key: cord-0852554-fijo89i9
authors: Valencia, Braulio M.; Cvejic, Erin; Vollmer-Conna, Ute; Hickie, Ian B.; Wakefield, Denis; Li, Hui; Pedergnana, Vincent; Rodrigo, Chaturaka; Lloyd, Andrew R.
title: The severity of the pathogen-induced acute sickness response is affected by polymorphisms in genes of the NLRP3 inflammasome pathway
date: 2021-01-09
journal: Brain Behav Immun
DOI: 10.1016/j.bbi.2021.01.005
sha: f5f88ba0ebb8ca644ce574995752089a49fb82da
doc_id: 852554
cord_uid: fijo89i9

The acute sickness response (ASR) is a stereotyped set of symptoms including fatigue, pain, and disturbed mood, which are present in most acute infections. The immunological mechanisms of the ASR are conserved, with variations in severity determined partly by the pathogen, but also by polymorphisms in host genes. ASR was characterised in three different serologically-confirmed acute infections in Caucasians (n=484) across four symptom domains or endophenotypes (termed ‘Fatigue’, ‘Musculoskeletal pain’, ‘Mood disturbance’, and ‘Acute sickness’). Correlations were sought with functional single nucleotide polymorphisms in the NLRP3 inflammasone pathway and severity of the endophenotypes. Individuals with severe Fatigue and Musculoskeletal pain endophenotypes were more likely to have prior episodes of significant fatigue (11.4 vs. 3.8%, p=0.07) or pain (14.3 vs. 1.2%, p=0.001), suggesting trait characteristics. The high functioning allele of the rs35829419 SNP in NLRP3 was more common in those with severe Fatigue (OR=13.3, 95% CI: 1.7-104), particularly in a dominant inheritance pattern (OR=13.4, 95% CI: 1.8-586.3). In a multivariable analysis assuming dominant inheritance, both rs35829419 and the rs4848306 SNP in Interleukin(IL)-1β, were independently associated with severe Fatigue (OR=29.6, 95% CI: 2.6-330.9 and OR=13, 95% CI: 2.7-61.8, respectively). The severity of fatigue in acute infection is influenced by genetic polymorphisms in NLRP3 and IL-1β.

Individual infectious diseases are associated with a wide variety of sickness manifestations ranging from brief and minimally symptomatic illness, through to severe, life-threatening conditions. This is well exemplified by infection with the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which ranges in manifestations from asymptomatic infection through to severe acute respiratory distress syndrome (ARDS) 1 . This range in the severity of infectious disease manifestations is shaped by pathogen characteristics, and by the host immune response, which in turn is influenced by the genetic make-up of the host, as well as acquired co-morbidities, and environmental factors 2 .

The innate immune responses of the host are largely conserved regardless of the infecting pathogen.

Such responses are initiated via pathogen-associated molecular patterns (PAMPs) or tissue damageassociated molecular patterns (DAMPs), which are recognised by widely expressed pattern-recognition receptors (PRRs), such as the Toll-like receptor (TLR) family. Activation of PRRs induces type-1 Interferon (IFN) responses to rapidly control replication and prevent the dissemination of intracellular pathogens 3 . PRRs also trigger pyrogenic responses via the induction of pro-inflammatory cytokines required for the recruitment and activation of additional immune cells 3 . One such innate pathway is mediated by a nucleotide-binding oligomerization domain (NOD)-like receptor pyrin domain-containing 3 (NLRP3) inflammasome which activates caspase 1 (CASP1) through caspase recruitment domains (CARD) 4 . CASP1 enhances the biological activity of several pro-inflammatory cytokines, particularly Interleukin-1β (IL-1β) and Interleukin-18 (IL-18), which promote both innate and adaptive responses 4 . NLRP3 responses are triggered by the recognition of viral RNA, bacterial products or sub-particles, as well as by tissue injury factors such as release of extracellular ATP, or the efflux of K + 3 . Interestingly, single nucleotide polymorphisms (SNPs) in different components, or regulators, of this pathway have been associated with heightened innate immune responses in different infective conditions, including dengue 5 , hepatitis C 6 , and tuberculosis 7 , but also auto-immune conditions as rheumatoid arthritis 8, 9 . 4 In this study functional SNPs in the NLRP3 pathway were examined in relation to the severity and duration of the acute sickness response (ASR) to several different infectious diseases.

The clinical manifestations generated by these innate responses are also highly conserved. Fever, fatigue, musculoskeletal pain, disturbed mood, cognitive impairment, loss of appetite and increased sleep requirements are all typical manifestations during acute infections, and are collectively defined as the ASR 10, 11 . Previous studies in the cohort reported here have identified associations between overall severity of the ASR, and functional single nucleotide polymorphsims (SNPs) in the pro-and antiinflammatory cytokines, IFN-γ, tumour necrosis factor (TNF)-α, IL-6, and IL-10 12 . In particular, the highproducing allele rs2430561 of pro-inflammatory cytokine IFN-ɣ, and the low-producing allele rs1800872 SNP of the anti-inflammatory cytokine IL-10 were associated with greater severity and prolonged duration of illness among n=296 individuals individuals affected by Epstein-Barr virus (EBV; which causes infectious mononucleosis), Ross River virus (RRV; a mosquito borne infection causing polyarthritis), and Coxiella burnetii, the causative agent of Q fever (QF) 13 . As C burnetii is a Gram negative, lipolpolysaccaride (LPS)-containing intracellular bacterium polymorphisms in the LPS receptors, TLR-2 and TLR-4 were also previously explored in n=85 individuals affected by QF, although no association with illness severity or course was identified 14 .

Although consistently present in the ASR, observations in this cohort revealed that individual manifestations often differed in intensity and duration among individuals infected with the same pathogen 12 . Accordingly, to better understand the genetic determinants of the ASR, these symptom manifestations have been grouped here into symptom domains, or endophenotypes, which encompass the set of self-reported symptoms of fatigue (e.g. feeling weak or tired, waking up tired), pain (e.g. myalgias, arthralgias), mood disturbance (e.g. irritability, or low mood), or acute sickness (e.g. fevers, anorexia) 12 . As originally described in psychiatry 15 , the endophenotype concept stipulates that the 5 characteristic should be reliably measurable, state independent (i.e., the characteristic is an underlying trait rather than simply being a feature of the current disease state), and heritable. In the psychiatric context, even if the endophenotype is genetically determined such as suicide risk 16 , it is clear that environmental factors (such as life stressors) modulate the expression. Accordingly, the present study aimed to explore the heritable nature of ASR endophenotypes and their association with nine SNPs in the NLRP3 inflammasome pathway among individuals acutely infected by EBV, RRV, or QF.

Longitudinally collected clinical datasets and DNA from participants of the Dubbo Infection Outcomes Study (DIOS) were used 17, 18 . The DIOS cohort enroled n=484 adults of Caucasian descent, who were acutely affected by a febrile illness which was provisionally diagnosed in primary care as being EBV, RRV or QF infection on the basis of screening by pathogen-specific IgM enzyme-linked immunosorbent assays ( Figure 1 ). These subjects were prospectively followed-up at 2-3 weeks, 4-6 weeks, and then three-monthly until 12 months post-infection. The EBV, RRV, or QF infections were confirmed by acute and convalesecent serological testing for pathogen-specific IgG antibodies. At baseline, a structured medical history interview was recorded to capture pre-morbid (i.e. pre-infection) physical and mental health status. At each follow-up visit, self-report questionnaires were administered to record the presence and severity of physical and psychological symptoms related to the ASR. For the purposes of the analysis reported here, demographic data was used in addition to that captured by the 34-item Somatic and Psychological Health Report (SPHERE) 19 This genetics study was approved by the institutional Human Research Ethics Committee at UNSW (HC190102). In DIOS, all subjects provided written informed consent for prospective clinical assessment and blood sampling for DNA storage and analysis under (HC 04257). These studies have adhered to the principles set out in the Declaration of Helsinki.

Data from the 53 individual items of the SPHERE and PSC obtained at the baseline were used to derive indices of the endophenotypes of the ASR using a principal component analysis (PCA) with orthogonal rotation 12 . SPHERE (34 items) 18 and PSC (19 items) 19 are self-administered questionnaires recording the prevalence ''in the past few weeks" of a wide range of physical and psychological symptoms of acute infection (see Table S1 for details of the question items), recorded as "none of the time or some of the time" (scored as 0), "a good part of the time" (scored as 1), and "most of the time" (scored as 2). Six items duplicated between the questionnaires were removed. Only the first four principal components were retained because they satisfied the eigenvalue-one criteria, and each provided an independent contribution greater than 3% to the overall variance 20 . These four PCs were labelled on the basis of the dominant theme captured by the items as ( Figure 1a ): 'Fatigue', 'Musculoskeletal pain', 'Mood disturbance', and 'Acute sickness' 18 . The loading scores for each questionnaire item in these indices were calculated from the baseline dataset and then applied to the data from each of the follow-up assessments (see Supplementary material, Table 1 ).

The SOMA subscale of the SPHERE questionnaire was used in the orginal study to designate the timepoint when the overall acute illness had resolved (with a score <3 being indicative of illness resolution) 12 . Using data from the 12 month follow-up timepoint from subjects (n=327) whose acute illness had resolved within 4 weeks of initial enrolment, a 'threshold' for background in the PC score for 7 each endophenotype was designated as the 99 th percentile of the recovered group. This threshold value was then used to identify the timepoint at which each subject first recorded scores below the threshold for each endophenotype. The midpoint in days between the last above-threshold and the first below-threshold timepoint was used to determine the duration of illness within each endophenotype from symptom onset to resolution (i.e. below background).

As trait characteristics are intrinsic and lifelong modulated by environmental factors 21, 22 , the medical history interview was scrutinised for prior episodes of: severe acute infections (as a surrogate for a propensity to severe Acute sickness), severe and prolonged fatigue or pain (as a surrogate for a propensity to severe Fatigue or Musculoskletal pain), and for premorbid mood disorder (as a surrogate for a propensity to severe Mood disturbances).

Individuals with mild or severe manifestations were selected from the lower and upper quartiles of the severity scores across all endophenotypes (see Figure 1b ) to examine the relationship between acute phase severity and duration of illness. A reverse genetics approach, which starts from a trait without genetic information and then works backward to find a genetic variant, was then applied 23 . To do this, of those subjects reporting mild or severe symptoms for each endophenotype, only who had a symptom onset less than 42 days prior to enrolment were retained in the analysis to ensure temporal proximity with the acute illness phase for reliable reporting of symptoms ( Figure 1c ). Finally, for comparisons of allelic frequencies, cases with severe endophenotypes were matched groupwise by age (within 10 years), gender, and the pathogen causing the acute illness to controls (Figure 1d ). illness groups were analysed by the logrank test comparing persistence of symptoms among each 9 endophenotype according to baseline severity. These comparisons were adjusted by age, gender, and infecting pathogen by Cox proportional hazards. Right censoring for the complete dataset was defined at 90 days after the initiation of symptoms. Chi-square tests for independence and odds ratio (OR) calculations with their respective 95% confidence intervals (95% CI) were used to infer whether the alleles and genotypes were associated with severe endophenotypes in the case-control groups using PLINK (http://zzz.bwh.harvard.edu/plink/) 26 .. A p-value of 0.05 was considered the threshold for statistical significance. p-values of allelic associations were subject to Bonferroni correction. All SNPs were in Hardy-Weinberg equilibrium in the study population.

Participants in the cohort (n=484) were predominantly female (n=248/484; 51%), and their median age was 32 years (interquatile range 18-42, range 16-69). EBV was the most frequent infection (n=144/484; 30%), followed by RRV (n=98/484; 20%), and QF (n=84/484; 17%). The pathogen provisionally implicated in the acute infectious disease was not serologically confirmed in 33% of participants (n=158/484). Endophenotype scores were not available at some longitudinal timepoints in a minority of individuals, hence the Fatigue endophenotype subgroup was composed of n=358 individuals, the Musculoskeletal pain endophenotype by n=390, the Mood disturbance endophenotype by n=394, and the Acute sickness endophenotype by n=387 participants (Figure 1 ). The baseline characteristics of endophenotype subgroups were all similar to the entire cohort (Table 1 ). 

The kinetics of resolution of illness in relation to baseline illness severity (Figure 1a 

Subjects with more severe acute phase illness, reflected by high scores for the individual endophenotypes (Figure 1b) , were also more likely to have pre-morbid illness reports suggesting these 13 same features, reflecting prior vulnerability or trait characteristics ( Figure 1b . 16 

As these differences between severe and mild endophenotypes were also potentially attributable to other features of the groups including age, gender and causative pathogen, the severe and mild comparison sub-groups for each endophenotype were identified from those with six weeks from illness onset at enrolment (Figure 1c ), and then matched by age, sex, and aetiological agent, for case-control testing of genetic associations (Figure 1d ). Demographic characteristics for these sub-groups were comparable across all endophenotypes, with no differences between cases and controls in key features as age, aetiological agent as well as the number of days between the initiation of symptoms and the baseline assessment (Table S2) Those with severe symptoms in the Acute sickness endophenotype also had more days in bed due to illness (7 vs. 2 days, p=0.0007).

The severity of the Fatigue endophenotype was associated with the minor allele of NLRP3 rs35829419

( respectively) assuming a dominant inheritance mode. This effect was not found for the Acute sickness endophenotype.

Given the strong association between severity and duration of the endophenotypes, the potential role of these genetic variants in determining illness duration was also investigated. No genetic association with sex, age, or infection type was identified for duration of illness for any of these endophenotypes.

Similarly, no associations with two SNPs associated with acute illness severity -NLRP3 rs35829419

and IL-1β rs4848306, were found with Fatigue duration (Cox proportional hazards model; p>0.2).

The severity of infectious diseases has been traditionally linked with pathogen virulence factors, as well as co-morbid conditions in the host, including immunosuppressed states and non-communicable conditions such as diabetes, obesity, or cardiovascular disease 27 . Despite the major advances in genetics including high throughput sequencing technologies, the contribution of genetic variation to infectious diseases outcomes has been explored for relatively few pathogens. Using the notion that endophenotypes in the ASR are similarly triggered by different infections, we have shown that these symptom domains appear to represent underlying vulnerabilities repetitively affecting susceptible individuals. Further in a case-control design with a reverse genetics approach, two functional SNPs in the NLRP3 inflammasome pathway were found to be associated with the severity of fatigue, independent of the pathogen causing the infection or demographic characteristics. Thus, our study supports the concept of endophenotypes applied to the ASR, the notion of genetically-determined 20 contributions to these symptom domains, and implicates the NLRP3 inflammasome in the genesis of the ASR.

The causative pathogens in this study (EBV, RRV and QF) generally cause self-limiting acute infections which are seldom associated with hospitalisation or mortality [28] [29] [30] [31] . Hence, the study sought to better understand genetic determinants of the varied severity of the typical acute illness manifesations. The generation of endophenotype scores using PCA provided a quantitative characterisation of otherwise qualitative symptoms such as fatigue. These endophenotypes were constructed from individual symptom items (Table S1 ), but also captured their frequency and individual contribution within each broad domain, generating a solid phenotype conceptualisation. These scores additionally highlighted differences in independent measures of the functional impact of the different symptom domains including the number of days out of the usual role, and days in bed in the acute phase of the illness 17, 32, 33 . It should be noted that the original concept of endophenotypes from psychiatry research stipulated that the endophenotype should be 'state' independent (that is consistently present whether the particular illness is present or in remission) 15 . In the context of acute infection, this may be taken to imply that individuals who experience severe fatigue associated with one of the infections in this study may similarly experience the same pattern with another acute infection or inflammatory insult. We corroborated this assumption with evidence that individuals experiencing severe pain or mood disturbance endophenotypes reported a higher frequency of similar episodes in the past than those with mild manifestations. It is also possible that the endophenotype may cross boundaries (i.e. be transdiagnostic), so that the propensity to mood disorder associated with acute infection may predict the advent of mood disorder consequent upon more traditional psychosocial stressors. Further investigations are required to test these notions as well as to explore their usefulness in clinical settings

where the identification of individuals with at-risk endophenotypes could preceed the development of severe manifestations. 21 In the analysis presented here, the participants in the lower and upper severity quartiles were selected for the matched case-control comparison. This approach was adopted firstly as the identification of risk or protective factors in case-control studies is more efficient when confounding factors are removed or controlled 34 . This may be considered the main limitation of genome-wide association studies (GWAS)

which often survey an enormous number of SNPs in subjects with broad and relatively heterogenous phenotypes, requiring very large sample sizes to manage confounding factors 35 . Secondly, the phenotype-to-genotype, or reverse genetics approach, and the candidate pathway chosen here for investigation, rely upon an evidence base of biological plausibility that the pathway is implicated in the pathophysiology of the illness 36 . The NLRP3 and pro-inflammatory cytokine pathway chosen for investigation here and the process of selection of functional SNPs meets this expectation.

Fatigue is reported as a dominant element of the ASR 12, 17, 18 , but is also prevalent in other conditions, including in association with cancer 37 , and autoimmune diseases 38 been all been associated with the severity of fatigue 37, 39, 40 . Several immunological therapies are also recognised to have fatigue and mood disturbances as a prominent adverse effects, with the most notable being IFN- treatments which were previously used commonly for chronic hepatitis C infection 41 . In the current SARS-CoV-2, fatigue and myalgias are among the most frequent symptoms reported 42 . Interestingly, a high frequency of post-infective fatigue has been reported after SARS-CoV-2 infection, independent of acute phase illness severity 43 .

The findings reported here that the NLRP3 rs35829419 is associated with severity of the fatigue endophenotype, and has a synergistic interaction with IL-1β rs4848306 extend the previous genetic association findings. The A allele of NLRP3 rs35829419 is a missense mutation associated with an overactive NLRP3 inflammasome generating increased CASP1 activity and excessive IL-1β and IL-18 levels 44 . Although sometimes considered as a SNP of uncertain clinical significance, an increase in plasma IL-1β levels were found in association with the A allele of NLRP3 rs35829419 and the T allele of CARD8 rs2043211 45 . Further, NLRP3 rs35829419 has been associated with multiple auto-immune conditions 46 . The IL-1β rs4848306 is located in an NF- binding site in the promote region, and the A allele has been associated with transcriptional factor binding and reduced IL-1β mRNA expression 47, 48 .

In the data presented here an increased risk of severe fatigue was putatively associated with increased levels of IL-1β induced by an overactive NLRP3 inflammasome due to the A allele in the rs35829419, and this effect could be magnified by the presence of a G allele of the IL-1β rs4848306, which enhances IL-1β mRNA expression.

Even though these associations were significant, the study has limitations. Firstly, there were wide confidence intervals for each relevant association involving rs35829419. Wide confidence intervals are multi-factorial, but are generally with inadequate sample sizes and co-linear interactions among variables 49 . We consider sample size is the main determinant in the data reported here. Secondly, although the endophenotypes may be influenced by psychosocial factors, these were not considered in the case-control matching -largely due to limitations in sample size. Accordingly, the findings require corroboration and replication in prospective studies with larger sample sizes, and careful characterisation of endophenotypes and outcomes of interest. 23 

This study has identified the fatigue endophenotype as a likely genetic determinant of the severity of the ASR, particularly implicating the NLRP3 inflammasome and the pro-inflammatory cytokine, IL-1.

These findings warrant further longitudinal studies of the phenotype to understand whether it is a transdiagnostic trait, and to explore other genetic contributions to the phenotype. was labelled to best reflect the majority of loading items.

The different clinical characteristics of corona virus disease cases between children and their families in China -the character of children with COVID-19. Emerging microbes & infections 2020

Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations

The induced responses of innate immunity

The NLRP3 inflammasome in health and disease: the good, the bad and the ugly

Association of genetic polymorphisms of

IL6 -174 G>C, IL10 -819 C>T and TNFα -308 G>A, involved in symptomatic patients with dengue in Brazil

Inflammasome Genes' Polymorphisms in Egyptian Chronic Hepatitis C Patients: Influence on Vulnerability to Infection and Response to Treatment

Polymorphisms in CARD8 and NLRP3 are associated with extrapulmonary TB and poor clinical outcome in active

Association of polymorphic variants in IL1B gene with secretion of IL-1β protein and inflammatory markers in north Indian rheumatoid arthritis patients

Polymorphisms and expression of inflammasome genes are associated with the development and severity of rheumatoid arthritis in Brazilian patients

Biological basis of the behavior of sick animals

Twenty years of research on cytokine-induced sickness behavior

Genetic associations of fatigue and other symptom domains of the acute sickness response to infection

Cytokine polymorphisms have a synergistic effect on severity of the acute sickness response to infection

Polymorphisms in Toll-like receptors-2 and -4 are not associated with disease manifestations in acute Q fever

Endophenotype Concept in Psychiatry: Etymology and Strategic Intentions

Endophenotypes as a measure of suicidality

Contribution of individual psychological and psychosocial factors to symptom severity and time-to-recovery after naturally-occurring acute infective illness: The Dubbo 26

Post-infective and chronic fatigue syndromes precipitated by viral and non-viral pathogens: prospective cohort study

Development of a simple screening tool for common mental disorders in general practice

A step-by-step approach to using SAS for factor analysis and structural equation modeling

The impact of gene-environment interaction on alcohol use disorders

Genetic Variation in the Social Environment Contributes to Health and Disease

Applications of Recombinant DNA Technology

An introduction to genetic analysis

The mutational constraint spectrum quantified from variation in 141

PLINK: a tool set for whole-genome association and population-based linkage analyses

Epidemiologic Principles

s Principles and Practice of Infectious Diseases

Epstein-Barr Virus (Infectious Mononucleosis, Epstein-Barr Virus-Associated Malignant Diseases, and Other Diseases)

s Principles and Practice of Infectious Diseases

Natural history of Ross River virus-induced epidemic polyarthritis

Ross River virus transmission, infection, and disease: a crossdisciplinary review

Coxiella burnetii (Q Fever)

s Principles and Practice of Infectious Diseases

The relationship between fatigue, psychological and immunological variables in acute infectious illness. The Australian and New Zealand journal of psychiatry

Neurocognitive disturbances associated with acute infectious mononucleosis, Ross River fever and Q fever: a preliminary investigation of inflammatory and genetic 28 correlates

Why match? Investigating matched case-control study designs with causal effect estimation

Benefits and limitations of genome-wide association studies

Forward genetics of infectious diseases: immunological impact

Cancer-related fatigue--mechanisms, risk factors, and treatments

Central pathways causing fatigue in neuro-inflammatory and autoimmune illnesses

Fatigue sensation following peripheral viral infection is triggered by neuroinflammation: who will answer these questions?

Interferon-free regimens improve health-related quality of life and fatigue in HIV/HCV-coinfected patients with advanced liver disease: A retrospective study

Correlations of Clinical and Laboratory Characteristics of COVID-19: A Systematic Review and Meta-Analysis

Persistent fatigue following SARS-CoV-2 infection is common and independent of severity of initial infection

The Q705K polymorphism in NLRP3 is a gain-offunction alteration leading to excessive interleukin-1β and IL-18 production

Cytokine Profile in a Cohort of Healthy Blood Donors Carrying Polymorphisms in Genes Encoding the NLRP3 Inflammasome

NLRP3 rs35829419 polymorphism is associated with increased susceptibility to multiple diseases in humans

Single nucleotide polymorphisms in the human interleukin-1B

gene affect transcription according to haplotype context

Associations between functional polymorphisms in the

NFκB signaling pathway and response to anti-TNF treatment in Danish patients with inflammatory bowel disease

Abnormally wide confidence intervals in logistic regression: interpretation of statistical program results

The support of the general practitioners, the public health unit, and the diagnostic pathology services in the Dubbo region, NSW, Australia are gratefully acknowledged, as well as the enduring cooperation of the subjects in this research. 

-Acute sickness manifestations are shared across different infectious diseases -Severity of symptoms as fatigue, pain, or mood disturbance are varied between individuals -Genetic polymorphisms could explain these varied manifestations -Polymorphisms in the inflammasome are associated with severity of fatigue