Lactic acid bacteria that activate immune gene expression in the Caenorhabditis elegans can antagonize the Campylobacter jejuni infection in nematodes, chickens and mice

Campylobacter jejuni is the major microbacillary pathogen responsible for human coloenteritis. Lactic acid bacteria have been shown to protect against Campylobacter infection. But LAB that showed a good ability to inhibit the growth of C. jejuni in vitro are less effective in antagonising C. jejuni in animals and animal models have the disadvantages of high cost, a long cycle, cumbersome operation and insignicant immune response indicators. Caenorhabditis elegans is increasingly used to screen probiotics for anti-pathogenic property. However, no research on the use of C. elegans to screen for probiotic candidates antagonistic to C. jejuni has been conducted to date. the group infected with C. jejuni alone. These data indicated that the ability of LAB to protect nematodes against C. jejuni-induced death was correlated with their inuence on the levels of transcription of immune genes. prolonged the life-span of C. elegans decreased the C. jejuni load in mice with T. gondii


Background
Infection by pathogenic Campylobacter may result in symptoms such as bloody diarrhoea, abdominal pain and fever. In many countries, Campylobacter species is the major microbacillary pathogen responsible for human coloenteritis [1]. In under developed countries, diarrhoeic disease is 10 times more likely to result from Campylobacter infection than from infection with Escherichia coli O157: H, Shigella species or Salmonella species [2,3]. Some peripheral neuropathies, such as Guillain-Barré syndrome and Miller Fieher syndrome, are long-term consequences of Campylobacter infection [2]. Campylobacter jejuni (C. jejuni) is also one of the primary reasons for microbacillary food-borne disease in some developed countries [3]. To date, all therapies for Campylobacter infection involved antibiotics, especially in poultry industry. There is an urgency to develop alternative approaches due to a gradual increase in antibiotic-resistant Campylobacter [4]. It is also necessary to illustrate the mechanisms underlying the functions of such alternatives to support their development and implementation.
Research has increasingly shown that lactic acid bacteria (LAB) colonise the human gastrointestinal tract and play a vital role in maintaining intestinal function and well being of the host. Among the bene cial effects of LAB are bacteriostatic activities targeting pathogens such as Escherichia coli (E. coli), Salmonella and Listeria monocytogenes [5,6]. LAB have been shown to protect against Campylobacter infection. Nishiyama reported that LAB inhibited Campylobacter and colonisation of this pathogen was reduced by Lactobacillus gasseri SBT2055 isolated from healthy chicken [7]. Another study reported a decrease in C. jejuni invasion in the gut of turkey poults after treatment with Lactobacillus salivarius NRRL B-30514, suggesting that competitive exclusion can play a role [8]. Wagner et al. conducted a simulation experiment with immunode cient and immunocompetent mice, they found that bi dobacteria and lactobacilli can increase colonisation resistance. Speci cally, the results showed that bi dobacteria and lactobacilli can resist C. jejuni enteric persistence in the human gut [9]. To date, studies of antagonism against C. jejuni by LAB have mainly been conducted in vitro or in vivo. Unfortunately, some of the strains of LAB that showed a good ability to inhibit the growth of C. jejuni in vitro are less effective in antagonising C. jejuni in animals. Whether poultry or mice, all animal models have the disadvantages of high cost, a long cycle, cumbersome operation and insigni cant immune response indicators, and do not permit rapid large-scale screening for LAB capable of effective antagonising C. jejuni.
The small, free-living (non-parasitic) soil worm Caenorhabditis elegans (C. elegans) has been widely used in biological studies as a model in vivo, due to its short generation time, diminutive form and clear genetic background. To date, C. elegans has been used to study a series of pathogenic microorganisms, such as Salmonella enterica, Pseudomonas aeruginosa, Enterococcus faecalis and Staphylococcus aureus [10][11][12][13][14]. C. elegans is also increasingly used to screen for antimicrobials and probiotics [15]. In addition, institutional studies of the effects of LAB can be processed using C. elegans. However, no research on the use of C. elegans to screen for probiotic candidates antagonistic to C. jejuni has been conducted to date. Moreover, the molecular mechanisms underlying the protective mechanism of LAB have yet to be established. In this study, a C. elegans life-span experimental model was developed to investigate the response of nematodes to C. jejuni infection, enabling rapid evaluation of the protective effects of LAB and understanding of microbe-host interactions. A Toxoplasma gondii (T. gondii) induced acute ileitis model abrogating the colonisation resistance in mice and chicken to C. jejuni infection was also used to validate the C. elegans model [16].

C. jejuni intake shortened life-span of C. elegans
The harm of foodborne pathogens to C. elegans was usually re ected in the life-span of the nematodes. A C. elegans life-span experimental model was used to measure the responses of the worm to C. jejuni infection ( Fig. 2A). E. coli OP50 was used as the food that normally sustains nematodes on reaching the L4 stage. As shown in Fig. 2B, C. jejuni killed about 50% of the worms within 5 days of their transferral to the lawn of the pathogen (C/day0), whereas 80% of the worms fed OP50 were still alive after 8 days (E/day0). All the nematodes in the E/day0 group died within 21 days; the equivalent period in the C/day0 group was 13 days. When the L4 stage nematodes were fed OP50 for 3 days before C. jejuni treatment (E/day0 + C/day3), they proved more resistant to C. jejuni: nearly 50% of these worms were still alive on day 10. However, their life-span was shorter than that of the worms fed OP50 only (E/day0). In addition, the initial number of C. jejuni cells recovered from the worms in the E/day0 + C/day3 group was smaller than that of the C. jejuni cells recovered from the worms in the C/day0 group (Fig. S1). This indicated that C. elegans at day 0 of the L4 stage was most sensitive to C. jejuni infection. Therefore, worms at day 3 of the L4 stage were chosen for the life-span assay due to the moderate life-span of the worm, pathogenicity of C. jejuni at this time point and available time space for LAB intervention.
The body size of C. elegans directly re ects the growth and development of nematodes which is closely related to their energy intake. In addition to pathogenicity, pathogens may also affect the life-span of nematodes due to their inability to be metobolized by C. elegans. To determine whether worms' life-span shorten by C. jejuni was due to caloric intake or pathogenicity of C. jejuni, the body size of the nematodes fed C. jejuni was compared with that of the worms in the E/day0 group (Fig. 2C). The (E + C)/day0 group nematodes were fed equal amounts of E. coli OP50 and C. jejuni concurrently from day 0 of L4 stage. For 8 days, the nematodes in the (E + C)/day0 group and the E/day0 + C/day3 group were nearly the same size as those in the E/day0 group, which indicated that the three groups had almost the same caloric intake during this period. Furthermore, although the intestinal load of C. jejuni in the nematodes in the E/day0 + C/day3 group was much larger than the load in the (E + C)/day0 group, the load of C. jejuni in each of the two groups remained stable (Fig. 2D), which indicated that the nematodes showed no preference for ingesting C. jejuni versus E. coli. Therefore, substituting C. jejuni for E. coli did not lead to fasting, which may have resulted in worm death.
Some LAB strains prolonged the life-span of C. elegans treated with C. jejuni A prolonged nematodes' life-span after infection is a common indicator of antibacterial ability in C. elegans. To enable rapid evaluation of the defensive effects of LAB, forty-four LAB were assessed on their ability to protect C. elegans from C. jejuni infection mediated death. On day 13, C. elegans fed only E. coli OP50 displayed 50% survival (LD50) whereas the survival of nematodes fed only C. jejuni was only 20%. As shown in Fig. 3, S2 and Table 1, S2, the LAB isolates varied in their ability to protect the live worms, with survival rates ranging from 15-47%. Among tested isolates, Z5, 13M2, N9, L103, G20, 1132 and 13 − 7 provided high levels of protection (each producing an approximately 40% worm survival rate), whilst 422, B and Z6 did not show signi cant protection.
Some LAB strains decreased the C. jejuni load in the intestine of C. elegans Diminishing pathogenic bacterial colonisation in the intestinal tract is probably one of underlying mechanisms of LAB in prolonging nematodes' life-span. Also, the live LAB cells colonized in the C. elegans intestine played a role in decreasing the number of pathogenic bacteria. To investigate whether association exist between the life-span of C. elegans and the load of different bacteria in nematodes' intestine, the number of LAB, C. jejuni and E. coli OP50 in intestine were counted. Twenty-six LAB strains which showed various levels of protection were evaluated on their ability to persist in the worm intestine from day 2 to day 6 ( Fig. 4 and S3). The loads of LAB strains such as Z5, 427, N34 and X13, which showed a strong ability to protect C. elegans against death (37.52%-40.36% survival rate) exceeded log10 5 CFU/worm during the assay. In contrast, the loads of B, G14, ZX7 and Z7 were found to be low (log10 4 CFU/worm), also showed low levels of nematode protection in the life-span assay. The C. jejuni loads in these worms from day 2 to day 6 were also checked. Similarly, different LAB strains were associated with different levels of C. jejuni loads (log10 3.5-4.5 CFU/worm) in the nematode intestine during the assay ( Fig. 5 and S4). The C. jejuni load in C. elegans treated with Z5, X13, 13 − 7, 427, G20 and N9 (offering high levels of protection in terms of worm life-span, 38.02%-43.28% survival rate) was almost 1.5 orders of magnitude smaller than that in the control group (E/day0 + C/day3, log10 4.5 ), whereas B, Z6, LGG and 422 (offering low levels of protection in terms of worm life-span) did not signi cantly decrease the C. jejuni load. The results of correlation analysis showed that the LAB load was highly positively correlated and conversely the C. jejuni load was highly negatively correlated with the survival of C. elegans. In addition, the load of LAB was moderately correlated with the load of C. jejuni in C. elegans (Table 2). It is worth noting that, there were no obvious differences in OP50 load among the E/day0 + C/day3 group and other LAB intervention groups (log10 1.2-1.5 CFU/worm) on day 6 ( Fig. S5).
Longevity effects of LAB on C. elegans treated with C.jejuni were not due to caloric reduction The body size of C. elegans directly re ects the growth and development of nematodes, which is closely related to their energy intake. Except the body size, pharyngeal pumping represents the feeding capacity of C. elegans and is a key index for the physiological activities. Also, it is an important parameter to evaluate the toxicity of drugs in C. elegans.To determine whether the longevity effects of LAB were the result of caloric reduction, the body size ( 26 LAB strains and C. jejuni (L/day0 + C/day3 groups) was compared with that of nematodes in the E/day0 + C/day3 group. The worms in the E/day0 + C/day3 group and those in the LAB intervention groups showed little differences in body size, with the exceptions of the worms treated with N34, 675, 427 and 676, which were smaller. However, the N34, 675, 427 and 676 strains had varying protective effects (37.52%, 22.67%, 38.02% and 27.05% survival rate on day 13, Table S2). These were not the top survival rates in the life-span assay. The same situation also appeared in the determination of nematodes' pharynx pumping. The pharynx pumping of worms was at the range of 50 to 58 per 30 s in the E/day0 + C/day3 group and most LAB intervention groups except for N34, 675, 427 and 676 treated groups. There were no signi cant differences in body size and number of pharynx pumping of nematodes treated by these LAB which showed varied longevity effects on C. elegans treated with C. jejuni.
In uence on C. jejuni growth by co-culturing E. coli OP50 To investigate whether the substance produced by E. coli OP50 such as certain bacteriocins would kill C. jejuni, the viability of C. jejuni cultured with or without live E. coli OP50 was measured. As showed in Fig. S8, the number of C. jejuni elevated in both tests after a 24-h incubation did not show signi cant difference. Furthermore, the pathogen was unin uenced after the live E. coli OP50 were added to the growing C. jejuni after 48 h of incubation.
Some LAB strains upregulated immune gene transcription to protect C. elegans against C. jejuni infection Regulation on the host's immune system through pivotal signalling pathways is also an underlying mechanism by LAB on prolonging nematodes' life-span against pathogen infection. To determine the mechanism by which LAB protected C. elegans against C. jejuni infection, the transcription levels of the 14 immune genes (tir-1, nsy-1, sek-1, pmk-1, spp-1, clec-85, abf-2, clec-60, lys-7, daf-16, age-1, dbl-1, skn-1 and bar-1) of C. elegans on day 6 were compared between the E/day0, E/day0 + C/day3 and L/day0 + C/day3 groups. As shown in Fig. 7 and S9, the transcription of tir-1, pmk-1 and bar-1 (respectively MAPK signalling pathway genes and an antioxidant gene) was to some extent increased when nematodes were infected with C. jejuni. In addition, slight increases were observed in some of the defence immune genes of C. elegans infected with C. jejuni, such as daf-16 and age-1 (Daf-16 signalling pathway genes) and dbl-1 (a TGF-β signalling pathway gene).
The 14 immune genes of C. elegans were also examined after treatment with 11 LAB strains, which showed variation in the protection against C. jejuni-induced worm death (Fig. 7, S9 and Table S3). It was found that the immune genes of C. elegans did not change signi cantly after LAB intervention without C. jejuni infection, which indicated that LAB strains were safe to healthy host (Table S3). As shown in Fig. 7 and S9, when nematodes were treated with LAB offering low levels of protection for survival, the transcription levels of their defence genes were almost identical to those of nematodes infected by C. jejuni alone. For these LAB strains, such as PC-T7 and B, only some genes (such as tir-1 and skn-1) were transcribed at a slightly higher rate than their counterparts in the group infected with C. jejuni alone, and the expression of spp-1 and bar-1 was even inferior to that in the control group. The transcription levels of some genes (such as tir-1, pmk-1 and bar-1) were much lower in some of the groups treated with poorly protective LAB (422, G14) than in the group infected with C. jejuni only. On the contrary, for those LAB strains, which protected the nematodes against C. jejuni-induced death in the life-span assay, the transcription levels of genes assayed were increased considerably on day 6. For example, treatment with 13 − 7, N9 and Z5 signi cantly enhanced the transcription of the MAPK signalling pathway genes nsy-1, sek-1 and pmk-1, the antioxidant gene skn-1 and the Daf-16 signal pathway genes age-1 and daf-16. The levels of transcription of the antibacterial peptide genes spp-1, clec-85 and lys-7 in these LAB intervention groups were 3-4 times higher than those in the group infected with C. jejuni alone. These data indicated that the ability of LAB to protect nematodes against C. jejuni-induced death was correlated with their in uence on the levels of transcription of immune genes.
LAB strains which prolonged the life-span of C. elegans decreased the C. jejuni load in mice with T. gondii induced acute ileitis LAB that showed protection of nematodes against bacterial infection might not applied to mammals due to obvious differences between the two organisms. To investigate whether the LAB screened from C. elegans model had the same effects in mammal, 11 LAB strains with different effects on C. elegans immune gene transcription were further investigated, to determine their capabilities to decrease the load of C. jejuni in mice. Seven days after T. gondii infection, the mice developed acute ileitis and likely to die, so the C. jejuni loads were checked on day 5. The load of C. jejuni in the faeces of mice in the C. jejuni infected group reached 10 8 CFU/g (Fig. 8A). LAB strains 13 − 7, Z5 and G20, which had already shown an outstanding ability to protect the nematodes, also exerted superior suppressive effect on C. jejuni load (less than 10 6 CFU/g faeces) in the mouse intestinal tract. Strains 422 and G14, which showed poor protective effects in terms of worm life-span, correspondingly played an inconspicuous role in suppressing C. jejuni load (around 10 8 CFU/g faeces). In addition, PC-T7 and Z6, which showed slight depressive effects on C. jejuni load (10 7 -10 8 CFU/g faeces) had previously offered moderate and low protection, respectively, to the worms. Meanwhile, N9 and 430, which had moderate depressive effects on C. jejuni load (10 6 -10 7 CFU/g faeces), had previously offered the worms high and moderate levels of protection, respectively. Therefore, the effects of all the 11 LAB strains apart from B and 427 were consistent across the C. elegans and mice samples.
Correlation analysis was conducted to examine the relationship between the ability of the 11 LAB strains to clear C. jejuni in mice and the survival rate of C. jejuni-infected nematodes. The relative index R 2 , which reached 0.79093, indicated that the C. jejuni-antagonistic activity of LAB in C. elegans was signi cantly corelated to their C. jejuni-antagonistic activity in mice (Fig. 8B).
LAB strains which prolonged the life-span of C. elegans decreased the C. jejuni load in chicken Outbreaks of campylobacteriosis can occur if humans ingest undercooked poultry contaminated by the C. jejuni. LAB applied in fodder could reduce Campylobacter colonization in poultry and stop the disease outbreak at its source. To investigate whether the LAB screened from C. elegans had the same effects in poultry, 11 LAB strains with different effects on C. elegans immune gene transcription were investigated to determine their abilities to clear C. jejuni in chicken. The inhibitory effect of LABs on C. jejuni colonization in chicks' cecum was examined. Approximately 24 h after hatching, chicks were inoculated orally with C. jejuni, and then LAB was administered daily for two weeks. CFU of C. jejuni in chicken cecum in all groups were evaluated on day 23. The average value of C. jejuni increased to 10 8 CFU/g cecal content in the C. jejuni infected group (Fig. 9A). Z5 and 427, exerted most signi cant suppressive effects on C. jejuni colonisation in the chicken cecum, which resulted in C. jejuni loads fell to below 10 4 CFU/g cecal content in these two groups. Correspondingly, Z5 had demonstrated an outstanding ability to protect the nematodes and decrease C. jejuni load in the mouse intestinal tract, while the antagonizing ability by 427 in mice was reversed. Meanwhile, 430, B and G14 showed poor abilities to clear C. jejuni in chicken cecum. The C. jejuni loads in the last three groups were higher than 10 7 CFU/g cecal content. The effects of these 3 strains on protecting the nematodes were similar to that of elimination of C. jejuni in chicken. However, 430 and B exhibited moderate strength of C. jejuni antagonization in mice, which was different from their performance in the chicken. In addition, 13 − 7, N9, G20, PC-T7, 422 and Z6 showed moderate scavenging activity on C. jejuni in the chicken cecal contents with the load of C. jejuni at the range from 10 4.8 to 10 5.5 CFU/g faeces. It is worth noting that a few strains (422 and Z6) showed a certain degree of inconsistency in the C. jejuni antagonism in different models..
Correlation analysis was conducted to examine the relationship between the ability of the 11 LAB strains to clear C. jejuni in chicken and the survival rate of C. jejuni-infected nematodes. The relative index R 2 , which reached 0.50071, indicated that the C. jejuni-antagonistic activity of LAB in C. elegans was related to their C. jejuni-antagonistic activity in chicken (Fig. 9B).

Discussion
The model organism C. elegans has been treated with numerous pathogenic microorganisms from animals and humans to examine the relationships between pathogenic bacteria and their hosts [10][11][12][13][14]. In this study, a C. elegans life-span experimental model was rst established to examine the response of the nematode to C. jejuni infection. Our experiments revealed that LAB strains varied in their ability to defend the nematodes against infection with C. jejuni NCTC 11168. Also, the antagonism effects of LAB isolates on C. jejuni in worms were veri ed in both mice and chickens. The worm life-span experimental model not only provided a useful way of screening for LAB candidates with the potential to mitigate C. jejuni infection, but also helped to explain the mechanism of the defence imparted by the LAB isolates. In addition, the ndings indicated that LAB defended C. elegans not only by inhibiting C. jejuni colonisation in the intestine, but also by activating the nematodes' defence immune genes.
The natural physiological activities of nematodes are usually halted by the colonisation of invading pathogens [17][18][19]. Some LAB (including L. rhamnosus [20], L. acidophilus [21], L. fermentum [13], L. gasseri and L. plantarum [22] have been reported to show high shielding e ciency to nematodes' from pathogens and to diminish the risk of bacterial colonisation in the intestinal tract and prolonging life-span. The intestinal colonisation of C. jejuni may occur in humans or animals, and is often observed in poultry. The binding of mbrial adhesins to the host's intestinal tract is a condition for C. jejuni nosogenesis, which leads to diseases such as diarrhoea [23,24]. Many investigators have showed that C. jejuni colonisation and growth in the host intestinal tract can be reduced by sectional strains of LAB [25,26]. The current article con rmed that the level of C. jejuni in the nematode intestinal tract was affected differentially by LAB strains, which offered strong to weak levels of protection. The C. jejuni load in the nematodes was highly correlated with the nematodes' lifespan, which indicated that the inhibition of C. jejuni colonisation in the intestinal tract was one of the mechanisms through which LAB offered protection to the worms. As reported, some foodborne pathogens, such as Salmonella typhimurium [21] and enteroinvasive Escherichia coli [27], colonized in the nematode intestinal tract to form a persistent lethal infection. The colonized LAB in the intestine produce antibacterial products continuously, they also occupy the adhesion sites of pathogens, which jointly exert antibacterial effects. However, in this study, LAB load and C. jejuni load in C. elegans were only moderately correlated, indicating that some LAB may alleviate the damage caused by infected pathogens in other ways, such as through worms' immune gene expression regulation.
In non-mammalian taxa, mice and some primates, calorie restriction is normally used to increase longevity and ease the consequences of aging [28][29][30]. For C. elegans, researchers investigated the caloric intake as well in C. elegans intervened by LAB through measurement of body size and pharyngeal pumping rate [10][11][12][13][14]. The growth curves obtained in this study for most of the LAB intervention groups did not support the assumption that LAB is a lower-calorie food than E. coli. The LAB-induced morphological and swallowing ability changes in the nematodes were not related to the life-span extension of C. elegans infected by C. jejuni. Therefore, the prolongation of worms' life-span could not be attributed to calorie restriction by the substitution E. coli with LAB strains.
With the growth and reproduction of LAB in a host's intestine, the host's immune system is regulated through pivotal signalling pathways with LAB involvement [31]. The host's immune response to pathogens may be induced by LAB [32]. LAB strains with protective e cacy, such as L. rhamnosus GG [33], L. plantarum [34] and L. delbrueckii [35], enhance the host immune defence against E. coli, Salmonella typhimurium and Streptococcus pyogenes infection by exciting the MAPK signalling pathway and through toll-like receptors. LAB may also defend the host by adjusting gene expression through cytokine and chemokine activity [36,37]. Another possible mechanism is the enhancement of membrane barrier [38,39]. C. elegans has two main signalling pathways, including the P38-MAPK and TGF-β signalling pathways. Its speci c defence system, including antimicrobial responses, falls short of an adaptive immune system [40]. Nevertheless, the upregulation of tir-1, nys-1, sek-1 and pmk-1, MAPK pathway genes enabling it to resist various microbial infections. The ndings demonstrated the activation of the MAPK pathway through LAB intervention [31]. In this study, the LAB strains offering good levels of protection for C. elegans against C. jejuni infection, and signi cantly enhanced the transcription of the MAPK pathway genes. Besides, as a downstream molecule regulated by the MAPK signalling pathway, the forkhead family transcription factor DAF-16 has been shown to adjust genes to improve dauer formation in the larval phase of nematodes, and to increase resistance and longevity when mature [41][42][43]. In this study, it is consistent with the above ndings, suggesting that these LAB strains increase worms' defence by upregulating DAF-16 via the MAPK signalling pathway and simultaneously prolonging the worms' life-span.
In addition, the cluster of C-type lectins has been argued to mask bacterial attachment and enhance resistance to microbial infection [44]. The increased levels of clec-60, clec-85, spp-1 and abf-2 upon LAB pre-treatment in this study also testify to the ability of some LAB to protect nematodes against C. jejuni infection. Skn-1 and bar-1 may also play critical roles in defending the host and prolonging its survival. The increased level of skn-1 in the pre-treated nematodes was due rst to antioxidant defence and second to the worms' improved survival [45][46][47]. Partly consistent with this claim, skn-1 transcription in the current study was signi cantly enhanced by LAB strains protective of nematode life-span. However, the same enhancement was also observed in worms treated with B and PC-T7. The transcription of bar-1, which re ects a response to the stress caused by C. jejuni infection, was unaffected by LAB strains with protective effects but down-regulated by some un-protective LAB strains, such as B, 422, G14 and Z6. It was con rmed that un-protective LAB strains could not increase the expression of skn-1 and bar-1 to extend the life-span of nematodes.
Although the TGF-β pathway is a major defence system in C. elegans, dbl-1 showed almost no obvious changes in transcription level across the LAB intervention groups, which indicated that this gene may not play an essential role in protecting against C. jejuni infection. These signalling pathways mentioned above which were highly conserved could be found among numerous homologus genes in mammals and played a pivotal role on growth promotion, acclimatization, resistance to pathogens and adaptation of external stress in C. elegans. Based on this, C. elegans were more widely used as the hosts of zoonotic pathogens by LAB intervention to study the change of virulence factors and antagonism to pathogenic mechanisms. But not limited to these, antibacterial substances, competition and repulsion, adhesion barrier and immunomodulatory from LAB also showed the characteristics of antagonizing pathogenic bacteria in host. It is worth noting that the immune genes of C. elegans without C. jejuni infection did not change signi cantly after LAB strains intervention. So the LAB strains did not make a signi cant effect on healthy body which means LAB is safe to healthy host. LAB only played a signi cant role when the nematodes received external infection.
As the growth temperature of C. elegans is much lower than the temperature in the human gut, and the microbiota of C. elegans is much simpler than that of a mammal, the reliability of this C. elegans life-span assay in testing the e ciency of LAB antagonism against C. jejuni colonisation in mammals should be con rmed. In addition, C. jejuni contamination is a serious problem in poultry production. The C. jejuni-infected chicken model has been widely used in screening of probiotics for feed with C. jejuni antagonism, which could be applied as antibiotic substitutes [16]. In mammal model, a few LAB strains in the current study yielded the opposite results in terms of the survival of C. elegans and the pathogen load in mice, which may be due to host species differences or to the interference of gut microbiota in mice. Nevertheless, the signi cant correlation between the ability of LAB to clear C. jejuni in mice and their ability to enhance the survival rate of C. jejuniinfected nematodes indicates that the life-span model of C. elegans infected with C. jejuni can to some extent be applied to C. jejuni-infected mammals. The short cycle of acute infection in mice model represents a research limitation, the chicken model was thus developed to investigate the long-term effects in this study. Although not every LAB strain showed the same antagonistic ability against C. jejuni in both nematodes and chickens, the signi cant correlation between the ability of LAB in clearing C. jejuni in chicken and that in elevating the survival rate of C. jejuni-infected nematodes still indicated that C. elegans is a good model organism for screening for C. jejuni-resistive LAB strains on a large scale.

Conclusion
This study established a C. elegans life-span assay capable of measuring the response of worms to C. jejuni. Different LAB had different effects on the response of C. elegans to infection with C. jejuni. The inhibition of C. jejuni intestinal colonisation may have been one of the mechanisms through which LAB protected C. elegans, and the protection offered by LAB may also have derived partly from their activation of the nematode's defence immune genes. This C. elegans life-span model can be used to screen for C. jejuni-antagonistic LAB on a large scale.
Methods C. elegans, LAB and C. jejuni C. elegans N2 Bristol wild-type strain (Caenorhabditis Genetics Center, Minnepolis, University of Minnesota) was used in the study. The C. elegans was maintained and cultivated at 20℃. S medium, M9 buffer and nematode growth medium (NGM) were used to cultivate the nematode and to conduct life-span experimental studies. The procedures for the cultivation, maintenance and synchronization of the nematode have been reported [48]. E. coli OP50 was grown at 37℃ for 24 h in Luria-Bertani medium to a bacterial concentration of 10 8 colony-forming units (CFU) per mL. It was used as food for C. elegans.
Six LAB were purchased from the American Type Culture Collection (ATCC) or the Japan Collection of Microorganisms and 38 were isolated from human faeces from different habitats and traditional fermented food. Samples of healthy human faeces and traditional fermented foods were collected and the LAB were enriched in sorbitol GM 17 medium at 35℃ for 12 h. After gradient dilution, the enriched samples were coated on GM 17 medium plate with 0.02% cresol violet and cultured for 24h. Some single bacterial colonies in line with LAB morphology were selected, and their Gram property determined by Gram staining. The selected LAB were identi ed through 16S gene sequencing. All the isolates were deposited at the Culture Collection of Food Microorganisms at Jiangnan University (Table 1 and S1). In this study, no human experiments was conducted.
The human volunteers were not expected to encounter risk or discomfort in the process of faecal sampling and written informed consent to handle faecal samples for public health purposes was obtained from the volunteers or, where relevant, their legal guardians. The whole genome sequence of 12 isolates, namely N8, N9, 422, 427, 430, Z5, L103, X13, JS-SZ-1-5, JS-WX-9-1, 9-5 and H27-1L were aligned and identi ed as new strains of the corresponding species reported. The other 26 isolates are assumed tentative strains, for they were isolated from samples of different origin, geographical location and span over seven years. All of the LAB were cultivated in deMan, Rogosa and Sharpe (MRS) agar at a 2% (v/v) inoculum size at 37°C for 18 to 20 h. The isolates were kept at -80°C in 30% glycerol for long-term storage. The LAB were sub-cultured twice before being used in experiments at 2% (v/v) inoculum size.
The C. jejuni strain NCTC 11168 was purchased from the ATCC. A C. jejuni selective supplement (Oxoid) and 5% sterile sheep blood were added to Columbia blood agar base plates (Oxoid, UK). C. jejuni strains were grown in this medium under special gas conditions (5% O 2 , 10% CO 2 , 85% N 2 ) for 48 h at 37°C.

Life-span experimental analysis of C. elegans
The nematodes were synchronised as previously described [27]. After synchronisation, the eggs were placed on E. coli OP50 NGM plates for 72 h until they reached the L4 stage. Most of the experiments lasted 25 days. All experiments were carried out at 20℃, and at least three independent replications were performed for each assay.
To build a model for the experimental analysis of worm death induced by C. jejuni, 10 9 CFU/mL of C. jejuni was prepared from the L4 stage worms at 25 days incubation. Each group had 80-100 worms. As shown in Fig. 1A, worms fed only with E. coli OP50 formed the negative control group (E/day0 group). In an experiment undertaken to evaluate the role of LAB in protecting the worms against death caused by C. jejuni, the worms were either fed 10 8 CFU/mL E. coli OP50 or 10 9 CFU/mL LAB strains for the rst 3 days. After 72 h of incubation, the worms were moved to new 6 cm plates with C. jejuni at a concentration of 10 9 CFU/mL. The worms that were rst fed E. coli OP50 (72 h) and next C. jejuni formed the C. jejuni reference group (E/day0+C/day3 group); those treated with LAB (72 h) before C. jejuni were regarded as the LAB protection groups (L/day0+C/day3 groups). A worm was considered dead when it failed to respond to gentle touch with a worm picker. The numbers of live worms were recorded, and the probability of their survival was calculated as described previously [27].

Examination ofbacterial load in the intestine of C. elegans
The numbers of E. coli OP50, Lactobacillus and C. jejuni in the nematodes' intestine were determined with some modi cation of the method described previously [27]. Worms were incubated with E. coli OP50, a Lactobacillus, or C. jejuni, and sampling (50 worms per sample) was done every 2 days. After surface sterilization, the worms were mashed mechanically with a pellet pestle motor, re-suspended in the M9 buffer, and inoculated onto eosin-methylene blue medium (EMB medium), MRS or Columbia blood agar for counting of E. coli OP50, Lactobacillus and C. jejuni, respectively. At least three independent replications were performed for each assay.

Measurement of body size and pharynx pumping
The live worms were examined for their body size measurements every 2 days and their pharynx pumping on the third day until they were infected with C. jejuni. Images of adult nematodes were taken with a VCT-VBIT digital microscope (Shimadzu, Kyoto, Japan) and analyzed using the ImageJ software. In this system, the area of the worm's projection was estimated automatically and used as an index of body size. And the worm's pharynx pumping was measured per 30s.
Evaluation of the effects on the growth of C. jejuni by co-culturing E. coli OP50 The growth of C. jejuni co-cultured with E. coli OP50 was determined by the following method as previously described. The C. jejuni cells (10 7 CFU/mL) suspended in antibiotic-free brain heart infusion broth (BHIB) containing 5% serum were incubated under microaerophilic conditions for 48 h at 37°C in the presence of a 10% volume of live E. coli OP50 (10 7 CFU/mL). The viability of C. jejuni was evaluated from the number of viable CFUs in C.jejuni culture as described above on C. jejuni-selective plates. At least three independent replications were performed for this assay.

RNA extraction, reverse transcription and quantitative real-time PCR analysis
The whole RNA of C. elegans and of the bacteria used in the life-span experimental was extracted. About 100 worms were prepared for the lysates. The RNA was extracted as previously described [27].
The transcription of mitogen-activated protein kinase (MAPK) pathway genes (tir-1, nsy-1, sek-1 and pmk-1), antimicrobial peptide genes (spp-1, clec-85, abf-2, clec-60 and lys-7), Daf-16 pathway genes (daf-16 and age-1), a TGF-β pathway gene (dbl-1) and antioxidant genes (skn-1 and bar-1) in C. elegans was determined by quantitative polymerase chain reaction (qPCR) [49]. GapA, as a housekeeping gene, was used to determine the levels of mRNA transcription of the C. elegans immune genes and to normalize the input amounts of RNA. PCR primers speci c to each of the genes were experimentally validated and used in the RT-qPCR assay. The annealing temperature of the RT-qPCR assay was 56°C. The delta Ct method was used to analyse the RT-qPCR data and to determine the relative abundance of the target genes (fold changes) (fold changes) [50,51].

Induction of acute iletis for C. jejuni infection and LAB intervention in mice
Three-week-old female C57BL/6 mice obtained from Shanghai Laboratory Animal Center (Shanghai, China) were used in the experiments. Eight mice were randomly housed in each cage, with a 12-h light-dark cycle and a controllable environment (humidity, 45% ± 5%; temperature, 22°C ± 2°C). All the experimental procedures To induce ileitis, the C57BL/6 mice were infected orally with 100 T. gondii cysts (ME49 strain; obtained from the Jiangsu Institute of Parasitic Diseases, Remington, Wuxi, China) in 0.3 mL phosphate-buffered saline (PBS, pH 7.4) by gavage, as described previously [52]. As shown in Fig. 1B, for the 4 days thereafter, the mice were successively (at 1 h intervals) treated with LAB (10 9 CFU in 0.3 mL or 0.3 mL PBS) and C. jejuni NCTC 11168 (10 9 CFU in 0.3 mL) by gavage on 2 consecutive days. The C. jejuni loads in their faeces were checked 3 days later. The faeces were resuspended in sterile PBS and serially diluted. The diluted samples were spread on Columbia blood agar with a C. jejuni selective supplement and incubated at 5% oxygen concentration at 37°C for 48 h. After incubation, the numbers of C. jejuni in the samples were counted. The mice were euthanized with CO 2 after experiment. For each mice, the treatment based on the different LAB, the blood collection and execution and analysis of indicators were all done by different investigator. The rst investigator was the only person aware of the treatment group allocation.

C. jejuni infection and LAB intervention in chicken
White leghorn chicken eggs (Jinan Baizhun Biologic Inspection Company, Ltd., China) were maintained in an egg incubator until the chicks hatched. About 24 hours after hatching, 8 chicks were randomly assigned to several groups. All the experimental procedures were approved by the Animal Care and Use Committee at Jiangsu Nannong Hi-technology company, LTD. All the experiments adhered to the Ministry of Science and Technology of China's Guide for the Care and Use of Laboratory Animals.
As shown in Fig. 1C, bacterial cells were washed and resuspended in ice-cold PBS prior to inoculation. All birds were administered 10 8 CFU of C. jejuni NCTC 11168 in a 0.3 mL suspension by oral gavage. Twenty-four hours after oral gavage, LAB (10 8 CFU in 0.3 mL) were orally administered daily to 11 groups (in total 88 birds) of C. jejuni-inoculated birds for two weeks. PBS was administered to the remaining one C. jejuni group of birds (8 birds). Chicks were euthanized with CO 2 at 23 days post-inoculation, and the cecal contents were diluted in icecold PBS to 0.1 g/mL. Ten-fold serial dilutions of each sample were prepared and then plated on Columbia blood agar with a C. jejuni selective supplement and incubated at a 5% oxygen concentration at 37°C for 48 h. For each mice, the treatment based on the different LAB, the blood collection and execution and analysis of indicators were all done by different investigator. The rst investigator was the only person aware of the treatment group allocation.

Statistical analysis
GraphPad Prism 5.0 and Origin 9.0 were used to perform the statistical analysis, and SPSS Statistics 20.0 was used for the signi cance analysis. The data were expressed as means ± standard deviations (SDs). Kaplan-Meier survival analysis was used to assess the survival rate of C. elegans. Correlation analysis of the C. elegans groups was conducted using SPSS Statistics 20.0 and Origin 9.0. The groups were compared using a two-tailed Student's t-test, and a two-sided p value of less than 0.05 was considered statistical signi cance.
Mean values with different superscript letters over the bars are signi cantly different (p < 0.05).

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analysed during this study are included in this published article [and its supplementary information les].