Evidence of Trained Immunity in a Fish: Conserved Features in Carp Jules Petit, Carmen W. E. Embregts, Maria Forlenza and Geert F. Wiegertjes This information is current as of October 2, 2021. J Immunol 2019; 203:216-224; Prepublished online 24 May 2019; doi: 10.4049/jimmunol.1900137 http://www.jimmunol.org/content/203/1/216 Downloaded from

References This article cites 49 articles, 9 of which you can access for free at: http://www.jimmunol.org/content/203/1/216.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on October 2, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Author Choice Freely available online through The Journal of Immunology Author Choice option Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2019 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Evidence of Trained Immunity in a Fish: Conserved Features in Carp Macrophages

Jules Petit,* Carmen W. E. Embregts,* Maria Forlenza,* and Geert F. Wiegertjes*,†

Trained immunity is a form of innate immune memory best described in mice and humans. Clear evidence of the evolutionary conservation of trained immunity in teleost fish is lacking. Given the evolutionary position of teleosts as early vertebrates with a fully developed immune system, we hypothesize that teleost myeloid cells show features of trained immunity common to those observed in mammalian macrophages. These would at least include the ability of fish macrophages to mount heightened responses to a secondary stimulus in a nonspecific manner. We established an in vitro model to study trained immunity in fish by adapting a well-described culture system of head kidney–derived macrophages of common carp. A soluble NOD-specific ligand and a soluble b-glucan were used to train carp macrophages, after which cells were rested for 6 d prior to exposure to a secondary stimulus. Unstimulated trained macrophages displayed evidence of metabolic reprogramming as well as heightened phagocytosis and increased expression of the inflammatory cytokines il6 and tnf-a. Stimulated trained macrophages showed heightened production Downloaded from of reactive oxygen and nitrogen species as compared with the corresponding stimulated but untrained cells. We discuss the value of our findings for future studies on trained immunity in teleost fish. The Journal of Immunology, 2019, 203: 216–224.

raditionally, innate immune responses have been viewed as Initial experiments showed that Rag2/2 miceexposedtoa rapid, relatively nonspecific, and lacking immunological sublethal Candida albicans infection were better protected against T memory. New insights have challenged this view, intro- a subsequent lethal dose of these yeasts, whereas - http://www.jimmunol.org/ ducing a novel concept referred to as trained immunity, which is deficient mice did not show this increased protection, suggesting defined as a heightened response to a secondary infection that can be a form of memory present in the myeloid compartment. Further- exerted toward both homologous and heterologous microorganisms more, in vitro re-exposure of human PBMC to C. albicans induced (1). Typical criteria of trained immunity include the following: 1) a significant IL-6 and TNF-a response, even after a resting period induction upon primary infections or immunizations and subsequent of 6 d following primary exposure to C. albicans. Activation of protection against a secondary infection in a T and B lymphocyte– the Dectin-1/RAF proto-oncogene serine/threonine–protein kinase independent manner, 2) a response that is less specific than an (Raf-1) pathway in PBMCs by b-glucans present in the cell adaptive immune response but that still confers increased resistance wall of C. albicans proved crucial for inducing trained immu- by guest on October 2, 2021 upon reinfection of the host and, 3) the involvement of innate cell nity (2, 3). Concurrently, nonspecific protection induced by bacillus types, such as NK cells and macrophages, involved in improved Calmette-Gue´rin(BCG)vaccinationalsoresultedinincreased pathogen recognition and an increased inflammatory response. cytokine production in upon secondary exposure to unrelated pathogens in a T and B cell–independent manner. This process, active for prolonged periods of up to 3 months after vac- *Cell Biology and Immunology Group, Wageningen University & Research, 6708 cination, proved dependent on recognition by nucleotide-binding WD Wageningen, the Netherlands; and †Aquaculture and Fisheries Group, Wageningen oligomerization domain–containing protein 2 (NOD2) and signal- University & Research, 6708 WD Wageningen, the Netherlands ing via receptor-interacting serine/threonine–protein kinase 2 (Rip2) ORCIDs: 0000-0003-1114-4125 (J.P.); 0000-0003-4710-9225 (C.W.E.E.); 0000- (4). Further investigation of receptor-specific induction of trained 0001-9026-7320 (M.F.). immunity revealed that, depending on the dose, membrane-bound Received for publication February 1, 2019. Accepted for publication April 22, 2019. TLR-activation could also induce trained immunity (5). Addition- This work was supported by and research leading to this review was funded by the Netherlands Organisation for Scientific Research and Sa˜o Paulo Research Founda- ally, it was shown that inhibitions of MAP kinases and histone tion, Brazil (FAPESP) as part of the Joint Research Projects BioBased Economy methylation and acetylation could abolish the trained immunity Netherlands Organization for Scientific Research–FAPESP Programme (Project cytokine profile. These studies highlighted that the pathways to 729.004.002). induce trained immunity could be redundant and that the down- Address correspondence and reprint requests to Prof. Geert F. Wiegertjes, Aquaculture and Fisheries Group, Wageningen University & Research, De Elst 1, Post Office Box stream activation is more determinant for the induction of trained 338, 6708 WD Wageningen, the Netherlands. E-mail address: [email protected] immunity. Nevertheless, the Dectin-1/Raf-1 (b-glucans) and NOD2/ Abbreviations used in this article: BCG, bacillus Calmette-Gue´rin; cRPMI, RPMI Rip2 (BCG) could be considered as the best-characterized pathways 1640 culture medium with 25 mM HEPES, supplemented with L-glutamine (2 mM), associated with trained immunity in mammalian monocytes. penicillin G (100 U/ml), streptomycin sulfate (50 mg/ml); cRPMI-1.5, cRPMI sup- plemented with heat-inactivated pooled carp serum (1.5% v/v); H3, histone 3; LAL, Further studies that sought to unravel the underlying mechanisms Limulus amebocyte lysate; LSD, least significant difference; NOD2, nucleotide- of trained immunity revealed crucial roles for epigenetic modifi- binding oligomerization domain–containing protein 2; PBS–EDTA, ice-cold PBS cations and metabolic reprogramming. Trained immunity, induced containing 0.02% (w/v) EDTA; PGN, soluble sonicated peptidoglycan from Escherichia b coli K12 resuspended in endotoxin-free LAL water; Raf-1, RAF proto-oncogene serine/ by exposure of monocytes to -glucans or BCG, was associated threonine–protein kinase; Rip2, receptor-interacting serine/threonine–protein kinase 2; with long-lived epigenetic modifications in the form of increased ROS, reactive oxygen species; SVCV, spring viremia of carp virus. trimethylation of histone 3 (H3) lysine 4 and acetylation of H3 This article is distributed under The American Association of Immunologists, Inc., lysine 27, both activation markers (2, 6, 7), and with decreased Reuse Terms and Conditions for Author Choice articles. trimethylation of H3 lysine 9, a repressor marker (8). These epi- Copyright Ó 2019 by The American Association of Immunologists, Inc. 0022-1767/19/$37.50 genetic modifications, positioned at promotor sites of immune www.jimmunol.org/cgi/doi/10.4049/jimmunol.1900137 The Journal of Immunology 217 genes, helped explain the heightened IL-6 and TNF-a response in phagocytosis, elevated constitutive expression of several immune- PBMC after a resting phase of 6 d following primary exposure and and glycolysis-related genes, as well as a metabolic shift from appear key to the inflammatory response associated with trained oxidative phosphorylation toward glycolysis, the latter measured immunity. Epigenetic modifications were also noted at promotor as increased production of lactate. Stimulated trained macrophages sites of metabolic genes, among which the mechanistic target of typicallyshowedanincreasedinflammatoryresponsemeasured rapamycin kinase (mTOR) and hexokinase 2 (HK2), introducing as increased production of reactive oxygen species (ROS) and NO. metabolic reprogramming as a mechanism underlying trained im- Altogether, our data provide evidence that innate immune cells, munity. Indeed, crucial to the onset of trained immunity appears such as macrophages, of teleost fish possess the ability to express to be a metabolic shift from oxidative phosphorylation toward specific features of trained immunity. glycolysis, orchestrated via the protein kinase B (Akt1)/mTOR/ hypoxia inducible factor 1a (HIF1A) pathway (6, 7). Transcriptome Materials and Methods analysis in mice, comparing constitutive expression of trained Animals and untrained monocytes, revealed significantly higher expres- 3 sion of several genes associated with glycolysis, among others, European common carp (Cyprinus carpio L.) of the R3 R8 strain were used, which results from crossing the Hungarian R8 strain and the Polish aldehyde dehydrogenase (Aldh2), ADP-dependent glucokinase R3 strain (30). Carp were bred and raised in the Aquatic Research Facility (Adpgk), and bisphosphoglycerate mutase (Bpgm) (7). Further, of Carus of Wageningen University at 20–23˚C in recirculating UV-treated activation of the cholesterol synthesis pathway was also noted water and fed pelleted dry food (Carpe-F; Skretting) twice daily. No as an important mechanism underlying trained immunity (9). The yeast-derived b-glucans were in the feed ingredients. All experiments were performed with the approval of the animal experiment committee metabolic shift resulted in the accumulation of several metabolites,

of Wageningen University (Ethical Committee documentation number Downloaded from such as fumarate, a substrate of the TCA cycle, and mevalonate, an 2017.W-0034). intermediate in the cholesterol synthesis pathway. Both metabolites could by themselves induce epigenetic modifications and induce culture and training stimulus cytokine profiles associated with trained immunity (8–10). In Carp were euthanized with 0.3 g/L tricaine methanesulfonate (Crescent addition, blocking metabolic shifts toward glycolysis or gluta- Research Chemicals) in aquarium water buffered with 0.6 g/L sodium minolysis abolished these epigenetic modifications and cytokine bicarbonate and bled via the caudal vein. Head kidney was isolated and

total head kidney leukocytes were separated on a Percoll (GE Healthcare, http://www.jimmunol.org/ profiles, illustrating the strong connections between epigenetic Thermo Fisher Scientific) density gradient. Macrophages were obtained modifications, metabolic reprogramming, and trained immunity. by culturing head kidney leukocytes in complete NMGFL15 medium at Little is known about the conservation of aspects of trained 17.5 3 106 cells per flask (75 cm2; CORN430725U; Corning) for 6 d at immunity in fish, including the aspects described above. Two 27˚C without CO2, as previously described (21). From this point onwards, 6-d cultured, head kidney–derived macrophages will be referred to as recent studies in zebrafish have touched upon such aspects, in- 2/2 “macrophages.” cluding observations on an increased antiviral state in rag Culture flasks were placed on ice for 15 min, macrophages were har- zebrafish associated with increased transcription of innate im- vested by gentle scraping, washed once with ice-cold PBS, and centrifuged mune genes (11), and a study describing the lack of effects of at 450 3 g for 10 min. NMGFL15 is formulated to promote proliferation; pre-exposure to different pathogen-associated molecular patterns therefore, subsequent culturing was performed in a different culture medium. Following centrifugation, cells were resuspended in RPMI 1640 by guest on October 2, 2021 on subsequent viral challenges (12). Given the evolutionary culture medium with 25 mM HEPES, supplemented with L-glutamine position of teleost fish as early vertebrates with a fully developed (2 mM), penicillin G (100 U/ml), streptomycin sulfate (50 mg/ml) (cRPMI), immune system (13), it is likely that the innate immune cells of and cRPMI supplemented with heat-inactivated pooled carp serum fish, such as macrophages, should possess the ability to express (2.5% v/v). For selection of the optimal concentration of training stimulus, pro- specific features of trained immunity. In fact, given that nu- duction of ROS and NO were measured according to the protocol described merous examples exist of a relatively prominent role of innate below. As the initial trained immunity experiments on human monocytes immunity in fish (14–18), trained immunity could be considered found that sublethal dose of C. albicans induced trained immunity, an highly relevant to this group of cold-blooded vertebrates. Re- optimal concentration was defined as a concentration that induces a clear cently, a review summarized potential benefits and constraints of but not maximal innate immune response, measured in this study by NO and ROS production. For NO production, macrophages were seeded at exploiting trained immunity to our benefit in larval aquaculture a density of 5 3 105 cells per well in 96-well culture plates (CORN3596; (19). We already summarized long-lived effects of b-glucans in fish Corning) and stimulated for 24 h with various concentrations of the (20) and hypothesized these might well be explained by features NOD-ligand soluble sonicated peptidoglycan from Escherichia coli K12 of trained immunity. (PGN-ECndss; InvivoGen) resuspended in endotoxin-free Limulus amebocyte lysate (LAL) water, hereafter referred to as “PGN.” For ROS production, In this study, we sought to investigate the hypothesis that trained macrophages were seeded at a density of 5 3 105 cells per well in white immunity is conserved in a teleost fish, the common carp. To this 96-well plates (CLS3912; Corning) and stimulated for 2 h with various end, a well-established in vitro culture of macrophages derived concentrations of PGN. from the head kidney (21), the fish equivalent to mammalian bone Optimization of the training period marrow, was adapted to study the mechanisms of trained im- munity in the common carp. Recent studies have shown trained Macrophages were obtained as described above, and the training stimulus immunity in monocytes and in bone marrow–derived macrophages (1 mg/ml PGN) was added directly to the culture flask (20 ml 1 mg/ml solution in 20 ml medium) for 2 h at 27˚C without CO2. Cells were har- (3, 22) and even comparable trained immunity profiles between vested as described above and washed 33 in ice-cold PBS containing these two cell populations (23). These findings support the use of 0.02% (w/v) EDTA (PBS–EDTA) to remove any unbound stimulus. Cells a heterogeneous myeloid primary cell culture, such as the head were then seeded at a density of 5 3 105 cells per well in 96-well culture kidney–derived macrophage culture in carp. Given that the exis- plates and either incubated in cRPMI supplemented with heat-inactivated tence of a true Dectin-1/Raf-1–like pathway in fish is still elusive pooled carp serum (1.5% v/v) (cRPMI-1.5) or stimulated for additional 24 h with 1 mg/ml PGN. NO production in the cell culture supernatants (24, 25), but that a NOD2/Rip2–like pathway was proven to exist was measured as described below. in zebrafish (26–28), we used a soluble NOD-specific ligand to train carp macrophages. Following a resting period of 6 d, like In vitro setup for studying trained immunity the experimental setup used to study trained immunity in human Macrophages were trained by adding PGN directly to the flask at a final PBMC (29), unstimulated trained macrophages showed increased concentration 1 mg/ml and incubated for 2 h at 27˚C in the absence of CO2. 218 b-GLUCAN–INDUCED TRAINED IMMUNITY IN FISH MACROPHAGES

Alternatively, laminarin (tlrl-lam; InvivoGen) at a concentration 50 mg/ml plates [CORN3596; Corning]) were incubated with fluorescent beads was added (100 ml 10 mg/ml solution in LAL water, in 20 ml medium). (PSF-001UM Red, MagSphere; cell/bead ratio of 1:10) for 120 min at As a control, cells were exposed to LAL water in a volume equal to the 27˚C in the presence of 5% CO2. Subsequently, macrophages were treated training stimulus. Cells were then harvested as described above and with 0.25% (v/v) trypsin–EDTA (11560626; Life Technologies, Thermo washed 33 in ice-cold PBS–EDTA to remove any unbound stimulus. Fisher Scientific) for 10 min, resuspended in ice-cold FACS buffer The obtained cells will be referred to as “trained macrophages” or (0.5% [w/v]) BSA [Roche], 0.01% [w/v] NaN3 in PBS), washed twice with “untrained macrophages.” Subsets of trained and untrained macrophages ice-cold FACS buffer, and centrifuged at 450 3 g for 5 min. Finally, were always analyzed for cell viability with trypan blue exclusion and macrophages were resuspended in FACS buffer, and phagocytosis was for reactivity with an NO assay, as described in section Nitrogen radicals quantified using a FACSCanto A (BD Biosciences); data were analyzed production. The remaining cells were seeded at 2.5 3 106 macrophages using FlowJo v10 (BD Biosciences). Phagocytic activity was calculated as in 20 ml cRPMI supplemented with heat-inactivated pooled carp serum the relative proportion of cells that ingested at least one bead. Phagocytic (2.5% v/v) in 75-cm2 culture flasks (T-75 TC treated, 0030711114; Eppendorf). capacity was calculated as the relative proportion of cells that ingested Cells were incubated at 27˚C for additional 6 d (resting period) before 1, 2, or $3 beads. Fold changes were calculated as phagocytic activity harvesting as described above. Subsequently, trained and untrained mac- or phagocytic capacity of trained-unstimulated macrophages relative to rophages were washed once with ice-cold PBS, centrifuged at 450 3 g for untrained-unstimulated controls. 10 min, and resuspended in cRPMI. Cell growth, viability, and morphol- ogy of trained macrophages after the resting period were comparable to Gene expression analysis that of untrained macrophages. A schematic overview of the in vitro setup 3 6 is depicted in Fig. 1. Gene expression was analyzed by directly lysing 1.5 10 trained or untrained macrophages (n = 3 independent cultures) in RLT buffer, ROS production immediately after harvest on day 12. Total RNA was isolated using the RNeasy Kit (QIAGEN), including on-column DNase treatment according Production of ROS was determined by a real-time luminol-based ECL to the manufacturer’s instructions, and stored at 280˚C. Prior to cDNA assay, as previously described (31). Trained or untrained macrophages were synthesis, 500 ng total RNA was treated with DNase I, Amplification Downloaded from 5 seeded at a density of 5 3 10 cells per well in white 96-well plates Grade (InvivoGen), and cDNA was synthesized using random primers (CLS3912; Corning) and stimulated with one of the following: zymosan (300 ng) and Superscript III First-Strand Synthesis for RT-PCR (InvivoGen). (tlrl-zyd, 250 mg/ml, heterologous stimulus; InvivoGen), PGN (10 mg/ml; cDNA samples were diluted in nuclease-free water prior to real-time homologous stimulus), cRPMI (unstimulated control), or PMA (P8139, quantitative PCR using the primers listed in Table I. Gene expression was 1 mg/ml, stimulated control; Sigma-Aldrich). The PGN concentration was measured by real-time quantitative PCR analysis using ABsolute QPCR, increased from 1 to 10 mg/ml to induce a maximal response. Chemilu- SYBR Green Mix (Thermo Fisher Scientific) in a Rotor-Gene 6000 minescence emission was measured in real time (every 2 min for 120 min) (Corbett Research), and fluorescence data were analyzed using Rotor-Gene with a FilterMax F5 Multi-Mode Microplate Reader at 27˚C and expressed Analysis software version 1.7. The relative gene expression of unstimu- http://www.jimmunol.org/ as area under the curve. Fold changes were calculated as the area under the lated trained versus untrained macrophages was measured immediately curve of trained- or untrained-stimulated macrophages relative to untrained- after the resting period (day 12). The relative expression ratio was calcu- unstimulated control (cRPMI). lated according to the Pfaffl method (34) based on the take-off deviation of the unstimulated trained sample versus each of the unstimulated untrained Nitrogen radicals production controls at the same time point and normalized relative to the s11 protein Production of NO was determined as nitrite accumulation using the Griess of the 40s subunit as reference gene (Table I). reaction, as previously described (32). Trained or untrained macrophages were seeded at a density of 5 3 105 cells per well in 96-well culture plates Lactic acid production (CORN3596; Corning) and stimulated with the one of the following: zy- Production of lactic acid was measured using a lactate colorimetric assay mosan (250 mg/ml; heterologous stimulus), PGN (10 mg/ml; homologous by guest on October 2, 2021 m (Kit II K627; BioVision), including an optional filtration step (Amicon 10K stimulus), cRPMI (unstimulated control), or LPS (L2880, 50 g/ml, spin column; Z677108-96EA; Sigma-Aldrich) according to the manufac- stimulated control; Sigma-Aldrich). After 15 h at 27˚C in the presence turer’s instructions. Briefly, macrophages (5 3 105 per well on a 96-well of 5% CO2, nitrite production was measured at OD540, using a FilterMax culture plate [CORN3596; Corning]) were incubated in 150 ml cRPMI-1.5 F5 Multi-Mode Microplate Reader and quantified using a sodium nitrite for 24 h at 27˚C plus 5% CO , after which supernatants from triplicate (NaNO ) standard curve. Fold changes were calculated as production of 2 2 wells (from n = 5 independent cultures) were pooled and filtered. For each nitrite by trained- or untrained-stimulated macrophages relative to untrained- pooled sample, a volume of 10 ml was diluted 53 in lactate assay buffer unstimulated control (cRPMI). and transferred to 96-well plates. Subsequently, 50 ml reaction mix com- Phagocytosis analysis posed of lactate substrate mix (2 ml), lactate enzyme mix (2 ml), and lactate assay buffer (46 ml), was added to each sample and incubated for Analysis of phagocytic capacity was performed by flow cytometry (FACS) 30 min at room temperature. OD was measured at 450 nm and concen- as previously described (33), with minor modifications. Briefly, trained tration of extracellular lactate was calculated based on a lactic acid cali- or untrained macrophages (5 3 104 cells per well in 96-well cell culture bration curve supplied by the manufacturer. Fold changes were calculated

FIGURE 1. Schematic representation of the in vitro experimental setup to obtain trained macrophages. On day 0, leukocytes are collected from common carp head kidney and cultured for 6 d to allow for differentiation into macrophages (differentiation period). On day 6, macrophages are or are not exposed to the training stimulus for 2 h in the culture flask. Subsequently, cells are harvested and washed three times in ice-cold PBS–EDTA. Cells are transferredto new culture flasks at a fixed density per flask and cultured for another 6 d (resting period). On day 12, trained macrophages are harvested and used for subsequent analysis. Macrophages not exposed at day 6 were treated similarly and served as untrained controls. The Journal of Immunology 219

as lactic acid production of trained-unstimulated macrophages relative to untrained- unstimulated controls. Intracellular fumarate accumulation Accumulation of intracellular fumarate was measured using a fumarate colorimetric assay (K633; BioVision) according to the manufacturer’s instructions, including an optional filtration step (Amicon 10K spin col- umns; as above). Briefly, macrophages (5 3 105 per well, 96-well culture plate [CORN3596; Corning]) were incubated in 150 ml cRPMI-1.5 for 24 h at 27˚C, after which supernatants were removed, and the macrophages were lysed in 50 ml fumarate assay buffer (n = 4 independent cultures). Cell lysates from duplicate wells were pooled and filtered as mentioned above. For each pooled sample, 25 ml were diluted 23 in fumarate assay buffer and transferred to a 96-wells plate. Subsequently, 100 ml reaction mix composed of fumarate developer mix (8 ml), fumarate enzyme mix (2 ml), and fumarate assay buffer (90 ml) was added to each sample and incubated for 30 min at room temperature. OD was measured at 450 nm, and concentration of intracellular fumarate was calculated based on a fu- marate calibration curve supplied by the manufacturer. Fold changes were calculated as fumarate accumulation in trained-unstimulated macrophages relative to untrained-unstimulated controls. Statistical analysis Downloaded from Statistical analyses were performed using SPSS (v23.0), and differences were considered significant if p # 0.05. Data presented as fold change were tested for significance after log transformation. Data used for phagocytosis activity and capacity were first logit-transformed. All data, after testing for Gaussian distribution using the Shapiro–Wilk test, were analyzed as paired data to eliminate interference caused by high variability between indi- vidual cultures. Optimization of training period was tested with a Friedman http://www.jimmunol.org/ test on untransformed NaNO2 values (micromolars), followed by a Dunn post hoc test. Comparison between untrained and trained macrophages of ROS and NO production and comparison of constitutive gene expression were performed with a linear mixed model, followed by a least signifi- cant difference (LSD) post hoc test. Comparison of phagocytic activity and phagocytic capacity was performed with a paired samples t test. Com- parison of lactate production and fumarate accumulation was performed with an independent samples t test. For multiple comparisons of the stimulatory effect of PGN versus laminarin, a multivariate analysis followed by LSD post hoc test was used. by guest on October 2, 2021 Results Soluble peptidoglycan can be used as primary stimulus to induce trained immunity in carp macrophages Given that induction of trained immunity in human and mouse monocytes could be achieved via stimulation with a NOD2 ligand, we used the PGN as a ligand to stimulate carp macrophages. To determine the optimal concentration, on day 6 of culture (Fig. 1), macrophages were stimulated with various concentrations of PGN. Induction of NO was measured in cell supernatants after 24 h (Fig. 2A), whereas ROS production was measured in real time for 2 h and expressed as relative light units (Fig. 2B). A dose-dependent response was observed for both assays. The concentration of 10 mg/ml induced a maximal response, judging from both cumu- lative NO production over 24 h and immediate oxidative burst measured in the 2 h following stimulation. The concentration of 0.1 mg/ml induced a substantial but not maximal NO production FIGURE 2. PGN induces a dose-dependent response in carp macro- and almost no oxidative burst. As the concentration of 1 mg/ml phages. After the differentiation period (day 6), carp macrophages were induced a substantial but not maximal response for both cumula- 5 harvested, seeded in 96-well plates (5 3 10 cells per well) and stimu- tive NO production over 24 h and oxidative burst immediately after A lated with cRPMI or with the indicated concentration of PGN. ( )Dose- stimulation, we chose this concentration as optimal dose for train- dependent analysis of PGN on NO production. NO production was ing. Subsequent experiments were performed with a concen- measured as NaNO in the supernatant 24 h poststimulation. Bars indicate 2 tration of 1 mg/ml PGN as primary stimulus to induce trained mean + SD of triplicate measurements from one representative experiment out of (n = 3) performed independently. (B) Dose-dependent analysis of PGN on ROS production. Real-time ROS production was measured imme- diately following stimulation. Lines indicate acquisition of light emission in PGN (2 h plus 24 h). NO production was measured as NaNO2 in the one representative experiment out of (n = 3) performed independently. (C) supernatant 24 h postseeding. Bars indicate mean + SEM of (n =5) Time course analysis of PGN on NO production. Cells were stimulated for independently performed experiments. Asterisk (*) indicates significant 2 h by directly adding PGN or RPMI to the flask. Cells were then seeded difference (p , 0.05) relative to the cRPMI control as assessed with a and either incubated in cRPMI-1.5 or stimulated for additional 24 h with Friedman test, followed by a Dunn post hoc test (C). 220 b-GLUCAN–INDUCED TRAINED IMMUNITY IN FISH MACROPHAGES immunity because this dose stimulated a robust but not maximal was selected to further investigate trained immunity in carp response. macrophages. After having determined the optimal concentration of the training stimulus, we next investigated the duration of the primary stimu- Trained macrophages show heightened innate lation required to train macrophages. Previous studies reported immune responses that exposure for 24 h to the training stimulus followed by a 6-d Next, we investigated whether the response of trained macrophages resting period, was required for optimal training of human or mouse differed from that of untrained macrophages upon restimulation macrophages (29). In our case we tested whether 24 h or an even with either the training stimulus (PGN, homologous) or with a shorter period would be suitable to train carp macrophages. When different stimulus (zymosan, heterologous) (for experimental setup carp macrophages were stimulated for a total of 26 h (2 h plus see Figs. 1, 3A). Stimulation with zymosan resulted in comparable 24 h) or for only 2 h with PGN, in all cases a significantly higher kinetics of ROS production between trained and untrained NO production was measured relative to the unstimulated un- macrophages, but the production of ROS was significantly higher trained cells (cRPMI) with no difference between the two du- in trained macrophages (Fig. 3B). Constitutive ROS production rations of treatment (Fig. 2C). This suggested that a stimulation was not different between untrained and trained macrophages period as short as 2 h was sufficient to stimulate carp macro- (Fig. 3C, cRPMI), whereas it was significantly higher in trained phages. Altogether, soluble PGN induced a dose-dependent pro- macrophages exposed to zymosan or PGN (Fig. 3C). Similarly, duction of NO and ROS in carp macrophages, and a 2 h primary exposure to the receptor-independent stimulus PMA resulted in a stimulation with 1 mg/ml PGN followed by a 6-d resting period significantly higher production of ROS in trained as compared Downloaded from http://www.jimmunol.org/ by guest on October 2, 2021

FIGURE 3. Trained carp macrophages show increased innate immune functions. (A) Experimental in vitro setup. Head kidney leukocytes were dif- ferentiated into macrophages for 6 d. On day 6, macrophages were stimulated with 1 mg/ml PGN (trained) or vehicle control (untrained), followed by a resting period of 6 d. On day 12, trained or untrained macrophages were harvested and used for subsequent analyses. (B–D) Trained or untrained mac- rophages were seeded in 96-well plates (5 3 105 cells per well) and stimulated with cRPMI, zymosan (250 mg/ml), PGN (10 mg/ml), or PMA (1 mg/ml). The concentration of PGN was increased to 10 mg/ml to ascertain the induction of a maximal response. (B) Real-time ROS production was measured immediately following stimulation. Lines indicate acquisition of light emission measured in one representative experiment out of (n = 11) independently performed experiments. (C and D) Total ROS production relative to unstimulated, untrained macrophages (cRPMI); bars indicate mean + SEM of (n = 11) experiments performed independently. (E) Trained or untrained macrophages were seeded in 96-well plates (5 3 105 cells per well) and stimulated with cRPMI, LPS (50 mg/ml), zymosan (250 mg/ml), or PGN (10 mg/ml). The concentration of PGN was increased to 10 mg/ml to ascertain the induction of a maximal response. NO production was measured as NaNO2 in the supernatant 15 h poststimulation. NO production is expressed relative to unstimulated untrained macrophages (cRPMI); bars indicate mean + SEM of (n = 7) experiments performed independently. (F) Unstimulated trained or unstimulated untrained macrophages were analyzed for their phagocytic activity (left, y-axis) and phagocytic capacity (right, y-axis) after 2 h incubation with fluorescent beads (1 cell/10 beads). Fold change is expressed relative to the corresponding untrained cells; bars indicate mean + SEM of (n = 6) experiments performed independently. For phagocytic capacity, only the proportion of cells phagocytosing $3 beads is shown. Asterisk (*) indicates significant difference (p , 0.05) relative to the corresponding untrained sample as assessed by a linear mixed model one-way ANOVA, followed by an LSD post hoc test (C–E)orpaired samples t test (F). The Journal of Immunology 221 with untrained macrophages (Fig. 3D). With respect to production glycolysis was obtained upon measurement of extracellular lactate of NO, constitutive production was not different between untrained concentrations and intracellular accumulation of the metabolite and trained macrophages (Fig. 3E, cRPMI). Exposure to zymosan fumarate. Trained macrophages showed significantly higher lac- but not to LPS or PGN resulted in a significantly higher NO tate values and a trend (p value = 0.077) toward higher accumu- production in trained as compared with untrained macrophages lation of fumarate as compared with untrained cells (Fig. 4C). (Fig. 3E). Phagocytic activity and phagocytic capacity of macro- To validate epigenetic reprogramming and elucidate kinetics of phages was compared between unstimulated trained and unsti- reprogramming, further analysis would be required. Nevertheless, mulated untrained macrophages. A significantly higher number the increased expression of both immune and metabolic genes in of cells with internalized beads, as well as higher number of beads unstimulated but trained macrophages after a 6-d resting period per cell, was observed for trained as compared with untrained suggests immunologic and metabolic reprogramming. macrophages, indicative of a heightened phagocytic activity as Soluble b-glucan laminarin is a potent training stimulus well as phagocytic capacity of trained macrophages (Fig. 3F). The increase in phagocytic activity and capacity may not necessary In human PBMC, stimulation with BCG or b-glucans could induce result in downstream proinflammatory responses. Nevertheless, trained immunity. We therefore investigated whether, besides PGN, the increased ROS production observed after stimulation with the soluble b-glucan laminarin could also induce trained immunity in zymosan could be considered reflective of both phagocytosis carp macrophages (Fig. 5). Training with laminarin but not PGN led to and increased proinflammatory responses because our analysis a marginal but significant increase in constitutive production of ROS of zymosan-induced ROS production is phagocytosis based (35). as compared with untrained cells (cRPMI). Training with laminarin Altogether, measurement of ROS and NO production, as well as and subsequent exposure to zymosan or PMA resulted in significantly measurement of phagocytosis, showed heightened innate immune heightened production of ROS, comparable to macrophages trained Downloaded from functions of trained carp macrophages. with PGN. Altogether, measurement of ROS production proved par- ticularly informative to identify ligands able to train carp macro- Trained macrophages show immune and phages and revealed that not only NOD-ligands but also b-glucans metabolic reprogramming can serve as suitable training stimuli of fish macrophages. Studies on trained human PBMC revealed a differential expression of IL-6 and TNF-a in response to microbial stimulation, which Discussion http://www.jimmunol.org/ introduced metabolic reprogramming toward glycolysis as an im- In the current study, we adapted a well-described in vitro culture portant mechanism underlying trained immunity. Sensitive, validated system of head kidney–derived macrophages to investigate con- cytokine ELISAs to detect fish cytokines are not commercially servation of trained immunity in teleost fish. A 2-h in vitro ex- available, and in-house development has been relatively unsuc- posure to a soluble NOD-specific ligand or to soluble b-(1,3/1,6) cessful (36), we therefore measured cytokine responses as gene glucan resulted in carp macrophages that displayed typical fea- expression. After having assessed that trained macrophages dis- tures of trained immunity for a period of at least 6 d. The use of play heightened responses to stimulation (Fig. 3), we analyzed soluble ligands allowed for thorough washing and removal of whether trained macrophages would display differences in consti- traces of ligands, a procedure we considered crucially important by guest on October 2, 2021 tutive gene expression of inflammatory or metabolic genes (Table I). to exclude continuous restimulation of macrophages. Typical fea- Trained macrophages showed moderate but significantly increased tures of trained macrophages of carp included heightened phago- expression of inflammatory genes, including il-6 and tnf-a, each cytosis and inflammatory responses following stimulation with present in multiple copies (paralogs) in carp (Fig. 4A). A trend (homologous or heterologous) microbial stimuli. The inflamma- toward increased il-1b expression in trained macrophages was tory profile displayed heightened production of ROS or nitrogen also observed (p value = 0.06), whereas il-10 and inosb expression species, increased constitutive gene expression of selected im- remained unchanged (data not shown). With respect to the expres- mune and metabolic genes, as well as an increased constitutive sion of metabolic genes, trained macrophages showed significantly lactate production, illustrative of a metabolic shift. Measurement increased expression of specific paralogs of adpgk, aldh2.1,and of the production of ROS proved particularly informative to bpgm (Fig. 4B), suggestive of activation of the glycolysis path- identify ligands able to train carp macrophages. way compared with untrained cells. Further evidence of meta- The present study provides in vitro evidence for conservation of bolic reprogramming in unstimulated trained macrophages toward several key features of trained immunity in macrophages of carp, a

Table I. Overview of real-time quantitative PCR primers used for in the current study

Primer Forward (59 –39) Reverse (59 –39) GenBank Accession Numbera 40s 59-CCGTGGGTGACATCGTTACA-39 59-TCAGGACATTGAACCTCACTGTCT-39 AB012087 il6a 59-CAGATAGCGGACGGAGGGGC-39 59-GCGGGTCTCTTCGTGTCTT-39 KC858890 il6b 59-GGCGTATGAAGGAGCGAAGA-39 59-ATCTGACCGATAGAGGAGCG-39 KC858889 Tnf-aa1 59-GAGCTTCACGAGGACTAATAGACAGT-39 59-CTGCGGTAAGGGCAGCAATC-39 AJ311800 Tnf-aa2 59-CGGCACGAGGAGAAACCGAGC-39 59-CATCGTTGTGTCTGTTAGTAAGTTC-39 AJ311801 Tnf-ab1 59-GAAGACGATGAAGATGATACCAT-39 59-AAGTGGTTTTCTCATCCTCAA-39 cypCar_00029601, LHQP01065580 Tnf-ab2 59-CTTGGACGAAGCCGATGAAGAC-39 59-ATCTTGTGACTGGCAAACA-39 cypCar_00023012, LHQP01037150 adpgk-1 59-GGCACCACTGAACTTCT-39 59-GCGTGACCTCTGAAAACAG-39 cypCar_00013411, LHQP01005743 adpgk-2 59-GCAAGCCGTGGATATTACA-39 59-GCGTGAGATGGAAGGA-39 cypCar_00024520, LHQP01021894 aldh2.1-1 59-TCCAGAACTTTCCCACAA-39 59-GCAGATAACCTCACCAGT-39 cypCar_00011521, LHQP01009285 aldh2.1-2 59-GATTCCTGCCCCGAGTC-39 59-TTCTCCACATCCGCCTTC-39 cypCar_00046381, LHQP01040595 bpgm-1 59-CGCCACCCCCCATTGAGGAGA-39 59-GCAGAGATGAGGACTGTTTG-39 cypCar_00001430, LHQP01009643 bpgm-2 59-CTAAACGAGCGGCACTAC-39 59-GGGCAGTTCCTCCTTT-39 cypCar_00018360, LHQP01029542 acypCar numbers identify open reading frames in the draft carp genome (BioProject: PRJNA73579) that were also confirmed by RNA sequencing; LHQP number refers to the accession number of the associated scaffold. 222 b-GLUCAN–INDUCED TRAINED IMMUNITY IN FISH MACROPHAGES

stimulants not specifically associated with antiviral immunity provide at least circumstantial in vivo evidence for trained immu- nity in zebrafish. All of the above-described studies clearly indicate the complexity of in vivo experiments, not unique to experiments in fish, in which it is not easy to exclude the involvement of adaptive immunity and the possibility of continuous stimulation as opposed to a single training event, a clear advantage of our in vitro system based on soluble ligands and rigorous washing steps. Zebrafish could prove especially informative for in vivo studies on trained immunity owing to the availability of rag2/2 strains. Increased survival of rag2/2 zebrafish upon lethal challenge with Edwardsiella ictaluri 8 wk postexposure to an attenuated nonvirulent strain of this bacterium (42) could be considered in vivo evidence of trained immunity, although persistence of attenuated bacteria and absence of nonspecific cross-protection against another bacterium, Yersinia ruckeri, could not fully exclude the involvement of other protective mechanisms. Only recently, rag2/2 zebrafish were shown to exhibit a constitutively height- ened innate immune activity, characterized by an increased an- tiviral state and associated resistance to a viral challenge with Downloaded from SVCV (11). In the latter study, NK cell markers, such as cd8 and nklysin, were also constitutively higher expressed in rag2/2 than in wild-type zebrafish. This observation is of particular interest in view of the BCG-induced (43) and virus-induced (44, 45) memory-like NK cells, recently highlighted as cell types asso- ciated with trained immunity in mammals. Further study could http://www.jimmunol.org/ thus be relevant for NK-like cells in relation to virus-induced trained immunity in the rag2/2 zebrafish model and to explore genes associated with an antiviral state as novel markers for trained immunity. Teleost fish (e.g., carp, zebrafish) are poikilotherms, which al- lows for manipulation of temperature and, thus, allows for studies on temperature-associated effects on trained immunity in vitro and in vivo. Adaptive immunity, more than innate immunity, is con- by guest on October 2, 2021 sidered sensitive to temperature change, reflected by reduced IgM serum concentrations and suppression of T cell responses at lower FIGURE 4. Trained carp macrophages show immune and metabolic temperatures (46, 47). Thus, the use of poikilothermic animals reprogramming. Trained or untrained macrophages were obtained as de- opens the possibility to knockdown adaptive immune responses 2 2 scribed in Fig. 3. (A and B) Gene expression of il6 and tnf-a paralogs (A) in animals for which rag / strains are not available and to or selected metabolic genes and their paralogs (B) in unstimulated trained study, in vivo, aspects of trained immunity, including the du- macrophages relative to unstimulated untrained controls; bars indicate ration of cross-protection against a secondary infection in a mean + SEM of (n = 3) experiments performed independently. (C) Lactate T and B lymphocyte–independent manner. Temperature-mediated and Fumarate accumulation. Unstimulated trained or unstimulated un- knockdown of adaptive immune responses in teleost fish might trained macrophages were seeded in 96-well plates (5 3 105 cells per thus allow for unraveling mechanisms underlying long-lived well), and accumulation of extracellular lactate and intracellular fumarate was measured 24 h later; bars indicate mean + SEM of (n = 5, lactate) and protection likely mediated by innate immune cells that also remain (n = 4, fumarate) experiments performed independently. Asterisk (*) in- active at lower temperatures. dicates significant difference (p , 0.05) relative to the corresponding One of the determining and underlying mechanisms of trained untrained sample as assessed by a linear mixed model one-way ANOVA, immunity is based on long-lived epigenetic modifications that followed by an LSD post hoc test (A and B) or independent samples test (C). persist even after removal of the training stimulus [as reviewed by (48)]. For example, histone modifications, such as trimethylation of H3K4 and mono acetylation of H3K27, have been associated representative teleost fish but provides no in vivo evidence for with b-glucan–induced trained immunity of human PBMC. These trained immunity. However, with BCG, a known stimulant epigenetic changes were strongly correlated with differential gene of trained immunity in mice/human monocytes, has already expression induced by b-glucan training in monocytes (2). In the been shown to provide cross-protection against Mycobacterium spe- same study, comparable epigenetic changes were observed in cies infections in several fish species (37–39), including zebrafish peritoneal macrophages isolated from C. albicans–infected mice, (40), a well-known animal model closely related to carp. The linking the observed in vitro histone modifications to an in vivo nonspecific protection provided by BCG injection also suggests model of trained immunity. The onset of these epigenetic modi- that in fish in vivo evidence for trained immunity already exists. In fications is strongly connected to metabolic reprogramming and addition, zebrafish i.p.-injected with b-glucans and subsequently dependent on specific metabolites, such as fumarate and mevalonate challenged with spring viremia of carp virus (SVCV) either showed (8, 10). In our in vitro model, we investigated epigenetic modifi- a significant but minor increase in survival at seven but not 35 d cations in trained macrophages by measuring (increased) consti- posttreatment (12) or a clear increase in survival at 14 d post- tutive expression of several immune- and glycolysis-related genes treatment (41). Thus, relatively long-lived effects of immune and a metabolic shift from oxidative phosphorylation toward The Journal of Immunology 223

FIGURE 5. Exposure to laminarin leads to trained immunity in carp macrophages. On day 6, macrophages were trained with laminarin (20 mg/ml) or PGN (1 mg/ml) or left untrained (vehicle control), followed by a resting period of 6 d. On day 12, trained or untrained macrophages were harvested and seeded in 96-well plates (5 3 105 cells per well). ROS production was measured immediately following stimulation with either cRPMI, zymosan (250 mg/ml), or PMA (1 mg/ml) and expressed relative to the unstimulated untrained control (cRPMI); bars indicate mean + SEM of (n = 6) experiments performed independently. Asterisk (*) indicates significant differences (p , 0.05) relative to the corresponding untrained sample as assessed by a multivariate analysis, followed by LSD post hoc test.

et al. 2014. mTOR- and HIF-1a-mediated aerobic glycolysis as metabolic basis glycolysis. To study in more detail epigenetic modifications un- Downloaded from for trained immunity. [Published erratum appears in 2014 Science 346: aaa1503.] derlying trained immunity in fish, future studies could build on, Science 345: 1250684. for example, a recent chromatin immunoprecipitation sequencing 8. Arts, R. J. W., A. Carvalho, C. La Rocca, C. Palma, F. Rodrigues, R. Silvestre, study performed on SVCV-infected zebrafish (49). Another in- J. Kleinnijenhuis, E. Lachmandas, L. G. Gonc¸alves, A. Belinha, et al. 2016. Immunometabolic pathways in BCG-induced trained immunity. Cell Rep. 17: teresting aspect of trained immunity that could be studied in vivo, 2562–2571. at least in oviparous fish, is transgenerational epigenetic reprog- 9. Bekkering, S., R. J. W. Arts, B. Novakovic, I. Kourtzelis, C. D. C. C. van der ramming. Because primordial F2 germ cells are not present upon Heijden, Y. Li, C. D. Popa, R. Ter Horst, J. van Tuijl, R. T. Netea-Maier, et al. 2018. Metabolic induction of trained immunity through the mevalonate pathway. http://www.jimmunol.org/ exposure of F0 individuals to potential training stimuli, one less Cell 172: 135–146.e9. generation is needed to prove transgenerational effects in ovipa- 10. Arts, R. J., B. Novakovic, R. Ter Horst, A. Carvalho, S. Bekkering, E. Lachmandas, rous fish (50). In conclusion, the constitutive antiviral state of F. Rodrigues, R. Silvestre, S. C. Cheng, S. Y. Wang, et al. 2016. Glutaminolysis and 2/2 fumarate accumulation integrate immunometabolic and epigenetic programs in rag zebrafish, the cold-blooded nature of teleost fish allowing trained immunity. Cell Metab. 24: 807–819. natural knockdown of the , and the suit- 11. Garcı´a-Valtanen, P., A. Martı´nez-Lo´pez, A. Lo´pez-Mun˜oz, M. Bello-Perez, R. M. Medina-Gali, M. D. Ortega-Villaiza´n, M. Varela, A. Figueras, V. Mulero, ability of oviparous fish for transgenerational experiments all provide B. Novoa, et al. 2017. Zebra fish lacking adaptive immunity acquire an antiviral arguments in favor of studying conserved but also unique aspects alert state characterized by upregulated gene expression of apoptosis, multigene of trained immunity in teleost fish. families, and interferon-related genes. Front. Immunol. 8: 121. 12. A´ lvarez-Rodrı´guez, M., P. Pereiro, F. E. Reyes-Lo´pez, L. Tort, A. Figueras, and B. Novoa. 2018. Analysis of the long-lived responses induced by immunosti- by guest on October 2, 2021 Acknowledgments mulants and their effects on a viral infection in zebrafish (Danio rerio). Front. We acknowledge Raphael Barbetta de Jesus and Fabiana Pilarski for dis- Immunol. 9: 1575. 13. Magor, B. G., and K. E. Magor. 2001. Evolution of effectors and receptors of cussions on the potential impact of trained immunity on aquaculture prac- innate immunity. Dev. Comp. Immunol. 25: 651–682. tice and Cassandra van Doorn for contribution to the initial cell culture 14. Rebl, A., and T. Goldammer. 2018. Under control: the innate immunity of fish optimization. from the inhibitors’ perspective. Fish Shellfish Immunol. 77: 328–349. 15. Tokunaga, Y., M. Shirouzu, R. Sugahara, Y. Yoshiura, I. Kiryu, M. Ototake, T. Nagasawa, T. Somamoto, and M. Nakao. 2017. Comprehensive validation of Disclosures T- and B-cell deficiency in rag1-null zebrafish: implication for the robust innate The authors have no financial conflicts of interest. defense mechanisms of teleosts. Sci. Rep. 7: 7536. 16. Aoki, T., T. Takano, M. D. Santos, H. Kondo, and I. Hirono. 2008. Molecular innate immunity in teleost fish: review and future perspectives. In Fisheries References for Global Welfare and Environment, Memorial Book of the 5th World Fisheries Congress. K. Tsukamoto, and T. Kawamura, eds. Terrapub, Tokyo, 1. Netea, M. G., J. Quintin, and J. W. van der Meer. 2011. Trained immunity: a Japan, p. 263–276. memory for innate host defense. Cell Host Microbe 9: 355–361. 17. Magnado´ttir, B. 2006. Innate immunity of fish (overview). Fish Shellfish 2. Quintin, J., S. Saeed, J. H. A. Martens, E. J. Giamarellos-Bourboulis, D. C. Ifrim, Immunol. 20: 137–151. C. Logie, L. Jacobs, T. Jansen, B. J. Kullberg, C. Wijmenga, et al. 2012. Candida 18. Wcisel, D. J., and J. A. Yoder. 2016. The confounding complexity of innate immune albicans infection affords protection against reinfection via functional reprog- receptors within and between teleost species. Fish Shellfish Immunol. 53: 24–34. ramming of monocytes. Cell Host Microbe 12: 223–232. 19. Zhang, Z., H. Chi, and R. A. Dalmo. 2019. Trained innate immunity of fish is a 3.Ifrim,D.C.,L.A.B.Joosten,B.J.Kullberg,L.Jacobs,T.Jansen, viable approach in larval aquaculture. Front. Immunol. 10: 42. D.L.Williams,N.A.R.Gow,J.W.M.vanderMeer,M.G.Netea,and 20. Petit, J., and G. F. Wiegertjes. 2016. Long-lived effects of administering b-glucans: J. Quintin. 2013. Candida albicans primes TLR cytokine responses through indications for trained immunity in fish. Dev. Comp. Immunol. 64: 93–102. a Dectin-1/Raf-1-mediated pathway. J. Immunol. 190: 4129–4135. 21. Joerink, M., C. M. Ribeiro, R. J. Stet, T. Hermsen, H. F. Savelkoul, and 4. Kleinnijenhuis, J., J. Quintin, F. Preijers, L. A. Joosten, D. C. Ifrim, S. Saeed, G. F. Wiegertjes. 2006. Head kidney-derived macrophages of common carp C. Jacobs, J. van Loenhout, D. de Jong, H. G. Stunnenberg, et al. 2012. Bacille (Cyprinus carpio L.) show plasticity and functional polarization upon differential Calmette-Guerin induces NOD2-dependent nonspecific protection from rein- stimulation. J. Immunol. 177: 61–69. fection via epigenetic reprogramming of monocytes. Proc. Natl. Acad. Sci. USA 22. Walachowski, S., G. Tabouret, M. Fabre, and G. Foucras. 2017. Molecular 109: 17537–17542. analysis of a short-term model of b-glucans-trained immunity highlights the 5. Ifrim, D. C., J. Quintin, L. A. B. Joosten, C. Jacobs, T. Jansen, L. Jacobs, N. A. accessory contribution of GM-CSF in priming mouse macrophages response. R. Gow, D. L. Williams, J. W. M. van der Meer, and M. G. Netea. 2014. Trained Front. Immunol. 8: 1089. immunity or tolerance: opposing functional programs induced in human 23. Saz-Leal, P., C. Del Fresno, P. Brandi, S. Martı´nez-Cano, O. M. Dungan, monocytes after engagement of various pattern recognition receptors. Clin. J. D. Chisholm, W. G. Kerr, and D. Sancho. 2018. Targeting SHIP-1 in my- Vaccine Immunol. 21: 534–545. eloid cells enhances trained immunity and boosts response to infection. Cell Rep. 6. Saeed, S., J. Quintin, H. H. Kerstens, N. A. Rao, A. Aghajanirefah, F. Matarese, 25: 1118–1126. S. C. Cheng, J. Ratter, K. Berentsen, M. A. van der Ent, et al. 2014. Epigenetic 24. Zelensky, A. N., and J. E. Gready. 2004. C-type lectin-like domains in Fugu programming of monocyte-to-macrophage differentiation and trained innate rubripes. BMC Genomics 5: 51. immunity. Science 345: 1251086. 25. Petit, J., E. C. Bailey, R. T. Wheeler, C. A. F. de Oliveira, M. Forlenza, and 7. Cheng, S. C., J. Quintin, R. A. Cramer, K. M. Shepardson, S. Saeed, V. Kumar, G. F. Wiegertjes. 2019. Studies into b-glucan recognition in fish suggests a key E. J. Giamarellos-Bourboulis, J. H. A. Martens, N. A. Rao, A. Aghajanirefah, role for the C-type lectin pathway. Front. Immunol. 10: 280. 224 b-GLUCAN–INDUCED TRAINED IMMUNITY IN FISH MACROPHAGES

26.Maharana,J.,B.R.Sahoo,A.Bej,I.Jena,A.Parida,J.R.Sahoo, 38. Kato, G., K. Kato, K. Saito, Y. Pe, H. Kondo, T. Aoki, and I. Hirono. 2011. Vaccine B.Dehury,M.C.Patra,S.R.Martha,S.Balabantray,etal.2015.Struc- efficacy of Mycobacterium bovis BCG against Mycobacterium sp. infection in tural models of zebrafish (Danio rerio) NOD1 and NOD2 NACHT domains amberjack Seriola dumerili. Fish Shellfish Immunol. 30: 467–472. suggest differential ATP binding orientations: insights from computational 39. Kato, G., H. Kondo, T. Aoki, and I. Hirono. 2012. Mycobacterium bovis BCG modeling, docking and molecular dynamics simulations. PLoS One 10: vaccine induces non-specific immune responses in Japanese flounder against e0121415. Nocardia seriolae. Fish Shellfish Immunol. 33: 243–250. 27. Oehlers, S. H., M. V. Flores, C. J. Hall, S. Swift, K. E. Crosier, and P. S. Crosier. 40. Oksanen, K. E., N. J. Halfpenny, E. Sherwood, S. K. Harjula, M. M. Hammare´n, 2011. The inflammatory bowel disease (IBD) susceptibility genes NOD1 and M. J. Ahava, E. T. Pajula, M. J. Lahtinen, M. Parikka, and M. Ra¨met. 2013. An NOD2 have conserved anti-bacterial roles in zebrafish. Dis. Model. Mech. 4: adult zebrafish model for preclinical vaccine development. Vaccine 832–841. 31: 5202–5209. 28. Zou, P. F., M. X. Chang, Y. Li, N. N. Xue, J. H. Li, S. N. Chen, and P. Nie. 2016. 41. M Medina-Gali, R., M. D. M. Ortega-Villaizan, L. Mercado, B. Novoa, J. Coll, NOD2 in zebrafish functions in antibacterial and also antiviral responses via NF- and L. Perez. 2018. Beta-glucan enhances the response to SVCV infection in kB, and also MDA5, RIG-I and MAVS. Fish Shellfish Immunol. 55: 173–185. zebrafish. Dev. Comp. Immunol. 84: 307–314. 29. Bekkering, S., B. A. Blok, L. A. Joosten, N. P. Riksen, R. van Crevel, and 42. Hohn, C., and L. Petrie-Hanson. 2012. Rag1-/- mutant zebrafish demonstrate M. G. Netea. 2016. In vitro experimental model of trained innate immunity in specific protection following bacterial re-exposure. PLoS One 7: e44451. human primary monocytes. [Published erratum appears in 2017 Clin. Vaccine 43. Kleinnijenhuis, J., J. Quintin, F. Preijers, L. A. Joosten, C. Jacobs, R. J. Xavier, Immunol. 24: e00096–e00117.] Clin. Vaccine Immunol. 23: 926–933. J. W. van der Meer, R. van Crevel, and M. G. Netea. 2014. BCG-induced trained immunity 30. Irnazarow, I. 1995. Genetic variability of Polish and Hungarian carp lines. in NK cells: role for non-specific protection to infection. Clin. Immunol. 155: 213–219. Aquaculture 129: 215. 44.Sun,J.C.,J.N.Beilke,andL.L.Lanier. 2009. Adaptive immune features of 31. Piazzon, M. C., H. S. Savelkoul, D. Pietretti, G. F. Wiegertjes, and M. Forlenza. natural killer cells. [Published erratum appears in 2009 Nature 457: 1168.] 2015. Carp Il10 has anti-inflammatory activities on phagocytes, promotes pro- Nature 457: 557–561. liferation of memory T cells, and regulates B cell differentiation and 45. Schlums, H., F. Cichocki, B. Tesi, J. Theorell, V. Beziat, T. D. Holmes, H. Han, secretion. J. Immunol. 194: 187–199. S. C. Chiang, B. Foley, K. Mattsson, et al. 2015. Cytomegalovirus infection 32. Saeij, J. P., W. B. Van Muiswinkel, A. Groeneveld, and G. F. Wiegertjes. 2002. drives adaptive epigenetic diversification of NK cells with altered signaling and Immune modulation by fish kinetoplastid parasites: a role for nitric oxide. effector function. Immunity 42: 443–456. Parasitology 124: 77–86. 46. Magnadottir, B. 2010. Immunological control of fish diseases. Mar. Biotechnol. 33. Li, J., D. R. Barreda, Y. A. Zhang, H. Boshra, A. E. Gelman, S. Lapatra, L. Tort, (NY) 12: 361–379. Downloaded from and J. O. Sunyer. 2006. B lymphocytes from early vertebrates have potent 47. Bowden, T. J. 2008. Modulation of the immune system of fish by their envi- phagocytic and microbicidal abilities. Nat. Immunol. 7: 1116–1124. ronment. Fish Shellfish Immunol. 25: 373–383. 34. Pfaffl, M. W. 2001. A new mathematical model for relative quantification in real- 48. Dominguez-Andres, J., and M. G. Netea. 2019. Long-term reprogramming of the time RT-PCR. Nucleic Acids Res. 29: e45. . J. Leukoc. Biol. 105: 329–338. 35. Scharsack, J. P., K. Koch, and K. Hammerschmidt. 2007. Who is in control of the 49. Medina-Gali, R., M. Bello´-Pe´rez, A. Martı´nez-Lo´pez, A. Falco´,M.M.Ortega- stickleback immune system: interactions between Schistocephalus solidus and Villaizan, J. A. Encinar, B. Novoa, J. Coll, and L. Perez. 2018. Chromatin its specific vertebrate host. Proc. Biol. Sci. 274: 3151–3158. immunoprecipitation and high throughput sequencing of SVCV-infected

36. Dixon, B., D. R. Barreda, and J. O. Sunyer. 2018. Perspective on the develop- zebrafish reveals novel epigenetic histone methylation patterns involved in http://www.jimmunol.org/ ment and validation of ab reagents to fish immune proteins for the correct as- antiviral immune response. Fish Shellfish Immunol. 82: 514–521. sessment of immune function. Front. Immunol. 9: 2957. 50. Best, C., H. Ikert, D. J. Kostyniuk, P. M. Craig, L. Navarro-Martin, L. Marandel, 37.Kato,G.,H.Kondo,T.Aoki,andI.Hirono.2010.BCGvaccineconfers and J. A. Mennigen. 2018. Epigenetics in teleost fish: from molecular mecha- adaptive immunity against Mycobacterium sp. infection in fish. Dev. Comp. nisms to physiological phenotypes. Comp. Biochem. Physiol. B Biochem. Mol. Immunol. 34: 133–140. Biol. 224: 210–244. by guest on October 2, 2021