key: cord-0329353-j4do0y7r
authors: Feliz-Norberto, Maria; Michael, Cassia; de Oliveira, Sofia
title: Neutrophil reverse migration from liver fuels neutrophilic inflammation to tissue injury in Nonalcoholic Steatohepatitis
date: 2021-10-04
journal: bioRxiv
DOI: 10.1101/2021.10.03.462893
sha: ec8fd2d2b448be0742cc4cbfb1a555595da61bfc
doc_id: 329353
cord_uid: j4do0y7r

Inflammation is a hallmark in the progression of nonalcoholic-fatty liver disease (NAFLD) to non-alcoholic steatohepatitis (NASH). Patients with NAFLD are characterized by a chronic low-grade systemic metabolic inflammation (i.e., metainflammation), which contributes to exacerbated however dysfunctional immune response. Neutrophils play an important pathological role in NAFLD progression to NASH; however, how NASH and associated chronic systemic inflammation impact overall the neutrophil response to injury is completely unexplored. Here, we investigated how neutrophil response to tissue injury is altered by the presence of NASH. We used a diet-induced NASH zebrafish model combined with tailfin transection in transgenic zebrafish larvae to study neutrophilic inflammation. Live non-invasive confocal microscopy was used to investigate neutrophil recruitment to tailfin injury through time. Photoconvertion of neutrophils at the liver area followed by time-lapse microscopy was performed to evaluate migration of neutrophils from liver to tailfin injury. Metformin and Pentoxifylline were used to pharmacologically reduce NASH and liver inflammation. We found that larvae with NASH display systemic inflammation and increased myelopoiesis. NASH larvae display a dysfunctional and exacerbated neutrophil response to tailfin injury, characterized by increased neutrophil recruitment, and delayed resolution of inflammation. Interestingly, we showed that neutrophils undergo reverse migration from the NASH liver to the wounded tailfin area. Finally, pharmacological treatment of NASH with Pentoxifylline and Metformin significantly reduced systemic chronic inflammation and the exacerbated recruitment of neutrophils to tissue injury. Taken together, our findings suggest that NASH exacerbates neutrophilic inflammation probably via neutrophil priming at the liver, which can further undergo reverse migration and respond to secondary inflammatory triggers such as tissue injury. Reverse migration of primed neutrophils from the liver might be an important mechanism that fuels the exacerbated neutrophil response observed in NASH conditions and associated metainflammation contributing to poor prognosis and increasing death in patients with metabolic syndrome.

62 Nonalcoholic fatty liver disease (NAFLD) is the hepatic manifestation of metabolic syndrome 63 (i.e., high blood pressure, high blood sugar, excess body fat around the waist, and abnormal 64 cholesterol levels) and is a major health issue and economic burden in western societies affecting 65 around 25-30% of overall population 1, 2 . The rising incidence of NAFLD correlates strongly with 66 the prevalence of metabolic syndrome, type 2 diabetes and obesity 1 . Consumption of calorie rich generating immune imbalances, from cellular to cytokine levels that predisposes patients with 72 NASH and associated metabolic diseases to chronic inflammation and infections 4, 5 . The high 73 incidence of such complications drastically impacts this high-risk group, both socially and 74 economically, and often leads to disability or death -as shown recently with COVID-19 pandemic 6, 75 7 . The lack of efficient therapeutic approaches to decrease such impact in this high-risk population 76 is a clear indicator that we do not fully understand how metabolic syndrome, nutrient excess or 77 overnutrition, and associated metainflammation are altering and regulating the overall 78 inflammatory response towards "secondary" inflammatory triggers such as tissue injury. 79 Neutrophils are first line responders to injury that rely in distinct tiers of arsenals to counter threats 80 including phagocytosis, protease secretion, and neutrophil extracellular traps (NETs) 8 . Such 81 mechanisms are not only protective but can also be destructive to tissues, therefore neutrophil 82 production, trafficking, and clearance need to be tightly regulated 9 . Neutrophils have a double-83 edge sword function being crucial for effective tissue repair but can also contribute for further 84 damaged in case a dysfunctional response is triggered 8 . Neutrophils have a crucial role on NAFLD 85 pathophysiology; with circulating neutrophils from patients with NASH exhibiting an activated 86 and immunosuppressive phenotype 10 . Multiple reports have also found that circulating neutrophils 87 in NASH have enhanced reactive oxygen species (ROS) production upon inflammatory stimulus, 88 and undergo spontaneous NETs formation (i.e., NETosis) 10-12 , based on this evidence we decided 89 to explore how neutrophilic inflammation to tissue injury was impacted in a NASH background. 90 The small vertebrate animal model, the zebrafish, with its unparallel transparency and genetic 91 similarity with humans provides a unique opportunity to explore neutrophilic inflammation in a 92 whole-animal context using non-invasive live imaging 13 . Here, we used a diet-induced NASH 93 zebrafish model 14, 15 by exposing transgenic zebrafish larvae with fluorescently-tagged neutrophils 94 to a high cholesterol diet for one week and performed tailfin transection 16 . Next, we investigated 95 neutrophil recruitment to tailfin injury by live non-invasive confocal microscopy. We 96 demonstrated that zebrafish larvae with NASH have systemic chronic inflammation, and increased 97 myelopoiesis, which resulted in increased number of neutrophils and macrophages. Importantly, 98 we found that NASH larvae have an exacerbated neutrophil response to tailfin injury, characterized 99 by increased neutrophil recruitment and delayed resolution of inflammation. We also found that 100 neutrophils undergo reverse migration from the NASH liver to the tailfin injury. Finally, we 101 demonstrated that pharmacological treatment of NASH with Pentoxifylline and Metformin 102 significantly reduced systemic chronic inflammation and the exacerbated recruitment of 103 neutrophils to tissue injury. Our findings suggest that NASH exacerbates neutrophilic 104 inflammation probably via neutrophil priming at the liver, which can further undergo reverse 105 migration and respond to secondary inflammatory triggers such as tissue injury. Adult zebrafish and embryos up to 5 days post-fertilization (dpf) were maintained as described 111 previously 17 . At 5 dpf, larvae were transferred to feeding containers and kept in E3 embryo Larvae diets were prepared as previously described 14, 15, 18 using Golden Pearl Diet 5-50 nm -120 Active Spheres (Brine Shrimp Direct). At 5 days post fertilization (dpf), zebrafish larvae were 121 separated into treatment groups in E3 without methylene blue as described before 14 . Briefly, 5 dpf 122 larvae were separated into different feeding tanks corresponding to normal diet (ND) and 10% 123 high cholesterol diet (HCD) and fed 0.1 mg of food per larvae per day. In general, 60-80 larvae 124 were placed in a breeding container with 400 mL of E3 without methylene blue and fed 6-8 mg of 125 ND or HFD daily. Feeding boxes were cleaned and E3 was replaced daily. Short-term feeding was 126 performed from 5 to 12 dpf. Before experimental procedure, larvae were fasted for 18 hours to 127 decrease intestine autofluorescence. At 13 dpf, larvae were anesthetized and screened for Confocal Microscopy Imaging-zWEDGI 145 All imaging was performed using a zWEDGI device as previously described 24 . Briefly, an 146 anesthetized larva was loaded into a zWEDGI chamber for time-lapse imaging. The loading 147 chamber was filled with 1% low melting point agarose (Sigma-Aldrich) in E3 to retain the larvae 148 in the proper position. Additional E3 supplemented with 0.16 mg/ml Tricaine was added as needed 149 to avoid dryness and provide required moisture to zebrafish larvae during imaging acquisition. All 150 images were acquired on a spinning disk confocal microscope (CSU-X; Yokogawa) with a 151 confocal scanhead on a Zeiss Observer Z.1 inverted microscope equipped with a Photometrics 152 Evolve EMCCD camera, and an EC Plan Neofluar NA 0.3/10 x air objective, z-stacks, 5 μm optical 

To quantify number of neutrophils and macrophages 13 dpf larvae were fixed and whole-larvae or 188 tailfin images were acquired as described previously. For quantification of the number of 189 neutrophils and macrophages in whole-larvae or at different regions, acquired whole-larvae images 190 were reconstructed on IMARIS Bitplane software (Version 9.5/9.6) rendering mode and total 191 number of neutrophils and macrophages were automatically counted in the whole larvae using 192 IMARIS spots function. Spots were defined as particles with 5 μm and 10 μm of X/Y and Z diameter, respectively. To quantify the number of neutrophils and macrophages at different regions 194 of the zebrafish larvae, a surface for each region/area was created, then a mask for the neutrophil 195 and macrophage signals were generated setting to "zero" signal outside the surface. Cells were 196 automatically counted using the IMARIS spots function and defined as particles with 5 μm and 10 197 μm of X/Y and Z diameter, respectively. To quantify neutrophil recruitment to a tailfin wound At 12 dpf, drugs were washed and replaced with fresh E3 without methylene blue 24 hours prior 254 to tailfin amputation to avoid direct effect of neutrophil recruitment during wounding assay. At 255 13 dpf, tailfin amputation was performed in larvae for all treatments. They were fixed at 4 hpw 256 and imaged as described. Automatic quantification of neutrophils at wound area, systemic chronic 257 inflammation (SCI), and neutrophil CHT depletion scorings were performed as described. lyzC:H2B-mCherry)] fed with normal or HCD. We observed that NASH larvae fed with HCD 280 have a significant increase in the total number of neutrophils and macrophages, 29% and 60% 281 respectively, when compared to control larvae that were fed with the normal diet ( Fig.1 A-C) . Fig. 3) . We also observed a cytokine imbalance with several inflammatory mediators 290 displaying altered gene expression levels in NASH conditions (data not shown). Interestingly, 291 analysis of the caudal hematopoietic tissue (CHT) region, the main hematopoietic niche at this 292 developmental stage, show drastic neutrophil depletion in NASH larvae fed with HCD ( Fig 1A   293 and E; Suppl. Fig.3B ). We further decided to evaluate, how many days we needed to expose larvae 294 to the HCD to induce SCI. Collecting larvae from 1 to 7 days of feeding (corresponding to 6 days-295 post-fertilization to 12 days-post-fertilization) we found that SCI starts to be observed at 4 days 296 of feeding with severe SCI becoming the main phenotype observed by 7 days (Suppl. Fig. 4) , these 297 observations matched when we also start to observe severe steatosis and infiltration of neutrophils 

The consumption of Western-type diets trigger myelopoiesis and transcriptional reprogramming 303 of myeloid precursor cells 27 . Next, we decided to address if the increased number of neutrophils 304 and macrophages found in NASH larvae were associated with increased proliferation at the caudal 305 hematopoietic tissue (CHT) and the kidney, the hematopoietic niches where myelopoiesis occurs 306 at this developmental stage 28 . To do so we incubated zebrafish larvae with EDU and measured the 307 number of cells proliferating in the CHT, kidney, and thymus (Fig. 2) . We found an increase in 308 total proliferation in larvae fed with HCD compared to normal diet (Fig. 2 A and B) . As expected, 309 we also observed an increase in the number of proliferating cells at CHT and kidney, but not at the 310 thymus ( Fig. 2 A and C) . In addition, we also observed an increased number of hematopoietic stem 311 cells in larvae fed with the HCD, which was assessed by automatic quantification CD41+ cells in 312 whole larvae (Suppl. Fig. 5 ). This data suggests that NASH larvae have increased myelopoiesis, for 8 hours, showed that neutrophils migrate at a higher speed at the wound and vicinity in NASH 326 larvae ( Fig. 4 A-B) . Interestingly, at this developmental stage a vast majority of the neutrophils 327 that arrive to the tailfin area use vessels, particularly in NASH larvae. We could observe this in 

Resolution of inflammation is a critical phase in the inflammatory response, where neutrophil 335 clearance from the injured area needs to occur so that the tissue repair machinery can be fully 336 activated. Once resolution of inflammation is dysregulated, progression from acute to chronic 337 inflammation occurs and tissue damage and disease results 31 . The kinetics of the recruitment curve 338 from time-lapse microscopy were different in NASH larvae compared to control suggesting that 339 neutrophil clearance from the wound is delayed and therefore possibly impacting resolution of 340 inflammation (Fig. 3B) . To test whether resolution phase is affected in NASH, we quantified the 341 number of neutrophils at the tailfin wounded area at 4 hpw, the peak of neutrophil recruitment, 342 and later at 24 hpw, when resolution phase is expected to have started in larvae fed with HCD 343 versus normal diet 16, 32-34 . As expected, we observed a reduction in the number of neutrophils at 344 the wound from 4 to 24 hpw in control larvae fed with normal diet (Fig. 4 C-D) . However, in 345 NASH larvae fed with HCD the number of neutrophils at the injury site were almost the same at 346 4 and 24 hpw (Fig. 4 C-D) . Collectively, these findings show that NASH impairs resolution of 347 inflammation in tissue injury.

Neutrophils from NASH larvae have increased ROS production at injury sites. 350 Next, we decided to evaluate if exacerbated neutrophil chemotaxis and impaired resolution of 351 inflammation were associated with increased ROS production in neutrophils at the tissue injury 352 site. To do so we stained larvae expressing a neutrophil cytoplasmic marker, Tg(mpx:mCherry), 353 with CellROX Deep Green to label intracellular H2O2, and performed colocalization analysis on 354 confocal microscopy images of the amputated tailfin (Fig. 5) . We observed that larvae fed with 355 HCD have a significantly higher number of neutrophils at the wound site that generate ROS at 4 356 hpw, as shown by the quantification of the number of colocalized voxels between neutrophil 357 signals (mCherry) and CellROX (Deep Green) (Fig. 5 A and B) . In addition, we also found that 358 neutrophils from larvae fed with HCD generate higher amount of H2O2, as shown by quantification 359 of the percentage of mCherry/Deep Green signal that colocalizes (Fig. 5 C) . Collectively, these 360 results indicate that NASH enhances the ability of neutrophils to produce a higher amount of ROS 361 at inflammatory sites, suggesting that NASH induces neutrophil priming. after photoconversion, a tail fin amputation was performed followed by non-invasive time-lapse 373 confocal microscopy imaging from 1-6 hpw. We observed that larvae exposed to HCD had a higher 374 number and higher percentage of photoconverted neutrophils at the wound from 1-6 hpw (Fig. 6) . 375 This data suggests that neutrophils from NASH liver can undergo reverse migration and massively 376 respond to a "secondary" inflammatory stimulus. Such effect was also observed in our whole- Neutrophil exacerbated response to tissue injury is alleviated by NASH pharmacological 380 intervention. 381 In our previous studies, metformin treatment reduced steatosis, and overall inflammation at the 382 liver that could be observed by reduced neutrophil infiltration 14 . In addition, tumor necrosis factor-383 alpha (TNFα) has been reported as a main inflammatory molecule upregulated in NASH, as we 384 have shown previously shown in our NASH model 14 . Inhibition of TNFα secretion with 385 pentoxifylline 36 , was found to be effective on improving liver function and histological changes 386 in patients with NASH To test whether pharmacological treatment of NASH with metformin and 387 pentoxifylline could revert the SCI as well the exacerbated neutrophil response, we first let larvae 388 to develop NASH and SCI by feeding larvae from 5 days post fertilization (dpf) to 9 dpf with HCD 389 (Suppl. Fig. 2) , followed by a 2-day treatment (10 dpf-12 dpf) with metformin or pentoxifylline 390 to revert this effect. Drug treatments were removed and replaced by E3 at least 16h before any 391 intervention in larvae to avoid direct effect on neutrophil recruitment to tailfin injury. We observed 392 that both these drugs partially rescued diet-induced SCI and decreased the hyper-responsiveness 393 of neutrophils to tailfin injury (Fig. 7) , suggesting that indeed NASH and associated liver 394 inflammation contributes to priming of neutrophils and that NASH-pharmacological intervention 395 can alleviate the adverse effect on exacerbated neutrophilic inflammation. that neutrophil response to tissue injury is drastically exacerbated with neutrophils from the liver 410 undergoing massive reverse migration upon sensing a "secondary" inflammatory stimulus, 411 functioning as a main source of primed neutrophils that fuels neutrophilic inflammatory response 412 to injury sites. We also probe that pharmacological treatment of NASH reverted the observed 413 exacerbated neutrophil response to tissue injury. Our work shows that zebrafish models provide 414 the perfect platform to study the pathophysiological mechanisms involved on how diet, NASH and 415 associated metainflammation impact neutrophils and inflammatory response. Overall, our findings 416 support the idea that under NASH conditions, the liver serves as a source of primed neutrophils 417 that reverse migrate and respond massively to "secondary" inflammatory triggers, identifying a 418 potential therapeutic target to reduce the adverse complications observed in patients with NASH 419 and associated comorbidities due to hyperactive neutrophil response.

The liver is a vital organ with high regenerative capacity that plays more than 500 functions, with 421 a central role in metabolic activities, nutrient storage, detoxification 40 . The liver is also the stage 422 for complex immunological actions; being exposed to dietary and commensal bacterial products 423 from the gut with inflammatory potential that routinely challenge the diverse population of resident 424 immune cells. Interestingly, the liver not just facilitates the removal and degradation of 425 immunogenic molecules from the gut 40 , but it is also target of inflammatory macromolecules from 426 the brain via a drainage mechanism that results in macrophage and neutrophil infiltration 41 . To 427 exert its functions, the liver tolerates those challenges but at same time triggers homeostatic 428 inflammation, a process that is continuously being activated and resolved to support tissue these findings make us speculate that NASH liver is an active source of primed neutrophils that 482 massively reverse migrate towards "secondary" inflammatory stimulus such as tissue injury; 483 therefore, the neutrophil pool at NASH liver might be a potential target to reduce the adverse 484 effects caused by dysfunctional and hyperactive neutrophil response observed in patients with 485 NASH and associated comorbidities that often lead to disability and death. Therefore, another important question that this study raises is at what cellular stage (e.g., 506 progenitor level, immature, or mature), where (e.g., hematopoietic tissues, liver or other tissues 507 and organs) and how neutrophils are being altered in NASH. Future studies will need to be 508 performed to specifically address these questions and the diet-induce NASH zebrafish model is a 509 unique model to visualize and investigate the cellular and molecular mechanisms that drive 510 neutrophil hyperactive response in NASH.

In summary, our data suggest that NASH exacerbates neutrophilic inflammation to tissue injury 512 probably via neutrophil priming at the liver, which can further undergo reverse migration and 513 respond to secondary inflammatory triggers. In the future, reverse migration of neutrophils from 514 the liver might be an important mechanism to target to diminish neutrophil response, improve 515 prognosis, and reduce disability and death in patients with NASH. 

Modeling the epidemic of 600 nonalcoholic fatty liver disease demonstrates an exponential increase in burden of disease

Non-alcoholic fatty liver disease and the interface between 603 primary and secondary care

Developmental origins of NAFLD: a 605 womb with a clue

Immune Imbalances in Non-Alcoholic Fatty Liver Disease: From General 607 Biomarkers and Neutrophils to Interleukin-17 Axis Activation and New Therapeutic Targets. 608 Front Immunol

NAFLD and Infection, a Nuanced Relationship

How Patients With 612 Chronic Liver Diseases Succeed to Deal With COVID-19

COVID-19 and non-alcoholic 614 fatty liver disease: Two intersecting pandemics

Neutrophils as emerging therapeutic targets

Neutrophils as protagonists and targets in 619 chronic inflammation

Circulating Neutrophils of Nonalcoholic Steatohepatitis Patients Show an 621

Activated Phenotype and Suppress T Lymphocytes Activity

Induced M4-Like Macrophages Recruit Neutrophils and Induce NETosis

Neutrophils undergo switch of apoptosis to NETosis during murine fatty liver 627 injury via S1P receptor 2 signaling

Zebrafish as a model for the study of 629 neutrophil biology

Metformin modulates innate immune-mediated inflammation and early 631 progression of NAFLD-associated hepatocellular carcinoma in zebrafish

Dietary cholesterol directly induces acute inflammasome-dependent 634 intestinal inflammation

Cxcl8 (IL-8) mediates neutrophil recruitment and behavior in the zebrafish 636 inflammatory response

Differential regulation of protrusion and polarity by PI3K during neutrophil 638 motility in live zebrafish

Analysis of Liver Microenvironment during 640

Early Progression of Non-Alcoholic Fatty Liver Disease-Associated Hepatocellular Carcinoma in 641 Zebrafish

Distinct inflammatory and wound healing responses to complex caudal fin 643 injuries of larval zebrafish

The role of microtubules in neutrophil polarity and migration in live zebrafish

Macrophages mediate 647 flagellin induced inflammasome activation and host defense in zebrafish

Live imaging reveals distinct modes of 650 neutrophil and macrophage migration within interstitial tissues

Redox and Src family kinase 653 signaling control leukocyte wound attraction and neutrophil reverse migration

zWEDGI: Wounding and Entrapment Device for Imaging Live Zebrafish Larvae

DnaJ-PKAc fusion induces liver 658 inflammation in a zebrafish model of fibrolamellar carcinoma

A Comprehensive Study of High Cholesterol Diet-Induced Larval Zebrafish Model: A 660

In Vivo Screening Method for Non-Alcoholic Fatty Liver Disease Drugs

Western Diet Triggers NLRP3-Dependent Innate Immune Reprogramming

Bloody Zebrafish: Novel Methods in Normal and Malignant 665 Hematopoiesis

Western Diet and the Immune System: An Inflammatory 667 Connection

Targeting immunometabolism as an anti-669 inflammatory strategy

Resolution of Inflammation: 671 What Controls Its Onset?

Resolution of inflammation by retrograde chemotaxis of neutrophils in 673 transgenic zebrafish

Activation of hypoxia-inducible factor-1alpha (Hif-1alpha) delays inflammation 675 resolution by reducing neutrophil apoptosis and reverse migration in a zebrafish inflammation 676 model

Neutrophil reverse migration becomes transparent with 678 zebrafish

Neutrophil migration in infection and wound 680 repair: going forward in reverse

Effects of pentoxifylline on nonalcoholic fatty liver disease: a 682 meta-analysis

Overnutrition Favour Lipotoxicity and Pro-Inflammatory Fatty Liver Disease

Lipotoxicity and the gut-liver axis in NASH pathogenesis

Neutrophils in liver diseases: pathogenesis and therapeutic targets

Liver immunology and its role in inflammation and 691 homeostasis

Drainage of inflammatory macromolecules from the brain to periphery targets the 693 liver for macrophage infiltration

The expanding role of fish models in understanding 695 non-alcoholic fatty liver disease

Systemic inflammation in nonalcoholic fatty liver disease is characterized 697 by elevated levels of CCL2

Adipose tissue macrophages induce hepatic neutrophil recruitment and 699 macrophage accumulation in mice

The CXCL12/CXCR4 Signaling Axis Retains Neutrophils at Inflammatory Sites in 701

CXCL12a/CXCR4b acts to retain neutrophils in caudal hematopoietic 703 tissue and to antagonize recruitment to an injury site in the zebrafish larva

Chemokine Signaling and the Regulation of Bidirectional Leukocyte Migration in 706 Interstitial Tissues

Genetic and pharmacological inhibition of CDK9 drives neutrophil apoptosis 708 to resolve inflammation in zebrafish in vivo

A zebrafish compound screen reveals modulation of neutrophil reverse 710 migration as an anti-inflammatory mechanism

ATP modulates acute inflammation in vivo through dual oxidase 1-derived 712 H2O2 production and NF-kappaB activation

Zebrafish tissue injury causes upregulation of interleukin-1 and caspase-714 dependent amplification of the inflammatory response

IL-1beta and reactive oxygen species differentially regulate neutrophil directional 716 migration and Basal random motility in a zebrafish injury-induced inflammation model

Lyn is a redox sensor that mediates 719 leukocyte wound attraction in vivo

A tissue-scale gradient of hydrogen 721 peroxide mediates rapid wound detection in zebrafish

Modeling inflammation in 723 the zebrafish: how a fish can help us understand lung disease

Dyslipidemic Diet Induces Mobilization of Peripheral Neutrophils and Monocytes That 725

Exacerbate Hemorrhagic Brain Injury and Neuroinflammation

Inflammatory chemokines direct and restrict leukocyte migration within live 728 tissues as glycan-bound gradients

Concepts of GPCR-controlled navigation in the immune 730 system

Neutrophil kinetics in health and disease

Primed polymorphonuclear leukocytes constitute a 733 possible link between inflammation and oxidative stress in hyperlipidemic patients: effect of 734 statins

Evaluation of oxidative 736 stress in patients with hyperlipidemia

Accumulation of Lipids in the Leukocytes of Rats Fed Atherogenic 738 Diets

Oxidized low-density 740 lipoprotein activates migration and degranulation of human granulocytes

Hypercholesterolemia increases the production of leukotriene B4 in neutrophils by 743 enhancing the nuclear localization of 5-lipoxygenase

Inflammasomes, neutrophil extracellular traps, and cholesterol

Single-cell transcriptome profiling reveals neutrophil heterogeneity in homeostasis 747 and infection