Bifidobacterium breve CNCM I-4035, Lactobacillus paracasei CNCM I-4034 and Lactobacillus rhamnosus CNCM I-4036 Modulate Macrophage Gene Expression and Ameliorate Damage Markers in the Liver of Zucker-Leprfa/fa Rats

Non-alcoholic fatty liver disease (NAFLD) has reached pandemic proportions worldwide. We have previously reported that the probiotic strains Bifidobacterium breve CNCM I-4035, Lactobacillus paracasei CNCM I-4034 and Lactobacillus rhamnosus CNCM I-4036 exert anti-inflammatory effects in the intestine of Zucker-Leprfa/fa rats. In this work, we focused on their hepatic effects. M1 macrophages are related to inflammation and NAFLD pathogenesis, whereas M2 macrophages release anti-inflammatory mediators. We evaluated the effects of these 3 strains on macrophage polarization, inflammation and liver damage of Zucker-Leprfa/fa rats. The animals received either a placebo or 1010 CFU of probiotics orally for 30 days. Nos2 and Cd86 mRNA levels were determined as markers of M1 macrophages, and Cd163 and Arg1 as M2 markers, respectively, by qRT-PCR. Liver damage was determined by lipid peroxidation, leukocyte infiltration and myeloperoxidase activity. We evaluated a panoply of circulating chemokines, the hepatic ratio P-Akt/Akt, NF-kB and P-NF-kB protein levels. All 3 probiotic strains modulated macrophage polarization in liver and circulating levels of inflammation-related mediators. L. paracasei CNCM I-4034 increased the ratio P-Akt/Akt and NF-kB protein levels. B. breve CNCM I-4035, L. paracasei CNCM I-4034 and L. rhamnosus CNCM I-4036 decreased both pro-inflammatory macrophage gene expression and leukocyte infiltration in the liver.


Introduction
Non-alcoholic fatty liver disease (NAFLD) has reached pandemic proportions worldwide [1]. This disease is characterized by fat accumulation in the form of micro and

Histology
Paraffin-embedded liver samples were sliced onto 5 µm-thick sections and stained with hematoxylin and eosin for histological examination. Four animals per group and 5 sections per animal were microscopically analyzed to determine leukocyte infiltrate with ImageJ software using gray-scale.

Lipid Peroxidation
Malondialdehyde (MDA) was determined by the thiobarbiturate reaction [25]. For this purpose, 1 mL of trichloroacetic acid (20%) was added to 1 mL of homogenate in phosphate buffer 0.067 M at pH 7.4. After mixing and centrifuging, 1 mL of thiobarbiturate (0.67%) was added to the supernatant and boiled for 60 min. After cooling, optical density at 530 nm was assayed.

Neutrophil Infiltration
Hepatic myeloperoxidase (MPO) activity was measured as a biomarker of neutrophil infiltration and activation. Two hundred µg of cryolyophilized liver was resuspended in 1.5 mL of 0.5% hexadecyltrimethylammonium bromide in 50 mM phosphate buffer, pH 6, and mechanically lysed using a vortex for 15 min. Homogenates were subjected to 3 cycles of freezing in liquid nitrogen/thawing at room temperature. Samples were then incubated during 2 h at 60 • C and centrifuged at 4000× g for 12 min. Supernatants were collected for MPO activity evaluation. Solutions containing 10 µL of supernatant, 10 µL of tetramethylbenzidine (work concentration 1.6 mM) in dimethyl sulfoxide, and 70 µL of H 2 O 2 (work concentration 3.0 mM) in 80 mM phosphate buffer pH 4.5 were measured photometrically at 630 nm every 15 s for a total of 4 min [26].

Hepatic LPS
LPS was measured in liver homogenates using a competitive inhibition enzyme immunoassay (CEB526Ge, Cloud-CloneCorp., Houston, TX, USA) following the manufacturer's protocol.

Statistical Analysis
One-way ANOVA followed by Dunnett's multiple comparisons test was performed using GraphPad Prism version 8.0.1 for Windows (GraphPad Software, San Diego, CA, USA). p < 0.05 was considered statistically significant. All results were expressed as mean ± standard error of the mean (SEM).

Results
In this work, we used Zucker-Lepr fa/fa rats as a model of insulin resistance syndrome (IRS) [24]. These rats manifest the liver component of IRS, NAFLD, among others [23]. At time 0, rats exhibited a lean phenotype with a weight of 179.9 ± 2.2 g, whereas after 30 days of intervention with the placebo the weight was 294.4 ± 5.7 g. Probiotic treatments did not modify rats' body weight. Body weights were 282.1 ± 11.5 g in the B. breve group, 292.4 ± 10.8 g in the L. paracasei group, and 295.8 ± 13.0 g in the L. rhamnosus group. Liver inflammation is regulated by chemokines, which modulate the migration and activities of hepatocytes, Kupffer cells, hepatic stellate cells, endothelial cells, and circulating immune cells [27]. Serum levels of GM-CSF, RANTES, MIP-1 α, and MCP-1 were determined ( Figure 1). Both MIP-1 α and MCP-1 increased in the serum of Zucker-Lepr fa/fa rats that received L. paracasei CNCM I-4034 for 30 days in comparison with the placebo group ( Figure 1C,D). Similar findings in these two chemokines and GM-CSF were induced by L. rhamnosus CNCM I-4036, although the differences with the placebo group were not statistically significant ( Figure 1A,C,D).
Serum levels of two adhesion molecules, E-selectin and sICAM-1, were also measured ( Figure 2). E-selectin is a specific endothelial adhesion receptor induced by proinflammatory stimuli. It mediates the adhesion of leukocytes to endothelial cells allowing their extravasation into inflammed tissues [28]. NAFLD is associated with elevated circulating levels of E-selectin and sICAM [29]. All groups exhibited similar values of E-selectin ( Figure 2A). However, circulating sICAM-1 significantly increased in the rats treated with L. paracasei CNCM I-4034 ( Figure 2B). Serum levels of two adhesion molecules, E-selectin and sICAM-1, were also measured ( Figure 2). E-selectin is a specific endothelial adhesion receptor induced by pro-inflammatory stimuli. It mediates the adhesion of leukocytes to endothelial cells allowing their extravasation into inflammed tissues [28]. NAFLD is associated with elevated circulating levels of E-selectin and sICAM [29]. All groups exhibited similar values of E-selectin ( Figure 2A). However, circulating sICAM-1 significantly increased in the rats treated with L. paracasei CNCM I-4034 ( Figure 2B).  Serum levels of two adhesion molecules, E-selectin and sICAM-1, were also measured ( Figure 2). E-selectin is a specific endothelial adhesion receptor induced by pro-inflammatory stimuli. It mediates the adhesion of leukocytes to endothelial cells allowing their extravasation into inflammed tissues [28]. NAFLD is associated with elevated circulating levels of E-selectin and sICAM [29]. All groups exhibited similar values of E-selectin ( Figure 2A). However, circulating sICAM-1 significantly increased in the rats treated with L. paracasei CNCM I-4034 ( Figure 2B). inflammatory cytokines, IL-4 increased in the L. paracasei group, although the difference with the placebo group did not reach statistical significance (p = 0.07, Figure 4A). IL-10 results are noteworthy ( Figure 4C): although no statistical significance was reached, this IL tended to be higher with L. paracasei CNCM-I 4034 and L. rhamnosus CNCM-I 4036 treatments, the reason probably being that the number of rats (out of 7-8/group) where this cytokine was detected was much higher in the latter groups (control: 1 of 8 rats; placebo: 5 of 8; B. breve: 2 of 8; L. paracasei: 7 of 8; L. rhamnosus: 4 of 8).   inflammatory cytokines, IL-4 increased in the L. paracasei group, although the difference with the placebo group did not reach statistical significance (p = 0.07, Figure 4A). IL-10 results are noteworthy ( Figure 4C): although no statistical significance was reached, this IL tended to be higher with L. paracasei CNCM-I 4034 and L. rhamnosus CNCM-I 4036 treatments, the reason probably being that the number of rats (out of 7-8/group) where this cytokine was detected was much higher in the latter groups (control: 1 of 8 rats; placebo: 5 of 8; B. breve: 2 of 8; L. paracasei: 7 of 8; L. rhamnosus: 4 of 8).   Due to the central role of leukocytes in the development and propagation of inflammation, we evaluated the infiltration of leukocytes in the rats' liver to determine the potential effect of the three probiotic strains ( Figure 5). Liver tissue from the control group was normal, while Zucker-Lepr fa/fa rats liver parenchyma clearly showed the accumulation of lipids in cytoplasmic vacuoles characteristic of macrovesicular hepatic steatosis ( Figure 5, upper panels). All three probiotic strains diminished leukocyte infiltration in the liver of the Zucker-Lepr fa/fa rats, as shown in the graph representing quantitative data obtained from all stained liver sections ( Figure 5, lower panel).

Lactobacillus paracasei CNCM I-4034, Bifidobacterium breve CNCM I-4035 and Lactobacillus rhamnosus CNCM I-4036 Decreased Leukocyte Infiltration in the Liver of Zucker-Lepr fa/fa Rats
Due to the central role of leukocytes in the development and propagation of inflammation, we evaluated the infiltration of leukocytes in the rats' liver to determine the potential effect of the three probiotic strains ( Figure 5). Liver tissue from the control group was normal, while Zucker-Lepr fa/fa rats liver parenchyma clearly showed the accumulation of lipids in cytoplasmic vacuoles characteristic of macrovesicular hepatic steatosis ( Figure  5, upper panels). All three probiotic strains diminished leukocyte infiltration in the liver of the Zucker-Lepr fa/fa rats, as shown in the graph representing quantitative data obtained from all stained liver sections ( Figure 5, lower panel). We also evaluated MPO activity in both liver tissue and serum, an enzyme that has been used as a marker of neutrophil infiltration and activation, and liver MDA as an indirect measure of lipid peroxidation ( Figure 6). Liver MPO activity significantly increased after 30 days of feeding with the placebo ( Figure 6A). Conversely, administration of L. paracasei CNCM I-4034 and L. rhamnosus CNCM I-4036 reduced MPO activity, reaching values of lean rats at time 0 ( Figure 6A). Liver MPO tended to decrease with B. breve CNCM I-4035 (p = 0.112). MDA content decreased in the liver of rats fed L. paracasei CNCM I-4034 and tended to decrease with L. rhamnosus CNCM I-4036 (p = 0.0724) for 30 days ( Figure 6C).
It is known that LPS may trigger the inflammatory process [30]. The loss of LPS-binding protein (LBP) is related to an attenuation of the liver-mediated inflammation [31]. We have described that treatment of Zucker-Lepr fa/fa rats with these three probiotic strains diminished the LBP concentration in serum [23]. We, therefore, determined the hepatic LPS content but found no differences among groups ( Figure 6D). We also evaluated MPO activity in both liver tissue and serum, an enzyme that has been used as a marker of neutrophil infiltration and activation, and liver MDA as an indirect measure of lipid peroxidation ( Figure 6). Liver MPO activity significantly increased after 30 days of feeding with the placebo ( Figure 6A). Conversely, administration of L. paracasei CNCM I-4034 and L. rhamnosus CNCM I-4036 reduced MPO activity, reaching values of lean rats at time 0 ( Figure 6A). Liver MPO tended to decrease with B. breve CNCM I-4035 (p = 0.112). MDA content decreased in the liver of rats fed L. paracasei CNCM I-4034 and tended to decrease with L. rhamnosus CNCM I-4036 (p = 0.0724) for 30 days ( Figure 6C).
It is known that LPS may trigger the inflammatory process [30]. The loss of LPSbinding protein (LBP) is related to an attenuation of the liver-mediated inflammation [31]. We have described that treatment of Zucker-Lepr fa/fa rats with these three probiotic strains diminished the LBP concentration in serum [23]. We, therefore, determined the hepatic LPS content but found no differences among groups ( Figure 6D).

Administration of Lactobacillus paracasei CNCM I-4034, Bifidobacterium breve CNCM I-4035 and Lactobacillus rhamnosus CNCM I-4036 Modulated Macrophage Polarization in Liver
Macrophages are recognized to exist as two distinct subtypes, M1 pro-inflammatory and M2 anti-inflammatory [6]. We investigated whether L. paracasei CNCM I-4034, B. breve CNCM I-4035 and L. rhamnosus CNCM I-4036 could affect the distribution of both subtypes of macrophages in the rats' livers. We measured Cd86 and Nos2 mRNA levels as markers of M1 macrophages ( Figure 7A,B), and Cd163 and Arg1 as M2 markers ( Figure  7C,D) by qRT-PCR. Administration of the placebo resulted in an induction of Cd86 mRNA in the liver ( Figure 7A), and feeding with all three probiotic strains for 30 days significantly down-regulated Cd86 expression ( Figure 7A). We did not find differences in Nos2 ( Figure 7B) and Cd163 ( Figure 7C) among groups. Arg1 mRNA was up-regulated with B. breve, whereas administration of L. paracasei CNCM I-4034 and L. rhamnosus CNCM I-4036 down-regulated Arg1 in the liver ( Figure 7D).

Administration of Lactobacillus paracasei CNCM I-4034, Bifidobacterium breve CNCM I-4035 and Lactobacillus rhamnosus CNCM I-4036 Modulated Macrophage Polarization in Liver
Macrophages are recognized to exist as two distinct subtypes, M1 pro-inflammatory and M2 anti-inflammatory [6]. We investigated whether L. paracasei CNCM I-4034, B. breve CNCM I-4035 and L. rhamnosus CNCM I-4036 could affect the distribution of both subtypes of macrophages in the rats' livers. We measured Cd86 and Nos2 mRNA levels as markers of M1 macrophages ( Figure 7A,B), and Cd163 and Arg1 as M2 markers ( Figure 7C,D) by qRT-PCR. Administration of the placebo resulted in an induction of Cd86 mRNA in the liver ( Figure 7A), and feeding with all three probiotic strains for 30 days significantly downregulated Cd86 expression ( Figure 7A). We did not find differences in Nos2 ( Figure 7B) and Cd163 ( Figure 7C) among groups. Arg1 mRNA was up-regulated with B. breve, whereas administration of L. paracasei CNCM I-4034 and L. rhamnosus CNCM I-4036 down-regulated Arg1 in the liver ( Figure 7D).
Activation of Akt is induced by different types of stress and provides a cell survival signal [32]. In addition, increased production of butyrate by probiotics has been associated with an amelioration of NAFLD through, among others, the activation of Akt [33]. Moreover, whereas NF-kB is known to promote inflammation, probiotics have been reported to counteract its activation [34]. We evaluated the ability of the probiotic strains to activate Akt and NF-kB by determining phosphorylared-Akt/Akt ratio, NF-kB and phosphorylated-NF-kB protein levels by Western blotting (Figure 8). Whereas B. breve CNCM I-4035 and L. rhamnosus CNCM I-4036 had no effects on P-Akt/Akt, P-NF-kB or NF-kB, L. paracasei CNCM I-4034 significantly increased the ratio P-Akt/Akt and NF-kB protein levels.

Discussion
Probiotics are live microorganisms that, when consumed in adequate amounts, confer a health effect on the host. Beneficial effects of probiotics have been reported in allergy, intestinal-related diseases, chronic liver disease, urinary tract infections, and respiratory infections, among others. Lactobacilli and bifidobacteria are the genera most frequently

Discussion
Probiotics are live microorganisms that, when consumed in adequate amounts, confer a health effect on the host. Beneficial effects of probiotics have been reported in allergy, intestinal-related diseases, chronic liver disease, urinary tract infections, and respiratory infections, among others. Lactobacilli and bifidobacteria are the genera most frequently used as probiotics, and they exert their benefits through a variety of mechanisms [23].
We have reported that the administration of Lactobacillus paracasei CNCM I-4034, Bifidobacterium breve CNCM I-4035 and Lactobacillus rhamnosus CNCM I-4036 to Zucker-Lepr fa/fa rats attenuates the accumulation of fat in the rats' liver and exerts anti-inflammatory effects such as lower serum concentrations of tumor necrosis factor alpha (TNF-α), IL-6 and bacterial LPS [24]. Also, expression of three genes (Adamdec1, Ednrb and Ptgs1/Cox1) was up-regulated in the intestinal mucosa of the obese rats compared with the lean rats [23]. This effect was in part mediated by a decrease in both macrophage and dendritic cell populations. Probiotic treatment also increased secretory IgA content and diminished the LBP concentration [24].
Inflammation is a hallmark of liver disease. The immune response plays a key role in initiation, progression and resolution of hepatic inflammation [35]. A great body of evidence points that regulation of the phenotype of resident hepatic macrophages, Kupffer cells, is related to the progression of liver diseases such as alcoholic liver disease, NAFLD, non-alcoholic steatohepatitis (NASH), acute liver injury, alcoholic hepatitis, acute liver failure and hepatocellular carcinoma [35].
In this work we used the NAFLD model of Zucker-Lepr fa/fa rats to investigate the immune response associated with the administration of these three bacterial strains in hepatic macrophage modulation and their effects in liver damage. Overall, after 30 days of intervention and in comparison with a placebo, we observed an amelioration of hepatic leukocyte infiltration and damage markers that might be associated, at least in part, with their ability to modulate the immune response and macrophage distribution, although the effects were strain-dependent.
Secreted chemokines have a high affinity to glycosaminoglycans bound to extracellular matrix and endothelial surface. This property favors the local immobilization and retention of chemokines, creating a concentration gradient that allows a coordinated trafficking of leukocytes toward injured tissue [27]. In this regard, MCP-1 is associated with progression of simple steatosis to NASH, while RANTES is mainly involved in migration of T cells, monocytes, neutrophils, and dendritic cells through binding to its cognate transmembrane receptors [10], and is involved in several chronic immune-inflammatory diseases [36]. RANTES and MIP-1 α, among others, are up-regulated in fibrotic livers [37].
We found increased serum MIP-1 α and MCP-1 levels after L. paracasei CNCM I-4034 administration compared with the placebo group. It is worthy to note that, in spite of the ameliorated NAFLD-associated liver injury, we found increased pro-inflammatory mediators in the L. paracasei group, which seems to disagree with the improvement in liver damage and steatosis (the B. breve group was the one that exhibited the lower leukocyte hepatic infiltration). However, this is not the first time that high pro-inflammatory markers are described for other probiotic strains. For instance, similar findings have been reported for L. paracasei CNCM I-1518 [38]. Likewise, we have previously described elevated proinflammatory factors (TNF-α, IL-8 and RANTES) in dendritic cells cultured in the presence of B. breve CNCM I-4035 [39], which implies that the balance of both types of cytokines is essential to promote an adequate immune response. Thus, a massive recruitment of immune cells in the lungs due to administration of L. paracasei CNCM I-1518 strain and prior to influenza infection led to increased pro-inflammatory cytokine release [38]. This preactivation state of the immune system seems to be responsible for the faster clearance of the infection, supporting the hypothesis that an early beneficial induction of the proinflammatory response before influenza infection, and a lower inflammatory response after this infection, is necessary to counteract an overactive immune response.
Administration of L. paracasei CNCM I-4034 induced NF-KB protein levels but no differences were found in its activated form. L. paracasei CNCM I-4034 also activated the Akt pathway, which is related to cell survival after different types of stress [32]. Furthermore, butyrate-producing probiotics have been associated with a reduction in the progression of fatty liver disease in rats through, among others, activation of Akt [33].
Specific adhesion glycoproteins are involved in the binding of leukocytes to endothelial cells [40]. NAFLD is associated with elevated circulating levels of E-selectin and sICAM [29,41]. We did not find differences in either E-selectin or hepatic LPS among the experimental groups, which suggests that changes in hepatic liver inflammation were not associated with changes in hepatic LPS. Circulating sICAM1 increased in the L. paracasei group. Although supposedly deleterious, the rise in this adhesion molecule together with the rise in the aforementioned chemokines may have been overcome by other protective mechanisms and for that reason the global effect in the liver was an improvement in this group.
We also found a modulatory effect on the expression of hepatic macrophages after probiotic administration. Cd86 mRNA levels were down-regulated by all three probiotic strains. Decreased expression of M1 pro-inflammatory marker Cd86 was accompanied by the improvement in liver damage evaluated by MPO activity and leukocyte infiltration. These findings are consistent with the reductions in IL-6 and, TNF-α, and the improvement in steatosis that we have previously reported for these probiotic strains [24]. Also, B. breve CNCM I-4035 in particular up-regulated Arg1 mRNA expression, a marker of the M2 anti-inflammatory phenotype.
In conclusion, here we report that B. breve CNCM I-4035, L. paracasei CNCM I-4034 and L. rhamnosus CNCM I-4036 modulated liver macrophage gene expression in the Zucker-Lepr fa/fa rat NAFLD model, which was accompanied by an improvement in hepatic leukocyte infiltration. Effects on hepatic lipid peroxidation and damage was strain-dependent. In particular L. paracasei CNCM I-4034 administration resulted in improvements in hepatic damage and peroxidation. Although the link between macrophage shift and hepatoprotection by probiotics should be further investigated, its confirmation would support the idea that probiotics may be useful as a coadjuvant treatment for pathologies such as NAFLD.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.