Release of HMGB1 and Toll-like Receptors 2, 4, and 9 Signaling Are Modulated by Bifidobacterium animalis subsp. lactis BB-12 and Salmonella Typhimurium in a Gnotobiotic Piglet Model of Preterm Infants

Gnotobiotic (GN) animals with defined microbiota allow us to study host–microbiota and microbiota–microbiota interferences. Preterm germ-free (GF) piglets were mono-associated with probiotic Bifidobacterium animalis subsp. lactis BB-12 (BB12) to ameliorate/prevent the consequences of infection with the Salmonella Typhimurium strain LT2 (LT2). Goblet cell density; expression of Toll-like receptors (TLRs) 2, 4, and 9; high mobility group box 1 (HMGB1); interleukin (IL)-6; and IL-12/23p40 were analyzed to evaluate the possible modulatory effect of BB12. BB12 prevented an LT2-induced decrease of goblet cell density in the colon. TLRs signaling modified by LT2 was not influenced by the previous association with BB12. The expression of HMGB1, IL-6, and IL12/23p40 in the jejunum, ileum, and colon and their levels in plasma were all decreased by BB12, but these changes were not statistically significant. In the colon, differences in HMGB1 distribution between the GF and LT2 piglet groups were observed. In conclusion, the mono-association of GF piglets with BB12 prior to LT2 infection partially ameliorated the inflammatory response to LT2 infection.


Introduction
Preterm birth (PTB) is a birth that occurs before 37 weeks of gestation, and its incidence affects about 11% of pregnancies [1]. Various reasons can trigger PTB, and the inflammatory process is one of them [2]. In addition, preterm infants have a low birth weight and underdeveloped organ systems, making them more susceptible to many lifethreatening comorbidities [2]. These factors and their possible concurrence result in special requirements for preterm infants that need supportive care in the neonatal intensive care unit (NICU) [3]. However, despite this particular regimen, the preterm infant suffers from increased morbidity that is inversely proportional to the length of gestation [4].
The initial colonization of vaginally born infants occurs immediately during childbirth by the mother's vaginal and fecal microbiota. This settlement starts with pioneer settlers that create suitable conditions for their followers [5]. However, colonization occurs more rapidly than was believed [6], because strictly anaerobic bifidobacteria are present in the feces of some infants on the first day after birth [7]. The establishment of balanced microbiota (eubiosis) in newborns is negatively influenced by antibiotic treatment of the

Clinical Signs of Enterocolitis
The non-infected piglets (GF and BB12 groups) did not show any signs of enterocolitis. In contrast, the Salmonella-infected piglets (LT2) were sleepy and had anorexia with nonbloody diarrhea and fever. The piglets associated with BB12 and one-week later infected with LT2 (BB12 + LT2) showed milder diarrhea than the piglets infected with LT2 only.

Goblet Cells in the Ileum
The GF ( Figure 1A) and BB12 ( Figure 1C) piglets showed long villi with many vacuolated enterocytes and mucin-producing, blue-stained goblet cells among enterocytes. Piglets infected with S. Typhimurium ( Figure 1B) had shortened and damaged villi with desquamated epithelial cells in the lumen. Colonization with BB12 did not fully protect ileal villus morphology against injury induced by S. Typhimurium infection ( Figure 1D). The number of goblet cells was reduced in the LT2 group, but this decrease was not statistically significant ( Figure 1E).

Clinical Signs of Enterocolitis
The non-infected piglets (GF and BB12 groups) did not show any signs of enterocolitis. In contrast, the Salmonella-infected piglets (LT2) were sleepy and had anorexia with non-bloody diarrhea and fever. The piglets associated with BB12 and one-week later infected with LT2 (BB12 + LT2) showed milder diarrhea than the piglets infected with LT2 only.

Goblet Cells in the Ileum
The GF ( Figure 1A) and BB12 ( Figure 1C) piglets showed long villi with many vacuolated enterocytes and mucin-producing, blue-stained goblet cells among enterocytes. Piglets infected with S. Typhimurium ( Figure 1B) had shortened and damaged villi with desquamated epithelial cells in the lumen. Colonization with BB12 did not fully protect ileal villus morphology against injury induced by S. Typhimurium infection ( Figure 1D). The number of goblet cells was reduced in the LT2 group, but this decrease was not statistically significant ( Figure 1E). 2 in the ileum of the one-week-old piglets: germ-free (GF; (A)), infected with S. Typhimurium LT2 for 24 h (LT2; (B)), associated with B. animalis subsp. lactis BB-12 (BB12; (C)), and associated with BB12 and infected with LT2 for 24 h (BB12 + LT2; (D)). Six samples from each group were analyzed, and statistical differences were calculated by a two-way ANOVA with Tukey's multiple comparison post hoc test. The values are presented as mean + SEM, and p < 0.05 among groups are denoted by different letters above the columns (E). A scale bar (D) depicts 200 μm.

Goblet Cells in the Colon
S. Typhimurium infection significantly reduced the number of goblet cells in the colon in the LT2 groups ( Figure 2B,E) compared to the other groups ( Figure 2A,C-E). Prior association with BB12 (BB12 + LT2; Figure 2C) prevented a decrease in the goblet cell counts, as is comparable to the GF ( 2 in the ileum of the one-week-old piglets: germ-free (GF; (A)), infected with S. Typhimurium LT2 for 24 h (LT2; (B)), associated with B. animalis subsp. lactis BB-12 (BB12; (C)), and associated with BB12 and infected with LT2 for 24 h (BB12 + LT2; (D)). Six samples from each group were analyzed, and statistical differences were calculated by a two-way ANOVA with Tukey's multiple comparison post hoc test. The values are presented as mean + SEM, and p < 0.05 among groups are denoted by different letters above the columns (E). A scale bar (D) depicts 200 µm.    BB12 did not increase the expression of TLR2 mRNA in the ileum compared to the GF piglet control ( Figure 3A). In contrast, both groups infected with S. Typhimurium (LT2 and BB12 + LT2) had significantly higher expression. The presence of BB12 (BB12 + LT2) did not significantly influence the TLR2 expression compared to S. Typhimurium infection alone (LT2). Similar changes of mRNA expression were found in TLR4 ( Figure 3B). An opposite trend was found in TLR9 mRNA expression ( Figure 3C). TLR9 mRNA expression was comparable in GF and BB12 groups but was downregulated by Salmonella infection. The prior association with BB12 (BB12 + LT2) did not ameliorate the LT2-inducted downregulation of TLR-9 expression. BB12 did not influence MyD88 mRNA expression, but it was significantly upregulated by Salmonella ( Figure 3D). The presence of BB12 enhanced this expression, so the expression in BB12 + LT2 group was significantly higher than in the LT2 group. In contrast, Salmonella downregulated the expression of TRIF mRNA ( Figure 3E). Salmonella infection downregulated RAGE mRNA expression compared to GF, but this downregulation was only significant for LT2, and not for BB12 + LT2 ( Figure 3F).  , and associated with BB12 and infected with LT2 for 24 h (BB12 + LT2). Six samples from each group were analyzed, and statistical differences were calculated by a two-way ANOVA with Tukey's multiple comparison post hoc test. The values are presented as mean + SEM, and a p < 0.05 among groups is denoted by different letters above the columns.

TLR2, TLR4, TLR9, MyD88, TRIF, and RAGE mRNA in Colon
BB12 downregulated TLR2 mRNA expression in the colon, but this downregulation in comparison to GF and LT2 groups was statistically non-significant ( Figure 4A). However, a previous association with BB12 (BB12 + LT2) caused significant upregulation in comparison to the BB12 group alone. Simultaneously, this upregulation was not statistically significant in comparison to the GF and LT2 groups. Salmonella significantly upregulated TLR4 mRNA expression in both infected groups (LT2 and BB12 + LT2) ( Figure 4B). TLR9 mRNA expression was comparable among all groups ( Figure 4C). MyD88 mRNA was significantly upregulated in the BB12 + LT2 group only ( Figure 4D). TRIF mRNA was , and RAGE (F) mRNA in the ileum of the one-week-old gnotobiotic piglets: germ-free (GF), infected with S. Typhimurium LT2 for 24 h (LT2), associated with B. animalis subsp. lactis BB-12 (BB12), and associated with BB12 and infected with LT2 for 24 h (BB12 + LT2). Six samples from each group were analyzed, and statistical differences were calculated by a two-way ANOVA with Tukey's multiple comparison post hoc test. The values are presented as mean + SEM, and a p < 0.05 among groups is denoted by different letters above the columns.
2.5. TLR2, TLR4, TLR9, MyD88, TRIF, and RAGE mRNA in Colon BB12 downregulated TLR2 mRNA expression in the colon, but this downregulation in comparison to GF and LT2 groups was statistically non-significant ( Figure 4A). However, a previous association with BB12 (BB12 + LT2) caused significant upregulation in comparison to the BB12 group alone. Simultaneously, this upregulation was not statistically significant in comparison to the GF and LT2 groups. Salmonella significantly upregulated TLR4 mRNA expression in both infected groups (LT2 and BB12 + LT2) ( Figure 4B). TLR9 mRNA expression was comparable among all groups ( Figure 4C). MyD88 mRNA was significantly upregulated in the BB12 + LT2 group only ( Figure 4D). TRIF mRNA was significantly downregulated in both Salmonella-infected groups compared to both non-infected groups ( Figure 4E). RAGE mRNA expression was downregulated in the Salmonella-infected groups (LT2 and BB12 + LT2) ( Figure 4F).  , and associated with BB12 and infected with LT2 for 24 h (BB12 + LT2). Six samples from each group were analyzed, and statistical differences were calculated by a two-way ANOVA with Tukey's multiple comparison post hoc test. The values are presented as mean + SEM, and a p < 0.05 among groups is denoted by different letters above the columns.

TLR2, TLR4, TLR9, MyD88, TRIF, and RAGE mRNA in Mesenteric Lymph Nodes
Salmonella significantly upregulated TLR2 mRNA expression in MLN ( Figure 5A). The same trend was observed for TLR4 mRNA, but only induction in the LT2-infected groups (LT2 and BB12) was significant against the GF group only ( Figure 5B). The opposite trend was found for TLR9, but the suppression by Salmonella was not statistically significant ( Figure 5C). MyD88 mRNA was significantly induced by Salmonella in both infected groups ( Figure 5D), but in the case of TRIF mRNA, this significant upregulation was observed in the LT2 group only ( Figure 5E). Finally, no effect of LT2 or BB12 on RAGE mRNA expression in MLN was found ( Figure 5F). , and RAGE (F) mRNA in the colon of the one-week-old gnotobiotic piglets: germ-free (GF), infected with S. Typhimurium LT2 for 24 h (LT2), associated with B. animalis subsp. lactis BB-12 (BB12), and associated with BB12 and infected with LT2 for 24 h (BB12 + LT2). Six samples from each group were analyzed, and statistical differences were calculated by a two-way ANOVA with Tukey's multiple comparison post hoc test. The values are presented as mean + SEM, and a p < 0.05 among groups is denoted by different letters above the columns.
2.6. TLR2, TLR4, TLR9, MyD88, TRIF, and RAGE mRNA in Mesenteric Lymph Nodes Salmonella significantly upregulated TLR2 mRNA expression in MLN ( Figure 5A). The same trend was observed for TLR4 mRNA, but only induction in the LT2-infected groups (LT2 and BB12) was significant against the GF group only ( Figure 5B). The opposite trend was found for TLR9, but the suppression by Salmonella was not statistically significant ( Figure 5C). MyD88 mRNA was significantly induced by Salmonella in both infected groups ( Figure 5D), but in the case of TRIF mRNA, this significant upregulation was observed in the LT2 group only ( Figure 5E). Finally, no effect of LT2 or BB12 on RAGE mRNA expression in MLN was found ( Figure 5F). , and associated with BB12 and infected with LT2 for 24 h (BB12 + LT2). Six samples from each group were analyzed, and statistical differences were calculated by a two-way ANOVA with Tukey's multiple comparison post hoc test. The values are presented as mean + SEM, and a p < 0.05 among groups is denoted by different letters above the columns.

HMGB1 Expression in the Colon
In the colon of GF piglets ( Figure 6A), the nuclear protein HMGB1 was localized in both the nucleus and the cytoplasm, whereas in LT2 piglets ( Figure 6B), HMGB1 was localized mainly in the cytoplasm.

Intestinal Levels of HMGB1, IL-6, and IL-12/23p40
HMGB1 release was significantly induced in the jejunum with Salmonella infection ( Figure 7A) compared to the GF group. The association with BB12 did not induce HMGB1 release in the jejunum. The previous association of the piglets with BB12 in the (BB12 + , and associated with BB12 and infected with LT2 for 24 h (BB12 + LT2). Six samples from each group were analyzed, and statistical differences were calculated by a two-way ANOVA with Tukey's multiple comparison post hoc test. The values are presented as mean + SEM, and a p < 0.05 among groups is denoted by different letters above the columns.

HMGB1 Expression in the Colon
In the colon of GF piglets ( Figure 6A), the nuclear protein HMGB1 was localized in both the nucleus and the cytoplasm, whereas in LT2 piglets ( Figure 6B), HMGB1 was localized mainly in the cytoplasm.

HMGB1 Expression in the Colon
In the colon of GF piglets ( Figure 6A), the nuclear protein HMGB1 was localized in both the nucleus and the cytoplasm, whereas in LT2 piglets ( Figure 6B), HMGB1 was localized mainly in the cytoplasm.

Intestinal Levels of HMGB1, IL-6, and IL-12/23p40
HMGB1 release was significantly induced in the jejunum with Salmonella infection ( Figure 7A) compared to the GF group. The association with BB12 did not induce HMGB1 release in the jejunum. The previous association of the piglets with BB12 in the (BB12 +  HMGB1 release was significantly induced in the jejunum with Salmonella infection ( Figure 7A) compared to the GF group. The association with BB12 did not induce HMGB1 release in the jejunum. The previous association of the piglets with BB12 in the (BB12 + LT2) group prevented the significant induction of HMGB1 ( Figure 7A). However, differences between LT2 and BB12 + LT2 were non-significant. The IL-6 levels in the Salmonellainfected piglets were significantly increased ( Figure 7B). IL-12/23p40 was also induced by the infection ( Figure 7C). The previous association of piglets with BB12 significantly suppressed this increase.
As was observed in the jejunum and ileum, HMGB1 BB12 suppressed HMGB1 concentrations in the colon ( Figure 7G), but the difference was not statistically significant. IL-6 levels were significantly induced by the infection with Salmonella ( Figure 7H). The previously applied BB12 non-significantly suppressed IL-12/23p40 levels in the colon ( Figure  7I).
All biomarkers showed low levels in the jejunum ( Figure 7A-C). They were highly increased in the ileum ( Figure 7D-F) and colon ( Figure 7G-I) of both LT2-infected groups, and these values were comparable in both organs. The same ratio of the y-axis in individual biomarkers was used to clearly depict this trend throughout the intestine. In the ileum, HMGB1 levels showed a similar trend as in the jejunum; that is, they were significantly induced in the LT2 group and previously associated with BB12 nonsignificantly reduced this increase ( Figure 7D). The IL-6 levels in the ileum were upregulated by the Salmonella infection, but suppression by the previous association with BB12 resulted in a non-significant increase against the control GF group (Figure 7E). At the same time, the suppression was not statistically significant compared to that of the LT2 group. A similar trend of induction/suppression in the ileum was also found in IL-12/23p40 group ( Figure 7F).
As was observed in the jejunum and ileum, HMGB1 BB12 suppressed HMGB1 concentrations in the colon ( Figure 7G), but the difference was not statistically significant. IL-6 levels were significantly induced by the infection with Salmonella ( Figure 7H). The previously applied BB12 non-significantly suppressed IL-12/23p40 levels in the colon ( Figure 7I).
All biomarkers showed low levels in the jejunum ( Figure 7A-C). They were highly increased in the ileum ( Figure 7D-F) and colon ( Figure 7G-I) of both LT2-infected groups, and these values were comparable in both organs. The same ratio of the y-axis in individual biomarkers was used to clearly depict this trend throughout the intestine.
2.9. HMGB, IL-6, and IL-12/23p40 in Plasma The levels of plasmatic HMGB1 ( Figure 8A), IL-6 ( Figure 8B), and IL-12/23p40 ( Figure 8C) reflect the situation in the intestine. It means that Salmonella induced these levels, and the previous association with BB12 suppressed them, but this suppression was not statistically significant.

HMGB, IL-6, and IL-12/23p40 in Plasma
The levels of plasmatic HMGB1 ( Figure 8A), IL-6 ( Figure 8B), and IL-12/23p40 ( Figure  8C) reflect the situation in the intestine. It means that Salmonella induced these levels, and the previous association with BB12 suppressed them, but this suppression was not statistically significant.

Animal Models of Immunocompromised Host and Probiotics
Most reports refer to probiotics as beneficial and safe for preterm infants [23]. However, it is necessary to consider that preterm neonates have underdeveloped immune systems and compromised intestinal barrier integrity [30]. It simplifies bacterial translocation, and these infants are at increased risk of probiotic-caused sepsis [31]. The experimental work with human volunteers is limited [32], and neither two-dimensional (2D) nor three-dimensional (3D; organoids) cell culture systems [33] sufficiently simulate a complex reaction of the whole organism. Thus, suitable translational animal models play a vital role in human disease research [34][35][36][37].

Animal Models of Immunocompromised Host and Probiotics
Most reports refer to probiotics as beneficial and safe for preterm infants [23]. However, it is necessary to consider that preterm neonates have underdeveloped immune systems and compromised intestinal barrier integrity [30]. It simplifies bacterial translocation, and these infants are at increased risk of probiotic-caused sepsis [31]. The experimental work with human volunteers is limited [32], and neither two-dimensional (2D) nor three-dimensional (3D; organoids) cell culture systems [33] sufficiently simulate a complex reaction of the whole organism. Thus, suitable translational animal models play a vital role in human disease research [34][35][36][37].
In our experiments, we infected the one-week-old preterm GF piglets with S. Typhimurium strain LT2 for 24 h [46]. This Salmonella strain was avirulent for one-week-old conventional (CV) piglets [52] but lethal for term GF piglets, which died 36-48 h postinfection [53]. Thus, its virulence is influenced by the presence of a microbiota and, expectedly, also the microbiome composition. Moreover, the virulence of the S. Typhimurium serovars for the GF piglets depends on the form completeness of its LPS. For example, Gram-negative bacteria secrete smooth LPS chemotype (S-LPS), which is more virulent than rough (R-LPS) chemotype mutants [47,53]. Colonization resistance, presence of maternal immunoglobulins and immune cells [54], and stimulation of innate immunity [55] in the CV piglets are probably responsible for their resistance to LT2 infection [9,13,29,34,52,56,57].

Intestinal Barrier
A single epithelial cell layer creates a specific barrier between the bacteria-rich intestinal lumen and the host's organism. Adjacent enterocytes are joined in their apical part with tight junction proteins, e.g., claudins and occludin, and create a semipermeable interface that protects the host against penetration of harmful dietary antigens and invading pathogens and their toxins [58]. The epithelial layer is covered with mucin composed of a lumen-oriented movable upper layer and an enterocyte-touched fixed lower layer [59]. The disruption of the mucus layer causes intestinal inflammation and facilitates bacterial translocation [60,61]. Mucins are produced and secreted from goblet cells which are specialized enterocytes and colonocytes in the small intestine and colon, respectively [58]. The main intestinal mucin in humans and mice is acid mucin 2 (MUC2) [59]. An impaired MUC2 synthesis predisposed preterm CV piglets to develop necrotizing enterocolitis [62], and a defect in the production of MUC2 dramatically increased the sensitivity of mice to infection with S. Typhimurium [63]. The presence of bacteria stimulates the production of mucin and GF animals show thinner mucin layer compare to CV ones [58].
We measured acid mucin-producing goblet cell density in the ileum and colon of the preterm GN piglets. In our previous study with the term piglets [61], we found that the term GN piglets showed a higher number of goblet cells in the ileum. In the current study, preterm GN piglets had a comparable goblet cell density in the ileum and colon. In contrast, comparable numbers of the acid and neutral mucin-containing goblet cells in the distal small intestine, but lower in the colon, were found in the preterm versus term CV piglets [61,64,65]. Mucin degradation allows for easier penetration of harmful bacteria, and the absence of mucin-degrading activities is a safety criterion for probiotic candidates [63,66].
In previous studies, BB12 alone or in combination with LT2 did not weaken the intestinal barrier or increase bacterial translocation [46], as was shown in mucinolytic B. boum RP36 [61]. Thus, we believe that BB12 did not provoke an adverse effect in the ileum of the preterm GN piglets. In the colon of the term GN piglets, which has a higher goblet cell density than the ileum, a negative effect of Salmonella was shown without an influence of either B. boum strain studied [61]. This significant Salmonella-induced downregulation of goblet cell density in the colon was also found in preterm piglets; however, this effect was ameliorated by probiotic BB12, supporting a beneficial effect on the host. Differences in goblet cell count in piglets can also be influenced by the formula used, as was found in formula-fed compared to colostrum-fed CV preterm piglets [62]. In our experiment, all GN piglet groups were fed the same way , excluding the possible effect of the diet.
Gram-positive BB12 alone did not upregulate TLR2 expression in the ileum, colon, and mesenteric lymph nodes. However, Gram-negative LT2 upregulated it in the ileum and mesenteric lymph nodes and combination with BB12 in all three observed organs. This finding is seemingly controversial. However, TLR2 recognizes shared patterns of both Gram-positive and Gram-negative bacteria, e.g., lipoproteins [69] and some lipopolysaccharides [67]. Moreover, this signaling pathway uses the CD14 molecule, which is mainly known as a co-receptor of the TLR4/MD-2 signaling pathway [69].
TLR4 was significantly upregulated in all organs of piglets infected with S. Typhimurium. The previous association with BB12 alleviated the upregulation in the mesenteric lymph nodes, but not in the intestine. The released LPS is bound to the LPS-binding protein, trapped by the CD14 molecule, and transported to TLR4/MD-2 complex [69]. The activation of the TLR4 pathway depends on the LPS structure and completeness of LPS. LT2 isogenic ∆rfa mutants with truncated R-LPS were shown to be less effective in the activation of TLR4/MD-2 signaling pathway and induction of local and systemic inflammatory cytokine levels than wild-type LT2 with S-LPS [53]. While released LPS causes life-threatening endotoxin shock [70], non-typhoidal avirulent Salmonella serovars with R-LPS can induce an inflammatory reaction that protects GN piglets against the subsequent infection with S-LPS virulent S. Typhimurium [71][72][73].
TLR4 is only one of the TLRs that use both MyD88 and TRIF adaptor molecules in cell surface and endosomal TLR4 signaling, respectively [67]. MyD88-dependent and TRIF-dependent signaling consequent in different spectrums of produced inflammatory cytokines [13]. Similar profiles of TLR4 and MyD88 in the piglet groups in the ileum and mesenteric lymph nodes attest to MyD88 as the main adaptor molecule mediated inflammatory signaling in the Salmonella infection [74]. In contrast, this trend did not appear in the colon, and the downregulation of TRIF in Salmonella infection was obvious. Similar trends in preterm groups infected with S. Typhimurium were observed in the ileum, colon, and mesenteric lymph nodes in term GN piglets, independent of previous association with pig commensal Lactobacillus amylovorus, Lactobacillus mucosae, or probiotic Escherichia coli Nissle 1917 [75]. In contrast to direct contact with Salmonella and other bacteria with host intestinal tissue, Salmonella only translocated to mesenteric lymph nodes in GN piglets associated with mucinolytic B. boum [53,61].

Cytokines
Monocytes/macrophages and neutrophil granulocytes are the first-line sentinel cells of the innate immune response that are early prenatally developed [76][77][78]. Thus, a broad spectrum of inflammatory mechanisms is available for a non-specific immune response immediately after birth. However, their excessive production is known as a "cytokine storm" [79] and can cause multiple organ dysfunction [80]. Possible discrimination between physiological and pathological levels predetermines inflammatory cytokines as members of sepsis biomarkers [81]. Commonly used interleukins (IL)-8, IL-10, and tumor necrosis factor (TNF)-α were also found to be valuable markers of enteric infections in GN piglets [82], and higher levels in intestinal tissue and plasma were found in Salmonella-infected preterm GN piglets [46]. IL-6 and IL-12/23p40 are other cytokines that go together with IL-8, IL-10, and TNF-α as biomarkers of prenatal and postnatal inflammation [83,84] and neonatal sepsis [85][86][87][88].
We found that infection with Salmonella excessively upregulated ileal, colonic, and plasma IL-6 and IL-12/23/p40 levels within the acute phase of the immune response. Their excessive levels attest to the deleterious effect of the Salmonella infection on the immunocompromised preterm GN piglets. The previous colonization of piglets with BB12 prevented a significant increase in IL-6 and IL-12/23p40 intestinal and plasma levels after infection with Salmonella compared to control GF piglets. BB12 ameliorated the cytokine storm [79] and the subsequent multiple dysfunction syndrome as its consequences [80].

HMGB1 Protein Expression in the Ileum and Its Intestinal and Plasmatic Levels
HMGB1 is a DNA-binding nuclear protein crucial for transcription that orchestrates responses to tissue damage and repair [17]. Released HMGB1 is also an inflammatory mediator with cytokine activity that emphasizes the production of inflammatory cytokines of intestinal inflammation associated with endotoxemia and NEC [89]. Intestinal HMGB1 was described as an inflammatory bowel disease (IBD) marker in children and a marker of the severity of enteric infections in GN piglets [90]. The part of GN piglets infected with necrotoxigenic E. coli O55 that relatively thrived showed low levels of plasmatic and intestinal HMGB1, but the piglets that suffered from the infection showed highly increased levels [91]. Our present finding confirms HMGB1 participation in a cytokine storm [79] with its detrimental effect in the Salmonella-infected GN piglets.
HMGB1 levels can be increased by the active secretion of immune cells or its passive release from necrotic cells [17]. Previously, we showed histopathological changes in the intestine in the Salmonella-infected preterm piglets [46]. In this work, we presented changes in the localization of HMGB1 in the enterocytes, justifying that increased levels in the Salmonella-infected piglets originated from both its stimulated secretion and necrotic release.

Conclusions
Exaggerated levels of cytokines within a cytokine storm [79] have systemic effects due to the damage of vital organs [92]. A modification of microbiota and renewal of its balance can be a therapeutic way of preventing or modulation of MOD in sepsis-suffered patients and increasing the ratio of patients that thrive [93]. Mono-associated GN piglets are the first step of bacterial interference studies and their consequences for the immunocompromised host. The GN piglets associated with a defined synthetic microbiota will be the logical next step in our future research.

Gnotobiotic Piglets
Preterm miniature germ-free (GF) piglets were derived by hysterectomy on day 104 of pregnancy and reared in fiberglass isolators with a partially heated floor; they were fed 6-7 times per day with cow-milk-based formula. Their microbiological state was tested as described elsewhere [29]. Piglets (n = 24) were divided into four groups, with six piglets per group (Figure 9), and orally colonized/infected with BB12 (BB12), LT2 (LT2), and their combination (BB12 + LT2), as we showed ( Figure 9) and described previously [46]. The bacteria were orally administered in 5 mL of the milk diet, and the control piglets (GF) received 5 mL of milk without bacteria. At the end of the experiment, the piglets were euthanized by exsanguination via cardiac puncture under isoflurane anesthesia.

Mucin and Goblet Cells in the Ileum and Colon
Acid mucin-producing cell density per area of the tunica mucosa was assessed as described elsewhere [61]. Briefly, Carnoy's fluid-fixed terminal ileum and colon were dehydrated and embedded in paraffin, and 5 µm cross-sections were stained with Alcian Blue and post-stained with Nuclear Fast Red. The specimens were examined under an Olympus BX 40 microscope with an Olympus Camedia C-2000 digital camera (Olympus, Tokyo, Japan).

Mucin and Goblet Cells in the Ileum and Colon
Acid mucin-producing cell density per area of the tunica mucosa was assessed as described elsewhere [61]. Briefly, Carnoy's fluid-fixed terminal ileum and colon were dehydrated and embedded in paraffin, and 5 μm cross-sections were stained with Alcian Blue and post-stained with Nuclear Fast Red. The specimens were examined under an Olympus BX 40 microscope with an Olympus Camedia C-2000 digital camera (Olympus, Tokyo, Japan).

Intestinal Lavage and Blood Plasma
Sections (40 cm) of proximal jejunum and the whole ileum with distal jejunum segments were filled with 2 mL of Dulbecco's PBS (DPBS; TPP, Pasching, Austria), gently kneaded, and rinsed. The colon was cut into small pieces and lavaged in 4 mL of DPBS. The lavages were briefly vortexed and centrifuged at 2500× g for 30 min at 8 °C, and supernatants were filtered through a 0.2 μm filter (Sartorius, Goettingen, Germany). Citrated blood was withdrawn by cardiac puncture and centrifuged at 1200× g for 10 min at 8 °C. A protease inhibitor cocktail (Roche Diagnostics, Manheim, Germany) was added to the lavage filtrates and plasma, and their aliquots were frozen and stored at −45 °C until processing.

RNA Purification and cDNA Synthesis
The terminal ileum and transverse colon cross-section slices and small pieces of mesenteric lymph nodes were put into RNAlater and stored at −20 °C. Later they were moved into the RTL buffer of RNeasy Mini Kit Plus (Qiagen) and homogenized with 2 mm zirconia beads (BioSpec Products, Bartlesville, OK, USA) in TissueLyser LT beadbeater (Qiagen). The total RNA was isolated according to the manufacturer's protocol. A total of 500 ng of total RNA was reverse transcribed by QuantiTect Reverse Transcription kit (Qiagen). The prepared cDNA was 1/10 diluted by PCR quality water (Life Technologies, Carlsbad, CA, USA), and this cDNA template was stored at −25 °C till quantitative PCR was performed.

Intestinal Lavage and Blood Plasma
Sections (40 cm) of proximal jejunum and the whole ileum with distal jejunum segments were filled with 2 mL of Dulbecco's PBS (DPBS; TPP, Pasching, Austria), gently kneaded, and rinsed. The colon was cut into small pieces and lavaged in 4 mL of DPBS. The lavages were briefly vortexed and centrifuged at 2500× g for 30 min at 8 • C, and supernatants were filtered through a 0.2 µm filter (Sartorius, Goettingen, Germany). Citrated blood was withdrawn by cardiac puncture and centrifuged at 1200× g for 10 min at 8 • C. A protease inhibitor cocktail (Roche Diagnostics, Manheim, Germany) was added to the lavage filtrates and plasma, and their aliquots were frozen and stored at −45 • C until processing.

RNA Purification and cDNA Synthesis
The terminal ileum and transverse colon cross-section slices and small pieces of mesenteric lymph nodes were put into RNAlater and stored at −20 • C. Later they were moved into the RTL buffer of RNeasy Mini Kit Plus (Qiagen) and homogenized with 2 mm zirconia beads (BioSpec Products, Bartlesville, OK, USA) in TissueLyser LT beadbeater (Qiagen). The total RNA was isolated according to the manufacturer's protocol. A total of 500 ng of total RNA was reverse transcribed by QuantiTect Reverse Transcription kit (Qiagen). The prepared cDNA was 1/10 diluted by PCR quality water (Life Technologies, Carlsbad, CA, USA), and this cDNA template was stored at −25 • C till quantitative PCR was performed.

Real-Time PCR
A total of 2 µL of the cDNA template was added into 18 µL of the FastStart Universal Probe Master (Roche Diagnostics), with 500 nM each of forward and reverse primers and 100 nM locked nucleic acid (LNA) probe (Universal ProbeLibrary; Roche Diagnostics). The PCR systems for the reference genes β-actin and cyclophilin A, as well as for the genes of interest, TLR2, TLR4, TLR9, MyD88, and TRIF, were listed elsewhere [29]. The PCR amplification was performed in duplicates in 45 cycles (95 • C for 15 s and 60 • C for 60 s) and run on an iQ cycler (Bio-Rad, Hercules, CA, USA). The evaluation of relative mRNA expression (fold changes) was described elsewhere [46].

Intestinal and Plasmatic HMGB1 Levels
HMGB1 levels in the intestinal lavages (jejunum, ileum, and colon) and blood plasma were measured by ELISA kit (Abbexa, Cambridge, UK), according to the producer's