Priming of innate antimycobacterial immunity by heat-killed Listeria monocytogenes induces sterilizing response in the adult zebrafish tuberculosis model

RESEARCH ARTICLE P iming o inna e an imycobac e ial immuni y by hea -killed Lis e ia monocy ogenes induces s e ilizing esponse in he adul zeb a ish ube culosis model Hanna Luukinen 1 , Milka Ma ju Hamma en 1, * ,‡ , Leena-Maija Vanha-aho 1, *, Aleksand a S o jo a 1 , Lau a Kan anen 1 , Sampsa Ja  inen 1 , B uno Vincen Luukinen 1 , E ic Du ou 1,2 ,MikaRa me 1,2,3,4 , Vesa Pekka Hy o nen 1,2,5 and Ma aleena Pa ikka 1,6 ABSTRACT Mycobac e ium ube culosis emains one o he mos p oblema ic in ec ious agen s, owing o i s highly de eloped mechanisms o e ade hos immune esponses combined wi h he inc easing eme gence o an ibio ic esis ance. Hos -di ec ed he apies aiming o op imize immune esponses o imp o e bac e ial e adica ion o o limi excessi e in lamma ion a e a new s a egy o he ea men o ube culosis. In his s udy, we ha e es ablished a zeb a ish- Mycobac e ium ma inum na u al hos -pa hogen model sys em o s udy induced p o ec i e immune esponses in mycobac e ial in ec ion. We show ha p iming adul zeb a ish wi h hea -killed Lis e ia monocy ogenes (HKLm) a 1 day p io o M. ma inum in ec ion leads o signi ican ly dec eased mycobac e ial loads in he in ec ed zeb a ish. Using ag1 −/− ish, we show ha he p o ec i e immuni y con e ed by HKLm p iming can be induced h ough inna e immuni y alone. A 24 h pos -in ec ion, HKLm p iming leads o a signi ican inc ease in he exp ession le els o mac ophage- exp essed gene 1 (mpeg1), umo nec osis ac o α( n a) and ni ic oxide syn hase 2b (nos2b), whe eas supe oxide dismu ase 2 (sod2) exp ession is down egula ed, implying ha HKLm p iming inc eases he numbe o mac ophages and boos s in acellula killing mechanisms. The p o ec i e e ec s o HKLm a e abolished when he injec ed ma e ial is p e ea ed wi h nucleases o p o einase K. Impo an ly, HKLm p iming signi ican ly inc eases he equency o clea ance o M. ma inum in ec ion by e oking s e ilizing immuni y (25 s 3.7%, P=0.0021). In his s udy, immune p iming is success ully used o induce s e ilizing immuni y agains mycobac e ial in ec ion. This model p o ides a p omising new pla o m o elucida ing he mechanisms unde lying s e ilizing immuni y and o de elop hos - di ec ed ea men o p e en ion s a egies agains ube culosis. This a icle has an associa ed Fi s Pe son in e iew wi h he i s au ho o he pape . KEY WORDS: Lis e ia monocy ogenes,Mycobac e ium ma inum, Mycobac e ial in ec ion, S e ilizing immuni y, Tube culosis, Zeb a ish INTRODUCTION Tube culosis (TB) is an ai bo ne espi a o y disease caused by he in acellula bac e ium Mycobac e ium ube culosis (M b). As ew as one o i e bac e ia can lead o an in ec ion (Raja am e al., 2014; Cambie e al., 2014).The ou come o TB is highly a iable, anging om apid clea ance by inna e immune mechanisms o de elopmen o ac i e disease o he o ma ion o a la en in ec ion ha can be ac i ely con ained inside g anulomas bu no e adica ed. Acco ding o Cen e s o Disease Con ol and P e en ion (CDC) es ima es, e en one hi d o he wo ld popula ion is in ec ed wi h M b. Howe e , only 5-10% o his popula ion de elops ac i e, p ima y TB. Commonly, in ec ion wi h M b leads o a la en , asymp oma ic disease wi h he inhe en abili y o eac i a e and dissemina e in o an ac i e disease e en decades a e ini ial exposu e, o example in he case o immunosupp ession. In 2015, 1.4 million people died o TB and a o al o 10.4 million new cases we e epo ed along wi h an inc easing numbe o mul id ug- esis an s ains (Wo ld Heal h O ganiza ion, 2016; h p://www. who.in / b/publica ions/2016/en/). Despi e a ailable mul id ug he apies and he Bacille Calme e–Gué in (BCG) accine, TB emains one o he leading in ec ious kille s wo ldwide. Acco ding o a ecen s udy, he s anda d 6-mon h an ibio ic ea men agains TB is ine ec i e in he e adica ion o M b e en in pa ien s wi h a success ul ollow h ough o he an ibio ic ea men (Malhe be e al., 2016). As he cu en p e en i e and ea men s a egies ha e p o en insu icien , new app oaches o con ol he global TB epidemic a e u gen ly needed. Hos -di ec ed he apies o e a p omising app oach o imp o e he ou come o an i-TB ea men s. Hos -di ec ed he apies a e a o m o adjunc i e he apy ha aim o modula e he hos immune esponses o e adica e o limi mycobac e ial in ec ion (Tobin, 2015). Mycobac e ia a e especially success ul in e ading immune esponses. Mac ophages a e known o limi mycobac e ial g ow h in ea ly in ec ion o some ex en (Clay e al., 2007). Howe e , in many cases, he ea ly e en s o mycobac e ial in ec ions a e cha ac e ized by bac e ial dominance. Pa hogenic mycobac e ia a e able o a oid ecogni ion by pa e n- ecogni ion ecep o s and can lu e mycobac e ium-pe missi e mac ophages o he si es o in ec ion (Cambie e al., 2014). Upon phagocy osis, hey block he usion o phagosomes wi h lysosomes (Russell, 2011), ansloca e Recei ed 7 Augus 2017; Accep ed 21 No embe 2017 1 Facul y o Medicine and Li e Sciences, FI-33014 Uni e si y o Tampe e, Tampe e, Finland. 2 BioMediTech Ins i u e, FI-33014 Uni e si y o Tampe e, Tampe e, Finland. 3 PEDEGO Resea ch Uni , and Medical Resea ch Cen e Oulu, FI-90014 Uni e si y o Oulu, Oulu, Finland. 4 Depa men o Child en and Adolescen s, Oulu Uni e si y Hospi al, FI-90220 Oulu, Finland. 5 Fimlab Labo a o ies, Pi kanmaa Hospi al Dis ic , FI-33520 Tampe e, Finland. 6 O al and Maxillo acial Uni , Tampe e Uni e si y Hospi al, FI-33521 Tampe e, Finland. *These au ho s con ibu ed equally o his wo k ‡ Au ho o co espondence (milka.[email p o ec ed]) M.M.H., 0000-0001-9076-8782; B.V.L., 0000-0003-3578-782X; M.P., 0000- 0001-5555-3815 This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License (h p://c ea i ecommons.o g/licenses/by/3.0), which pe mi s un es ic ed use, dis ibu ion and ep oduc ion in any medium p o ided ha he o iginal wo k is p ope ly a ibu ed. 1 © 2018. Published by The Company o Biologis s L d | Disease Models & Mechanisms (2018) 11, dmm031658. doi:10.1242/dmm.031658 Disease Models & Mechanisms o he cy oplasm (Simeone e al., 2012; Houben e al., 2012) and neu alize ni ic oxide (NO) species (Flynn and Chan, 2003), allowing hem o su i e wi hin mac ophages. As onishingly, mycobac e ia a e e en capable o exploi ing mac ophages o issue dissemina ion (Clay e al., 2007). In addi ion o a oiding inna e killing mechanisms, mycobac e ia also inhibi anspo a ion o mycobac e ial an igens o lymph nodes (Wol e al., 2008; Reiley e al., 2008; Gallegos e al., 2008), he eby hampe ing he ini ia ion o adap i e esponses (Chacke ian e al., 2002). Agg ega es o inna e and adap i e immune cells, called g anulomas, a e o med o con ain he bac e ia and o localize he in ec ion o a limi ed a ea wi hou e adica ing he bac e ia. Depending on he immune s a us o he hos , ei he an ac i e in ec ion o a la en in ec ion wi h a li e- long isk o eac i a ion ensues (Ba y e al., 2009). Despi e M b being good a e ading hos immune esponses and ha ing he abili y o cause agg essi e ac i e o pe sis en la en in ec ions, some people a e known o be na u ally p o ec ed agains TB. The e a e signi ican di e ences in he abili y o indi iduals o esis mycobac e ial in ec ion, e lec ing he he e ogenic na u e o he human popula ion. Acco ding o epidemiological da a, a 7-43% p opo ion o hea ily exposed indi iduals a e able o clea he in ec ion be o e he onse o adap i e immuni y, esul ing in nega i e ube culin skin es s and in e e on-gamma (I nγ) elease assays ( e iewed in Ve all e al., 2014). Wi h his in mind, i should be possible o shi he balance o hos -pa hogen in e ac ions in a o o he hos by di ec ing he immune esponse o he igh immune ac i a ion a he ea ly s ages o in ec ion, when he bac e ial loads a e a he small. Op imal immune ac i a ion could p e en mycobac e ial e asion s a egies, enhance killing o mycobac e ia and ideally lead o s e iliza ion o in ec ion. Howe e , o ou knowledge, s e ilizing an imycobac e ial immuni y has no been success ully induced in i o. In his s udy, we used he zeb a ish model o s udy p o ec i e immune esponses agains mycobac e ia a he ea ly s ages o in ec ion. The zeb a ish has ecen ly become a well-accep ed gene ically ac able e eb a e model o human TB pa hogenesis. Zeb a ish a e na u ally suscep ible o Mycobac e ium ma inum,a close gene ic ela i e o M b. M. ma inum causes a disease ha sha es he main pa hological and his ological ea u es o human TB, including he o ma ion o mac ophage agg ega es and g anulomas ( e iewed in Meije , 2016; an Leeuwen e al., 2014). As he basic mechanisms o inna e and adap i e immuni y a e conse ed om zeb a ish o humans, he inna e immune esponses o M. ma inum can be s udied in zeb a ish la ae (Ramak ishnan, 2013), while he adul zeb a ish has been p o en an applicable model o s udy also he adap i e esponses in TB (Hamma én e al., 2014; Pa ikka e al., 2012; Oksanen e al., 2013). E en he ansi ion o an acu e p ima y in ec ion o la ency and i s eac i a ion can be modeled in he adul zeb a ish (Pa ikka e al., 2012), which has been di icul in o he models. Ou p e ious s udies ha e shown ha an in ec ion wi h a low dose o M. ma inum causes a la en mycobac e ial disease wi h s eady bac e ial coun s in he majo i y o he ish popula ion, whe eas, in a small p opo ion o he ish (<1.5%), p ima y ac i e disease leads o mo ali y (Pa ikka e al., 2012). A ound 10% o he ish a e able o clea he mycobac e ial in ec ion (Hamma én e al., 2014). The spec um o di e en disease ou comes in he M. ma inum-zeb a ish model hus esembles ha o human TB. He e, we ha e used he zeb a ish model o es whe he he numbe o indi iduals s e ilizing he in ec ion can be inc eased h ough injec ion wi h di e en p iming agen s o ci cum en mycobac e ial i ulence s a egies, which gene ally lead o pe sis en , la en in ec ions (Pa ikka e al., 2012). Ou s udy shows ha he p iming o zeb a ish wi h hea -killed Lis e ia monocy ogenes esul s in he s e iliza ion o M. ma inum in ec ion in one ou h o indi iduals. The p o ec i e e ec is caused by a p o ein and/o nucleic acid componen o L. monocy ogenes and is accompanied by he induc ion o umo nec osis ac o α( n α)andni ic oxide syn hase 2b (nos2b), and down egula ion o supe oxide dismu ase 2 (sod2). He eby, we show ha he adul zeb a ish is a easible model o deciphe ing he mechanisms o s e ilizing immuni y, knowledge o which is c ucial o he de elopmen o new p e en i e s a egies and adjunc i e he apies agains TB. RESULTS Immune ac i a ion by hea -killed L. monocy ogenes leads o lowe mycobac e ial bu dens in adul zeb a ish In his s udy, we se ou o in es iga e p o ec i e immune esponses a he ea ly s ages o an M. ma inum in ec ion. Ou hypo hesis is ha he igh immune ac i a ion a he ea ly s ages o an in ec ion p e en s mycobac e ial e asion s a egies and leads o p o ec i e immune esponses, inc eased killing o mycobac e ia o e en clea ing o mycobac e ia. To s udy ou hypo hesis, we wan ed o di ec he zeb a ish immune esponse owa ds an op imal an i-TB esponse be o e he onse o an M. ma inum in ec ion. To ind a ac o ha p omo es a p o ec i e immune esponse in he ea ly s ages o an in ec ion in adul zeb a ish, ish we e in ape i oneally (i.p.) injec ed wi h di e en p iming agen s 1 day be o e a low-dose i.p. M. ma inum in ec ion (Fig. 1A,B). Mycobac e ial loads o he ish we e de e mined wi h quan i a i e eal- ime PCR (qPCR) om in e nal o gans a 7 weeks pos -in ec ion (wpi). We es ed eigh di e en agen s, including Toll-like ecep o (TLR) ligands, accine adju an s and hea -killed bac e ia, namely: hea -killed M. ma inum (HKMm), L. monocy ogenes (HKLm), S ep ococcus iniae (HKSi) and Esche ichia coli (HKEc), lipopolysaccha ide (LPS), pacli axel, mu amyl dipep ide (MDP) and zymosan. In his p elimina y expe imen , he es ed p iming agen s had a iable e ec s on mycobac e ial loads. P iming o he adul ish wi h HKLm led o a clea dec ease in median mycobac e ial loads a 7 wpi compa ed o phospha e-bu e ed saline (PBS) con ols. Howe e , wi h mos o he o he es ed p iming agen s, including HKMm, he educ ion in mycobac e ial loads was no as d ama ic as wi h HKLm (Fig. 1B). Impo an ly, HKLm p iming was he mos e icien way o induce bac e ial clea ance. The equency o clea ance inc eased by 4.8- old compa ed o PBS-p imed con ols. Fo pacli axel, p iming led o an inc ease in mycobac e ial loads. Wi h he igh dosing, pacli axel can induce p oin lamma o y esponses (B acci e al., 2014) bu can also be oxic o immune cells (Tang e al., 2017), which migh explain he inc ease in mycobac e ial load by pacli axel. Ou esul s indica e ha ea ly immune p iming a ec s mycobac e ial loads a 7 wpi and ha HKLm is he mos p omising p iming agen wi h ou expe imen al se up. To ensu e ha he educ ion in mycobac e ial loads in adul ish was no due o di ec bac e icidal e ec s o HKLm p iming, we incuba ed M. ma inum in 7H9 medium oge he wi h di e en concen a ions o HKLm (Fig. S1). Based on in i o cul u ing, HKLm does no kill M. ma inum, indica ing ha he lowe ed mycobac e ial bu dens in zeb a ish induced by HKLm p iming is no due o di ec killing o he mycobac e ia by HKLm, bu is media ed h ough e ec s on he hos immune sys em. To cha ac e ize he du a ion o he e ec a e HKLm p iming, we injec ed ish a ei he 1 o 7 days p io o in ec ion wi h M. ma inum. The mycobac e ial loads we e de e mined al eady a 4 wpi because, a his s age o he in ec ion, bac e ial coun s ha e gene ally eached a s eady s a e and hey a e easily measu able. P iming 1 day p io o in ec ion was mos e icien , inc easing he equency o a clea ing 2 RESEARCH ARTICLE Disease Models & Mechanisms (2018) 11, dmm031658. doi:10.1242/dmm.031658 Disease Models & Mechanisms esponse by 3.7- old compa ed o he con ol g oup (7 days: 2.3- old) in his p elimina y expe imen (Fig. 1C). Based on hese esul s, we con inued wi h HKLm p iming a 1 day be o e M. ma inum in ec ion. P iming wi h HKLm inc eases he equency o clea ance o mycobac e ial in ec ion in adul zeb a ish To e i y ou inding om he p elimina y p iming expe imen s, zeb a ish we e p imed wi h an injec ion o HKLm (Fig. 2A) 1 day p io o M. ma inum in ec ion, and PBS was used as a nega i e con ol. Compa ed o con ol ea men s, p iming wi h HKLm 1 day be o e in ec ion consis en ly led o a signi ican dec ease (14.1- old dec ease, on a e age) in mycobac e ial loads in adul zeb a ish a 4 wpi (Fig. 2A shows a ep esen a i e esul o ou sepa a e expe imen s). F om ou p e ious s udy, we know ha , wi hou p iming, app oxima ely 10% o M. ma inum-in ec ed zeb a ish a e able o na u ally clea he mycobac e ial in ec ion (Hamma én e al., 2014). Combining he esul s om ou di e en expe imen s shows ha p iming wi h HKLm 1 day p io o M. ma inum in ec ion inc eases he equency o s e ilizing esponse by 6.8- old in he wild- ype ish a 4 wpi [PBS 3.7% (n=54) s HKLm 25% (n=56), P=0.0021] (Fig. 2B). Howe e , indi iduals ha ha e been unable o clea he in ec ion also bene i om HKLm p iming in e ms o lowe ed bac e ial loads (Fig. S2). In a wild- ype popula ion wi h sus ained in ec ion, HKLm educed he median bac e ial load by 11.1- old (P<0.0001). Inna e immune mechanisms media e HKLm-induced p o ec ion agains mycobac e ial in ec ion Ou esul s show a bene icial e ec o HKLm p iming on mycobac e ial loads and clea ance a 7 wpi (Fig. 1B) and 4 wpi (Fig. 2A), as well as a 2 wpi (Fig. 2C), all o which a e la e enough s ages o bo h inna e and adap i e esponses o be ac i e, al hough i is known ha i ulen mycobac e ia cause a delay in he ac i a ion o T-cell esponses (Gallegos e al., 2008). To es whe he HKLm p iming has he same educing e ec on mycobac e ial loads in he absence o an adap i e immune esponse, we used ag1 −/− mu an ish. Fish lacking ag1 do no unde go V(D)J ecombina ion, which is essen ial o he p oduc ion o he ull a ie y o T- and B-cell ecep o s. The ag1 −/− mu an s he e o e ely solely on hei inna e immune mechanisms o p o ec ion agains in ec ions (Wienholds e al., 2002). When ag1 −/− mu an s we e p imed wi h HKLm 1 day be o e M. ma inum in ec ion and bac e ial loads we e measu ed wi h qPCR a 4 wpi, HKLm p iming caused a simila educ ion in bac e ial loads in ag1 −/− mu an s (30- old, P=0.0089) as in wild- ype ish (9.7- old, P=0.0013) (Fig. 2D,A). Also in ag1 −/− mu an s, p iming wi h HKLm inc eased he equency o s e ilizing esponse om 0% (n=26) o 17% (n=23) (P=0.0418) (Fig. 2E). Looking a he ish wi h a sus ained in ec ion (unable o clea he in ec ion), HKLm p iming educed he bac e ial load in ag1 mu an s by 7.2- old (Fig. S2B, no s a is ically signi ican ). HKLm p iming did no a ec he cumula i e mo ali y o low-dose-in ec ed wild- ype adul ish Fig. 1. The e ec o p iming agen s on M. ma inum loads in adul zeb a ish. (A) Ou line o he s udy. Adul wild- ype zeb a ish we e p imed wi h 0.5×10 7 - 1×10 7 colony- o ming uni s (c u) pe ish o hea -killed bac e ia o wi h o he p iming agen s (13.5 µg pe ish, excep MDP 4.5 µg) 1 o 7 days p io o M. ma inum in ec ion (−1/7 day) wi h an i.p. injec ion. S e ile 1× PBS was used as an injec ion con ol. The ollowing day (0 day), a low dose o M. ma inum was injec ed i.p. in o he zeb a ish. In e nal o gans we e collec ed ei he 2, 4 o 7 weeks pos in ec ion (wpi), DNA was ex ac ed and he bac e ial coun s we e measu ed wi h M. ma inum-speci ic qPCR. (B) P iming wi h hea -killed L. monocy ogenes (HKLm) educes mycobac e ial loads a 7 wpi. Zeb a ish we e p imed wi h ei he PBS (n=8), hea -killed M. ma inum (HKMm; n=9), HKLm (n=10), hea -killed S. iniae (HKSi; n=8), hea -killed E. coli (KHEc; n=9), lipopolysaccha ide (LPS; n=8), pacli axel (n=5), mu amyl dipep ide (MDP; n=9) o zymosan (n=10) 1 day p io o M. ma inum in ec ion (16±4 c u). O gans we e collec ed a 7 wpi. P iming wi h HKLm led o a bigge dec ease in he mycobac e ial loads in adul zeb a ish compa ed o o he es ed p iming agen s. Pacli axel inc eased he mycobac e ial loads in adul zeb a ish. Medians a e shown in he igu e. (C) P iming wi h HKLm 1 day p io o M. ma inum in ec ion inc eases he equency o clea ance. The old change in he pe cen age o ish ha we e able o clea he M. ma inum in ec ion was highe in he g oup ha was p imed wi h HKLm 1 day p io (3.7- old; PBS: n=12, HKLm: n=13) o in ec ion compa ed o g oup ha was p imed 7 days (2.3- old; PBS: n=12, HKLm: n=12) be o e he in ec ion. 3 RESEARCH ARTICLE Disease Models & Mechanisms (2018) 11, dmm031658. doi:10.1242/dmm.031658 Disease Models & Mechanisms (PBS: 11.7%; HKLm: 11.2%) bu caused a end o educed mo ali y in ag1 −/− mu an ish (PBS: 24.5%; HKLm: 11.1%) (Fig. 2F,G). Toge he , hese esul s imply an impo an ole o inna e immuni y in he o ma ion o p o ec i e HKLm-induced immune esponses agains mycobac e ial in ec ion. Impo an ly, hey show ha an op imal esponse induced by HKLm p iming signi ican ly inc eases he equency o clea ance o mycobac e ial in ec ion also in he absence o adap i e immuni y. P o ec i e e ec s o HKLm p iming a e no seen in he la al M. ma inum-in ec ion model As ou esul s showed ha he p o ec i e e ec s o HKLm p iming can be media ed h ough inna e immune esponses alone, we nex used zeb a ish la ae o es he e ec o HKLm p iming on M. ma inum in ec ion. Du ing he i s weeks o de elopmen , zeb a ish lack adap i e immune esponses and ely solely on inna e immune esponses, making i possible o s udy mechanisms o mycobac e ial in ec ion ha a ise om inna e immuni y. To his end, zeb a ish la ae we e p imed wi h HKLm o induce a p o ec i e immune esponse agains M. ma inum in ec ion. Zeb a ish we e p imed by in a enous injec ion ei he a 1 day pos - e iliza ion (dp ) (Fig. S3A) o a 2 dp (Fig. S3B). M. ma inum in ec ion was ca ied ou in a enously a 2 dp . Wi hin his expe imen al se ing, mycobac e ial loads we e no educed in zeb a ish la ae a e p iming wi h HKLm (Fig. S3), possibly due o he imma u i y o immune esponses in he young la ae o due o he ela i ely lowe HKLm dose ha could be deli e ed in o he la ae.As we we e unable o see any p o ec i e e ec in he zeb a ish la ae, we con inued using adul zeb a ish o subsequen expe imen s. P o ec i e immuni y agains M. ma inum in ec ion is media ed by a p o ein and/o nucleic acid componen o HKLm Based on he esul s o he p elimina y expe imen wi h di e en p iming agen s, HKLm seemed o con ain a speci ic componen esponsible o he induc ion o s e ilizing immuni y. To cha ac e ize Fig. 2. P iming wi h HKLm signi ican ly educes mycobac e ial loads in adul zeb a ish ia inna e esponses. (A) P iming o adul zeb a ish wi h 0.5×10 7 c u o HKLm 1 day p io o M. ma inum in ec ion (27±2 c u) led o a signi ican dec ease in mycobac e ial loads compa ed o con ol injec ion o s e ile 1× PBS. The g aph shows one ep esen a i e expe imen . Samples we e collec ed a 4 wpi (PBS: n=19, HKLm: n=19). (B) P iming wi h HKLm 1 day p io o M. ma inum in ec ion leads o s e iliza ion o M. ma inum in 25% o he wild- ype (WT) zeb a ish. Clea ance pe cen age in he WT PBS con ol g oup was 3.7%. The da a we e collec ed om ou independen expe imen s. M. ma inum in ec ion doses in he independen expe imen s we e 27±2 c u, 26±13 c u, 75±13 c u and 26±8 c u. PBS: n=54, HKLm: n=56. (C) P iming o adul zeb a ish wi h HKLm leads o a signi ican dec ease in mycobac e ial loads compa ed o PBS con ols al eady a 2 wpi. In ec ion dose: 48±8 c u. PBS: n=11, HKLm: n=12. (D) HKLm p iming 1 day p io o M. ma inum in ec ion signi ican ly educed mycobac e ial loads in ag1 −/− mu an ish compa ed o he PBS con ol g oup a 4 wpi, indica ing a ole o inna e immune esponses. The g aph con ains a combined esul om wo sepa a e expe imen s. M. ma inum in ec ion doses we e 48±8 c u and 27±2 c u. (E) P iming wi h HKLm 1 day p io o M. ma inum in ec ion leads o s e iliza ion o M. ma inum in 17% o he ag1 −/− mu an ish. Clea ance pe cen age in he PBS con ol g oup was 0%. Da a a e pooled om wo independen expe imen s. PBS: n=26, HKLm: n=23. (F,G) HKLm p iming did no a ec cumula i e mo ali y in WT adul ish (F; PBS: n=152, HKLm: n=170), bu caused a end o educed mo ali y in HKLm-injec ed ag1 −/− mu an ish (PBS: n=53, HKLm: n=55). P- alues in A, C and D we e calcula ed wi h a wo- ailed non-pa ame ic Mann–Whi ney es wi h G aphPad P ism. Medians o he indi idual expe imen s a e shown in he igu es. The P- alues in B, E, F and G we e calcula ed wi h Fishe ’s es using G aphPad (QuickCalcs) online so wa e. 4 RESEARCH ARTICLE Disease Models & Mechanisms (2018) 11, dmm031658. doi:10.1242/dmm.031658 Disease Models & Mechanisms he p o ec i e componen (s), we ca ied ou a se o expe imen s in which zeb a ish we e p imed wi h di e en p epa a ions o HKLm 1 day p io o M. ma inum in ec ion and bac e ial g ow h was analyzed a 4 wpi. P iming wi h spen L. monocy ogenes medium did no cause a signi ican educ ion in he mycobac e ialloads o he ish, indica ing ha a sec e ed ac o was no he causa i e agen behind he ac i a ion o an imycobac e ial immune esponses (Fig. 3A). We also ound ha , a e au ocla a ion o L. monocy ogenes, p iming wi h he insoluble ma e ial p o ided a p o ec i e e ec agains M. ma inum in ec ion, whe eas he soluble ma e ial was no p o ec i e (Fig. 3B). Howe e , we we e able o show ha ea men o he HKLm ex ac wi h ei he p o einase K o a cock ail o RNase and DNase abolished he p o ec i e e ec (Fig. 3C). These esul s indica e ha he p o ec i e componen is s able, esis an o high empe a u e and p essu e, non-sec e ed, insoluble, and likely consis s o bo h a p o ein and nucleic acid componen . The apeu ic po en ial o HKLm ea men To u he cha ac e ize he ea u es o HKLm ea men , a 30- old highe HKLm dose was injec ed 1 day p io o low-dose M. ma inum in ec ion. Wi h such a high dose, HKLm los i s bene icial e ec s (Fig. 4A) and he e was a end o inc eased mo ali y o he ish by 4 wpi (PBS: 22.7%; HKLm: 41.7%) (Fig. 4B). Acco ding o hese esul s, i was concluded ha , by inc easing he HKLm dose, i is no possible o u he boos he p o ec i e e ec bu a he ice e sa. The op imal dosage hus plays a c ucial ole in inducing a p o ec i e esponse. We hen es ed whe he p iming wi h HKLm 1 day p io o in ec ion could p o ec ish agains a high-dose in ec ion challenge wi h 4883±919 colony- o ming uni s (c u) o M. ma inum. In he con ex o a high-dose in ec ion, HKLm did no induce signi ican di e ence in he mycobac e ial loads, clea ance o cumula i e end-poin mo ali y a 4 wpi (Fig. 4C,D). We nex wan ed o es whe he HKLm can induce p o ec i e immune esponses i he mycobac e ial in ec ionis al eadyes ablished. We ha e p e iously shown in adul zeb a ish ha , a 2 wpi, M. ma inum in ec ion is well es ablished and g anulomas ha e al eady s a ed o o m (Pa ikka e al., 2012). We in ec ed adul zeb a ish wi h a low dose o M. ma inum, injec ed HKLm a 2 wpi and measu ed mycobac e ial loads, as well as de e mined he cumula i e mo ali y a 4 wpi. Based on ou esul s, a a ime poin in which he mycobac e ia ha e al eady mul iplied subs an ially and s a ed o ming g anulomas, HKLm injec ion was no able o lowe he bac e ial loads, inc ease bac e ial clea ance o a ec cumula i e mo ali y (Fig. 4E,F). P iming adul zeb a ish wi h HKLm induces mpeg1, n αand nos2 exp ession, and down egula es sod2, a he ea ly phase o mycobac e ial in ec ion We we e in e es ed in u he cha ac e izing he na u e o he immune esponse induced by HKLm p iming in adul zeb a ish. We hypo hesized ha he p o ec i e e ec could be media ed h ough enhanced killing o mycobac e ia a he ea ly phase o he in ec ion due o changes in he numbe s o ac i i y o inna e immune cells. To s udy he de ails o HKLm-induced immune ac i a ion by qPCR, he ish we e p imed wi h HKLm o PBS, in ec ed wi h a low dose o M. ma inum 1 day a e p iming and he o al RNA was ex ac ed om zeb a ish o gans collec ed 1 day pos -in ec ion (dpi). Fi s , he possible e ec s o HKLm ea men on he numbe o mac ophages and neu ophils we e assessed by measu ing he exp ession o mac ophage exp essed gene 1 (mpeg1) (Elle e al., 2011) and myeloid-speci ic pe oxidase (mpx) (Lieschke e al., 2001), espec i ely. In he HKLm g oup, mpeg1 exp ession was signi ican ly highe han in he con ol g oup (Fig. 5A; HKLm: 59.9- old; PBS: 39.3- old; P=0.0352), sugges ing ha he numbe o mac ophages was inc eased due o HKLm p iming. The exp ession o mpx was no a ec ed by HKLm (Fig. 5B). Fig. 3. P o ec i e immuni y agains M. ma inum is media ed by a p o ein and/o nucleic acid componen o HKLm. (A-C) Zeb a ish we e p imed wi h di e en componen s o HKLm and L. monocy ogenes 1 day p io o M. ma inum in ec ion and mycobac e ial loads we e de e mined wi h M. ma inum-speci ic qPCR a 4 wpi. (A) P o ec i e immuni y is no media ed by a sec e ed componen in he L. monocy ogenes g ow h medium. In ec ion dose: 26±6 c u; PBS: n=10, medium: n=16. (B) P o ec i e immuni y is media ed by a componen ha was ound in he insoluble phase. In ec ion dose: 33±11 c u; PBS: n=10, insoluble: n=12, soluble: n=12. (C) P o ec i e e ec o HKLm p iming was los when HKLm was ea ed wi h DNase and RNase, o p o einase K. In ec ion dose: 33±11 c u; PBS: n=10, DNase and RNase: n=7, p o einase K: n=8. P- alues o all expe imen s we e calcula ed wi h a wo- ailed non- pa ame ic Mann–Whi ney es wi h G aphPad P ism and co ec ed wi h he Bon e oni’s me hod. Medians o each expe imen a e shown. 5 RESEARCH ARTICLE Disease Models & Mechanisms (2018) 11, dmm031658. doi:10.1242/dmm.031658 Disease Models & Mechanisms To assess he HKLm-induced changes a he le el o immune cell ac i a ion, we measu ed he exp ession o a selec ion o ma ke s ela ed o he e ec i e an imycobac e ial unc ions o inna e immune cells 1 day a e M. ma inum in ec ion. Based on li e a u e on he mechanisms limi ing in acellula mycobac e ial g ow h, he genes chosen o analysis we e n α(Coba e al., 2015; Roca and Ramak ishnan, 2013), i nγ(Flynn e al., 1993) and nos2b (Nicholson e al., 1996; Thoma-Uszynski e al., 2001). Exp ession o i nγwas no di e en ially induced be ween PBS- and HKLm- ea ed g oups (Fig. 5C,D). The median exp ession le els o n α (Fig. 5F; HKLm 10.3- old s PBS 1.1- old, P=0.0043) and nos2b (Fig. 5G; HKLm 10.7- old s 0.6- old PBS, P=0.0001) we e signi ican ly inc eased in he HKLm-p imed g oup compa ed o he PBS con ol g oup. We also analyzed he exp ession le els o sod2, which encodes a mi ochond ial p o ein ha con e s he byp oduc s o oxida i e phospho yla ion o hyd ogen pe oxide and oxygen, leading o neu aliza ion o mi ochond ial eac i e oxygen species (ROS) (Pias e al., 2003) and a g1, which is an al e na i e mac ophage ac i a ion ma ke (Go don and Ma inez, 2010). In he HKLm g oup, sod2 exp ession was signi ican ly mo e down egula ed as compa ed o he PBS g oup, indica ing inc eased le els o ROS due o HKLm p iming (PBS 0.59- old s HKLm 0.27- old, P=0.0022) (Fig. 5H). A g1, howe e , was no induced in HKLm-p imed ish, sugges ing ha al e na i e ac i a ion does no ha e an impac on he ea ly mycobac e ial elimina ion (Fig. 5E). Toge he , hese esul s sugges ha HKLm induces M1- ype classical mac ophage ac i a ion leading o enhanced in acellula killing a he ea ly s ages o a mycobac e ial in ec ion. P iming o mac ophages wi h HKLm leads o dec eased oxygen consump ion in i o Classically ac i a ed M1 mac ophages ha e been shown o change hei me abolism upon immune ac i a ion (Cheng e al., 2014). To es whe he his is also he case wi h HKLm p iming, we se up a RAW264.7 cell cul u e and p imed hem ei he wi h LPS (50 ng/µl) o HKLm (MOI: 530). A 19-24 h la e , oxygen consumed by hese cells was measu ed using a Cla k elec ode. We used LPS as a posi i e con ol and showed ha bo h ea men s lowe he oxygen consump ion signi ican ly (Fig. 5I). Pooled esul om h ee independen expe imen s concluded ha HKLm- o LPS-p imed mouse mac ophages consume, espec i ely, 1.8- imes (P=0.008) and 2.3- imes (P=0.0042) less oxygen compa ed o un ea ed cells, in i o (Fig. 5I). This esul implies ha HKLm- ea ed mac ophages exhibi me abolic changes indica i e o diminished oxygen simila o hose obse ed du ing classical mac ophage ac i a ion. DISCUSSION Despi e subs an ial p og ess in he ield o medicine, TB s ill kills millions o people e e y yea and has been decla ed a global public heal h c isis (Wo ld Heal h O ganiza ion, 2016; h p://www.who. in / b/publica ions/2016/en/). The ad ances in he ba le agains TB ha e been hinde ed by he complex na u e o he disease and he limi a ions o ube culosis animal models. TB mani es s i sel in a wide spec um o disease, wi h mos a ec ed indi iduals being unable o e adica e he causa i e bac e ia (M b), leading o he de elopmen o la en TB in ec ion, which has a li e ime isk o eac i a ion in 5-10% o cases. The Wo ld Heal h O ganiza ion has Fig. 4. HKLm ea men does no p o ec agains high-dose o es ablished M. ma inum in ec ion. (A,B) P o ec i e e ec o HKLm p iming is los wi h high-dose p iming. Fish we e injec ed wi h a high dose o HKLm (15.6×10 7 c u, 30- old compa ed o p e ious dose) 1 day p io o M. ma inum in ec ion (34±11 c u). P iming wi h a high dose o HKLm did no educe mycobac e ial numbe s (A) and led o an inc ease in he mo ali y o he ish a 4 wpi (B). PBS: n=10, HKLm: n=10. (C,D) HKLm p iming does no p o ec om high- dose M. ma inum in ec ion. Fish we e p imed wi h HKLm 1 day p io o high-dose M. ma inum in ec ion (4883±919 c u). No e ec was obse ed on bac e ial loads (C) o cumula i e end-poin mo ali y (D). PBS: n=16, HKLm: n=17. (E,F) HKLm does no p o ec agains an es ablished M. ma inum in ec ion. Fish we e injec ed wi h HKLm 2 weeks a e an M. ma inum in ec ion (22±6 c u). No e ec on mycobac e ial loads (E) o cumula i e end-poin mo ali y (F) was obse ed. PBS: n=14, HKLm: n=16. P- alues o bac e ial loads we e calcula ed wi h a wo- ailed non-pa ame ic Mann–Whi ney es wi h G aphPad P ism (A,C,E). Medians o he expe imen s a e shown. 6 RESEARCH ARTICLE Disease Models & Mechanisms (2018) 11, dmm031658. doi:10.1242/dmm.031658 Disease Models & Mechanisms es ima ed ha 2-3 billion people a e la en ly in ec ed wi h M b (Wo ld Heal h O ganiza ion, 2016), c ea ing a huge pool o indi iduals wi h a po en ial o de elop an ac i e, ansmissi e disease. As cu en accina ion and an ibio ic schemes ha e p o en insu icien o con olling he global TB epidemic, hos -di ec ed he apies inducing p o ec i e immune esponses a e eme ging as a no el s a egy o ea TB (Tobin, 2015). E icien hos -di ec ed he apies used ei he alone o in combina ion wi h an ibio ics a e an app oach ha could po en ially lead o s e iliza ion o TB. Al hough genome-wide associa ion s udies ca ied ou in humans ha e gi en impo an clues on he na u e o p o ec i e immune esponses agains TB (Azad e al., 2012; Wilkinson e al., 2000; Coba e al., 2009, 2015), animal models a e essen ial o he execu ion o mo e mechanis ic s udies. The mouse, abbi and macaque ha e been widely used o s udy TB (Myllymäki e al., 2015). To he bes o ou knowledge, spon aneous o induced s e ilizing immuni y has no been obse ed a he o ganismal le el in mammalian animal models o TB. Clea ance o cul i able mycobac e ia occu s in he abbi model (Subbian e al., 2012). Howe e , s anda d bac e ial cul u ing me hods only de ec he ac i ely eplica ing mycobac e ial popula ions, bu no do man bac e ia (Chao and Rubin, 2010), and, in he abbi , he clea ance o cul i able mycobac e ia indica es he es ablishmen o a uly la en disease ins ead o s e iliza ion (Subbian e al., 2012). The zeb a ish has ecen ly become a well-accep ed gene ically ac able e eb a e model o human TB pa hogenesis o complemen he mo e adi ional mammalian models (Myllymäki e al., 2015). In ou p e ious s udy, using a qPCR-based me hod, we we e able o see spon aneous clea ance o mycobac e ial in ec ion in he zeb a ish- M. ma inum in ec ion model in app oxima ely 10% o he ish a 4 wpi (Hamma én e al., 2014). As we obse ed no clea ance a 2 wpi (Hamma én e al., 2014), spon aneous ea ly clea ance (likely induced by inna e mechanisms) in ou wild- ype zeb a ish popula ion is a a e e en . In his cu en s udy, we we e able o inc ease he equency o s e ilizing mycobac e ial in ec ion by p iming he inna e immune esponse p io o M. ma inum in ec ion. Ou hypo hesis was ha p iming o s imula ion o he immune esponse in he adul zeb a ish be o e M. ma inum in ec ion could lead o a s e ilizing immune esponse ins ead o a le hal p ima y ac i e disease o a la en in ec ion ha is p one o eac i a ion la e in li e (Pa ikka e al., 2012). By using an a ay o di e en p iming agen s, we wan ed o s udy whe he we could c ea e a clima e ha would allow he immune esponse o ci cum en hei e icien i ulence s a egies and e en e adica e he mycobac e ia. P iming app oaches ha e been success ul in inducing p o ec i e esponses agains a ious o he bac e ial in ec ions in he ui ly, D osophila melanogas e , which elies solely on inna e immuni y (Pham e al., 2007). Indeed, ou esul s showed ha s e ilizing immuni y can be induced in he M. ma inum zeb a ish model. In ou hands, p iming wi h HKLm induces a s e ilizing esponse in 25% o M. ma inum- Fig. 5. HKLm p iming induces mpeg, n αand nos2b exp ession, down egula es sod2 exp ession in adul zeb a ish and leads o dec eased oxygen consump ion in i o.(A-H) The exp ession le els o mpeg (A), mpx (B), i nγ1-1 (C), i nγ1-2 (D), a g1 (E), n α(F), nos2b (G) and sod2 (H) we e measu ed wi h qPCR om wild- ype ish p imed wi h HKLm o s e ile PBS bu e as a con ol. A 1 day a e he p iming, he ish we e in ec ed wi h a low dose (67±16 c u) o M. ma inum. Samples o qPCR analysis we e collec ed a 1 dpi. The esul s we e no malized o unin ec ed wild- ype baseline con ol. PBS: n=10, HKLm: n=11. (I) HKLm p iming leads o me abolic changes in i o. RAW264.7 cells we e p imed wi h LPS o HKLm and oxygen consump ion was measu ed 19-24 h a e p iming. HKLm p iming leads o a 1.8- old dec ease (P=0.008) in oxygen consump ion compa ed o con ol. LPS was used as a posi i e con ol (2.3- old dec ease, P=0.0042). PBS: n=9, LPS: n=7, HKLm: n=9. P- alues o all expe imen s we e calcula ed wi h a wo- ailed non-pa ame ic Mann–Whi ney es wi h G aphPad P ism. Medians o each expe imen a e shown. Bon e oni co ec ion was used in C. 7 RESEARCH ARTICLE Disease Models & Mechanisms (2018) 11, dmm031658. doi:10.1242/dmm.031658 Disease Models & Mechanisms in ec ed ish a 4 wpi, whe eas he pe cen age o spon aneous clea ance was 3.7% in he PBS-p imed con ol g oup. The M. ma inum-speci ic qPCR used o quan i y bac e ial loads is a sensi i e me hod, wi h a de ec ion limi o app oxima ely 100 bac e ial genomes in he en i e ish (Pa ikka e al., 2012; Hamma én e al., 2014). The majo ad an age o a qPCR-based me hod o e cul u ing me hods is ha qPCR de ec s all bac e ial genomes i espec i e o me abolic s a e so ha do man bac e ia a e also de ec ed. An ad an age o he zeb a ish is ha , due o i s small size, all in ec ion a ge o gans can be collec ed o de e mina ion o bac e ial load, which is no p ac ically easible in la ge animals. Based on he qPCR esul s, we a e able o say ha 25% o he wild- ype ish we e able o clea he in ec ion, indica ing ha HKLm p iming indeed inc eases he equency o s e ilizing immune esponse. This model p o ides ools o elucida ing he de ailed mechanisms behind s e ilizing immuni y agains TB. To s udy whe he he p o ec i e e ec s o HKLm p iming equi e a unc ional adap i e immune sys em, we used ag1 mu an zeb a ish. Al hough ag1 mu an zeb a ish a e known o be hype suscep ible o M. ma inum in ec ion due o a ailu e in he de elopmen o unc ional T and B cells (Swaim e al., 2006; Pa ikka e al., 2012), hei bac e ial loads we e signi ican ly lowe ed and e en clea ed by HKLm p iming, sugges ing ha he p o ec i e esponse can be media ed by inna e immuni y alone. The ole o inna e immuni y is also suppo ed by he signi ican HKLm- induced educ ion in bac e ial loads as ea ly as 2 wpi in wild- ype ish, by which ime adap i e esponses a e only s a ing o a ise in mycobac e ial in ec ions (Ande sen and Woodwo h, 2014). Despi e he clea in ol emen o inna e esponses in HKLm- induced p o ec i e esponses in adul ish, he esul s could no be ep oduced in young zeb a ish la ae ha ha e only inna e immuni y, p obably due o an imma u e immune esponse and echnical di icul ies o deli e high enough doses o HKLm by mic oinjec ion me hods. As adap i e esponses a e cen ally in ol ed in he pa hogenesis o mycobac e ial in ec ions, using he adul zeb a ish will likely be e model he e ec s o immunomodula o y ea men s on he immune esponse as a whole. On op o he p o ec i e e ec s o HKLm media ed h ough inna e immuni y, an addi ional le el o p o ec ion was obse ed in he p esence o a unc ional adap i e immune sys em in zeb a ish. Howe e , he obse a ion ha ac i a ion o inna e immuni y alone by immune p iming wi h HKLm in some cases was su icien o induc ion o a p o ec i e o s e ilizing immuni y opens new a enues o hos -di ec ed he apies o p e en i e s a egies e en in he absence o a ully unc ional adap i e immune sys em, such as in pa ien s wi h an HIV co-in ec ion. A he momen , he e is no immuno he apy ha could cu e an ongoing mycobac e ial in ec ion by simply boos ing he immune sys em. The e o e, we wan ed o s udy whe he HKLm could induce a p o ec i e esponse du ing an es ablished mycobac e ial in ec ion. A single dose o HKLm was injec ed 2 weeks a e M. ma inum in ec ion, when ea ly g anulomas ha e s a ed o o m and a e isible in a ious o gans (Pa ikka e al., 2012). Howe e , his single injec ion o HKLm did no dec ease mycobac e ial loads a 4 wpi (Fig. 4E,F). By his ime poin , M. ma inum has al eady had plen y o ime o exe i s ea ly i ulence s a egies leading o e ec i e a oidance o he p o-in lamma o y hos immune esponse (Elks e al., 2014; Que al e al., 2016; Bha e al., 2017; Cambie e al., 2014). Ha ing gained a oo hold wi hin i s hos , mycobac e ia a e no as p one o he e ec s caused by a single he apeu ic injec ion o HKLm as hey a e when en e ing a p imed hos . Adul zeb a ish ha e also been used o model ac i e ulminan TB by in ec ing indi iduals wi h a high M. ma inum dose (Pa ikka e al., 2012). Indi iduals in ec ed wi h a high M. ma inum dose did no bene i om HKLm p iming. The high in ec ion dose o a ew housand mycobac e ia leads o a disease s a e in which he capaci y o he immune sys em apidly becomes sa u a ed, allowing he bac e ia o g ow almos loga i hmically (Pa ikka e al., 2012) possibly due o he limi ed numbe o mac ophages (Pagán e al., 2015) compa ed o he low in ec ion dose; a si ua ion oo challenging o o e come e en in he p esence o HKLm p iming. Howe e , in ou p elimina y expe imen s, we saw ha , in addi ion o p iming 1 day p io o in ec ion, p o ec i e e ec s we e s ill isible when HKLm p iming was deli e ed 1 week p io o M. ma inum in ec ion (Fig. 1C), sugges ing ha his ype o p o ec i e esponse could be conside ed in designing new p e en i e s a egies. The p o ec i e e ec s o HKLm ea men deli e ed p io o in ec ion could be media ed h ough he induc ion o ained immuni y. I is known ha inna e immune cells can media e an enhanced immune esponse upon ein ec ion wi h he same pa hogen (Quin in e al., 2012). The inna e immune sys em can also c oss- eac wi h a new pa hogen acco ding o p e ious s imuli (Kleinnijenhuis e al., 2014). I has been epo ed ha some accines can p oduce du able c oss-p o ec ion ha canno be explained by adap i e esponses (Aaby e al., 2014). This ype o nonspeci ic inna e memo y is e e ed o as ained immuni y. Mechanisms o ained immuni y ha e also been shown o be esponsible o BCG- induced by-s ande p o ec ion agains Candida albicans in he mouse (Van’ Wou e al., 1992). Inna e memo y is media ed h ough e e sible epigene ic changes a he han i e e sible gene ic ecombina ion seen du ing he o ma ion o classical immunological memo y in adap i e immune cells and can las o weeks o mon hs ( e iewed by Ne ea e al., 2016). The e ec s o ained immuni y in he con ex o suscep ibili y o TB is an in e es ing a ea o esea ch and can yield new app oaches in he de elopmen o p e en i e s a egies based on inna e immuni y. I has been epo ed ha he numbe o mac ophages is c i ical o he disease ou come and ha mac ophage de iciency is connec ed o accele a ed p og ession o mycobac e ial in ec ion (Pagán e al., 2015). To assess he e ec o HKLm p iming on he numbe o mac ophages and neu ophils, we measu ed he exp ession o he commonly used ma ke s mpeg1 (Elle e al., 2011) and mpx (Lieschke e al., 2001), espec i ely. Based on he exp ession o hese ma ke s, he numbe o mac ophages was signi ican ly highe in HKLm-p imed ish (P=0.0352), whe eas he amoun o neu ophils emained unchanged (Fig. 5A,B). This change seen in mac ophages is po en ially media ing he p o ec i e esponse agains mycobac e ia. To deciphe he ype o p o ec i e immune esponse induced by HKLm p iming, we measu ed an a ay o genes ela ed o inna e immune ac i a ion in he o gans o M. ma inum-in ec ed zeb a ish a 1dpi.Inlinewi hin i o s udies on Lis e ia (Ba buddhe e al., 1998; Mi ko i ch e al., 2006), in i o p iming wi h HKLm caused a signi ican inc ease in nos2b and n αwi h a simul aneous dec ease in sod2 exp ession. nos2b is one o he NO syn hases in zeb a ish (Lepille e al., 2009). NO is known o be mycobac e iocidal (Nicholson e al., 1996) bu , as pa hogenic mycobac e ia ha e de eloped e asion s a egies o inhibi he p oduc ion o NO (Elks e al., 2014; Que al e al., 2016; Bha e al., 2017), he NO le els na u ally induced in M b-in ec ed mac ophages seem o be insu icien o lysing mycobac e ia (Jung e al., 2013). Thus, he addi ional p oduc ion o Nos caused by HKLm p io o in ec ion likely po en ia es he in acellula killing mechanisms. The bene icial e ec s o inc eased NO in neu ophils (Elks e al., 2013) as well as in 8 RESEARCH ARTICLE Disease Models & Mechanisms (2018) 11, dmm031658. doi:10.1242/dmm.031658 Disease Models & Mechanisms mac ophages (Cambie e al., 2014) du ing ea ly M. ma inum in ec ion ha e p e iously been demons a ed in zeb a ish la ae. Cambie and colleagues showed ha i ulen mycobac e ia a oid NO- media ed in acellula killing du ing he ea ly phase o in ec ion by hiding hei TLR ligands unde a ph hioce ol dimycoce ose a e coa (Cambie e al., 2014). Co-in ec ing zeb a ish la ae wi h M. ma inum and S aphylococcus au eus o Pseudomonas ae uginosa leads o a enua ion o mycobac e ial in ec ion (Cambie e al., 2014). Also, co-injec ion o li e M. ma inum wi h hea -killed M. ma inum o wi h a mu an wi h exposed TLR ligands caused simila a enua ion (Cambie e al., 2014). I is likely ha a leas pa o he p o ec i e e ec s caused by HKLm in he adul zeb a ish a e media ed h ough TLR ligands. The componen analysis o HKLm sugges ed ha nucleo ides we e impo an o HKLm-media ed p o ec ion agains M. ma inum. TLR9, which ecognizes double-s anded DNA and leads o he induc ion o Nos2 (I o e al., 2005), is a ecep o po en ially esponsible o he p o ec ion. Howe e , de ailed analysis o he signaling pa hways ac i a ed by HKLm ea men was beyond he scope o his s udy. Tn αis also known o media e in acellula killing o mycobac e ia by mac ophages (Roca and Ramak ishnan, 2013) and op imal le els o his cy okine ha e been p oposed o lead o ea ly clea ance o TB in humans (Coba e al., 2015). S udies in zeb a ish la ae ha e shown ha high Tn αle els alongside high ROS le els wi hin mac ophages, du ing he ea ly days o mycobac e ial in ec ion, is bac e icidal (Roca and Ramak ishnan, 2013). Sod2 is an enzyme ha ac s h ough neu aliza ion o mi ochond ial ROS (Pias e al., 2003). I s down egula ion by HKLm should hus cause an inc ease in mi ochond ial ROS, he high le els o which ha e been shown o enhance in acellula killing mechanisms wi hin mac ophages ( e iewed in Hall e al., 2014). Recen ly, in a human popula ion s udy, a gene ic a ian leading o educed ac i i y o Sod2 was ound o be associa ed wi h inc eased esis ance o lep osy, a disease caused by Mycobac e ium lep ae (Ramos e al., 2016). The e o e, a likely mechanism o clea ing he mycobac e ial in ec ion by HKLm p iming in he adul zeb a ish model is media ed h ough inc eased Tn αand dec eased Sod2 p oduc ion ha oge he lead o highe , mycobac e iocidal, le els o ROS wi hin mac ophages. Roca and Ramak ishnan also showed ha , when he exp ession o Tn αis endogenously high, con inuously high ROS le els a e he i s days o in ec ion lead o excessi e in lamma ion, nec osis and exace ba ion o he disease (Roca and Ramak ishnan, 2013). This p obably also explains he inc eased mo ali y wi h he highe HKLm dose. Wi h a single small dose o HKLm used in ou s udy, he e ec s o he ea men we e undoub edly posi i e, bu i mus be kep in mind ha excessi e o p olonged induc ion o Tn αand ROS can also ha e de imen al e ec s. The dosage o ea men as well as he geno ype o he hos , a ec ing he baseline p oduc ion o in lamma o y cy okines, will also need o be ca e ully conside ed in he de elopmen o hos -di ec ed immunomodula o y ea men s. Based on he gene exp ession da a, he changes in he inna e immuni y induced by HKLm in he adul zeb a ish seem o be media ed h ough an inc eased numbe and ac i a ion o M1- ype mac ophages. Recen esea ch on he me abolism o di e en inna e immune cells has shown ha M1 mac ophages ha e dec eased oxygen consump ion, and inc eased glycolysis and lac a e p oduc ion (Cheng e al., 2014). In ou expe imen s, he oxygen consump ion o mouse mac ophages was signi ican ly dec eased by HKLm p iming (Fig. 5I), p o iding a u he piece o e idence o M1 mac ophages playing a cen al ole in HKLm-media ed p o ec ion agains mycobac e ial in ec ion. The esul also implies ha he ypes o ac i a ion caused by HKLm in he zeb a ish a e simila o hose induced in mammalian mac ophages. O e all, we show ha p o ec i e and e en s e ilizing immune esponses can be induced in he zeb a ish model o TB by p iming wi h HKLm.The esponse is induced e en in he absence o adap i e immuni y and is accompanied by he inc ease in he numbe o mac ophages, he induc ion o n αand nos2b, and he down egula ion o sod2, likely leading o inc eased p oduc ion o adical ni ogen and oxygen species and enhanced in acellula killing o mycobac e ia. Based on ou esul s, i seems ha he ype o ac i a ion induced by HKLm ea men is only e ec i e when deli e ed a an ea ly enough ime poin p io o exposu e o pa hogenic mycobac e ia. The model p o ides a pla o m in which bo h inna e and adap i e mechanisms leading o s e iliza ion o mycobac e ial in ec ion can be eliably s udied. Such knowledge will con ibu e o he de elopmen o new accina ion s a egies as well as hos -di ec ed he apies aimed a p e en ion o ansmission and s e ilizing ea men o TB disease. MATERIALS AND METHODS Zeb a ish lines and housing Adul 5- o 10-mon h-old male and emale AB wild- ype zeb a ish (Danio e io) and ag1 −/− (hu1999) mu an zeb a ish ( om Zeb a ish In e na ional Resou ce Cen e , Uni e si y o O egon, OR, USA) we e used in he expe imen s. The ish we e housed in low- h ough wa e -ci cula ion sys ems wi h a 14 h/10 h ligh /da k cycle. E hics s a emen All expe imen s we e conduc ed acco ding o he Finnish Ac on Animal Expe imen a ion (62/2006) and he Ac on he P o ec ion o Animals Used o Scien i ic o Educa ional Pu poses (497/2013). ELLA (Eläinkoelau akun a; he Na ional Animal Expe imen Boa d in Finland unde he Regional S a e Adminis a i e Agency o Sou he n Finland) app o ed he Tampe e zeb a ish acili y and he animal expe imen s ca ied ou in his p ojec unde he licenses ESAVI/6407/04.10.03/2012, ESAVI/ 8245/04.10.07/2015 and ESAVI/10079/04.10.06/2015. Expe imen al M. ma inum in ec ions Mycobac e ium ma inum (ATCC 927) was i s p e-cul u ed on Middleb ook 7H10 pla es wi h OADC en ichmen (Fishe Scien i ic, NH, USA) a 29°C o 1 week. A e pla e cul u ing, M. ma inum was ans e ed in o Middleb ook 7H9 medium wi h ADC en ichmen (Fishe Scien i ic, NH, USA) wi h 0.2% Tween-80 (Sigma-Ald ich, MO, USA), cul u ed o 3-4 days, dilu ed 1:10 and cul u ed o a u he 2 days un il OD600 nm eached 0.460-0.650. Fo adul zeb a ish in ec ions, M. ma inum was i s ha es ed by cen i uging o 3 min a 10,000 gand was hen esuspended and dilu ed in s e ile 1× PBS wi h 0.3 mg/ml o Phenol Red (Sigma-Ald ich, MO, USA). A o al o 5 µl o he suspension (33±19 c u/ ish) was injec ed i.p. wi h an Omnican 100 30 G insulin needle (B aun, Melsungen, Ge many) unde 0.02% 3-aminobenzoic acid e hyl es e (pH 7.0) (Sigma- Ald ich, MO, USA) anes hesia. In ec ion doses we e e i ied by pla ing 5 µl o he injec ion suspension on a 7H10 pla e. Fo la al in ec ions, he M. ma inum pTEC15 s ain was used. This in-house-made M. ma inum wasabi- luo escen s ain was made by ans o ming a pTEC15 plasmid (Addgene plasmid #30174, deposi ed by Lali a Ramak ishnan; Takaki e al., 2013) in o he M. ma inum ATCC 927 s ain by elec opo a ion. Fo la al in ec ions, he M. ma inum pTEC15 s ain was cul u ed o 4-5 days in supplemen ed 7H9 medium wi h 75 µg/ ml hyg omycin (Me ck, Da ms ad , Ge many), dilu ed 1:10, cul u ed o 3 days un il he OD600 nm was 0.407-0.537 and ha es ed o in ec ion by cen i uga ion. Zeb a ish la al in ec ion expe imen s To s udy he e ec o HKLm p iming in zeb a ish la ae, 1 nl o HKLm (240 c u) o PBS con ol we e injec ed in o he caudal ein a 1 dp unde 9 RESEARCH ARTICLE Disease Models & Mechanisms (2018) 11, dmm031658. doi:10.1242/dmm.031658 Disease Models & Mechanisms