Tjitske van Engelen

Pneumonia Pathogenesis, biomarkers and clinical implications Tjitske van Engelen

Pneumonia Pathogenesis, biomarkers and clinical implications Tjitske van Engelen

Colofon Pneumonia: pathogenesis, biomarkers and clinical implications Copyright 2024 © Tjitske van Engelen ISBN 978-94-6506-760-5 All rights reserved. No parts of this thesis may be reproduced, stored in a retrieval system or transmitted in any form or by any means without permission of the author. Printing: Ridderprint | www.ridderprint.nl Layout and design: Henry Smaal, persoonlijkproefschrift.nl The printing of this thesis was kindly supported by the Amsterdam UMC.

Pneumonia: pathogenesis, biomarkers and clinical implications ACADEMISCH PROEFSCHRIFT ter verkrijging van de graad van doctor aan de Universiteit van Amsterdam op gezag van de Rector Magnificus prof. dr. ir. P.P.C.C. Verbeek ten overstaan van een door het College voor Promoties ingestelde commissie, in het openbaar te verdedigen in de Agnietenkapel op donderdag 23 januari 2025, te 16.00 uur door Tjitske Sibongile Rozanna van Engelen geboren te Amsterdam

Promotiecommissie Promotores: prof. dr. T. van der Poll Amsterdam UMC - UvA prof. dr. W.J. Wiersinga Amsterdam UMC - UvA Copromotor: prof. dr. J.M. Prins Amsterdam UMC - UvA Overige leden: prof. dr. O.L. Cremer UMC Utrecht dr. J. den Dunnen Amsterdam UMC - UvA prof. dr. A.H. Maitland-van der Zee Amsterdam UMC - UvA prof. dr. A. Verbon UMC Utrecht prof. dr. M. Prins Amsterdam UMC - UvA dr. C.W. Seymour University of Pittsburgh Faculteit der Geneeskunde

Voor mijn ouders

Table of contents 1. General introduction 9 Part I: Pathogenesis and host response 19 2. Pathogenesis of sepsis Handbook of Sepsis, 2018 21 3. Biomarkers in sepsis Critical Care Clinics, 2018 37 4. Association of hyperferritinemia with distinct host response aberrations in patients with community-acquired pneumonia Journal of Infectious Diseases, 2022 61 5. Plasma protein biomarkers reflective of the host response in patients developing Intensive Care Unit-acquired pneumonia Critical Care, 2023 83 Part II: Intermezzo – focus on the microbiome 129 6. Impact of Sepsis and Empiric Antimicrobial Therapy on the Gut Microbiome Journal of Antimicrobial Chemotherapy, 2019 131 7. Gut Microbiome Modulation by Antibiotics in Adult Asthma: A Human Proof-of-Concept Intervention Trial Clinical Gastroenterology and Hepatology, 2022 151 Part III: Towards improvement of diagnosis and clinical care 165 8. Classifying the diagnosis of study participants in clinical trials: a structured and efficient approach European Radiology Experimental, 2020 167 9. Ultra-low-dose CT versus chest X-ray for patients suspected of pulmonary disease at the emergency department: a multicentre randomised clinical trial Thorax, 2023 209 10. Limited clinical impact of Ultra-Low-Dose computed tomography in suspected community-acquired pneumonia Open Forum Infectious Diseases, 2023 243

Part IV: Discussion and future perspective 255 11. Towards precision medicine in sepsis: a position paper from the European Society of Clinical Microbiology and Infectious Disease Clinical Microbiology and Infection, 2018 257 12. General discussion 293 Part V: Summaries 303 13. Summary in English 305 14. Summary in Dutch - Nederlandse samenvatting 311 Addendum 317 Portfolio 318 List of contributors 322 Acknowledgements in Dutch – Dankwoord 331 About the author 333

Chapter 1 General introduction

10 Chapter 1 Pneumonia Lower respiratory tract infections are the leading infectious cause of death worldwide and the fifth-leading cause of death overall [1]. Pneumonia is an acute lower respiratory tract infection which arises when a pathogen successfully crosses the mucosal barrier, escapes the host defence system, and manages to multiply to ensure its survival [2]. Pneumonia is mostly seen in children younger than 5 years and in older adults with chronic comorbidities. Most often, pneumonia is caused by bacteria or viruses – less often by other pathogens. However, development of the disease is more driven by the immune response of the patient than by characteristics of the invading pathogen [3]. The diagnosis of pneumonia is based on a combination of acute signs and symptoms of a lower respiratory tract infection (such as cough, fever, breathlessness and expectoration) and a new pulmonary infiltrate on chest imaging [2,4,5]. In the absence of a clear viral aetiology, the treatment of pneumonia consists of rapid administration of empiric antibiotics. In most outpatients, diagnostic testing for bacteria is not necessary - testing for influenza and SARS-CoV-2 can be considered [2]. In hospitalized patients, comprehensive microbiologic testing is recommended to determine the appropriate pathogen-specific therapy [2,5]. Supportive care (such as fluids and oxygen) is administered if necessary [5,6]. Pneumonia can be further specified as communityacquired pneumonia and hospital-acquired pneumonia (which includes Intensive Care Unit (ICU)-acquired pneumonia and ventilator-associated pneumonia)[4,7], based on the location where the infection was acquired. Sepsis With insufficient control, pneumonia can deteriorate into sepsis. At the turn of the century, sepsis was predominantly understood as a consequence of an overwhelming inflammatory response following pathogen invasion [8]. Current consensus recognises two seemingly opposing host reactions in severe infections, characterized by both proinflammatory and anti-inflammatory features. Sepsis is a characterized by dysregulation of the immune response to infection with a failure to return to homeostasis, resulting in a life-threatening condition which often requires organ support in the ICU. Nearly a fifth of all global deaths are sepsis-related with an especially high health-related burden in Sub-Saharan Africa [9]. If ICU admission for sepsis is necessary, the sepsis syndrome is fatal for approximately one out of every four patients [10]. Pneumonia is the most common cause of sepsis [11,12]. Decades of research and many failed therapeutic sepsis trials have accentuated the high variability of this syndrome at the individual patient level [13,14]. The critical care community now seeks to utilize clinical and biologic features to identify subgroups that are more homogeneous within the heterogenous population of sepsis patients [15].

11 General introduction Biomarkers Biomarkers, which are naturally occurring molecules, genes, or other characteristics reflecting physiological or pathological processes, hold great promise for personalized medicine in pneumonia and sepsis [16]. Not all patients with pneumonia develop sepsis. Patient characteristics, clinical signs and disease severity scores only partly explain the variation in the course of disease. Biomarkers can shed light on the pathophysiological mechanisms underlying pneumonia and help us better understand how this disease evolves and why this varies amongst individuals. Biomarkers can be of diagnostic, prognostic or theragnostic value; the latter being the use of a biomarker to select and evaluate specific therapies. Biomarkers can be useful tools for clinicians, as additional information to the clinical evaluation, aiding in optimized patient care at the bedside [17]. They can help stratify patients into subgroups for targeted therapy and personalize follow-up of response to treatment. In recent years, there has been a surge of interest in biomarkers in the sepsis field. The challenge lies in the intricate nature of sepsis pathophysiology, making it unlikely that a single biomarker can accurately capture the complexity of the host response. In this context, novel technologies focussing on panels of biomarkers, biomarker trajectories over time, and the “omics” field of system biology hold promise in the management of patients with sepsis [16,17]. Microbiome The human gut microbiota, comprising diverse microbial communities in the intestine, plays a crucial role in protecting against harmful agents. Gut micro-organisms and their derived metabolites are associated with the susceptibility to sepsis as well as outcomes of sepsis [18]. The interaction between sepsis and the microbiome is a complex, bidirectional relationship. The disease state of sepsis can disrupt the microbiota, while clinical interventions, particularly antimicrobial therapy, can also affect microbial composition [18]. Furthermore, disruptions in gut microbiota-related immunological processes have been linked to pulmonary disease severity and treatment responses. Mechanistic data supporting this so-called gut-lung-axis are scarce but would offer potential in future management of patients with pulmonary diseases. Clinical trials This thesis describes research on the pathogenesis, host response, diagnostics, and clinical management of patients with pneumonia. Two large clinical trials, the OPTIMACT trial and the ASPIRE-ICU trial, and one smaller human proof-of-concept intervention trial, provided the infrastructure for our research questions and were the backbone of the studies discussed in this thesis (Figure 1). 1

12 Chapter 1 1. Can we classify the diagnoses of study participants reliable and valid? 2. Does the higher accuracy of ULDCT for pneumonia affect clinical management and outcomes? 3. Are circulating levels of ferritin amongst patients hospitalized with pneumonia associated with clinical features, outcomes and key host response pathways? OPTIMACT trial: Patients suspected of non-traumatic pulmonary disease at the Emergency Department n = 2418 Randomized Study questions Patients with pneumonia ULDCT n = 1208 CXR n = 1210 ULDCT n = 225 CXR n = 169 1. Are there host response differences shortly after ICU admission between those that do and those that do not develop ICU-acquired pneumonia? 2. Are there host response aberrations at the time of ICUacquired pneumonia? 3. What is the change in host response from ICU admission to the day of ICU-acquired pneumonia? ASPIRE-ICU trial: Patients in the Intensive Care Unit (ICU) who do or do not develop ICU-acquired pneumonia n = 1997 ICU-acquired pneumonia Study questions Patients at risk for ICU-acquired pneumonia Yes n = 316 No n = 1681 Hospitals n = 30 Countries n = 11 1. Does modulation of the gut microbiome by broad-spectrum antibiotics of adults with allergic asthma influence lung inflammation in a house dust mite plus lipopolysaccharide provocation model? Human proof-of-concept intervention trial: Microbiome modulation in adults with allergic asthma n = 20 Gut microbiome modulation Study question Type of antibiotics Yes n = 7 No n = 13 Ciprofloxacin Vancomycin Metronidazole Figure 1. Overview of clinical trials. The three clinical trials summarized in the figure were the backbone of the studies discussed in this thesis. The number of study participants (n) and study questions are provided per trial. ASPIRE-ICU = Advanced understanding of Staphylococcus aureus and Pseudomonas aeruginosa Infections in EuRopE - Intensive Care Units; CXR = chest X-ray; ICU = intensive care unit; OPTIMACT = OPTimal IMAging strategy in patients suspected of nontraumatic pulmonary disease at the ED: chest X-ray or CT; ULDCT = ultra-low dose computed tomography. First, the OPTimal IMAging strategy in patients suspected of non-traumatic pulmonary disease at the ED: chest X-ray or CT (OPTIMACT) was a multicentre randomized clinical trial (RCT) designed to study the value of a novel imaging technique – ultra-low-dose chest computed tomography (CT) – in 2418 patients presenting at the Emergency Department (ED) and suspected of pulmonary disease [19,20]. During randomly assigned periods of one calendar month, either ultra-low dose CT or conventional chest X-ray was used. Eligible for inclusion were ED patients aged 18 years and older, suspected of non-traumatic pulmonary disease and requiring chest X-ray according to the attending physician. Valid and reliable classification of the clinical diagnosis of study participants in this RCT was a prerequisite for the evaluation of outcome of both imaging strategies, as well as for biomarker research. Presented in this thesis, we developed a structured approach in which study participants were assigned diagnostic labels using a carefully designed reference standard, which could be used by a team of medical students, residents and an expert panel, and which we hypothesized to be efficient and valid for diagnostic classification in large clinical trials. Also, we performed a preplanned subgroup analysis, where we hypothesized that the known

13 General introduction higher diagnostic sensitivity of ultra-low-dose chest CT to detect pneumonia would affect clinical management and patients’ outcomes in patients suspected of communityacquired pneumonia. Furthermore, we designed a biomarker study where blood samples were taken from over 600 patients included in this large multicentre cohort to investigate the pathophysiology of pneumonia. In this biomarker study, we aimed to study the association of hyperferritinemia in patients with community-acquired pneumonia with their clinical presentation, outcome and aberrations in key host response pathways implicated in the immunopathology of pneumonia and sepsis. Second, we performed a study as part of the “Advanced understanding of Staphylococcus aureus and Pseudomonas aeruginosa Infections in EuRopE - Intensive Care Units” (ASPIRE-ICU) project, a study of 1997 adult ICU patients at 30 hospitals in 11 European countries aimed to report the incidence density and the risk factors of S. aureus and Ps. aeruginosa ICU pneumonia in Europe [21]. Presented in this thesis, our project was designed as a nested case-control study within the ASPIRE-ICU population aimed to obtain insight into host response protein differences between cases (patients who developed an ICU-acquired pneumonia) and controls (patients who did not develop an ICU-acquired pneumonia). We studied the host response prior to and during ICUacquired pneumonia, and the host response trajectory, i.e. the changes in host response over time. Last, to study gut microbiota-related immunological processes and its relation to pulmonary immunity, we designed and performed a proof-of-concept human intervention trial to address the question whether modulation of the gut microbiome of adults with allergic asthma influences lung inflammation in a house dust mite provocation model. Twenty patients with asthma and house dust mite allergy were recruited and randomized into two groups, receiving either broad-spectrum antibiotics orally for 7 days to disrupt the intestinal microbiota, or no treatment. After a wash-out period of antibiotics, all patients were challenged through house dust mite installation in one lung segment by bronchoscopy. Bronchoalveolar lavage fluid was obtained during a second bronchoscopy several hours later. The primary outcome was the influx of inflammatory cells into the bronchoalveolar space as a measure of pulmonary inflammation. Outline thesis Pneumonia remains a major cause of mortality globally, and its incidence is rising due to an aging population, increased use of immunosuppressive drugs, and antibiotic resistance. Bearing the complex pathophysiology of pneumonia and the potentially serious progression to sepsis in mind, the overall objective of this thesis is to contribute, if only modestly, to the understanding and improved management of this lifethreatening condition that continues to challenge the global healthcare community. Part I of this thesis focuses on the pathogenesis and host response in pneumonia and sepsis. Chapter 2 is a book chapter from the Handbook of Sepsis, written for educational purposes, to summarize the current understanding of the pathogenesis of sepsis. In 1

14 Chapter 1 Chapter 3 we describe the role of biomarkers in sepsis. Chapter 4 studies the incidence and pathophysiological implications of hyperferritinemia in patients with communityacquired pneumonia, and in Chapter 5 we analyse plasma protein biomarkers reflective of the host response in patients developing ICU-acquired pneumonia. Part II of this thesis is a brief intermezzo on the relation between the microbiome, sepsis and pulmonary immunity. Chapter 6 presents an overview of current knowledge of the effect of the most common empiric antibiotics utilized during sepsis on the composition of the gut microbiota, and the effects of sepsis itself on gut microbiota. In Chapter 7 we describe the outcomes of a human proof-of-concept intervention trial modulating the gut microbiome by antibiotics in patients with adult asthma. Part III of this thesis moves towards improvement of diagnosis and clinical care of patients with pneumonia. In Chapter 8 we address the challenge of proper classification of diagnoses in large (randomized) cohorts of study participants suspected of nontraumatic pulmonary disease, and we present a structured, efficient, and validated approach for diagnostic classification. Chapter 9 contains outcomes of a RCT aimed to study the value of ultra-low-dose chest CT in clinical management of patients suspected of non-traumatic pulmonary disease at the Emergency Department. Chapter 10 specifically aims to address the value of ultra-low-dose chest CT in patients suspected of pneumonia. Part IV can be considered the epilogue of this thesis and consists of a position paper regarding future perspectives on precision medicine in sepsis in Chapter 11, a general discussion of the work in this thesis in Chapter 12, and a summary of the main results of this thesis, both in English and Dutch, in Chapter 13 and Chapter 14.

15 General introduction References 1. Collaborators GL. Estimates of the global, regional, and national morbidity, mortality, and aetiologies of lower respiratory tract infections in 195 countries: a systematic analysis for the Global Burden of Disease Study 2015. Lancet Infect Dis [Internet]. 2017;17:1133–61. 2. File TMJ, Ramirez JA. Community-Acquired Pneumonia. N Engl J Med. United States; 2023;389:632–41. 3. Torres A, Cilloniz C, Niederman MS, Menéndez R, Chalmers JD, Wunderink RG, et al. Pneumonia. Nat Rev Dis Prim. England; 2021;7:25. 4. Prina E, Ranzani OT, Torres A. Community-acquired pneumonia. Lancet. 2015/08/19. 2015;386:1097–108. 5. Metlay JP, Waterer GW, Long AC, Anzueto A, Brozek J, Crothers K, et al. Diagnosis and Treatment of Adults with Community-acquired Pneumonia. An Official Clinical Practice Guideline of the American Thoracic Society and Infectious Diseases Society of America. Am J Respir Crit Care Med. 2019;200:e45–67. 6. American Thoracic S, Infectious Diseases Society of A. Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med [Internet]. 2005;171:388–416. 7. Melsen WG, Rovers MM, Groenwold RHH, Bergmans DCJJ, Camus C, Bauer TT, et al. Attributable mortality of ventilator-associated pneumonia: a meta-analysis of individual patient data from randomised prevention studies. Lancet Infect Dis. United States; 2013;13:665–71. 8. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer M, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). Jama. 2016/02/24. 2016;315:801–10. 9. Rudd KE, Johnson SC, Agesa KM, Shackelford KA, Tsoi D, Kievlan DR, et al. Global, regional, and national sepsis incidence and mortality, 1990-2017: analysis for the Global Burden of Disease Study. Lancet (London, England). England; 2020;395:200–11. 10. Shankar-Hari M, Harrison DA, Rubenfeld GD, Rowan K. Epidemiology of sepsis and septic shock in critical care units: comparison between sepsis-2 and sepsis-3 populations using a national critical care database. Br J Anaesth. England; 2017;119:626–36. 11. Shankar-Hari M, Phillips GS, Levy ML, Seymour CW, Liu VX, Deutschman CS, et al. Developing a New Definition and Assessing New Clinical Criteria for Septic Shock: For the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). Jama. 2016/02/24. 2016;315:775–87. 12. Mayr FB, Yende S, Angus DC. Epidemiology of severe sepsis. Virulence. 2014;5:4–11. 13. Stanski NL, Wong HR. Prognostic and predictive enrichment in sepsis. Nat Rev Nephrol. 2020;16:20–31. 14. Marshall JC. Why have clinical trials in sepsis failed? Trends Mol Med. 2014/03/04. 2014;20:195–203. 15. DeMerle KM, Angus DC, Baillie JK, Brant E, Calfee CS, Carcillo J, et al. Sepsis Subclasses: A Framework for Development and Interpretation. Crit Care Med. 2021;49:748–59. 16. Cajander S, Kox M, Scicluna BP, Weigand MA, Mora RA, Flohé SB, et al. Profiling the dysregulated immune response in sepsis: overcoming challenges to achieve the goal of precision medicine. Lancet Respir Med. England; 2024;12:305–22. 17. Póvoa P, Coelho L, Dal-Pizzol F, Ferrer R, Huttner A, Conway Morris A, et al. How to use biomarkers of infection or sepsis at the bedside: guide to clinicians. Intensive Care Med. United States; 2023;49:142–53. 18. Kullberg RFJ, Wiersinga WJ, Haak BW. Gut microbiota and sepsis: from pathogenesis to novel treatments. Curr Opin Gastroenterol. United States; 2021;37:578–85. 1

16 Chapter 1 19. van den Berk IA, Kanglie MM, van Engelen TS, Bipat S, Dijkgraaf MG, Bossuyt PM, et al. OPTimal IMAging strategy in patients suspected of non-traumatic pulmonary disease at the emergency department: chest X-ray or ultra-low-dose CT (OPTIMACT)—a randomised controlled trial chest X-ray or ultra-low-dose CT at the ED: design and rationale. Diagnostic Progn Res. 8 August 2. 2018. 20. Kanglie MMNP, Bipat S, van den Berk IAH, van Engelen TSR, Dijkgraaf MGW, Prins JM, et al. OPTimal IMAging strategy in patients suspected of non-traumatic pulmonary disease at the emergency department: chest X-ray or ultra-low-dose chest CT (OPTIMACT) trialstatistical analysis plan. Trials. 2020;21:407. 21. Paling FP, Troeman DPR, Wolkewitz M, Kalyani R, Prins DR, Weber S, et al. Rationale and design of ASPIRE-ICU: a prospective cohort study on the incidence and predictors of Staphylococcus aureus and Pseudomonas aeruginosa pneumonia in the ICU. BMC Infect Dis. 2017;17:643.

17 General introduction 1

Part I Pathogenesis and host response

Chapter 2 Pathogenesis of sepsis Tjitske S.R. van Engelen W. Joost Wiersinga Tom van der Poll Handbook of Sepsis, Springer International Publishing, 2018, p. 31-43

22 Chapter 2 Abstract Sepsis is a life-threatening organ dysfunction due to a dysregulated host response to infection. Both hyperinflammation and immune suppression ensue, to an extent that is harmful to the host. The inflammatory balance is disturbed and this is associated with a failure to return to homeostasis. All pathogens with sufficient load and virulence can cause sepsis, after they succeed to adhere and pass the mucosal barrier of the host. The host defense system can recognize molecular components of invading pathogens, called pathogen-associated molecular patterns (PAMPs), with specialized receptors known as pattern recognition receptors (PRRs). Through several signaling pathways, overstimulation of PRRs has pro-inflammatory and immune suppressive consequences. Hyperinflammation is characterized by activation of target genes coding for proinflammatory cytokines (leukocyte activation), inefficient use of the complement system, activation of the coagulation system and concurrent downregulation of anticoagulant mechanisms and necrotic cell death. The release of endogenous molecules by injured cells, called dangerassociated molecular patterns (DAMPs) or alarmins, leads to deterioration in a vicious cycle by further stimulation of PRRs. Features of immune suppression are massive apoptosis and thereby depletion of immune cells, reprogramming of monocytes and macrophages to a state of a decreased capacity to release pro-inflammatory cytokines and a disturbed balance in cellular metabolic processes.

23 Pathogenesis of sepsis Introduction Before the turn of the century, the pathogenesis of sepsis was considered to be driven by an abundant inflammatory response following the invasion of pathogens [1]. Current consensus acknowledges the occurrence of two opposite host reactions to severe infection with proinflammatory and anti-inflammatory features [2]. In sepsis, the normally careful inflammatory balance is disturbed and hyperinflammation together with immune suppression ensue. This dysregulated immune response to infection is associated with a failure to return to homeostasis and harms the host, resulting in the life-threatening condition called sepsis [3]. While insights in the pathogenesis of sepsis have rapidly grown, this complex syndrome is not yet fully understood and our increased understanding of pathophysiological mechanisms underlying sepsis has thus far failed to improve health outcome. This chapter provides a brief overview of the pathogenesis of sepsis (Figure 1). A B CBRAIN LUNG HEART LIVER KIDNEY INTESTINE BONEMARROW delirium encephalopathy ARDS high output distributive shock coagulopathy cholestatis AKI compromised intestinal barrier suppression cytopenia PRO-INFLAMMATORY RESPONSE IMMUNE SUPPRESSION PATHOGEN load virulence PAMPs e.g. LPS, lipopeptide, peptidoglycan, LTA, flagella, DNA, RNA Leukocyte activation Complement activation Coagulation activation Necrotic cell death Reprogramming of monocytes and macrophages Apoptosis immune cells Changes in cellular metabolism DAMPs e.g. HSPs, fibrinogen, hyaluronic acid, HMGB1 PRRs Figure 1. Pathogenesis of sepsis. (A) Sepsis is defined as a dysregulated host response to infection, leading to life-threatening organ dysfunction. The normally careful inflammatory balance is disturbed and this dysregulation is associated with a failure to return to homeostasis. Hyperinflammation and immune suppression ensue, to an extent that is detrimental to the host. (B) Once a pathogen has succeeded to cross the mucosal barrier of the host, it can cause sepsis depending on its load and virulence. The host defense system can recognize molecular components of invading pathogens (PAMPs) with specialized receptors (PRRs). Stimulation of PRRs has pro-inflammatory and immune suppressive consequences. It leads to activation of target genes coding for proinflammatory cytokines (leukocyte activation), inefficient use of the complement system, activation of the coagulation system and concurrent downregulation of anticoagulant mechanisms and necrotic cell death. This starts a vicious cycle with further progression to sepsis, due to the release of endogenous molecules by injured cells (DAMPs or 2

24 Chapter 2 alarmins), which can further stimulate PRRs. Immune suppression is characterized by massive apoptosis and thereby depletion of immune cells, reprogramming of monocytes and macrophages to a state of a decreased capacity to release pro-inflammatory cytokines and a disturbed balance in cellular metabolic processes. (C) Sepsis is by definition a disease with organ failure. The clinical manifestation can be heterogeneous. Clinicians use physical examination, laboratory testing and imaging techniques to determine the severity and origin of organ failure. Antimicrobial treatment is aimed to eliminate the causative pathogen, where supportive care is aimed to restore organ function. Abbreviations: ARDS: acute respiratory distress syndrome, AKI: acute kidney injury, DAMPs: danger-associated molecular patterns, DNA: deoxyribonucleic acid, HMGB1: high-mobility group box-1 protein, HSPs: heat shock proteins, LPS: lipopolysaccharide, LTA: lipoteichoic acid, PAMPs: pathogen-associated molecular patterns, PPRs: pattern recognition receptors, RNA: ribonucleic acid Pathogens and infection sites A successful pathogen must attach to and cross the mucosal barrier, escape the host defense system and multiply to ensure its own survival. All invading microorganisms with a sufficient load and virulence can cause sepsis. However, several pathogens are well known for their impressive arsenal to attack the host. In a point-prevalence study entailing 14,000 Intensive Care Units (ICUs) patients in 75 countries 62% of positive isolates were gram-negative bacteria, versus 47% gram-positive and 19% fungal [4]. The most common gram-negative isolates in sepsis patients are Escherichia coli, Klebsiella sp. and Pseudomonas aeruginosa; the most frequent gram-positive organisms Staphylococcus aureus and Streptococcus pneumoniae [5, 6]. The incidence of fungal infections as the cause of sepsis is rising, which is problematic due to the associated increased mortality. The most common site of infection is the respiratory tract with 63.5% of the culture-positive infections in the ICU, followed by abdominal infections (19.6%), bloodstream infections (15.1%), renal or urinary tract infections (14.3%), skin infections (6.6%), catheter-related infections (4.7%), infections of the central nervous system (2.9%) and others [4]. Host recognition of pathogens The host can recognize molecular components of invading pathogens, called pathogenassociated molecular patterns (PAMPs), with specific receptors. Examples of key bacterial PAMPs are lipopolysaccharide (LPS, also known as endotoxin, a cell wall component of gram-negative bacteria), peptidoglycan, lipopeptides (constituents of many pathogens), lipoteichoic acid (a cell wall component of gram-positive bacteria), flagellin (factor in the mobility of bacteria) and bacterial DNA [7]. In the early response to infection, pathogens or more specifically PAMPs, are recognized by a limited number of specialized host receptors, known as pattern recognition receptors (PRRs). PRRmediated pathogen recognition is an important defense mechanism of the host against invading pathogens and results in upregulation of inflammatory gene transcription and initiation of innate immunity [2, 7, 8]. However, if the innate immune system fails to eradicate the pathogen, overstimulation of PRRs by a growing bacterial load can result in dysregulation of the host response, which then no longer benefits the host but causes tissue injury, organ dysfunction and progression to sepsis. A contributing factor herein is that PRRs can also be stimulated by endogenous molecules released by injured cells, so-called danger-associated molecular patterns (DAMPs or alarmins)[9]. Examples of

25 Pathogenesis of sepsis DAMPs are heat shock proteins, fibrinogen, hyaluronic acid and high-mobility group box-1 protein (HMGB-1)[9]. Thus, PRRs recognize molecular components of both the pathogen (PAMPs) and the host (DAMPs), resulting in a vicious cycle and perpetuation of inflammation. Four main PRR families have been identified: Toll-like receptors (TLRs), C-type lectin receptors (CLRs), Retinoic acid-inducible gene (RIG)-I-like receptors (RLRs) and NOD-like receptors (NLRs)[7, 8]. TLRs comprise the most well-known family of PRRs [7]. They are expressed both extracellularly (TLR1, -2, -4, -5, -6) and intracellularly (TLR3, -7, -8, -9, -10, -11, -12, -13) in endosomes and lysosomes. Ten different TLRs have so far been identified in humans (TLR1 to -10); twelve are found in mice (TRL1 to -9, TLR11, -12, -13)[10]. TLRs are activated by a broad range of ligands presented by bacteria, viruses, parasites, fungi and the host itself. The signaling pathways of TLRs run via four adaptor proteins, namely myeloid differentiation primary-response protein 88 (MyD88), TIR-domain-containing-adaptorprotein (TIRAP), TIR-domain-containing-adaptor-protein-inducing-IFN-β (TRIF) and TRIF-related-adaptor-molecule (TRAM). This signaling eventually leads to the translocation of nuclear factor (NF-κB) into the nucleus which starts the transcoding of genes and is crucial for early activation of the immune system [8]. As an example of TLR signaling, TLR4 is stimulated through its ligand LPS, the virulence factor of gramnegative bacteria. It activates both the MyD88- and the TIRAP-dependent pathways for early-phase activation of NF-κB and results in late-phase activation of NF-κB via the TRIF-dependent pathway [7]. TLR3 is stimulated by dsRNA derived from viruses or virus-infected cells and activates the TRIF-dependent pathway [8]. NLRs are cytoplasmic proteins composed of a central nucleotide-binding domain and C-terminal leucine-rich repeats [11]. NLRs are an important factor in the initial immune response through their formation of multiprotein complexes called “inflammasomes”. These complexes activate caspase-1 leading to the maturation of proinflammatory cytokines interleukin 1β (IL-1β) and IL-18 [12]. RLRs are cytoplasmic proteins that can recognize the genomic RNA of RNA viruses [13, 14]. CLRs are transmembrane receptors with a carbohydrate-binding domain. CLR-mediated microbial recognition occurs through their ability to recognize carbohydrates on viruses, bacteria and fungi (Table 1). 2

26 Chapter 2 Table 1. Pattern recognition receptors and their ligands in humans Pattern recognition receptor Ligand Origin of ligand Toll-like receptors (TLRs) TLR1 Triacyllipoprotein (forms heterodimer with TLR2), soluble factors Bacteria TLR2 Lipoprotein (forms heterodimer with TLR1 and TLR6) Bacteria, viruses, fungi, self TLR3 Double-stranded RNA Viruses TLR4 Lipopolysaccharide, envelop proteins (syncytial viruses), glycoinositol phospholipids, HSPs 60 and 70, S100a8 (ligand from dying cells) Bacteria, viruses, self TLR5 Flagellin Bacteria TLR6 Diacyllipoprotein (forms heterodimer with TLR2) Bacteria, viruses TLR7 Single stranded RNA, synthetic compounds (e.g. imidazoquinolines) Bacteria, viruses, self TLR8 Single stranded RNA, small purine analog compounds (imidazoquinolines) Viruses TLR9 CpG-DNA, insoluble crystal hemozoin (Plasmodium falciparum) Bacteria, viruses, parasites, self TLR10 Unknown NOD-like receptors (NLRs) NOD1 Peptidoglycan (iE-DAP) Bacteria NOD2 Peptidoglycan (MDP) Bacteria C-type lectins (CLRs) Dectin-1 β-Glucan Fungi Dectin-2 β-Glucan Fungi MINCLE SAP130 Fungi, self Retinoic-acid inducible gene (RIG)-I-like receptors (RLRs) RIG-I Short double-stranded RNA, 5’triphosphate dsRNA Viruses MDA5 Long double-stranded RNA Viruses LGP2 Double-stranded RNA Viruses DDX3 Viral RNA Viruses The innate immune system recognizes pathogens by four main classes of pattern recognition receptors. The table shows the main receptors, their main ligands and the origin of these ligands. Note that some receptors also recognize “self” antigens, primarily in the context of injury, wherein self-antigens function as alarmins to the host. Abbreviations: CpG-DNA: cytosine-phosphateguanosine-DNA, DDX3: DEAD/H Box 3, iE-DAP: g-D-glutamyl-meso-diaminopimelic acid, LGP2: Laboratory of genetics and physiology-2, MDA5: melanoma differentiation-associated gene 5, MDP: muramyl dipeptide, MINCLE: macrophage-inducible C-type lectin, SAP130: Sin3Aassociated protein of 130kDa. Table adapted from references 8, 10 and 59.

27 Pathogenesis of sepsis Hyperinflammation Sepsis is associated with a strong activation of the immune system, by stimulation of PRRs by PAMPs and DAMPs, leading to the activation of target genes coding for proinflammatory cytokines such as tumor necrosis factor (TNF), IL-1β, IL-12 and IL18[2]. Cytokines are small proteins that can regulate the host response both locally and systemically, after their release from various cell types such as monocytes and neutrophils. These cells can further attribute to activation of the immune system by expression of the Triggering Receptor Expressed on Myeloid cells-1 (TREM-1) that amplifies TLR- and NLR-mediated inflammatory response [15]. Several mechanisms regulate the activation of PRRs to avoid overstimulation, including the negative regulators MyD88 short (MyD88s), ST2, single-immunoglobulin-interleukin (IL)-1 receptor-related-molecule (SIGIRR), toll-interacting protein (TOLLIP), suppressorof-cytokine signaling (SOCS), A20 and IRAK-M [16]. If the delicate balance between activation and inhibition of the inflammatory response is disturbed, the pleiotropic hyperinflammatory response in sepsis ensues. This includes activation of the complement and coagulation systems and disturbance of vascular permeability [2], which have been considered important factors in sepsis mortality. Complement system The complement system comprises over 40 components that, when activated, work as a cascade and contribute to the innate immune surveillance system [17, 18]. A close collaboration between the complement system and other proinflammatory stimuli such as cytokines is necessary: the complement system tags dangerous cells or pathogens, and phagocytic cells can respond more properly after activation by proinflammatory mediators. This teamwork is dysregulated in sepsis resulting in inefficient use of the complement system. The complement system contributes directly to the activation of the immune system by the release of anaphylatoxins C3a and C5a. Anaphylatoxins are proinflammatory molecules that activate surrounding cells when they reach a threshold concentration, can lead to the recruitment of other immune cells (macrophages, basophils, neutrophils, eosinophils and mast cells) and can activate endothelial and epithelial cells and platelets [17, 18]. The harmful role of C5a in sepsis has been linked to neutrophil dysfunction, apoptosis of lymphoid cells, exacerbation of systemic inflammation, cardiomyopathy, disseminated intravascular coagulation (DIC) and complications associated with multiple organ failure [19]. Several experimental sepsis studies have highlighted the beneficial effect of blockage of C5a signaling on outcome [20]. As such, C5a is considered a potential therapeutic target in sepsis. Coagulation system and vascular endothelium Activation of PRRs leads to upregulation of inflammatory mediators which results in a systemic inflammatory response, including activation of the coagulation system and concurrent downregulation of anticoagulant mechanisms [21]. Coagulation abnormalities can range from mild to clinically relevant fulminant coagulopathies. DIC is the most severe manifestation of disturbed hemostasis with microvascular thrombosis and, through consumption of clotting factors and platelets, simultaneous 2

28 Chapter 2 hemorrhage [22]. The most important initiator of coagulation in sepsis is tissue factor (TF). Indeed, inhibition of TF prevents DIC and improves survival in experimental sepsis [21]. TF is predominantly produced by macrophages and monocytes and its expression is enhanced by proinflammatory cytokines, exemplifying the close interaction between inflammation and coagulation [23]. Furthermore, TF can reside in micro particles that are formed by hematopoietic and endothelial cells. These micro particles play a significant role in both coagulation and inflammation [24]. In healthy hosts, coagulation is controlled by three main anticoagulant pathways: the antithrombin system, tissue factor pathway inhibitor (TFPI) and the protein C system. In septic patients all these pathways are impaired in their function, partially due to endothelial dysfunction, resulting in low levels of these coagulation inhibitors [25, 26]. The physiological function of the protein C system has been supported by investigations in which interventions inhibiting this pathway resulted in severe coagulopathy and death in otherwise nonlethal infection models. During the early stages of inflammation plasminogen activators are released to help breakdown fibrin. Sepsis is associated with high levels of plasminogen activator inhibitor type 1 (PAI-1), a main inhibitor of fibrinolysis, further facilitating microvascular thrombosis [27]. The interaction between inflammation and coagulation is not unilateral. Coagulation factors regulate inflammation in particular through proteolytic cleavage of protease activated receptors (PARs)[28]. Activated protein C (APC) influences inflammation, by reducing the expression of receptors for cytokines and chemokines [29], by downregulating the production of inflammatory mediators [30, 31] and by blockage of cytokine release and leukocyte activation [32]. During sepsis the vascular endothelium is involved in the disturbance of anticoagulant mechanisms. Glycosaminoglycans on the endothelial surface supports antithrombin mediated inhibition of thrombin formation and platelet adhesion. Sepsis reduces the production of glycosaminoglycans averting not only antithrombin function, but also that of TFPI with regard to inhibiting the main coagulation TF-factor VIIa complex. In healthy hosts endothelium generates APC from protein C through an interaction between thrombin and thrombomodulin (a receptor expressed by endothelial cells); formation of APC by the thrombomodulin-thrombin complex is accelerated by the endothelial protein C receptor (EPCR). APC inactivates coagulation cofactors Va and VIIIa by proteolysis, thereby inhibiting coagulation. In sepsis APC levels are reduced due to impaired production caused by downregulation of both thrombomodulin and EPCR on endothelial cells, as well as by increased consumption. Adhesion of cells to the endothelium is increased in sepsis. Physiologically, injured endothelium activates von Willebrand factor which forms multimers at the site of injury as a primary step in protective coagulation [25]. Von Willebrand multimers are cleaved by a proteolytic enzyme ADAMTS13 to control adhesion and prevent formation of large obstructive von Willebrand multimers. In sepsis there is a relative deficiency of ADAMTS13 leading to ultra-large von Willebrand multimers at injured sites, contributing to overwhelming platelet adhesion and microvascular thrombosis, and possibly eventually multiple organ dysfunction. Furthermore, activation of platelets because of vascular injury during sepsis starts a vicious cycle which leads to more activated endothelium and platelets which further increases coagulation [25].

29 Pathogenesis of sepsis Impaired vascular barrier function is a key pathogenic mechanism in sepsis, associated with protein leakage into the extravascular space, tissue edema and diminished microvascular perfusion [25]. Important regulators of vascular barrier function are sphingosine-1-phosphate (S1P) and angiopoietin-1 [25, 33]. SP1 activates the endothelial S1P receptor 1, thereby preserving vascular integrity [33]. Angiopoietin-1 activates TIE2, supporting barrier function. Angiopoietin-2 antagonizes angiopoietin-1 and a high angiopoietin-2/angiopoietin-1 ratio has been used as a marker for vascular barrier dysfunction in patients with sepsis [34]. Neutrophil extracellular traps Activation of the coagulation system and vascular injury are amplified by the release of neutrophil extracellular traps (NETs) by neutrophils [35]. NETs are composed of DNA, histones and neutrophil-derived proteinases and can protect the host by eliminating pathogens. However, NETs may also contribute to collateral damage and thrombosis in the dysregulated immune response in sepsis [35]. Immune suppression Much attention has been drawn to immune suppression in patients which sepsis, which in many patients can already be detected on admission to the ICU and is a prominent feature in those patients that remain in the ICU for extended periods of time [2, 36]. Targeted immune-enhancing therapy may be beneficial for selected patients with immune suppression [2, 36]. Transcriptomic analysis of peripheral blood leucocytes of septic patients recently resulted in the classifications of distinct sepsis endotypes with implications for main pathophysiological mechanisms and prognosis [37, 38]. These studies further confirmed the existence of subgroups of sepsis patients with a predominant immune suppressive phenotype [37, 38]. Apoptosis of immune cells Sepsis associated immune suppression involves several cell types. During sepsis massive apoptosis leads to depletion of immune cells, especially CD4+ and CD8+ T cells and B cells. This depletion is seen in lymphoid organs and body sites, such as spleen, thymus, lymph nodes and gut-associated lymphoid tissue [36, 39]. T regulatory (Treg) cells are more resistant to sepsis-induced apoptosis which, combined with the substantial apoptosis of CD4+ and CD8+ T cells and B cells, lead to a more immunosuppressive phenotype. Furthermore, surviving CD4+ and CD8+ T cells shift from a Th1 proinflammatory phenotype to the more immunosuppressive Th2 phenotype. Inhibition of lymphocyte apoptosis was associated with better outcomes in various experimental sepsis models, suggesting a causal relationship between lymphocyte apoptosis and sepsis mortality [2, 36]. A recently identified potential therapeutic target in sepsis is the programmed cell death 1 (PD1) – PD1 ligand (PDL1) pathway. Patients with sepsis showed enhanced expression of PD1 on CD4+ T cells together with increased expression of PDL1 on macrophages and endothelial cells [39]. Enhanced PD1 – PDL1 interaction 2

30 Chapter 2 is expected to impair T cell function, and in mice inhibition of this pathway conferred protection against lethality following experimentally induced sepsis [40]. Clinical trials seeking to inhibit PD1 – PDL1 signaling in sepsis patients are under way. Contrary to lymphocytes, apoptosis of neutrophils in sepsis is delayed [2, 36]. Furthermore, the bone-marrow releases immature neutrophils which together results in high numbers of circulating neutrophils in different stages of maturation. The function of neutrophils is impaired in sepsis, with reduced chemotaxis and reactive oxygen production. Reprogramming of monocytes and macrophages Sepsis is further characterized by profound changes in the function of antigen presenting cells [2, 36]. Monocytes and macrophages demonstrate a strongly decreased capacity to release pro-inflammatory cytokines upon stimulation with bacterial agonists (a feature commonly referred to as “endotoxin tolerance”) and reduced HLA-DR expression. Notably, monocytes/macrophages do not show a general unresponsiveness, but rather are reprogrammed: after stimulation with bacterial compounds they produce equal or even increased amounts of anti-inflammatory cytokines. Correspondingly, mRNA expression levels of genes encoding pro-inflammatory mediators have been reported downregulated upon stimulation with concurrent upregulation of mRNA’s of antiinflammatory mediators [2, 36]. HLA-DR expression on monocytes has been suggested as a biomarker to select sepsis patients for immune stimulatory therapy. Epigenetic regulation of gene function likely plays a significant role in the host response to infection through suppression of proinflammatory gene expression and/ or activation of anti-inflammatory genes, thereby contributing to immune suppression [41]. Protein expression can be regulated both at the pre- and post-transcriptional level. Pretranscriptional regulation takes place on chromatin, the complex formed by the DNA double helix packaged by histones. The gene loci on chromatin can be organized in transcriptionally active “euchromatin” or transcriptionally silent “heterochromatin”. The chromatin activation state is regulated by histone modifications due to acetylation, methylation, ubiquitination and phosphorylation. For example, acetylation of lysine residues within histones usually facilitates transcription [41]. “Endotoxin tolerance” in monocytes has been linked to reduced expression of marks of open chromatin such as histone H3 lysine 4 trimethylation (H3K4me3)[42], and “endotoxin tolerant” macrophages showed enhanced levels of the repressive histone modification H3K9 dimethylation (H3K9m2) at the promoter sites of the genes encoding the proinflammatory cytokines TNF and IL-1β [43]. One mechanism by which microbial stimuli induce epigenetic gene regulation is through increased expression of the histone lysine demethylase KDM6B via NF-κB activation [44]. KDM6B primes genes for transcription and it is postulated that this promotes IL-4 maturation. The latter is a potent cytokine to counteract various proinflammatory cytokines and contributes to immune suppression. This IL-4/KDM6B axis appears to be one of the important pathways in the epigenetic regulation of macrophage activation [41]. The immune suppressive effects of sepsis can remain for months, perhaps even longer. It is hypothesized that epigenetic imprints occur both on mature immune cells in the periphery as well as progenitor cells in the bone marrow, thereby contribution to this long lasting immune suppression [41].

31 Pathogenesis of sepsis Cellular metabolism Changes in cellular metabolism may contribute to immune suppression [45]. A metabolic shift from oxidative phosphorylation to glycolysis, known as the Warburg effect, is critical for cells to mount an inflammatory response when stimulated by LPS, and failure in this shift can result in reduced cellular responsiveness [2]. Consequently, an imbalance in cellular metabolism has been associated with the altered phenotype of monocytes in sepsis. However, the underlying mechanisms appear to be more complex than simple shifts between oxidative phosphorylation and glycolysis [2]. Unlike LPS, which triggers a classical Warburg effect, other bacterial stimuli have been observed to increase both glycolysis and oxidative phosphorylation in monocytes [46]. Likewise, in sepsis patients with immune suppression, monocyte metabolic dysfunction is not confined to glycolysis but extends to a broad suppression of metabolic pathways, including glycolysis, fatty acid oxidation, and oxidative phosphorylation [47]. Microbiome The microbiome consists of trillions of bacteria of which most are found in the gastrointestinal tract [48]. Dysbiosis of the microbiome (meaning a decreased microbial diversity) has been associated with altered immune responses (for instance altered cytokine production capacity of immune cells). Sepsis affects the composition of the intestinal microbiome, characterized by a loss of diversity, lower abundances of key commensal genera (such as Faecalibacterium, Blautia, Ruminococcus) and overgrowth of opportunistic pathogens [49]. Small studies show that the gut is overrun by a single bacterial genus in patients with sepsis, most notably by Clostridium difficile, Staphylococcus spp., Escherichia spp., Shigella spp., Salmonella spp., and Enterococcus spp. [50]. This overgrowth by one genus occurs in roughly one third of the septic patients, but increases with time spend on the ICU [51]. The underlying mechanism is not fully understood, but antibiotic treatment that is part of standard care in septic patients seems to have the most disruptive effect on the microbiome, possibly amplified by the use of (par)enteral feeding and gastric acid inhibitory drugs [52]. Murine studies support a role for the microbiome in regulation of granulocytosis, neutrophil homeostasis and host resistance to sepsis [53]. In pneumonia derived sepsis disruption of the gut microbiome impaired host defense; underlying mechanisms likely include a reduced responsiveness to microbial stimulation and an impaired phagocytosis capacity of alveolar macrophages [54]. In addition, neutrophils from microbiota-depleted mice demonstrated a diminished capacity to migrate into inflamed tissues [55]. The immune response can further be compromised when translocation of pathological microbes through disintegrated epithelial barriers results in systemic and lymphatic spreading of pathogens. Theories of connections between the gut microbiome and distant organ function, the so-called gut-organ axis, are rapidly developing. For instance, a recent study showed evidence of gut bacteria present in the lung microbiome in mice with experimental sepsis and humans with acute respiratory distress syndrome, supporting the existence of the gut-lung axis [56]. Research concerning the pathophysiological mechanism underlying these phenomena is growing rapidly [52, 57], as are studies regarding the microbiome as a therapeutic target in critically ill patients [58]. 2

RkJQdWJsaXNoZXIy MTk4NDMw