrhagic shock (HS) rats, HS swines, SS swines, and CS swines; iv) in-vitro cardiomyocytes shock models.
Analyses spanned all Omics technologies (metabolomics, proteomics, peptidomics, and transcriptomics) addressing the circulating blood/plasma and, in animals, post-mortem samples of the myocardial tissue. Omics outcomes were integrated with clinical indexes and parameters relevant to CV regulation from hemodynamic monitoring. Artificial intelligence and data-mining methods were applied to work out risk prediction algorithms.
Therapeutic advancements were pursued starting from the “autodigestion hypothesis”, which focuses the detrimental proteolytic effects from digestive enzymes leaking from the intestinal compartment via the wall damaged by the ongoing shock. This pushed to confirm previous experiments on rats both in the same species and in the swine shock models. The latter experiments asked to further develop technologies for intestinal infusion through the naso-gastric access beyond the pylorus by a dedicated infusion line and smart control pumps.
Plans for the future exploitation of the project outcomes address three areas: i) risk prediction algorithms to be integrated into existing cardiovascular monitors or addressing the Omics outcomes; ii) future development of diagnostic-kits addressing the potential biomarkers; iii) the drug delivery technologies into the intestinal lumen.
Project Context and Objectives:
GENERAL CONTEXT
Circulatory shock is a major complication in patients admitted to the ICU. After a sufficiently long impairment (order of hour/hours) of oxygen and nourishment transport to organs, damage and inflammation of vasculature and organs may start a cascade of adverse events, which may increase out of control and lead to fatal outcomes. Importantly, shock may occur despite the recovery of cardiovascular stability reached in the ICU by assistive and pharmacological treatment and resuscitation maneuvers.
Hence, new cutting edge knowledge is necessary to assist anesthesiologists, intensive care specialists and emergency physicians in the timely delivery of effective therapies to reduce the incidence of mortality and morbidity in critically ill patients in shock. The multiscale approach of the ShockOmics project was to gain an integrated view at both biomolecular and hemodynamic level aiming at more accurate and precocious biomarkers as well as novel therapeutic approaches.
The relevance of the disease is indicated by epidemiological data. The Italian clinical study on sepsis “Albios” has shown a mortality of 45% in septic shock patients (http://www.negrisud.it/albios/). Trauma resulting from accidental injuries is the leading cause of death in individuals aged 1–44 yr in the U.S.A. and hemorrhagic shock affects 36–39% of trauma victims. Most of shock patients develop multiple organ failure (MOF), or multiple organ dysfunction syndrome (MODS), the main cause of death in the ICU, a unique syndrome in which organs not directly injured in the original trauma become dysfunctional. Moreover, in the case of the heart, preserving its function is instrumental for ensuring hemodynamic stability and an adequate perfusion of all vital organs.
Four groups of shocks have been defined: hypovolemic shock (e.g. hemorrhagic shock (HS)), cardiogenic shock (CS) (which is developed following myocardial infarction), obstructive shock, and distributive shock (e.g. septic shock (SS)). All types of shock are characterized by low blood pressure and accelerated heart beat.
The relationship between shock, ischemia and reperfusion (I/R) injury, hemodynamic instability, systemic inflammatory response syndrome (SIRS), and MOF/MODS has been extensively investigated, but there is no consensus on the trigger mechanisms of tissue injury at the molecular level. Inflammation induced by shock has been described at the microvascular level, as a tissue repair reaction to injury, characterized by cellular dysfunctions, microvascular perfusion deficiencies, autoregulation impairment, capillary stasis and augmented permeability, coagulation, tissue edema, leukocyte-endothelial interaction, oxygen radical free formation, apoptosis, metabolic defects, hyperglycemia, insulin resistance. Biochemical markers for inflammation, e.g. the expression of cytokines, indicate a tissue repair reaction to injury.
Several therapeutic countermeasures against circulatory shock have been proposed, such as endotoxin or cytokine blockade, endotoxin binding protein receptor, blockade of leukocyte adhesion molecules, complement depletion, oxygen free radical scavengers, nitric oxide modulators, hypertonic resuscitation, designer crystalloids and hemoglobin carrying solutions that preserve the fluid shear stress on the capillary endothelium. However, none of the above approaches represents a clinically effective intervention. Current therapies are targeted to reduce symptoms of shock/MODS, but they are unable to target the root cause or to act at the “beginning of the cascade”, because of the lack of a model explaining the molecular basis of shock induced tissue injury and ensuing MODS. As a consequence, the effectiveness of anti-hypotensive interventions such as fluid resuscitation is limited and depends on the trade-off between their short-term beneficial effect and their potential long-term risk.
Fluid administration may restore blood pressure within minutes, but serious complications such as pulmonary edema may arise on a longer time scale as a consequence of patients’ unresponsiveness to excessive fluid infusion, and cause more severe stress to the patient’s hemodynamic stability, cardiorespiratory efficiency, and immune system.
OBJECTIVES
The main goal of ShockOmics was to investigate the molecular triggers of acute heart failure (HF) induced by shock and to identify inflammatory mediators and markers activated in shock, with a particular emphasis on the role of uncontrolled proteolytic activity.
To achieve this goal, a systematic analysis and multiscale integration of information from the gene (transcriptomics), peptide (peptidomics), protein (proteomics), and metabolite (metabolomics) levels was carried out. The ultimate aim of the proposed multiscale approach is to interpret the patho-physiological changes in system level variables, i.e. cardiovascular (CV) measurements, which are routinely available in emergency departments, ICUs, and operating rooms to describe the hemodynamics of patients, in function of the fundamental processes of disease.
System level symptoms of shock are well known. Hypotension and hemodynamic instability are the main issues that clinicians have to cope with when treating shock patients. In the context of low blood pressure, low flow, and reduced ventricular contractility typical of circulatory shock and hemodynamic instability, the in-depth monitoring of heart function is essential to prevent further hypoperfusion of organs. Thus, ShockOmics focused on acute heart failure due to shock, and explored the presence of biomarkers, which could be related (but not limited to) to the role of the intestine as a source of powerful inflammatory factors in shock, which are responsible for uncontrolled proteolytic activity.
Summarizing, the specific objectives of the ShockOmics project were:
– Identification of candidate biomarkers from pre-existing and prospective populations of patients, by means of cutting edge Omics techniques;
– Assessment of heart injury in response to uncontrolled proteolytic activity, which may cause cleavage of important receptors (e.g. insulin, adrenergic and VEGFR2 receptors), in animal models of shock, and in in vitro studies;
– Analysis of gene expression and protein expression by means of state-of the-art high-throughput techniques, such as Next Generation Sequencing for whole transcriptome analysis and shotgun proteomics, to characterize the cellular response to shock due to proteolysis in heart tissues and cells obtained from animal experiments and in vitro studies, from peripheral blood from in vivo studies, and from patient blood samples;
– Development of multiscale, systems biology based models to describe the relationship between hemodynamic measurements/waveforms available in ICU and the progression of the shock induced acute heart disease;
– Identification of novel targets for effective cardioprotective therapies to prevent/contrast shock induced cardiovascular instability;
Refinement and further validation of highly innovative technologies for the early identification of biomarkers of the disease, for the interpretation of hemodynamic alterations, and for a timely and efficient drug delivery for the prevention of hemodynamic instability due to shock induced HF.
CONSORTIUM
The ShockOmics Consortium involved 11 participants from 4 nations: Italy, Spain, Belgium, Switzerland and USA.
The partners have multidisciplinary integrated expertise:
1. Politecnico di Milano – POLIMI- Milan (Italy): Hemodynamics and multiscale modeling
2. Filarete Servizi srl – FIL- Milan (Italy): Transcriptomics and proteomics
3. Istituto di Ricerche Farmacologiche Mario Negri –IRFMN- Milan (Italy): Bio-banks, metabolomics, animal experiments on hemorrhagic shock
4. Neurozone s.r.l. –NZ- Milan (Italy): In vitro studies
5. Custom Software Electronics –CSE- Barcelona (Spain): Data base development, data mining
6. Hospital Universitari Mutua Terrassa –MDT- Terrassa (Spain): Access to bio-banks
7. Parc Científic Barcelona – PCB- Barcelona (Spain): Proteomics
8. Genève University Hospital – UNIGE- Geneva (Switzerland): Clinical study and animal experiments on cardiogenic shock
9. Université Libre de Bruxelles –ULB- Brussels (Belgium): Clinical study and animal experiments on cardiogenic shock
10. University of California San Diego –UCSD- San Diego (USA): Animal experiments, bio-signaling
11. Antlia –Antlia- Lausanne (Switzerland): development of pumps for drug delivery.
Figure 1 summarizes the main activities in which the partners were involved.
Project Results:
1 COLLECTED DATA AND BIO-SAMPLES
1.1 Clinical study in ICU patients
The ShockOmics clinical trial (ClinicalTrials.gov identifier NCT02141607) was a multicenter prospective observational study.
The main inclusion criteria were:
– Sequential Organ Failure Assessment (SOFA) score > 5
– First blood sample and first hemodynamic measurements available within 16 h from admission to the ICU
– Informed Consent available
The main exclusion criteria were:
– Risk of fatal illness and death within 24 h
– Patients already enrolled in other interventional studies
– More than 4 units of red blood cells transfused
– Patients receiving plasma or whole blood
– Active hematological malignancy
– Metastatic and/or active cancer
– Immunodepression, including transplant patients; patients infected by the human immunodeficiency virus (HIV+); constitutive immune system deficiency; any immunosuppressive therapy, including oral and parenteral corticosteroids (aerosols are allowed)
– Patients with pre-existing end stage renal disease needing renal replacement therapy. The introduction of continuous veno-venous hemofiltration from the day of admission onward was allowed
– Cardiac surgery in the previous ten days
– Child-Pugh C cirrhosis or acute liver failure
– Terminal illness
Data collection timeline was:
– Time 0 (T0): admission to the ICU with diagnosis of shock or time of shock diagnosis in patients initially admitted without shock symptoms;
– Time 1 (T1): time at which the first blood sample for analysis is collected, within 16 h after T0;
– Time 2 (T2): time at which the first blood sample for analysis is collected, at 48 h after T0. At this time point, the treatment has been administered for a long enough amount of time to evaluate its effects;
– Time 3 (T3): time at which the third blood sample for analysis is collected, on day 7 of the ICU stay of the patient or before discharge from the ICU in case of shorter stays or before discontinuing therapy (death).
– Time 4 (T4): follow up on ~ day 100 from Time 0 in non re-hospitalized patients; this follow up, including a further blood sampling, was severely limited.
Out of 529 patients screened for eligibility, 41 septic shock patients, 22 cardiogenic shock patients, 9 sepsis patients not in shock, 3 cardiac patients not in shock were enrolled in the study.
1.2 Experimental studies in swine shock models
Swine shock models addressed three different insults leading to shock onset: i) septic shock (ULB) was induced by cecal ligation puncture; ii) hemorrhagic shock (IRFMN) was induced by controlled bleeding; iii) cardiogenic shock (UNIGE) was induced either by sustained ventricular pacing or by coronary artery occlusion (CAO).
The following sampling times were defined in order to consider the different phases of each experiment comparable:
– T1: Baseline, before the start of the insult to induce shock;
– T2: Two hours after the beginning of shock and ahead of resuscitation maneuvers;
– T3: After fluid and vasopressor resuscitation;
– T4: After blood resuscitation.
Autopsy and myocardial tissue sampling followed the animal sacrifice.
As scheduled, 6 SS and 6 HS animals were collected with the complete set of data (blood samples, heart tissues and hemodynamics recordings) together with 6 sham animals.
Severe limitations were encountered in the CS experiments due to the high rate of either a spontaneous recovery from AHF or animal loss ahead of experiment completion. Sustained pacing never induced stable CS conditions, while CAO permitted the collection of 2 CS cases only, after which the experiments were stopped due to not renewal of ethical committee approval for further animals. Samples from 3 sham animals and 2 CS were available.
1.3 Treated swine shock models
Tranexamic acid (TXA) is well known in clinics as intravenous TXA is an effective adjunct after hemorrhagic shock (HS) because of its antifibrinolytic properties. TXA is also a serine protease inhibitor, and recent laboratory data demonstrated that intraluminal TXA into the small bowel inhibits digestive proteases and protects the gut. The developed techniques for pilot intestinal infusion (multi lumen infusion line, regurgitation prevention, pump control, etc.) are better described in the dedicated paragraph of results section.
3 SS and 5 HS animals were treated with TXA by enteral infusion; 3 further animals treated with enteral infusion of vehicle only were considered as controls. CS was not considered for the above mentioned reasons.
2 CARDIOVASCULAR REGULATION
We selected a subset of 21 septic shock patients enrolled at Genève University Hospital from October 2014 till December 2015. All participants gave prior informed consent. Strict inclusion criteria were observed during the recruitment phase, in order to avoid a too high inhomogeneity within the population.
The patients received initial therapy according to the standards immediately after shock diagnosis (time T0). Patients were analyzed at two relevant time points: within 16 hours from T0 when the inflammatory cascade has been just activated (T1), and within 48 hours after T0 time point (T2).
SOFA score (Sequential Organ Failure Assessment score) was used to classify patients into two groups according to their responsiveness to early therapy: responder patients (R, n=14) consisted in patients with a positive response to initial treatment. Those patients, who decreased their SOFA score of at least 5 points between T1 and T2(ΔSOFAT2-T1>5) or reached a SOFA score at T2 lower than 8 were classified as R. Non responder patients (NR, n=7) consisted in patients, who still had a SOFA score at T2 higher than 8 or ΔSOFAT2-T1<5.
All patients had similar severity of shock at the enrollment and received the same initial therapy, but the R group improved their condition within the first 2 days after ICU admission (i.e. between T1 and T2), whereas the NR patients did not improve or even worsened.
NR patients displayed a higher fluid balance during the first days after shock development due to a larger fluid infusion and a reduction in urine output
Arterial pressure (AP) parameters showed an increase between T1 and T2 in the R group: moreover, at T2, R patients had significantly higher values than NR in both mean AP (MAP) and systolic AP (SA). All AP parameters were slightly decreased in NR patients between the two time points. A significant decrease of heart rate (HR) mean value from T1 to T2 was reported in NRs.
In our study the feedback gain of the baroreflex (BR) was quantified by the BR sensitivity (BRS), based on the relationship between SAP and HR mediated by the nervous system. The BRS increased significantly at T2 in NRs. If cardiopulmonary (CP) afferent activation indirectly affects the BR control of heart and peripheral resistance, then the elevated loading (or stretching) of CP baroreceptors, due to large central volume pressures, could be a possible explanation of the different trend in BRS observed in NR patients with respect to R patients whose fluid balance was significantly decreased at T2.
Moreover, the overall variability in the NR group was significantly lower with respect to the R group. All patients were able to maintain their MAP over the targeted threshold of 65 mmHg, even if the trend between T1 and T2 in NR patients was slightly decreasing and they displayed lower values with respect to R patients both at T1 and T2. Looking at the international guidelines for acute shock resuscitation, all these patients seemed to be responsive to the therapy.
However, the hemodynamic analyses allowed to discriminate those patients who actively responded restoring the autonomic nervous system regulation of BP and HR, and those who recovered from hypotension without a marked improvement in cardiovascular autonomic control.
Finally, it is a common evidence that patients in SS receive excess fluids even after they are weaned from vasopressors. In our opinion, this happens because indices relating to CV autonomic control are not commonly used to guide fluid therapy and our work is meant to contribute in this direction. Hence, CV regulation indices offer acnon-invasive options to guide fluid therapy and permit to assess the hemodynamic response of patients to the therapy.
These findings were confirmed in HS swine experiments, showing that the evaluation of the autonomic nervous system could be valuable in determining the effectiveness of the therapy and the grade of recovery.
Briefly, baseline measurements were similar in both HS and sham pigs. During HS, concurrently with MAP reductions, there was a significant decrease in CO (1.7±0.3 L/min vs. 3.2±0.7 L/min, p<0.01) and LV end diastolic and systolic volumes (p<0.01). Heart rate (HR) showed a 3-fold increase compared to baseline (217±5 vs. 77±16 b/min, p<0.01) and pulse pressure variation (PPV) was > 35 mmHg (p<0.01). Arterial lactate increased above 8 mmol/L in each HS pig (p<0.01). After resuscitation with fluids and vasopressor support, hemodynamics and LV volumes improved, although did not return to baseline, i.e. MAP was 30% lower, HR 2-fold greater, and lactate was still above 6 mmol/L (p<0.05). Only after reinfusion of the shed blood, all the hemodynamic variables, CO, LV volumes, PPV, and lactate completely recovered. As expected, all the above variables remained stable throughout the experiment in the sham animals (p<0.01 vs. HS).
LF oscillations in diastolic arterial pressure (DAP) series in shock were lower than baseline and did not recover after the first resuscitation. Baroreflex absolute gain was greatly reduced with respect to baseline; the opposite trend was observed in the feedforward gain (i.e. the AP response to HR changes). The same results were found for the variation in BRS and feedforward gains after a vasoconstriction maneuver by a phenylephrine bolus. Finally, DAP variability explained by the BR control increased in shock condition, whereas the HRV component was reduced. Only after reinfusion of the shed blood all the hemodynamics parameters recovered to baseline values.
3 METABOLOMICS
3.1 Clinical study
The plasma metabolome of all patients enrolled in the ShockOmics project was successfully analyzed by mass spectrometry-based metabolomics strategies.
By combining mass spectrometry-based untarget and target metabolomics approaches , many dyseregulated metabolic pathways ranging from energy and lipid metabolism to amino acids synthesis were observed in shock patients..
We made a comprehensive metabolomics study of septic shock patients stratified as responders (R) and not responders (NR) to standard therapy on the basis of the changes in SOFA score in the first two days in the ICU.
Changes in the levels of metabolites over time were shown to discriminate positive response to therapy. Metabolomics data elaboration by datamining techniques revealed that the metabolic trajectory during the first 48 hours after ICU admission in NR patients was characterized by significant lipidome alterations.
The reduction in circulating lysophosphatidylcholines (LysoPC) in NR patients might reflect their enhanced conversion to lysophosphatidic acid, which induces a multitude of cellular responses through its action on immunologically relevant cells and might also promote an excessive immune response, with detrimental effect in NR patients. NR patients also showed a marked decrease in phosphatidylcholine species (PC), which originate in the liver. The imbalance of LysoPC/PCs cycling suggested that hepatic homeostasis and function is compromised even before any clinical manifestation.
Indeed the higher plasma level of alanine found in NR vs R may be a sign of lower hepatic capacity for conversion of alanine to glucose and suggests a different approach for monitoring hepatic function, which will be more specific than bilirubin.
Therefore these findings are really promising as they could pave the way for tailored therapies.
These results have been published in Cambiaghi et al. Scientific Reports 2017, Aug 29;7(1):9748. doi: 10.1038/s41598-017-09619-x.
Variation of metabolite plasma concentration between acute phase (T1) and short temporal window after diagnosis (T2) {Δ = T2-T1} was analyzed in cardiogenic shock patients. Again lipid species including LysoPC and PC were deregulated over time. Further statistical analyses were hampered by the small sample size of cardiogenic shock patients-
Nevertheless, machine-learning techniques were applied for the identification of metabolomics changes able to cluster the two types of shocks in patients during the first 24 hours of disease evolution.
The two shocks presented significant differences in the expression levels of PC, suggesting again that lipidome alteration is an important modulation in shock and the understanding of regulatory pathway of lipids is thus crucial for the development of an effective and tailored therapy.
3.2 Swine studies
Metabolomics analyses of heart and plasma samples from swine studies were successfully completed.
So far, the most soundness results were obtained elaborating the plasma metabolomics data from septic shock swine untreated model.
Again, circulating lysophosphatidylcholine species (LysoPC) had reduced level in septic shock swines compared to their control counterpart. Plasma alanine concentration increased in septic shock swines compared to controls.
These findings corroborate what we previously found in the clinical settings and highlight the role of LysoPCs level reduction that may promote an excessive immune response with detrimental effect.
Moreover, the confirmation that also in the septic shock swine model there is a significant enhance in circulating alanine, further supports the role of alanine in early dysregulated liver gluconeogenesis. Interestingly, the Fisher ratio, indicator of liver damage, is significantly decreased after resuscitation in septic shock swine only. Furthermore, citrulline concentration and the metabolic ratio citrulline/glutamine were significantly enhanced in the plasma of septic shock swine compared to their control counterpart. In the liver, citrulline functions as part of the urea cycle in the detoxification of ammonia and, because of the channeling of urea cycle intermediates, little or no citrulline escapes the liver. The citrulline produced in the small intestine, however, enters the portal vein and appears in the peripheral circulation serving as precursor for arginine synthesis. As such, these preliminary results converge towards an early dysregulated intestinal-liver axis.
In conclusion, significant results have been produced showing the same plasma metabolomic trajectories in Septic Shock patients and Septic Shock swine model, suggesting that changes in specific constellations of circulating metabolites might be early plausible indicators of (i) deranged immune response (i.e. lipidome alteration) and (ii) dysregulated intestinal-liver axis (i.e. alanine, citrulline concentration changes).
4 PROTEOMICS AND PEPTIDOMICS
4.1 Clinical study
Proteomics and peptidomics analyses were performed in both clinical and animal model samples.
The clinical study involved analysis of samples from all patients enrolled in ShockOmics, except hemolyzed samples that could not be included. We analyzed the first three time points (T1, T2 and T3) of septic shock (SS) and cardiogenic shock (CS) patients. Samples from healthy volunteers (H, n=9), septic controls (SSC, n=9) and cardiogenic controls (CSC, n=3) were used as controls.
As explained above, SS patients were classified according to their responsiveness to early therapy: responder (R, n=16) and non-responder (NR, n=13). For the cardiogenic study, the response criterion was also based on SOFA score, but excluding the cardiovascular score from total SOFA calculation. Thus, responders (R, n=6) were defined as individuals showing ΔSOFA (T1-T2>3) or (T1-T3>6). For CS the percentage of non-responders (NR, n=14) was higher than in SS.
The proteomics results obtained in the SS study showed a significant deregulation of proteins related to hydrolase and peptidase activities. Nevertheless, we did not detect significant differences between septic control and SS at T1, the first 16h after ICU admission. On the other hand, we observed a group of deregulated proteins that were common for SS groups (R and NR) at T1 and septic control, when compared to healthy condition. The network formed by these deregulated proteins is enriched in proteins related to proteolysis. In addition, deregulated proteins that were only observed in SS formed a network enriched in proteins related to response to stress. The comparison of R and NR groups showed a significant up-regulation of proteins related to the regulation of the immune system and complement and coagulation cascades in NR group.
We observed higher variability among individuals of the CS groups. A functional analysis performed in the proteins with significant changes between R and NR patients showed an enrichment in proteins related to heart diseases. Most of the deregulated proteins detected in the comparisons between R and NR patients of SS and SC are specific for the disease. From proteomics results, few overlap were observed between SS and SC, although more in deep statistical analysis are still in progress.
The peptidomics results for the SS study showed a clear increase in the number of peptides that can be correlated to the increment of proteolysis in the SS group at T1 when compared with sepsis controls and healthy controls.
For the peptidomics analysis we also classified the patients according to survival (Survivors, S, n=23; Non-survivors, NS, n=6) at day 28. The highest values in terms of number and abundance of peptides were observed at T2 in NS. In addition, the number of peptides at T3 was lower than at T1, except for NS. The peptidomics analysis also allowed differentiating patients according to the location of the infection. The patients with source of infection in the urinary tract presented lower mean abundance of proteins than those patients with abdominal or respiratory source of infection.
The peptidomics data from cardiogenic shock patients corroborates the septic shock data and points to a dysregulation in proteolysis in shock, which supports the ShockOmics hypothesis.
4.2 Animal studies
The animal study involved peptidomics and proteomics analyses of samples from rat and pigs.
Preliminary experiments on plasma obtained from control and hemorrhagic rat models were performed by using a shotgun label-free quantitative approach. Overall, the data obtained confirmed that massive proteolysis occurs as a fundamental degrading phenomenon induced by shock and that chymotrypsin like proteases are mainly involved. In particular a huge increase was observed in terms of peptides and proteins in shock in comparison to control.
Peptidomics was applied to verify if enteral protease inhibition delivery (EPID) is effective in preventing mortality and morbidity after HS. The peptidomic differences between animals EPID-treated and those that do not (untreated) allowed to determine those proteins and peptides that are associated and could be causal to increased morbidity and mortality. The main finding of this work is the protective effects against uncontrolled proteolytic activity of a protease inhibitor such as Tranexamic Acid. (TXA) The differences seen both in peptides number and in peptides intensity are particularly evident at the END stages, showing a marked decrease of peptides number and intensity values in subjects treated with TXA compared to untreated animals.
These very promising results led the EIAB steering committee to strongly recommend performing peptidomics analysis on pig and clinical plasma samples.
On plasma samples from hemorrhagic shock (HS) swine models the peptidomic results show that the increase of peptide number and intensity found in rat plasma samples upon HS is confirmed and that there is an evident difference between T1 and all the other time points.
In plasma samples from SS swine models, instead, the peptidomic data show that there was no a significant increase in peptide number upon SS opposite to what we observed in rat and pig in HS and in human patients in SS.
Other major findings using the peptidomic approach on plasma samples from rat and swine shock models were:
• no direct evidence of cleavage of specific membrane receptors, namely insulin receptors, β-adrenergic receptors and vascular endothelial growth factor receptor 2
• presence of potentially bioactive peptides with anti-hypertensive activity in rats and swine plasma samples
• the predominant activity of chymotrypsin–like proteases in shock plasma samples in rat and swine models.
The proteomic analysis carried out on plasma samples from rat and swine shock models allowed to obtain a direct experimental evidence of oxidative stress-induced post-translational modification (Tyr nitration) in the presence and in the absence of the protease inhibitor TXA in plasma samples from hemorrhagic shock while no direct evidence of oxidative stress-induced was detected in all the myocardic tissues analyzed (left ventriculum, right ventriculum and right atrium) from the same shock models. In the case of septic shock no evidence of oxidative stress could be detected either in plasma or in all the myocardic tissues analyzed (left ventriculum, right ventriculum and right atrium).
5 TRANSCRIPTOMICS
Transcriptomic analysis, using RNAseq technology, has been performed in all human and animal samples.
RNAseq is a powerful tool to analyze transcriptome, to quantify gene expression levels, to investigate and compare at molecular level specific pathological or physiological conditions, tissues, cell types, with the aim to identify genes differentially expressed.
5.1 Human study
We addressed two major aims. The first aim was the transcriptomic profiling in whole blood of septic shock patients along the clinical course (the three time points considered in the clinical protocol). This analysis showed at steady state, day 7 of the ICU stay, the downregulation of Toll-like receptors, C-type lectin receptors, IL-1 and IL18 receptors family proteins and alarmins suggesting the downregulation of innate immune response and acute inflammation. Conversely, at the same time point, the upregulation of genes of the adaptive immunity related to T and B lymphocytes activation was observed. A transcriptional regulation of genes involved in carbohydrate metabolism, lipid inflammatory pathway, transport of vesicles and protein synthesis was observed as well. These data have been presented in the Poster section at the 40th Meeting of Shock Society, Florida June 2017. Braga D. et al. Preliminary profiling of transcriptomic modifications in septic shock patients at multiple time points.
The second aim was the comparison of the transcriptomic profile in whole blood of septic shock patients defined as Responders or not Responders to early supportive therapy. This study showed that Responders specifically alter the expression of adaptive immune response genes (ICOS, LCK, VSIG4, upregulated genes related to functions of granulocytes and NK cells, and genes involved in pathogen lipids clearance (PCSK9, SORT1). Common functions were shared between R and NR, as the downregulation of genes involved in acute inflammation (alarmins, metalloproteases, C-type lectins, interferon receptors) and the upregulation of genes expressed in neutrophils and involved in bacterial killing (GZMM, GZMH, KLRG1). These results have been partially presented as poster at the Meeting of the Society of Complexity in acute illness (ICCAI), Milan, July 27-29, 2017 (Braga D. et al. Differential gene expression in septic shock patients according to the early supportive therapy response).
In cardiogenic shock patients, stratified as R or NR to early supporting therapy, the transcriptomic analysis showed that both groups modulate the expression of genes involved in neutrophils degranulation and immunoglobulin mediated immune response. Responders specifically modulate genes involved in lymphocytes differentiation and T cell activation, as observed in septic Rs. Instead, NRs specifically alter the expression of genes of the pattern recognition receptors signaling (PRRs), and genes involved in the inflammatory NFKB pathway.
5.2 Animal study
In septic shock untreated swines, most of the variability in whole blood gene expression was observed by comparing basal to shock condition.
In sepetic shock (SS), compared to the basal condition, genes that can counteract the infection were upregulated. These genes are involved in neutrophils chemotaxis, response to LPS or have antimicrobial activity (CD177, ALPL, LCN2, LTF, NPG1). A significant upregulation was observed for genes that act as negative regulators of protease activity and as inhibitors of serine protease (TNFAIP6, CSTA, RETN). MMP9 that is known to increase in the early phase of shock and whose levels correlate with severity and mortality in sepsis was significantly upregulated in the shock phase. At this time point the inflammatory mediators IL18 and IL5 were significantly downregulated. A downregulation was also observed for CD101 that plays a role in inhibition of T cells proliferation.
A pilot transcriptomic study has been carried on in whole blood of hemorrhagic shock (HS) rats. The transcriptome of shocked animals shows modulation of genes related to inflammation and immune response (Tlr13, Il1b, Ccl6, Lgals3), antioxidant functions (Mt2A, Mt1), tissue injury and repair pathways (Gpnmb, Trim72) and lipid mediators (Alox5ap, Ltb4r, Ptger2) compared with control animals. These findings are in line with previous observations reported in hemorrhagic shock using other Omics approaches: a change in expression of genes encoding lipid receptors and aminoacid metabolism is consistent with previous metabolomics studies, as well as the increased expression of genes with antioxidant function and of the complement pathway is in line with previous proteomic studies. For a more detailed description of these data refer to the paper “Braga D. et al. Preliminary profiling of blood transcriptome in a rat model of haemorrhagic shock. Experimental Biology and Medicine 2017;242:1462-1470.
6 INTEGRATED ANALYSES
In this activity we have analyzed the different datasets from the clinical studies conducted at the ICUs in Geneva and Brussels. It focused on clinical data, metabolomics, proteomics and transcriptomics with the primary objective of finding potential biomarkers for cardiogenic and septic shock. The secondary objective of these activities has been assessing the risk of death (RoD) of patients with shock.
In this regard, this activity has been mainly concerned with the methodological aspects of the Artificial Intelligence and Machine Learning techniques that we have used to discover hidden information (mainly interactions between clinical data, metabolites, proteins and transcripts) from the large amount of data generated during the project execution.
The methodologies used in our integrated analyses are mainly concerned with:
– Data visualization and dimensionality reduction.
– Kernel manifold learning (i.e. finding optimal hyperplanes/manifolds separating the two shocks considered in this project).
– Classification of shock and RoD from sequential data (i.e. from the different time points considered in the clinical study).
– Ensembles of networks (i.e. study the interactions between the different Omics data).
The integrated analyses have been organized as follows:
1. We developed the methodological framework for integrated data analysis.
2. Development of a pipeline for dimensionality reduction:
a. For analyses where the dimensionality is much greater (about two orders of magnitude) than the number of patients we used bagged ensembles of trees.
b. For other analyses where the dimensionality is not as great (about one order of magnitude) we used a pipeline based on kernel manifold learning.
3. After dimensionality reduction, we applied a generative kernel for data analysis (i.e. the simplified Fisher kernel). The main rationale for using such kernels is that they induce graphical models for the posterior assessment of interactions between attributes.
4. Once we had selected the most appropriate sub-set of clinical attributes / Omics data, we developed a baseline classifier for assessing RoD from clinical data. This approach is based on the generative kernel presented above.
5. We also assessed shock from a metabolomics, proteomics and transcriptomics. We also presented the results of assessing RoD for each of the omics approaches.
6. The main result of this activity is a graphical model including all the Omics data (i.e. from a Multi-Omics point of view) for assessing RoD in the ICU.
Besides the Multi-Omics graphical model, we also presented an alternative representation of sequence alignment based on algebraic statistics. This activity was mainly focused in the presentation of the Needleman-Wunch algorithm for sequence alignment from an algebraic point of view as well as its relation to graphical models and the Viterbi algorithm.
The methodologies used in this activity have been mainly concerned with:
– Min/Max or ‘Tropical’ algebra.
– Phylogenetic trees.
– Graphical models and hidden Markov models.
– The Viterbi algorithm.
The main research activities of this research are as follows:
1. Firstly, we have presented the required methodology for sequence alignment from an algebraic statistics point of view.
2. Secondly, we have evaluated the accuracy and efficiency of our alternative representation through some examples of alignments from the data obtained in our transcriptomics analyses.
4.7 Enteral Drug Delivery Technologies
The development of innovative enteral drug delivery technologies and protocols was motivated by one of the most important initial hypotheses of ShockOmics, i.e. the need to inhibit digest proteases in order to protect the integrity of the intestinal barrier, and as a consequence to limit the injury to distal organs.
A modified and enhanced enteral feeding catheter was developed and an international patent application was filed (International Patent Application PCT/US2017/028583). The main advantages of the system, comprising a multi-lumen infusion catheter, which represents an evolution and improvement of the currently available orogastric or nasogastric tubes for enteral feeding of acutely ill patients are:
• ability to sustain high infusion rates
• easy placement in the stomach
• viable placement in the duodenum, as recommended on the basis of experiments in rats and pigs, with the aid of a gastroscope as guide in large animals
• ability to prevent reflux in the esophagus and in the airways
In order to satisfy the need to achieve a rapid infusion into the intestine, and to prevent airway aspiration, an aspiration line is added for safety and placed in the esophagus for aspiration out of the mouth.
The rat experiments have demonstrated the effectiveness of the proposed infusion protocol, based on the insertion of the infusion catheter through the mouth of the animal (in humans, the access will be either via the mouth or nose) and placement in the duodenum. There is evidence that the optimal placement of the catheter tip for the purpose of delivering the drug to the small intestine is in the duodenum. The placement of the catheter in the stomach results in the inflation of the stomach, due to the high infusion rate and slower emptying rate of the stomach into the duodenum through the pylorus. The effectiveness of the infusion system was verified from the point of view of: a) filling of the small intestine in its entirety; b) preservation of the morphological integrity of the gut barrier.
A dedicated protocol for enteral infusion of protease inhibitor was designed, and its specifics have been determined in rats and pigs:
• infusion rate: 0.10 ml/min (in rats); 20 ml/min (in pigs);
• infusion volumes: 17.5 ml (in rats); 2.5 L (in pigs)
• optimal placement of the tip of the infusion catheter: post-pyloric, ~ 5 in into the duodenum.
The enteral delivery system can be integrated by sensors and control system to automatically detect backward flows and activate possible alarms to alert the critical care personnel and verify the risk for reflux and aspiration into the airways. An example of sensor which can be embedded in the delivery system is a pressure sensor to monitor intragastric pressure. To this purpose, one of the lumens of the multichannel catheter can be used as a fluid filled catheter and connected to a pressure transducer for continuous pressure monitoring. The elevation of pressure above certain thresholds can activate an alarm and interrupt the flow into the gut in order to prevent the possible reflux into the esophagus and from there into the airways.
8 IN VITRO VALIDATIONS
Aim of the in vitro validation study was to identify the most appropriate cell models and in vitro paradigms for the functional evaluation of the different shock challenges in vitro. Currently, no widely accepted suitable in vitro models for adult human cardiac biology exist, and several different approaches are being taken into consideration, ranging from standardized cell line approaches to differentiated human induced progenitor cells, for selected assays aimed at identifying several aspects of safety, pharmacology and toxicology assessment.
We have tested different cell lines currently used in screening activities, and characterized them versus primary cardiomyocytes taken from rat pups. Primary rat cardiomyocytes were the only ones showing multinucleated bodies, typical of cardiomyocytes in culture, and a clear presence of cardiac specific troponin marker staining. Moreover, only primary cardiomyocytes in culture were capable of beating in vitro, forming functional syncytia. A selected group of enzymes was identified, in order to mimic the different shock conditions, as follows: elastase, LPS and MMP9 for septic shock; alpha amylase, hsp70 for cardiogenic shock; oxygen glucose deprivation for cardiogenic shock.
A dose dependent time course was carried out in order to identify the optimal detrimental range for inducing a metabolic alteration of cardiomyocytes. In particular, parameters such as cell viability, metabolic activity and expression of selected genes of interest were quantified in order to obtain clear indications of the effect of the challenge.
For the septic shock model, elastase was shown to induce an increase of heart failure marker hsp70 as well as pro inflammatory cytokine IL6. On the other hand, LPS induced a more pronounced increase of several inflammatory markers such as hsp70, hsp40, MMP9, IL 6 and channels such as CACNA1C and KCNJ2. MMP9 exposure induced a significant increase in the expression of most of the genes tested in cell lines, with a particular emphasis on the expression of calcium and potassium channels, while no significant increase in the expression of inflammatory markers was observed in primary cardiomyocytes.
For the cardiogenic shock model, a time curve of oxygen and glucose deprivation (OGD) was carried out, and the percentage of dying cells at the different time points was quantified. The optimal experimental conditions were identified at 5 hours of OGD challenge followed by 2 hours of reperfusion in normoxic conditions.
The hemorrhagic shock model was mimicked using different stimuli, such as hsp70 or amylase exposure, but data did not yield significant results.
We thus concentrated our efforts on the most detrimental conditions, namely LPS exposure (for septic shock) and OGD exposure (for cardiogenic shock), and used these challenge in a more complex in vitro scenario. In particular, we took advantage of the possibility of using microfluidics devices previously developed in our lab (Bianco et al. Analytical Chemistry, 2012) in order to create micro-environmental scenario mimicking interaction of primary cardiomyocytes with immune cells, under the different experimental scenarios. Primary rat cardiomyocytes previously characterized were cultured on the microfluidic platform and exposed to macrophages isolated from rat spleen and liver. We established a correct protocol for macrophage isolation, and optimized conditions for culturing on a microscale.
In order to recreate the septic shock model, we challenged macrophages with LPS and then put them in microfuidic communication with primary cardiomyocytes. We monitored biochemical and functional parameters and compared the expression of the panel of genes previously identified in cardiomyocytes exposed to the three different key micro-environmental conditions: untreated cells, macrophages and activated macrophages. Activated macrophages induced a significant increase in cardiomyocytes cell death. Moreover, we showed that exposure of cardiomyocytes to macrophage-conditioned medium (either primed or not) is sufficient to induce a metabolic stress on primary cardiomyocytes.
Similarly to the septic shock model set up, in the cardiogenic shock model microenvironment set up we challenged the cells with previously characterized ischemia (OGD) insult, and compared the response of primary cardiomyocytes either exposed or not to ischemic conditions in microfluidic connection with macrophages. The presence of activated macrophages significantly increased cardiomyocyte cell death and remaining living cardiomyocytes show a significantly altered metabolism when exposed to macrophage-derived inflammatory storm.
Once the in vitro models were set up, we applied selected inhibitors in order to quantify their modulatory effect on the in vitro models. Given the results obtained from the Consortium ongoing “omic” studies revealed a key role for inflammatory pathways being activated by the shock conditions, resulting in the activation of many different pathways, (i.e. oxidative stress, mitochondrial damage, increase in intracellular calcium levels and modulation of viability), we selected three different inhibitors, targeting different mechanisms linked to different aspects observed to be altered in the challenged cardiomyocytes; namely, strophanthidin, ghrelin, and tranexamic acid (TXA). A series of biochemical and functional assays were carried out, either in the presence or absence of such inhibitors, including analysis of cell viability, cell metabolism, intracellular calcium dynamics, reactive oxygen species production and NO release, both for the septic as well as the cardiogenic models.
In the coculture septic model, exposure to LPS significantly increases the percentage of dying cells, but this effect was counteracted by concomitant exposure to strophanthidin and ghrelin. Similarly, when evaluating the efficacy in reducing nitric oxide release, we observed that strophanthidin and ghrelin were able to counteract LPS, while TXA did not modulate the LPS mediated damaging effect. On the other hand, ghrelin or TXA did restore metabolic parameters altered by LPS exposure such as metabolic activity and calcium levels, while strophanthidin did not. A significant increase in the production of total ROS was observed in LPS challenged cardiomyocytes. This increase was significantly inhibited by concomitant exposure to ghrelin, while no statistically significant variation was observed with either strophanthidin or TXA.
In the cardiogenic model, OGD exposure increased the percentage of dying cells, but this effect was counteracted by concomitant exposure to strophanthidin and ghrelin. Concerning metabolic activity, none of the tested inhibitors was able to counteract the 20% reduction observed after LPS exposure. On the other hand, LPS induced an increase in intracellular basal calcium levels and this effect was counteracted significantly by all of the tested inhibitors. For what concerns the oxidative stress, a significant increase in the production of total ROS was observed in LPS challenged cardiomyocytes. This increase was significantly inhibited by concomitant exposure to ghrelin, while no statistically significant variation was observed in cardiomyocytes challenged with LPS in the presence of either strophanthidin or TXA. Finally, for what concerns the LPS-induced increase in NO release, this effect was significantly inhibited by the presence of either strophanthidin or ghrelin.
Overall, the in vitro models developed in this study yielded several results comparable with the data obtained in the Omics studies of patients’ biopsies, such as activation of selected inflammatory pathways, vascular inflammation and vasodilation. Moreover, the platforms developed in this project represent a valuable tool for the in vitro characterization of the mechanism of action of novel molecules aimed at targeting selectively either cardiogenic or septic shock damage.
9 CONCLUSIONS
The ShockOmics Consortium faced several challenges specific to the addressed syndrome and to the need of integrating many disciplines. First of all the need of a timely activation of recruitment in emergency departments and soon after the unpredictable onset of shock. Secondly, the difficulties in a “stable” reproduction in animal experiment of a condition, which is by definition unstable. Finally, a tightly serial workflow from biosample collection and HD recording to the wet-labs for the Omics and the data processing units, and next to data mining and integration.
Despite the above, the ShockOmics project was able to reach most of the original goals, although with a wealth of bio-samples and data still to be processed. So, the overall ensemble of reached results and stored repositories should be considered as project achievements.
The integrated approach followed has cast new light both relevant to the mainstream hypothesis of autodigestion and collateral ones related to inflammatory pathways and those addressing the energetic metabolism. The latter ones opening a new frontier towards biomarkers for the early detection of vital organs (e.g. the liver) and to more effective therapeutic treatments. A major issue has been the congruency of such results given by a comparison of the outcomes from different Omics, looking at diverse steps of the biomolecular chains. This overall view was generally achieved starting from metabolomic findings and next focusing the correlates in transcriptomics and proteomics, as well as in-vitro experiments.
A further broad field opened concerns the massive circulating peptides analyzed by peptidomics. Firstly, the individuation of trypsin as main protein degradation source strongly supports our working hypothesis based on autodigestion triggered by the intestine damage, thus fostering efforts towards therapeutic protocols based on protease inhibitors such as TXA. Secondly, the bulk level of peptides was confirmed as a powerful tool for the evaluation of progressive organ damage and MOF. Finally, the issue is risen relevant to the bio-activity performed by the anomalous concentrations of specific peptides, which eventually can amplify the ongoing organ failures.
As underlined in the next section, the summarized results, together with the data processing and integration methods, and the progresses in therapy experiment relevant to intestinal delivery of protease inhibitors represent a consistent potential impact of the ShockOmics project outcomes towards enhanced risk prediction and timely therapeutic actions in the shock syndromes.
Potential Impact:
1 SOCIAL IMPACT
The ShockOmics consortium addressed new insight on CV shock, a potentially fatal syndrome for which current therapies are directed at symptoms, rather than at the prevention of the onset of the pathology.
Considering the starting context, it is obvious that acute heart disease in shock and ensuing hemodynamic instability are a major challenge for intensive care physicians, since current therapies fail when progressive shock is started. The approach proposed by ShockOmics addressed several fundamental questions that have not been answered so far thanks to cutting edge tools and a multidisciplinary team. Although the analyses of clinical datasets and the parallel experiments have been at a stage of basic biomedical research, still the translational potentials of the findings have been emphasized towards improved risk assessment tools and therapeutic strategies and techniques.
The ShockOmics consortium had been built to achieve in overcoming the main barriers limiting research in the field. The Omics methodologies for the experimental investigation of diseases and by means of the innovative biotechnologies nowadays available offered the technical support for a multiscale study that aimed at relating system level measurements, commonly available in clinics, to the basic biomolecular pathways of interest.
Our goals stemmed from emerging evidences, provided in the literature, for instance, by the “Autodigestion Hypothesis”, previously formulated by the group of Dr. Geert W. Schmid-Schönbein, UCSD, insured a strong theoretical background to formulate the fundamental hypotheses of the proposed studies. The ShockOmics consortium included scientists with very diverse and clearly complementary expertise, since the joint effort of medical doctors, biologists, computer engineers, bioengineers, and small entrepreneurs was necessary to address the complexity underlying the questions and the goals of ShockOmics.
State-of-the art techniques and methods for the study of biomolecular pathways and CV regulation as well as drug delivery technologies supported the achieved advancements in the knowledge and treatment of circulatory shock.
Several evidences demonstrate the relevance of shock and its associated complications, HF included. Trauma resulting from accidental injuries is the leading cause of death in individuals aged 1–44 yr in the United States of America, and hemorrhagic shock, for instance, affects 36–39% of trauma victims. Many develop multiple organ failure (MOF), the leading cause of death in the intensive care unit.
The management of shock patients and sepsis shock patients has been the object of a large clinical study(Caironi et al. Albumin replacement in patients with severe sepsis or septic shock. N. Engl. J. Med., 2014), which estimated a mortality rate up to about 45%. Such numbers confirm that clarifying the molecular mechanisms of shock is of paramount urgency in order to find effective therapies.
The above evidence is highlighted by the major medical societies; e.g.: the American Heart Association, the European Society of Critical Care Medicine, and the Society of Critical Care Medicine. Moreover, the costs of hemorrhagic shock in US have been estimated to spam from $51,000, for a trauma-patient with uncomplicated shock, to $321,000 for a severe hemorrhagic shock with MOF.
Besides the relevance of the acute cardiovascular illness per se, the social and financial costs of the disease are evident also considering the long term impact on shock survivors, who have displayed a high two year mortality (44.9%) in the case of septic shock and a lower quality of life, as long as 1.5 yrs. after their ICU discharge.
For all these reasons, the relevance of advancements in the comprehension of shock mechanisms and the consequent therapies has highly significant social and economical implications. Namely the following achievements are highlighted for the advancement in the field of personalized medicine and potential social impact:
– Omics technologies (metabolomics, proteomics, peptidomics, and transcriptomics) were tested for the timely identification/recognition of circulating biomarkers of pathogens potentially responsible for MOF progression and hemodynamic instability. The focused biomolecules and pathways still need further experimental and clinical research to reach the level of potential biomarkers and be challenged in clinical studies; nonetheless, a road map is open, starting for further analyses over the collected bio-banks and data sets.
– A new potential protocol for the timely and efficient delivery of therapy to prevent the evolution of proteolysis and systemic inflammation in shock starting from the intestinal compartment was tested. Namely, starting from previous trials in open abdomen small animals experiments of naso-gastric access in large animals were carried out, thus progressing towards a future application in the ICU.
– Software platforms were developed for the integration of new, systems biology based, multiscale models for the analysis of hemodynamic stability in critical care monitors. Models exploiting the multivariate beat-to-beat variability series, drawn from current ICU monitoring, were translated in algorithms providing indices of CV regulation impairment. Data-mining approaches on multiscale analysis outcomes (clinical, hemodynamic, and Omics) showed enhanced risk prediction value, compared to separate analyses.
– The ShockOmics database is available for further studies by the Consortium Partners and, upon request, by accredited scientific institutions. It will include the HD recordings and the derived beat-to-beat series, as well as the lists of biomolecular analyses (metabolites, proteins, peptides, transcripts). Storage will be maintained on a repository, while sub-sets will be replicated on public repositories of the specific disciplines (e.g. proteomics), when applicable. Both clinical (fully anonymized) and animal studies will be included.
– Benefits to the involved SMEs and to the European market where highlighted in three main areas:
1. Risk prediction algorithms based on: a) easily measurable endpoints, which could be integrated into existing cardiovascular monitors, b) integration of Omics analyses from blood/plasma samples and clinically available parameters;
2. Bases for the future development of diagnostic-kits targeting the potential biomarkers among the set of biomolecules shown to predict the risks related to shock and organ failure;
3. Adaptation of the existing implantable drug delivery technology to better target shock-therapy was pursued concerning: a) design of an infusion catheter (patented) specific to infusion in the small intestine lumen; b) insertion techniques in the intestinal lumen through naso-gastric way and the pylorus; c) design of smart pumps controlling both infusion flow and pressure.
– Finally, the international profile of ShockOmics through three European Community Countries plus Switzerland and the US, together with the high level of interdisciplinary synergies, fostered cultural and scientific integration, as testified by the personnel interchange and dissemination activities.
The theme of the intellectual property is of great importance for the translational implications of ShockOmics and to deliverables were respectively dedicated to market searches and to exploitation plans in the three above mentioned areas. The latter, in full agreement with the IPR management rules set in the Consortium Agreement.
2 DISSEMINATION ACTIVITIES
Several measures were adopted to maximize the dissemination of the scientific results obtained from the experimental studies carried out during the four year duration of ShockOmics, besides the traditional publications on peer reviewed journals and presentations at scientific conferences.
As regards the journal publications, the partners of the consortium were committed to ensuring the open access to their publications and data, in line with the previously listed impacts inherent to the open access and open source availability of data and mathematical resources and tools.
The presentation of preliminary results to the most important international medical, biological, and engineering conferences on the topic of interest of ShockOmics was also integrated by the organization of special sessions and mini-symposia within such congresses. Given the innovation of ShockOmics and the relatively unfocused research on the multiscale phenomena which characterize acute HF in shock, discussion was fostered within the scientific community. In this way, besides giving visibility to ShockOmics, other scientists interested in shock and acute heart disease were involved in the important research field explored by ShockOmics.
Dissemination through conferences and full papers on indexed international journals is still ongoing. So far, 41 conference presentations (poster or podium) were performed while the published full papers were 14 on the following journals: Medical and Biological Engineering and Computing, Shock (two papers), Briefings in Bioinformatics, Scientific Reports (two papers), Scandinavian Journal of Trauma, Resuscitation and Emergency Medicine, Journal of Clinical Monitoring and Computing (two papers), Computational and Mathematical Methods in Medicine, Physiological Measurements, Journal of Clinical Monitoring and Computing, Experimental Biology and Medicine.
ShockOmics also proposed a specific international workshop, embedded in a small though highly focused conference in the field; i.e. the International Conference in Complex Acute Illness (ICCAI, Milan, Italy, July 27-29, 2017, iccai.org ), which was the 16th annual conference of the homonymous society, SCAI. The ShockOmics coordinating unit was in charge of hosting the conference and this gave a unique opportunity in opening to the wide audience of the major societies in the fields of Shock and Intensive Care, to which the conference call was addressed. The invitation of renowned keynote speakers permitted the contact with major stakeholders. Moreover, the SCAI board of members and the ICCAI’17 scientific committee also included reference scientist in the field and the Editors of major journals such as Shock, Critical Care in Medicine, the Journal of Critical Care, and Physiological Measurements.
This event gave also wide visibility on the scientific press since the Journal of Critical Care kindly accepted to publish the 9 abstract presented in the ShockOmics workshop sessions preceded by an editorial providing a general overview of the project close to its end (Ferrario M, Aletti F, Baselli G, The EU ShockOmics Project International Workshop at ICCAI'17, J. Critical Care, 2017).
The website represented and will represent in the next future a unique way of communication with the public at large, beyond the borders of the scientific community, open access publications related to ShockOmics will be periodically updated.
Dissemination initiatives aimed at undergraduate and graduate students were undertaken including the recruitment to the laboratories involved in ShockOmics for internships, BSc projects, MSc and PhD thesis.
The researchers of ShockOmics consortium actively participated to other well-known initiatives, such as the White Night of Researchers, which regularly takes place across the Continent, to increase the awareness of European citizens towards the importance of research, development and education for enhancing the quality of lives.
List of Websites:
Project website: www.shockomics.org
Main contact: Project Coordinator - giuseppe.baselli@polimi.it
The management of hemodynamic stability in shock patients is of paramount importance in the intensive care unit (ICU). New, cutting edge knowledge is necessary to overcome the shortcomings of available therapies, in order to cope with the challenges that anesthesiologists, intensive care specialists and emergency physicians face when dealing with shock patients.
Current therapies are targeted to reduce symptoms of shock and multiple organ failure, but they are unable to target the root cause or to act at the “beginning of the cascade”, because of the lack of a model explaining the molecular basis of shock-induced tissue injury and ensuing multiple organ failure. Hence, the effectiveness of anti-hypotensive interventions such as fluid resuscitation is limited. Fluid may restore blood pressure within minutes, but complications such as pulmonary edema may arise on a longer time scale in patients who are unresponsive to fluid infusion, and cause more severe stress to the patient’s hemodynamic stability, respiratory efficiency, and immune system.
The ShockOmics project focused on the molecular triggers of acute heart failure in association to shock, in the presence of uncontrolled proteolytic activity, in order to identify inflammatory mediators and markers which are activated in shock, through a systematic analysis of expression levels of genes and their protein products, and novel targets for the delivery of new therapies.
In this framework, several biosample and data sources, integrated by clinical records and hemodynamic monitoring: i) plasma samples from existing biobanks; ii) clinical studies in septic shock (SS) and cardiogenic shock (CS) patients in ICU sampled at T1<16h, T2=48h, T3=7d; iii) animal shock models include hemorrhagic shock (HS) rats, HS swines, SS swines, and CS swines; iv) in-vitro cardiomyocytes shock models.
Analyses spanned all Omics technologies (metabolomics, proteomics, peptidomics, and transcriptomics) addressing the circulating blood/plasma and, in animals, post-mortem samples of the myocardial tissue. Omics outcomes were integrated with clinical indexes and parameters relevant to CV regulation from hemodynamic monitoring. Artificial intelligence and data-mining methods were applied to work out risk prediction algorithms.
Therapeutic advancements were pursued starting from the “autodigestion hypothesis”, which focuses the detrimental proteolytic effects from digestive enzymes leaking from the intestinal compartment via the wall damaged by the ongoing shock. This pushed to confirm previous experiments on rats both in the same species and in the swine shock models. The latter experiments asked to further develop technologies for intestinal infusion through the naso-gastric access beyond the pylorus by a dedicated infusion line and smart control pumps.
Plans for the future exploitation of the project outcomes address three areas: i) risk prediction algorithms to be integrated into existing cardiovascular monitors or addressing the Omics outcomes; ii) future development of diagnostic-kits addressing the potential biomarkers; iii) the drug delivery technologies into the intestinal lumen.
Project Context and Objectives:
GENERAL CONTEXT
Circulatory shock is a major complication in patients admitted to the ICU. After a sufficiently long impairment (order of hour/hours) of oxygen and nourishment transport to organs, damage and inflammation of vasculature and organs may start a cascade of adverse events, which may increase out of control and lead to fatal outcomes. Importantly, shock may occur despite the recovery of cardiovascular stability reached in the ICU by assistive and pharmacological treatment and resuscitation maneuvers.
Hence, new cutting edge knowledge is necessary to assist anesthesiologists, intensive care specialists and emergency physicians in the timely delivery of effective therapies to reduce the incidence of mortality and morbidity in critically ill patients in shock. The multiscale approach of the ShockOmics project was to gain an integrated view at both biomolecular and hemodynamic level aiming at more accurate and precocious biomarkers as well as novel therapeutic approaches.
The relevance of the disease is indicated by epidemiological data. The Italian clinical study on sepsis “Albios” has shown a mortality of 45% in septic shock patients (http://www.negrisud.it/albios/). Trauma resulting from accidental injuries is the leading cause of death in individuals aged 1–44 yr in the U.S.A. and hemorrhagic shock affects 36–39% of trauma victims. Most of shock patients develop multiple organ failure (MOF), or multiple organ dysfunction syndrome (MODS), the main cause of death in the ICU, a unique syndrome in which organs not directly injured in the original trauma become dysfunctional. Moreover, in the case of the heart, preserving its function is instrumental for ensuring hemodynamic stability and an adequate perfusion of all vital organs.
Four groups of shocks have been defined: hypovolemic shock (e.g. hemorrhagic shock (HS)), cardiogenic shock (CS) (which is developed following myocardial infarction), obstructive shock, and distributive shock (e.g. septic shock (SS)). All types of shock are characterized by low blood pressure and accelerated heart beat.
The relationship between shock, ischemia and reperfusion (I/R) injury, hemodynamic instability, systemic inflammatory response syndrome (SIRS), and MOF/MODS has been extensively investigated, but there is no consensus on the trigger mechanisms of tissue injury at the molecular level. Inflammation induced by shock has been described at the microvascular level, as a tissue repair reaction to injury, characterized by cellular dysfunctions, microvascular perfusion deficiencies, autoregulation impairment, capillary stasis and augmented permeability, coagulation, tissue edema, leukocyte-endothelial interaction, oxygen radical free formation, apoptosis, metabolic defects, hyperglycemia, insulin resistance. Biochemical markers for inflammation, e.g. the expression of cytokines, indicate a tissue repair reaction to injury.
Several therapeutic countermeasures against circulatory shock have been proposed, such as endotoxin or cytokine blockade, endotoxin binding protein receptor, blockade of leukocyte adhesion molecules, complement depletion, oxygen free radical scavengers, nitric oxide modulators, hypertonic resuscitation, designer crystalloids and hemoglobin carrying solutions that preserve the fluid shear stress on the capillary endothelium. However, none of the above approaches represents a clinically effective intervention. Current therapies are targeted to reduce symptoms of shock/MODS, but they are unable to target the root cause or to act at the “beginning of the cascade”, because of the lack of a model explaining the molecular basis of shock induced tissue injury and ensuing MODS. As a consequence, the effectiveness of anti-hypotensive interventions such as fluid resuscitation is limited and depends on the trade-off between their short-term beneficial effect and their potential long-term risk.
Fluid administration may restore blood pressure within minutes, but serious complications such as pulmonary edema may arise on a longer time scale as a consequence of patients’ unresponsiveness to excessive fluid infusion, and cause more severe stress to the patient’s hemodynamic stability, cardiorespiratory efficiency, and immune system.
OBJECTIVES
The main goal of ShockOmics was to investigate the molecular triggers of acute heart failure (HF) induced by shock and to identify inflammatory mediators and markers activated in shock, with a particular emphasis on the role of uncontrolled proteolytic activity.
To achieve this goal, a systematic analysis and multiscale integration of information from the gene (transcriptomics), peptide (peptidomics), protein (proteomics), and metabolite (metabolomics) levels was carried out. The ultimate aim of the proposed multiscale approach is to interpret the patho-physiological changes in system level variables, i.e. cardiovascular (CV) measurements, which are routinely available in emergency departments, ICUs, and operating rooms to describe the hemodynamics of patients, in function of the fundamental processes of disease.
System level symptoms of shock are well known. Hypotension and hemodynamic instability are the main issues that clinicians have to cope with when treating shock patients. In the context of low blood pressure, low flow, and reduced ventricular contractility typical of circulatory shock and hemodynamic instability, the in-depth monitoring of heart function is essential to prevent further hypoperfusion of organs. Thus, ShockOmics focused on acute heart failure due to shock, and explored the presence of biomarkers, which could be related (but not limited to) to the role of the intestine as a source of powerful inflammatory factors in shock, which are responsible for uncontrolled proteolytic activity.
Summarizing, the specific objectives of the ShockOmics project were:
– Identification of candidate biomarkers from pre-existing and prospective populations of patients, by means of cutting edge Omics techniques;
– Assessment of heart injury in response to uncontrolled proteolytic activity, which may cause cleavage of important receptors (e.g. insulin, adrenergic and VEGFR2 receptors), in animal models of shock, and in in vitro studies;
– Analysis of gene expression and protein expression by means of state-of the-art high-throughput techniques, such as Next Generation Sequencing for whole transcriptome analysis and shotgun proteomics, to characterize the cellular response to shock due to proteolysis in heart tissues and cells obtained from animal experiments and in vitro studies, from peripheral blood from in vivo studies, and from patient blood samples;
– Development of multiscale, systems biology based models to describe the relationship between hemodynamic measurements/waveforms available in ICU and the progression of the shock induced acute heart disease;
– Identification of novel targets for effective cardioprotective therapies to prevent/contrast shock induced cardiovascular instability;
Refinement and further validation of highly innovative technologies for the early identification of biomarkers of the disease, for the interpretation of hemodynamic alterations, and for a timely and efficient drug delivery for the prevention of hemodynamic instability due to shock induced HF.
CONSORTIUM
The ShockOmics Consortium involved 11 participants from 4 nations: Italy, Spain, Belgium, Switzerland and USA.
The partners have multidisciplinary integrated expertise:
1. Politecnico di Milano – POLIMI- Milan (Italy): Hemodynamics and multiscale modeling
2. Filarete Servizi srl – FIL- Milan (Italy): Transcriptomics and proteomics
3. Istituto di Ricerche Farmacologiche Mario Negri –IRFMN- Milan (Italy): Bio-banks, metabolomics, animal experiments on hemorrhagic shock
4. Neurozone s.r.l. –NZ- Milan (Italy): In vitro studies
5. Custom Software Electronics –CSE- Barcelona (Spain): Data base development, data mining
6. Hospital Universitari Mutua Terrassa –MDT- Terrassa (Spain): Access to bio-banks
7. Parc Científic Barcelona – PCB- Barcelona (Spain): Proteomics
8. Genève University Hospital – UNIGE- Geneva (Switzerland): Clinical study and animal experiments on cardiogenic shock
9. Université Libre de Bruxelles –ULB- Brussels (Belgium): Clinical study and animal experiments on cardiogenic shock
10. University of California San Diego –UCSD- San Diego (USA): Animal experiments, bio-signaling
11. Antlia –Antlia- Lausanne (Switzerland): development of pumps for drug delivery.
Figure 1 summarizes the main activities in which the partners were involved.
Project Results:
1 COLLECTED DATA AND BIO-SAMPLES
1.1 Clinical study in ICU patients
The ShockOmics clinical trial (ClinicalTrials.gov identifier NCT02141607) was a multicenter prospective observational study.
The main inclusion criteria were:
– Sequential Organ Failure Assessment (SOFA) score > 5
– First blood sample and first hemodynamic measurements available within 16 h from admission to the ICU
– Informed Consent available
The main exclusion criteria were:
– Risk of fatal illness and death within 24 h
– Patients already enrolled in other interventional studies
– More than 4 units of red blood cells transfused
– Patients receiving plasma or whole blood
– Active hematological malignancy
– Metastatic and/or active cancer
– Immunodepression, including transplant patients; patients infected by the human immunodeficiency virus (HIV+); constitutive immune system deficiency; any immunosuppressive therapy, including oral and parenteral corticosteroids (aerosols are allowed)
– Patients with pre-existing end stage renal disease needing renal replacement therapy. The introduction of continuous veno-venous hemofiltration from the day of admission onward was allowed
– Cardiac surgery in the previous ten days
– Child-Pugh C cirrhosis or acute liver failure
– Terminal illness
Data collection timeline was:
– Time 0 (T0): admission to the ICU with diagnosis of shock or time of shock diagnosis in patients initially admitted without shock symptoms;
– Time 1 (T1): time at which the first blood sample for analysis is collected, within 16 h after T0;
– Time 2 (T2): time at which the first blood sample for analysis is collected, at 48 h after T0. At this time point, the treatment has been administered for a long enough amount of time to evaluate its effects;
– Time 3 (T3): time at which the third blood sample for analysis is collected, on day 7 of the ICU stay of the patient or before discharge from the ICU in case of shorter stays or before discontinuing therapy (death).
– Time 4 (T4): follow up on ~ day 100 from Time 0 in non re-hospitalized patients; this follow up, including a further blood sampling, was severely limited.
Out of 529 patients screened for eligibility, 41 septic shock patients, 22 cardiogenic shock patients, 9 sepsis patients not in shock, 3 cardiac patients not in shock were enrolled in the study.
1.2 Experimental studies in swine shock models
Swine shock models addressed three different insults leading to shock onset: i) septic shock (ULB) was induced by cecal ligation puncture; ii) hemorrhagic shock (IRFMN) was induced by controlled bleeding; iii) cardiogenic shock (UNIGE) was induced either by sustained ventricular pacing or by coronary artery occlusion (CAO).
The following sampling times were defined in order to consider the different phases of each experiment comparable:
– T1: Baseline, before the start of the insult to induce shock;
– T2: Two hours after the beginning of shock and ahead of resuscitation maneuvers;
– T3: After fluid and vasopressor resuscitation;
– T4: After blood resuscitation.
Autopsy and myocardial tissue sampling followed the animal sacrifice.
As scheduled, 6 SS and 6 HS animals were collected with the complete set of data (blood samples, heart tissues and hemodynamics recordings) together with 6 sham animals.
Severe limitations were encountered in the CS experiments due to the high rate of either a spontaneous recovery from AHF or animal loss ahead of experiment completion. Sustained pacing never induced stable CS conditions, while CAO permitted the collection of 2 CS cases only, after which the experiments were stopped due to not renewal of ethical committee approval for further animals. Samples from 3 sham animals and 2 CS were available.
1.3 Treated swine shock models
Tranexamic acid (TXA) is well known in clinics as intravenous TXA is an effective adjunct after hemorrhagic shock (HS) because of its antifibrinolytic properties. TXA is also a serine protease inhibitor, and recent laboratory data demonstrated that intraluminal TXA into the small bowel inhibits digestive proteases and protects the gut. The developed techniques for pilot intestinal infusion (multi lumen infusion line, regurgitation prevention, pump control, etc.) are better described in the dedicated paragraph of results section.
3 SS and 5 HS animals were treated with TXA by enteral infusion; 3 further animals treated with enteral infusion of vehicle only were considered as controls. CS was not considered for the above mentioned reasons.
2 CARDIOVASCULAR REGULATION
We selected a subset of 21 septic shock patients enrolled at Genève University Hospital from October 2014 till December 2015. All participants gave prior informed consent. Strict inclusion criteria were observed during the recruitment phase, in order to avoid a too high inhomogeneity within the population.
The patients received initial therapy according to the standards immediately after shock diagnosis (time T0). Patients were analyzed at two relevant time points: within 16 hours from T0 when the inflammatory cascade has been just activated (T1), and within 48 hours after T0 time point (T2).
SOFA score (Sequential Organ Failure Assessment score) was used to classify patients into two groups according to their responsiveness to early therapy: responder patients (R, n=14) consisted in patients with a positive response to initial treatment. Those patients, who decreased their SOFA score of at least 5 points between T1 and T2(ΔSOFAT2-T1>5) or reached a SOFA score at T2 lower than 8 were classified as R. Non responder patients (NR, n=7) consisted in patients, who still had a SOFA score at T2 higher than 8 or ΔSOFAT2-T1<5.
All patients had similar severity of shock at the enrollment and received the same initial therapy, but the R group improved their condition within the first 2 days after ICU admission (i.e. between T1 and T2), whereas the NR patients did not improve or even worsened.
NR patients displayed a higher fluid balance during the first days after shock development due to a larger fluid infusion and a reduction in urine output
Arterial pressure (AP) parameters showed an increase between T1 and T2 in the R group: moreover, at T2, R patients had significantly higher values than NR in both mean AP (MAP) and systolic AP (SA). All AP parameters were slightly decreased in NR patients between the two time points. A significant decrease of heart rate (HR) mean value from T1 to T2 was reported in NRs.
In our study the feedback gain of the baroreflex (BR) was quantified by the BR sensitivity (BRS), based on the relationship between SAP and HR mediated by the nervous system. The BRS increased significantly at T2 in NRs. If cardiopulmonary (CP) afferent activation indirectly affects the BR control of heart and peripheral resistance, then the elevated loading (or stretching) of CP baroreceptors, due to large central volume pressures, could be a possible explanation of the different trend in BRS observed in NR patients with respect to R patients whose fluid balance was significantly decreased at T2.
Moreover, the overall variability in the NR group was significantly lower with respect to the R group. All patients were able to maintain their MAP over the targeted threshold of 65 mmHg, even if the trend between T1 and T2 in NR patients was slightly decreasing and they displayed lower values with respect to R patients both at T1 and T2. Looking at the international guidelines for acute shock resuscitation, all these patients seemed to be responsive to the therapy.
However, the hemodynamic analyses allowed to discriminate those patients who actively responded restoring the autonomic nervous system regulation of BP and HR, and those who recovered from hypotension without a marked improvement in cardiovascular autonomic control.
Finally, it is a common evidence that patients in SS receive excess fluids even after they are weaned from vasopressors. In our opinion, this happens because indices relating to CV autonomic control are not commonly used to guide fluid therapy and our work is meant to contribute in this direction. Hence, CV regulation indices offer acnon-invasive options to guide fluid therapy and permit to assess the hemodynamic response of patients to the therapy.
These findings were confirmed in HS swine experiments, showing that the evaluation of the autonomic nervous system could be valuable in determining the effectiveness of the therapy and the grade of recovery.
Briefly, baseline measurements were similar in both HS and sham pigs. During HS, concurrently with MAP reductions, there was a significant decrease in CO (1.7±0.3 L/min vs. 3.2±0.7 L/min, p<0.01) and LV end diastolic and systolic volumes (p<0.01). Heart rate (HR) showed a 3-fold increase compared to baseline (217±5 vs. 77±16 b/min, p<0.01) and pulse pressure variation (PPV) was > 35 mmHg (p<0.01). Arterial lactate increased above 8 mmol/L in each HS pig (p<0.01). After resuscitation with fluids and vasopressor support, hemodynamics and LV volumes improved, although did not return to baseline, i.e. MAP was 30% lower, HR 2-fold greater, and lactate was still above 6 mmol/L (p<0.05). Only after reinfusion of the shed blood, all the hemodynamic variables, CO, LV volumes, PPV, and lactate completely recovered. As expected, all the above variables remained stable throughout the experiment in the sham animals (p<0.01 vs. HS).
LF oscillations in diastolic arterial pressure (DAP) series in shock were lower than baseline and did not recover after the first resuscitation. Baroreflex absolute gain was greatly reduced with respect to baseline; the opposite trend was observed in the feedforward gain (i.e. the AP response to HR changes). The same results were found for the variation in BRS and feedforward gains after a vasoconstriction maneuver by a phenylephrine bolus. Finally, DAP variability explained by the BR control increased in shock condition, whereas the HRV component was reduced. Only after reinfusion of the shed blood all the hemodynamics parameters recovered to baseline values.
3 METABOLOMICS
3.1 Clinical study
The plasma metabolome of all patients enrolled in the ShockOmics project was successfully analyzed by mass spectrometry-based metabolomics strategies.
By combining mass spectrometry-based untarget and target metabolomics approaches , many dyseregulated metabolic pathways ranging from energy and lipid metabolism to amino acids synthesis were observed in shock patients..
We made a comprehensive metabolomics study of septic shock patients stratified as responders (R) and not responders (NR) to standard therapy on the basis of the changes in SOFA score in the first two days in the ICU.
Changes in the levels of metabolites over time were shown to discriminate positive response to therapy. Metabolomics data elaboration by datamining techniques revealed that the metabolic trajectory during the first 48 hours after ICU admission in NR patients was characterized by significant lipidome alterations.
The reduction in circulating lysophosphatidylcholines (LysoPC) in NR patients might reflect their enhanced conversion to lysophosphatidic acid, which induces a multitude of cellular responses through its action on immunologically relevant cells and might also promote an excessive immune response, with detrimental effect in NR patients. NR patients also showed a marked decrease in phosphatidylcholine species (PC), which originate in the liver. The imbalance of LysoPC/PCs cycling suggested that hepatic homeostasis and function is compromised even before any clinical manifestation.
Indeed the higher plasma level of alanine found in NR vs R may be a sign of lower hepatic capacity for conversion of alanine to glucose and suggests a different approach for monitoring hepatic function, which will be more specific than bilirubin.
Therefore these findings are really promising as they could pave the way for tailored therapies.
These results have been published in Cambiaghi et al. Scientific Reports 2017, Aug 29;7(1):9748. doi: 10.1038/s41598-017-09619-x.
Variation of metabolite plasma concentration between acute phase (T1) and short temporal window after diagnosis (T2) {Δ = T2-T1} was analyzed in cardiogenic shock patients. Again lipid species including LysoPC and PC were deregulated over time. Further statistical analyses were hampered by the small sample size of cardiogenic shock patients-
Nevertheless, machine-learning techniques were applied for the identification of metabolomics changes able to cluster the two types of shocks in patients during the first 24 hours of disease evolution.
The two shocks presented significant differences in the expression levels of PC, suggesting again that lipidome alteration is an important modulation in shock and the understanding of regulatory pathway of lipids is thus crucial for the development of an effective and tailored therapy.
3.2 Swine studies
Metabolomics analyses of heart and plasma samples from swine studies were successfully completed.
So far, the most soundness results were obtained elaborating the plasma metabolomics data from septic shock swine untreated model.
Again, circulating lysophosphatidylcholine species (LysoPC) had reduced level in septic shock swines compared to their control counterpart. Plasma alanine concentration increased in septic shock swines compared to controls.
These findings corroborate what we previously found in the clinical settings and highlight the role of LysoPCs level reduction that may promote an excessive immune response with detrimental effect.
Moreover, the confirmation that also in the septic shock swine model there is a significant enhance in circulating alanine, further supports the role of alanine in early dysregulated liver gluconeogenesis. Interestingly, the Fisher ratio, indicator of liver damage, is significantly decreased after resuscitation in septic shock swine only. Furthermore, citrulline concentration and the metabolic ratio citrulline/glutamine were significantly enhanced in the plasma of septic shock swine compared to their control counterpart. In the liver, citrulline functions as part of the urea cycle in the detoxification of ammonia and, because of the channeling of urea cycle intermediates, little or no citrulline escapes the liver. The citrulline produced in the small intestine, however, enters the portal vein and appears in the peripheral circulation serving as precursor for arginine synthesis. As such, these preliminary results converge towards an early dysregulated intestinal-liver axis.
In conclusion, significant results have been produced showing the same plasma metabolomic trajectories in Septic Shock patients and Septic Shock swine model, suggesting that changes in specific constellations of circulating metabolites might be early plausible indicators of (i) deranged immune response (i.e. lipidome alteration) and (ii) dysregulated intestinal-liver axis (i.e. alanine, citrulline concentration changes).
4 PROTEOMICS AND PEPTIDOMICS
4.1 Clinical study
Proteomics and peptidomics analyses were performed in both clinical and animal model samples.
The clinical study involved analysis of samples from all patients enrolled in ShockOmics, except hemolyzed samples that could not be included. We analyzed the first three time points (T1, T2 and T3) of septic shock (SS) and cardiogenic shock (CS) patients. Samples from healthy volunteers (H, n=9), septic controls (SSC, n=9) and cardiogenic controls (CSC, n=3) were used as controls.
As explained above, SS patients were classified according to their responsiveness to early therapy: responder (R, n=16) and non-responder (NR, n=13). For the cardiogenic study, the response criterion was also based on SOFA score, but excluding the cardiovascular score from total SOFA calculation. Thus, responders (R, n=6) were defined as individuals showing ΔSOFA (T1-T2>3) or (T1-T3>6). For CS the percentage of non-responders (NR, n=14) was higher than in SS.
The proteomics results obtained in the SS study showed a significant deregulation of proteins related to hydrolase and peptidase activities. Nevertheless, we did not detect significant differences between septic control and SS at T1, the first 16h after ICU admission. On the other hand, we observed a group of deregulated proteins that were common for SS groups (R and NR) at T1 and septic control, when compared to healthy condition. The network formed by these deregulated proteins is enriched in proteins related to proteolysis. In addition, deregulated proteins that were only observed in SS formed a network enriched in proteins related to response to stress. The comparison of R and NR groups showed a significant up-regulation of proteins related to the regulation of the immune system and complement and coagulation cascades in NR group.
We observed higher variability among individuals of the CS groups. A functional analysis performed in the proteins with significant changes between R and NR patients showed an enrichment in proteins related to heart diseases. Most of the deregulated proteins detected in the comparisons between R and NR patients of SS and SC are specific for the disease. From proteomics results, few overlap were observed between SS and SC, although more in deep statistical analysis are still in progress.
The peptidomics results for the SS study showed a clear increase in the number of peptides that can be correlated to the increment of proteolysis in the SS group at T1 when compared with sepsis controls and healthy controls.
For the peptidomics analysis we also classified the patients according to survival (Survivors, S, n=23; Non-survivors, NS, n=6) at day 28. The highest values in terms of number and abundance of peptides were observed at T2 in NS. In addition, the number of peptides at T3 was lower than at T1, except for NS. The peptidomics analysis also allowed differentiating patients according to the location of the infection. The patients with source of infection in the urinary tract presented lower mean abundance of proteins than those patients with abdominal or respiratory source of infection.
The peptidomics data from cardiogenic shock patients corroborates the septic shock data and points to a dysregulation in proteolysis in shock, which supports the ShockOmics hypothesis.
4.2 Animal studies
The animal study involved peptidomics and proteomics analyses of samples from rat and pigs.
Preliminary experiments on plasma obtained from control and hemorrhagic rat models were performed by using a shotgun label-free quantitative approach. Overall, the data obtained confirmed that massive proteolysis occurs as a fundamental degrading phenomenon induced by shock and that chymotrypsin like proteases are mainly involved. In particular a huge increase was observed in terms of peptides and proteins in shock in comparison to control.
Peptidomics was applied to verify if enteral protease inhibition delivery (EPID) is effective in preventing mortality and morbidity after HS. The peptidomic differences between animals EPID-treated and those that do not (untreated) allowed to determine those proteins and peptides that are associated and could be causal to increased morbidity and mortality. The main finding of this work is the protective effects against uncontrolled proteolytic activity of a protease inhibitor such as Tranexamic Acid. (TXA) The differences seen both in peptides number and in peptides intensity are particularly evident at the END stages, showing a marked decrease of peptides number and intensity values in subjects treated with TXA compared to untreated animals.
These very promising results led the EIAB steering committee to strongly recommend performing peptidomics analysis on pig and clinical plasma samples.
On plasma samples from hemorrhagic shock (HS) swine models the peptidomic results show that the increase of peptide number and intensity found in rat plasma samples upon HS is confirmed and that there is an evident difference between T1 and all the other time points.
In plasma samples from SS swine models, instead, the peptidomic data show that there was no a significant increase in peptide number upon SS opposite to what we observed in rat and pig in HS and in human patients in SS.
Other major findings using the peptidomic approach on plasma samples from rat and swine shock models were:
• no direct evidence of cleavage of specific membrane receptors, namely insulin receptors, β-adrenergic receptors and vascular endothelial growth factor receptor 2
• presence of potentially bioactive peptides with anti-hypertensive activity in rats and swine plasma samples
• the predominant activity of chymotrypsin–like proteases in shock plasma samples in rat and swine models.
The proteomic analysis carried out on plasma samples from rat and swine shock models allowed to obtain a direct experimental evidence of oxidative stress-induced post-translational modification (Tyr nitration) in the presence and in the absence of the protease inhibitor TXA in plasma samples from hemorrhagic shock while no direct evidence of oxidative stress-induced was detected in all the myocardic tissues analyzed (left ventriculum, right ventriculum and right atrium) from the same shock models. In the case of septic shock no evidence of oxidative stress could be detected either in plasma or in all the myocardic tissues analyzed (left ventriculum, right ventriculum and right atrium).
5 TRANSCRIPTOMICS
Transcriptomic analysis, using RNAseq technology, has been performed in all human and animal samples.
RNAseq is a powerful tool to analyze transcriptome, to quantify gene expression levels, to investigate and compare at molecular level specific pathological or physiological conditions, tissues, cell types, with the aim to identify genes differentially expressed.
5.1 Human study
We addressed two major aims. The first aim was the transcriptomic profiling in whole blood of septic shock patients along the clinical course (the three time points considered in the clinical protocol). This analysis showed at steady state, day 7 of the ICU stay, the downregulation of Toll-like receptors, C-type lectin receptors, IL-1 and IL18 receptors family proteins and alarmins suggesting the downregulation of innate immune response and acute inflammation. Conversely, at the same time point, the upregulation of genes of the adaptive immunity related to T and B lymphocytes activation was observed. A transcriptional regulation of genes involved in carbohydrate metabolism, lipid inflammatory pathway, transport of vesicles and protein synthesis was observed as well. These data have been presented in the Poster section at the 40th Meeting of Shock Society, Florida June 2017. Braga D. et al. Preliminary profiling of transcriptomic modifications in septic shock patients at multiple time points.
The second aim was the comparison of the transcriptomic profile in whole blood of septic shock patients defined as Responders or not Responders to early supportive therapy. This study showed that Responders specifically alter the expression of adaptive immune response genes (ICOS, LCK, VSIG4, upregulated genes related to functions of granulocytes and NK cells, and genes involved in pathogen lipids clearance (PCSK9, SORT1). Common functions were shared between R and NR, as the downregulation of genes involved in acute inflammation (alarmins, metalloproteases, C-type lectins, interferon receptors) and the upregulation of genes expressed in neutrophils and involved in bacterial killing (GZMM, GZMH, KLRG1). These results have been partially presented as poster at the Meeting of the Society of Complexity in acute illness (ICCAI), Milan, July 27-29, 2017 (Braga D. et al. Differential gene expression in septic shock patients according to the early supportive therapy response).
In cardiogenic shock patients, stratified as R or NR to early supporting therapy, the transcriptomic analysis showed that both groups modulate the expression of genes involved in neutrophils degranulation and immunoglobulin mediated immune response. Responders specifically modulate genes involved in lymphocytes differentiation and T cell activation, as observed in septic Rs. Instead, NRs specifically alter the expression of genes of the pattern recognition receptors signaling (PRRs), and genes involved in the inflammatory NFKB pathway.
5.2 Animal study
In septic shock untreated swines, most of the variability in whole blood gene expression was observed by comparing basal to shock condition.
In sepetic shock (SS), compared to the basal condition, genes that can counteract the infection were upregulated. These genes are involved in neutrophils chemotaxis, response to LPS or have antimicrobial activity (CD177, ALPL, LCN2, LTF, NPG1). A significant upregulation was observed for genes that act as negative regulators of protease activity and as inhibitors of serine protease (TNFAIP6, CSTA, RETN). MMP9 that is known to increase in the early phase of shock and whose levels correlate with severity and mortality in sepsis was significantly upregulated in the shock phase. At this time point the inflammatory mediators IL18 and IL5 were significantly downregulated. A downregulation was also observed for CD101 that plays a role in inhibition of T cells proliferation.
A pilot transcriptomic study has been carried on in whole blood of hemorrhagic shock (HS) rats. The transcriptome of shocked animals shows modulation of genes related to inflammation and immune response (Tlr13, Il1b, Ccl6, Lgals3), antioxidant functions (Mt2A, Mt1), tissue injury and repair pathways (Gpnmb, Trim72) and lipid mediators (Alox5ap, Ltb4r, Ptger2) compared with control animals. These findings are in line with previous observations reported in hemorrhagic shock using other Omics approaches: a change in expression of genes encoding lipid receptors and aminoacid metabolism is consistent with previous metabolomics studies, as well as the increased expression of genes with antioxidant function and of the complement pathway is in line with previous proteomic studies. For a more detailed description of these data refer to the paper “Braga D. et al. Preliminary profiling of blood transcriptome in a rat model of haemorrhagic shock. Experimental Biology and Medicine 2017;242:1462-1470.
6 INTEGRATED ANALYSES
In this activity we have analyzed the different datasets from the clinical studies conducted at the ICUs in Geneva and Brussels. It focused on clinical data, metabolomics, proteomics and transcriptomics with the primary objective of finding potential biomarkers for cardiogenic and septic shock. The secondary objective of these activities has been assessing the risk of death (RoD) of patients with shock.
In this regard, this activity has been mainly concerned with the methodological aspects of the Artificial Intelligence and Machine Learning techniques that we have used to discover hidden information (mainly interactions between clinical data, metabolites, proteins and transcripts) from the large amount of data generated during the project execution.
The methodologies used in our integrated analyses are mainly concerned with:
– Data visualization and dimensionality reduction.
– Kernel manifold learning (i.e. finding optimal hyperplanes/manifolds separating the two shocks considered in this project).
– Classification of shock and RoD from sequential data (i.e. from the different time points considered in the clinical study).
– Ensembles of networks (i.e. study the interactions between the different Omics data).
The integrated analyses have been organized as follows:
1. We developed the methodological framework for integrated data analysis.
2. Development of a pipeline for dimensionality reduction:
a. For analyses where the dimensionality is much greater (about two orders of magnitude) than the number of patients we used bagged ensembles of trees.
b. For other analyses where the dimensionality is not as great (about one order of magnitude) we used a pipeline based on kernel manifold learning.
3. After dimensionality reduction, we applied a generative kernel for data analysis (i.e. the simplified Fisher kernel). The main rationale for using such kernels is that they induce graphical models for the posterior assessment of interactions between attributes.
4. Once we had selected the most appropriate sub-set of clinical attributes / Omics data, we developed a baseline classifier for assessing RoD from clinical data. This approach is based on the generative kernel presented above.
5. We also assessed shock from a metabolomics, proteomics and transcriptomics. We also presented the results of assessing RoD for each of the omics approaches.
6. The main result of this activity is a graphical model including all the Omics data (i.e. from a Multi-Omics point of view) for assessing RoD in the ICU.
Besides the Multi-Omics graphical model, we also presented an alternative representation of sequence alignment based on algebraic statistics. This activity was mainly focused in the presentation of the Needleman-Wunch algorithm for sequence alignment from an algebraic point of view as well as its relation to graphical models and the Viterbi algorithm.
The methodologies used in this activity have been mainly concerned with:
– Min/Max or ‘Tropical’ algebra.
– Phylogenetic trees.
– Graphical models and hidden Markov models.
– The Viterbi algorithm.
The main research activities of this research are as follows:
1. Firstly, we have presented the required methodology for sequence alignment from an algebraic statistics point of view.
2. Secondly, we have evaluated the accuracy and efficiency of our alternative representation through some examples of alignments from the data obtained in our transcriptomics analyses.
4.7 Enteral Drug Delivery Technologies
The development of innovative enteral drug delivery technologies and protocols was motivated by one of the most important initial hypotheses of ShockOmics, i.e. the need to inhibit digest proteases in order to protect the integrity of the intestinal barrier, and as a consequence to limit the injury to distal organs.
A modified and enhanced enteral feeding catheter was developed and an international patent application was filed (International Patent Application PCT/US2017/028583). The main advantages of the system, comprising a multi-lumen infusion catheter, which represents an evolution and improvement of the currently available orogastric or nasogastric tubes for enteral feeding of acutely ill patients are:
• ability to sustain high infusion rates
• easy placement in the stomach
• viable placement in the duodenum, as recommended on the basis of experiments in rats and pigs, with the aid of a gastroscope as guide in large animals
• ability to prevent reflux in the esophagus and in the airways
In order to satisfy the need to achieve a rapid infusion into the intestine, and to prevent airway aspiration, an aspiration line is added for safety and placed in the esophagus for aspiration out of the mouth.
The rat experiments have demonstrated the effectiveness of the proposed infusion protocol, based on the insertion of the infusion catheter through the mouth of the animal (in humans, the access will be either via the mouth or nose) and placement in the duodenum. There is evidence that the optimal placement of the catheter tip for the purpose of delivering the drug to the small intestine is in the duodenum. The placement of the catheter in the stomach results in the inflation of the stomach, due to the high infusion rate and slower emptying rate of the stomach into the duodenum through the pylorus. The effectiveness of the infusion system was verified from the point of view of: a) filling of the small intestine in its entirety; b) preservation of the morphological integrity of the gut barrier.
A dedicated protocol for enteral infusion of protease inhibitor was designed, and its specifics have been determined in rats and pigs:
• infusion rate: 0.10 ml/min (in rats); 20 ml/min (in pigs);
• infusion volumes: 17.5 ml (in rats); 2.5 L (in pigs)
• optimal placement of the tip of the infusion catheter: post-pyloric, ~ 5 in into the duodenum.
The enteral delivery system can be integrated by sensors and control system to automatically detect backward flows and activate possible alarms to alert the critical care personnel and verify the risk for reflux and aspiration into the airways. An example of sensor which can be embedded in the delivery system is a pressure sensor to monitor intragastric pressure. To this purpose, one of the lumens of the multichannel catheter can be used as a fluid filled catheter and connected to a pressure transducer for continuous pressure monitoring. The elevation of pressure above certain thresholds can activate an alarm and interrupt the flow into the gut in order to prevent the possible reflux into the esophagus and from there into the airways.
8 IN VITRO VALIDATIONS
Aim of the in vitro validation study was to identify the most appropriate cell models and in vitro paradigms for the functional evaluation of the different shock challenges in vitro. Currently, no widely accepted suitable in vitro models for adult human cardiac biology exist, and several different approaches are being taken into consideration, ranging from standardized cell line approaches to differentiated human induced progenitor cells, for selected assays aimed at identifying several aspects of safety, pharmacology and toxicology assessment.
We have tested different cell lines currently used in screening activities, and characterized them versus primary cardiomyocytes taken from rat pups. Primary rat cardiomyocytes were the only ones showing multinucleated bodies, typical of cardiomyocytes in culture, and a clear presence of cardiac specific troponin marker staining. Moreover, only primary cardiomyocytes in culture were capable of beating in vitro, forming functional syncytia. A selected group of enzymes was identified, in order to mimic the different shock conditions, as follows: elastase, LPS and MMP9 for septic shock; alpha amylase, hsp70 for cardiogenic shock; oxygen glucose deprivation for cardiogenic shock.
A dose dependent time course was carried out in order to identify the optimal detrimental range for inducing a metabolic alteration of cardiomyocytes. In particular, parameters such as cell viability, metabolic activity and expression of selected genes of interest were quantified in order to obtain clear indications of the effect of the challenge.
For the septic shock model, elastase was shown to induce an increase of heart failure marker hsp70 as well as pro inflammatory cytokine IL6. On the other hand, LPS induced a more pronounced increase of several inflammatory markers such as hsp70, hsp40, MMP9, IL 6 and channels such as CACNA1C and KCNJ2. MMP9 exposure induced a significant increase in the expression of most of the genes tested in cell lines, with a particular emphasis on the expression of calcium and potassium channels, while no significant increase in the expression of inflammatory markers was observed in primary cardiomyocytes.
For the cardiogenic shock model, a time curve of oxygen and glucose deprivation (OGD) was carried out, and the percentage of dying cells at the different time points was quantified. The optimal experimental conditions were identified at 5 hours of OGD challenge followed by 2 hours of reperfusion in normoxic conditions.
The hemorrhagic shock model was mimicked using different stimuli, such as hsp70 or amylase exposure, but data did not yield significant results.
We thus concentrated our efforts on the most detrimental conditions, namely LPS exposure (for septic shock) and OGD exposure (for cardiogenic shock), and used these challenge in a more complex in vitro scenario. In particular, we took advantage of the possibility of using microfluidics devices previously developed in our lab (Bianco et al. Analytical Chemistry, 2012) in order to create micro-environmental scenario mimicking interaction of primary cardiomyocytes with immune cells, under the different experimental scenarios. Primary rat cardiomyocytes previously characterized were cultured on the microfluidic platform and exposed to macrophages isolated from rat spleen and liver. We established a correct protocol for macrophage isolation, and optimized conditions for culturing on a microscale.
In order to recreate the septic shock model, we challenged macrophages with LPS and then put them in microfuidic communication with primary cardiomyocytes. We monitored biochemical and functional parameters and compared the expression of the panel of genes previously identified in cardiomyocytes exposed to the three different key micro-environmental conditions: untreated cells, macrophages and activated macrophages. Activated macrophages induced a significant increase in cardiomyocytes cell death. Moreover, we showed that exposure of cardiomyocytes to macrophage-conditioned medium (either primed or not) is sufficient to induce a metabolic stress on primary cardiomyocytes.
Similarly to the septic shock model set up, in the cardiogenic shock model microenvironment set up we challenged the cells with previously characterized ischemia (OGD) insult, and compared the response of primary cardiomyocytes either exposed or not to ischemic conditions in microfluidic connection with macrophages. The presence of activated macrophages significantly increased cardiomyocyte cell death and remaining living cardiomyocytes show a significantly altered metabolism when exposed to macrophage-derived inflammatory storm.
Once the in vitro models were set up, we applied selected inhibitors in order to quantify their modulatory effect on the in vitro models. Given the results obtained from the Consortium ongoing “omic” studies revealed a key role for inflammatory pathways being activated by the shock conditions, resulting in the activation of many different pathways, (i.e. oxidative stress, mitochondrial damage, increase in intracellular calcium levels and modulation of viability), we selected three different inhibitors, targeting different mechanisms linked to different aspects observed to be altered in the challenged cardiomyocytes; namely, strophanthidin, ghrelin, and tranexamic acid (TXA). A series of biochemical and functional assays were carried out, either in the presence or absence of such inhibitors, including analysis of cell viability, cell metabolism, intracellular calcium dynamics, reactive oxygen species production and NO release, both for the septic as well as the cardiogenic models.
In the coculture septic model, exposure to LPS significantly increases the percentage of dying cells, but this effect was counteracted by concomitant exposure to strophanthidin and ghrelin. Similarly, when evaluating the efficacy in reducing nitric oxide release, we observed that strophanthidin and ghrelin were able to counteract LPS, while TXA did not modulate the LPS mediated damaging effect. On the other hand, ghrelin or TXA did restore metabolic parameters altered by LPS exposure such as metabolic activity and calcium levels, while strophanthidin did not. A significant increase in the production of total ROS was observed in LPS challenged cardiomyocytes. This increase was significantly inhibited by concomitant exposure to ghrelin, while no statistically significant variation was observed with either strophanthidin or TXA.
In the cardiogenic model, OGD exposure increased the percentage of dying cells, but this effect was counteracted by concomitant exposure to strophanthidin and ghrelin. Concerning metabolic activity, none of the tested inhibitors was able to counteract the 20% reduction observed after LPS exposure. On the other hand, LPS induced an increase in intracellular basal calcium levels and this effect was counteracted significantly by all of the tested inhibitors. For what concerns the oxidative stress, a significant increase in the production of total ROS was observed in LPS challenged cardiomyocytes. This increase was significantly inhibited by concomitant exposure to ghrelin, while no statistically significant variation was observed in cardiomyocytes challenged with LPS in the presence of either strophanthidin or TXA. Finally, for what concerns the LPS-induced increase in NO release, this effect was significantly inhibited by the presence of either strophanthidin or ghrelin.
Overall, the in vitro models developed in this study yielded several results comparable with the data obtained in the Omics studies of patients’ biopsies, such as activation of selected inflammatory pathways, vascular inflammation and vasodilation. Moreover, the platforms developed in this project represent a valuable tool for the in vitro characterization of the mechanism of action of novel molecules aimed at targeting selectively either cardiogenic or septic shock damage.
9 CONCLUSIONS
The ShockOmics Consortium faced several challenges specific to the addressed syndrome and to the need of integrating many disciplines. First of all the need of a timely activation of recruitment in emergency departments and soon after the unpredictable onset of shock. Secondly, the difficulties in a “stable” reproduction in animal experiment of a condition, which is by definition unstable. Finally, a tightly serial workflow from biosample collection and HD recording to the wet-labs for the Omics and the data processing units, and next to data mining and integration.
Despite the above, the ShockOmics project was able to reach most of the original goals, although with a wealth of bio-samples and data still to be processed. So, the overall ensemble of reached results and stored repositories should be considered as project achievements.
The integrated approach followed has cast new light both relevant to the mainstream hypothesis of autodigestion and collateral ones related to inflammatory pathways and those addressing the energetic metabolism. The latter ones opening a new frontier towards biomarkers for the early detection of vital organs (e.g. the liver) and to more effective therapeutic treatments. A major issue has been the congruency of such results given by a comparison of the outcomes from different Omics, looking at diverse steps of the biomolecular chains. This overall view was generally achieved starting from metabolomic findings and next focusing the correlates in transcriptomics and proteomics, as well as in-vitro experiments.
A further broad field opened concerns the massive circulating peptides analyzed by peptidomics. Firstly, the individuation of trypsin as main protein degradation source strongly supports our working hypothesis based on autodigestion triggered by the intestine damage, thus fostering efforts towards therapeutic protocols based on protease inhibitors such as TXA. Secondly, the bulk level of peptides was confirmed as a powerful tool for the evaluation of progressive organ damage and MOF. Finally, the issue is risen relevant to the bio-activity performed by the anomalous concentrations of specific peptides, which eventually can amplify the ongoing organ failures.
As underlined in the next section, the summarized results, together with the data processing and integration methods, and the progresses in therapy experiment relevant to intestinal delivery of protease inhibitors represent a consistent potential impact of the ShockOmics project outcomes towards enhanced risk prediction and timely therapeutic actions in the shock syndromes.
Potential Impact:
1 SOCIAL IMPACT
The ShockOmics consortium addressed new insight on CV shock, a potentially fatal syndrome for which current therapies are directed at symptoms, rather than at the prevention of the onset of the pathology.
Considering the starting context, it is obvious that acute heart disease in shock and ensuing hemodynamic instability are a major challenge for intensive care physicians, since current therapies fail when progressive shock is started. The approach proposed by ShockOmics addressed several fundamental questions that have not been answered so far thanks to cutting edge tools and a multidisciplinary team. Although the analyses of clinical datasets and the parallel experiments have been at a stage of basic biomedical research, still the translational potentials of the findings have been emphasized towards improved risk assessment tools and therapeutic strategies and techniques.
The ShockOmics consortium had been built to achieve in overcoming the main barriers limiting research in the field. The Omics methodologies for the experimental investigation of diseases and by means of the innovative biotechnologies nowadays available offered the technical support for a multiscale study that aimed at relating system level measurements, commonly available in clinics, to the basic biomolecular pathways of interest.
Our goals stemmed from emerging evidences, provided in the literature, for instance, by the “Autodigestion Hypothesis”, previously formulated by the group of Dr. Geert W. Schmid-Schönbein, UCSD, insured a strong theoretical background to formulate the fundamental hypotheses of the proposed studies. The ShockOmics consortium included scientists with very diverse and clearly complementary expertise, since the joint effort of medical doctors, biologists, computer engineers, bioengineers, and small entrepreneurs was necessary to address the complexity underlying the questions and the goals of ShockOmics.
State-of-the art techniques and methods for the study of biomolecular pathways and CV regulation as well as drug delivery technologies supported the achieved advancements in the knowledge and treatment of circulatory shock.
Several evidences demonstrate the relevance of shock and its associated complications, HF included. Trauma resulting from accidental injuries is the leading cause of death in individuals aged 1–44 yr in the United States of America, and hemorrhagic shock, for instance, affects 36–39% of trauma victims. Many develop multiple organ failure (MOF), the leading cause of death in the intensive care unit.
The management of shock patients and sepsis shock patients has been the object of a large clinical study(Caironi et al. Albumin replacement in patients with severe sepsis or septic shock. N. Engl. J. Med., 2014), which estimated a mortality rate up to about 45%. Such numbers confirm that clarifying the molecular mechanisms of shock is of paramount urgency in order to find effective therapies.
The above evidence is highlighted by the major medical societies; e.g.: the American Heart Association, the European Society of Critical Care Medicine, and the Society of Critical Care Medicine. Moreover, the costs of hemorrhagic shock in US have been estimated to spam from $51,000, for a trauma-patient with uncomplicated shock, to $321,000 for a severe hemorrhagic shock with MOF.
Besides the relevance of the acute cardiovascular illness per se, the social and financial costs of the disease are evident also considering the long term impact on shock survivors, who have displayed a high two year mortality (44.9%) in the case of septic shock and a lower quality of life, as long as 1.5 yrs. after their ICU discharge.
For all these reasons, the relevance of advancements in the comprehension of shock mechanisms and the consequent therapies has highly significant social and economical implications. Namely the following achievements are highlighted for the advancement in the field of personalized medicine and potential social impact:
– Omics technologies (metabolomics, proteomics, peptidomics, and transcriptomics) were tested for the timely identification/recognition of circulating biomarkers of pathogens potentially responsible for MOF progression and hemodynamic instability. The focused biomolecules and pathways still need further experimental and clinical research to reach the level of potential biomarkers and be challenged in clinical studies; nonetheless, a road map is open, starting for further analyses over the collected bio-banks and data sets.
– A new potential protocol for the timely and efficient delivery of therapy to prevent the evolution of proteolysis and systemic inflammation in shock starting from the intestinal compartment was tested. Namely, starting from previous trials in open abdomen small animals experiments of naso-gastric access in large animals were carried out, thus progressing towards a future application in the ICU.
– Software platforms were developed for the integration of new, systems biology based, multiscale models for the analysis of hemodynamic stability in critical care monitors. Models exploiting the multivariate beat-to-beat variability series, drawn from current ICU monitoring, were translated in algorithms providing indices of CV regulation impairment. Data-mining approaches on multiscale analysis outcomes (clinical, hemodynamic, and Omics) showed enhanced risk prediction value, compared to separate analyses.
– The ShockOmics database is available for further studies by the Consortium Partners and, upon request, by accredited scientific institutions. It will include the HD recordings and the derived beat-to-beat series, as well as the lists of biomolecular analyses (metabolites, proteins, peptides, transcripts). Storage will be maintained on a repository, while sub-sets will be replicated on public repositories of the specific disciplines (e.g. proteomics), when applicable. Both clinical (fully anonymized) and animal studies will be included.
– Benefits to the involved SMEs and to the European market where highlighted in three main areas:
1. Risk prediction algorithms based on: a) easily measurable endpoints, which could be integrated into existing cardiovascular monitors, b) integration of Omics analyses from blood/plasma samples and clinically available parameters;
2. Bases for the future development of diagnostic-kits targeting the potential biomarkers among the set of biomolecules shown to predict the risks related to shock and organ failure;
3. Adaptation of the existing implantable drug delivery technology to better target shock-therapy was pursued concerning: a) design of an infusion catheter (patented) specific to infusion in the small intestine lumen; b) insertion techniques in the intestinal lumen through naso-gastric way and the pylorus; c) design of smart pumps controlling both infusion flow and pressure.
– Finally, the international profile of ShockOmics through three European Community Countries plus Switzerland and the US, together with the high level of interdisciplinary synergies, fostered cultural and scientific integration, as testified by the personnel interchange and dissemination activities.
The theme of the intellectual property is of great importance for the translational implications of ShockOmics and to deliverables were respectively dedicated to market searches and to exploitation plans in the three above mentioned areas. The latter, in full agreement with the IPR management rules set in the Consortium Agreement.
2 DISSEMINATION ACTIVITIES
Several measures were adopted to maximize the dissemination of the scientific results obtained from the experimental studies carried out during the four year duration of ShockOmics, besides the traditional publications on peer reviewed journals and presentations at scientific conferences.
As regards the journal publications, the partners of the consortium were committed to ensuring the open access to their publications and data, in line with the previously listed impacts inherent to the open access and open source availability of data and mathematical resources and tools.
The presentation of preliminary results to the most important international medical, biological, and engineering conferences on the topic of interest of ShockOmics was also integrated by the organization of special sessions and mini-symposia within such congresses. Given the innovation of ShockOmics and the relatively unfocused research on the multiscale phenomena which characterize acute HF in shock, discussion was fostered within the scientific community. In this way, besides giving visibility to ShockOmics, other scientists interested in shock and acute heart disease were involved in the important research field explored by ShockOmics.
Dissemination through conferences and full papers on indexed international journals is still ongoing. So far, 41 conference presentations (poster or podium) were performed while the published full papers were 14 on the following journals: Medical and Biological Engineering and Computing, Shock (two papers), Briefings in Bioinformatics, Scientific Reports (two papers), Scandinavian Journal of Trauma, Resuscitation and Emergency Medicine, Journal of Clinical Monitoring and Computing (two papers), Computational and Mathematical Methods in Medicine, Physiological Measurements, Journal of Clinical Monitoring and Computing, Experimental Biology and Medicine.
ShockOmics also proposed a specific international workshop, embedded in a small though highly focused conference in the field; i.e. the International Conference in Complex Acute Illness (ICCAI, Milan, Italy, July 27-29, 2017, iccai.org ), which was the 16th annual conference of the homonymous society, SCAI. The ShockOmics coordinating unit was in charge of hosting the conference and this gave a unique opportunity in opening to the wide audience of the major societies in the fields of Shock and Intensive Care, to which the conference call was addressed. The invitation of renowned keynote speakers permitted the contact with major stakeholders. Moreover, the SCAI board of members and the ICCAI’17 scientific committee also included reference scientist in the field and the Editors of major journals such as Shock, Critical Care in Medicine, the Journal of Critical Care, and Physiological Measurements.
This event gave also wide visibility on the scientific press since the Journal of Critical Care kindly accepted to publish the 9 abstract presented in the ShockOmics workshop sessions preceded by an editorial providing a general overview of the project close to its end (Ferrario M, Aletti F, Baselli G, The EU ShockOmics Project International Workshop at ICCAI'17, J. Critical Care, 2017).
The website represented and will represent in the next future a unique way of communication with the public at large, beyond the borders of the scientific community, open access publications related to ShockOmics will be periodically updated.
Dissemination initiatives aimed at undergraduate and graduate students were undertaken including the recruitment to the laboratories involved in ShockOmics for internships, BSc projects, MSc and PhD thesis.
The researchers of ShockOmics consortium actively participated to other well-known initiatives, such as the White Night of Researchers, which regularly takes place across the Continent, to increase the awareness of European citizens towards the importance of research, development and education for enhancing the quality of lives.
List of Websites:
Project website: www.shockomics.org
Main contact: Project Coordinator - giuseppe.baselli@polimi.it
