Insulin and the immune response to critical illness
In order to survive, the human body has to protect itself against the continuous threat of external aggressions, as well as against endogenous stress caused by ageing, dysfunctional, mutated or damaged cells. In conditions of extreme stress, the organism will either recover quickly and survive, or will succumb. Thanks to innovations in intensive care medicine, patients are nowadays able to survive such acute life-threatening conditions. Unfortunately, not all patients will recover from this acuteinsult within a few days. A substantial proportion of the patients willthus enter a state of prolonged critical illness, which is characterised by, amongst others, profound disruption of endocrine and immune homeostasis.
Traditionally, the immune system has been divided into innate and acquired components. The innate immune system comprises the primitive, first-line defences against invasion by microorganisms and plays an important role in healing of tissue damage. Inflammation is the basic reaction that aims at removing the nocuous intruder and at damage repair. However, uncontrolled inflammation can cause additionaldamage. Critically ill patients are extremely vulnerable to complications that develop due to dysregulated inflammation, leading to prolonged dependency on intensive care, organ dysfunction and eventually death.
Critically ill patients typically develop hyperglycaemia during acute stress. This metabolic response has long been considered an adaptive phenomenon, that guarantees sufficient glucose supply to organs that depend on glucose for energy production but do not require insulin for its uptake. Unfortunately, during periods of prolonged stress, hyperglycaemia persists and fuels the inflammatory cascade, leading to increased morbidity and mortality. Our research group demonstratedin three large randomised studies that maintaining normoglycaemia with intensive insulin therapy (IIT) reduced morbidity and mortality of critically ill patients, as compared with tolerating stress hyperglycaemia (conventional insulin therapy, CIT). The randomised intervention reduced infectious complications and prevented excessive inflammation and organ dysfunction. Both prevention of hyperglycaemia as well as direct effects of insulin may have contributed to the beneficial effects of the therapy. Unexpectedly, besides a lowering of the C-reactive protein (CRP), no effect of IIT could be demonstrated on the classical pro-inflammatory cytokines that traditionally are put forward as central mediators of inflammation after trauma or infection. This raised the question whether there could be an interaction with other components of innate immunity.
The aim of this thesis was to increase insight in how criticalillness affects the innate immune system, in relation to the typical complications, and what the impact is of strict glycaemic control with insulin. Since the beneficial effects of the therapy could not be explainedby alterations in the classical pro-inflammatory cytokines, we hypothesised that other pathways are involved. We thus focused on the study of several components of the receptor for advanced glycation end-products (RAGE) axis, sCD163 as a marker of macrophage activation, and the lectinpathway of complement activation. In addition, we investigated the potential role of genetic predisposition in the acquisition of the typical complications of prolonged critical illness, an insight that might enableus in the future to make the right preventive and therapeutic decisionsfor each individual patient.
Engagement of the RAGE receptor by ligand binding leads to amplification and perpetuation of inflammation. NF-κB activation, amongst others, induces the production of pro-inflammatory cytokines and up-regulation of the RAGE receptor, generating a positive feedback loop of persistent inflammation. Thus, RAGE activation has been linked to inflammation, organ damage and adverse outcome in for instance sepsis, myocardial dysfunction and acute lung injury. Moreover, we know from the diabetes literature that hyperglycaemiaplays a pivotal role in the development of typical inflammatory complications, which amongst others is driven by hyperglycaemia-induced production of advanced glycation end-products (AGEs), in turn leading to increased oxidative stress and enhanced RAGE activation. Other RAGE ligands include high-mobility group box 1 (HMGB1, a ubiquitous nuclear protein released by cellular injury or cell death) and S100A12 (a typical danger-signal protein released by activated neutrophils). sRAGE is the cleaved,circulating isoform of RAGE, which possibly acts as a decoy receptor and would thus protect against further RAGE activation.
We studied the relation of sRAGE, HMGB1 and S100A12 with baseline characteristics and clinical outcome of a large heterogeneous group of prolonged critically ill patients who had undergone major surgery. We comparedthese parameters with CRP, which is the classical, routinely measured clinical parameter of inflammation. Levels of sRAGE, HMGB1, S100A12 and CRP were elevated upon admission to the ICU. HMGB1, S100A12 and CRP levels decreased during ICU stay, but remained above healthy control levels. In contrast, sRAGE decreased to levels that were significantly lower than those of healthy volunteers. Cardiac surgery was associated with a lower grade of inflammation, reflected by lower upon-admission CRP, HMGB1 and S100A12 levels, as compared with other types of surgery. Conversely, sepsis upon admission was associated with higher levels of CRP, HMGB1 and S100A12. sRAGE did not differ for these baseline characteristics. In particular a high upon-admission sRAGE level was associated with a higherrisk of hospital mortality and organ dysfunction (liver dysfunction, need for renal replacement therapy and hemodynamic support). Also in multivariable analysis correcting for baseline risk factors, sRAGE remained an independent predictor of organ dysfunction. Unexpectedly, the randomised IIT intervention did not affect circulating levels of sRAGE, HMGB1 and S100A12. From this study, we concluded that critical illness affects several components of the RAGE axis and that sRAGE is associated with clinical complications of critical illness. sRAGE likely reflects ongoing inflammation and shedding of the receptor.
CD163 is a receptor that is exclusively expressed by monocytes and macrophages and is considered to reflect macrophage activation. The receptor scavenges haemoglobin-haptoglobin complexes from the circulation and plays a role in the resolution of inflammation. Under inflammatory stimuli, CD163 is cleaved from the membrane surface. The shedded, circulating soluble CD163 thus reflects ongoing inflammation. sCD163 has been proposed as a predictor of morbidity and mortality in small uniform groups of patients with systemic inflammation (SIRS), pneumococcal bacteraemia and liver failure, and also in a number of chronic inflammatory diseases.
We studied sCD163 in a large group of critically ill patients with widely varying pathologies. We included short- and long-stay surgical and medical ICU patients. Critically ill patients presented with elevated sCD163 levels upon admission to ICU, as compared with healthy controls. High admission levels were associated with an increased risk of mortality, organ dysfunction (liver and kidney dysfunction) and a prolonged stay in ICU, but was not associated with infection (bacteraemia). The already elevated sCD163 levels upon admission further increased during ICU stay, especially in those patients who subsequently developed organ failure or who ultimately died. Preventing hyperglycaemia with IIT slightly attenuated the elevated sCD163 levels. The CD163 mRNA levels in postmortem liver samples from critically ill patients were approximately twice ashigh as those of healthy control patients (who underwent uncomplicated elective surgery) and correlated with the respective circulating levels.We concluded from this study that sCD163 levels are increased during critical illness, with the liver as potential source. This study underscores the association between increased sCD163 and organ damage and mortality in critical illness. Maintaining normoglycaemia with IIT decreased circulating sCD163 levels, supporting the anti-inflammatory effect of the therapy.
Most often, clinical studies in ICU exclude children, as they represent only a small population, spread over different developmental stages. Results obtained from adult studies are too often extrapolated to children. In this regard, a large randomised study onIIT in our paediatric ICU, represents a landmark. Children are admittedto the paediatric ICU after major or life-threatening surgery, extensive trauma or severe infection. Systemic inflammation, evoked by differentcauses, is a serious complication which is associated with high morbidity. Moreover, critically ill children are extremely prone to nosocomial infections. The question why some patients are more susceptible to developing these complications than others is intriguing and raises the possibility of genetic predisposition. Defects in complement activation have been linked to increased susceptibility to infections in young children with an immature adaptive immune system (like neonates) and in children with associated pathologies (like e.g. cancer). In these situations, combatting infection is predominantly based on the integrity of the innate immune system. The complement system has an important role in inflammation and repair of tissue damage, opsonisation and clearance of immune complexes, and in lysis of microorganisms. Insufficient as well as excessive activation of the complement system can be detrimental. Mannose-binding lectin (MBL) and ficolins are circulating receptor proteins that recognise and bind specific molecular patterns on e.g. microorganisms. They associate with a MBL-associated serine protease (MASP) or MBL-associated protein (MAp). MBL/MASP- or ficolin/MASP-complexes then activate the complement cascade by the so-called lectin-pathway of complement activation.
We studied MBL, MASP-1, MASP-2, MASP-3, MAp44, and M- and H-ficolin in all critically ill children included in the randomised study on IIT as well as in a cohort of healthy children. As alreadydescribed in healthy children, there was a certain age-related variation in the levels of these proteins in critically ill children. Serum levels of MASP-1, MASP-2, MASP-3 and MAp44 were lower in critically ill children as compared with age-matched healthy controls, while levels of M-ficolin appeared to be higher upon ICU admission. In particular a low MASP-3 level upon admission was independently associated with the risk of acquiring a new infection and with the risk of needing prolonged intensivecare. This association remained after correction for known risk factors. MASP-3 levels varied with age, severity of illness and with two of thestudied genetic polymorphisms in the MASP-1 gene, which encodes MASP-1, MASP-3 and MAp44. This genetic variation, however, was not related to outcome. The different alleles for these polymorphisms were equally distributed in patients and controls. This means that genetic variation inthe studied polymorphisms does not engender predisposition to critical illness, nor to adverse outcome of critical illness. We analysed profiles of the proteins in a subgroup of patients, in order to detect any effect of the randomised IIT intervention. However, levels of the measured components of the lectin pathway were comparable in CIT and IIT patients.Thus, we were able to demonstrate that low upon ICU-admission MASP-3 levels are independently associated with the risk of acquiring a new infection and with the risk of a prolonged ICU stay in critically ill children, while IIT did not appear to have any effect on circulating levels of several proteins of the lectin pathway of complement activation.
As indicated, preventing stress-related hyperglycaemia with IIT in critical illness reduces morbidity and mortality, with a decreasein new infections and an attenuation of excessive inflammation. However, this intervention did not affect circulating pro-inflammatory cytokines. In summary for this doctoral thesis, we can state that our studies did not provide a clear explanation to further understand this effect, as we only encountered a small effect on sCD163. However, the studied pathways captured important signals as several components of the studied inflammatory pathways appeared to be strongly associated with outcome. The biological implications of these findings are unclear at this time and need to be clarified in future research. Additional epidemiological and molecular research is indeed needed to further unravel these insights in arelatively underexplored area of very complex processes in the critically ill patient.