LF feeding deficiency during suckling period increased depression risk in adult mice
Male and female mice that consumed normal milk and LF-free milk during the suckling period were fed a normal diet until 9 weeks of age, after which, the CUMS model was established for 4 weeks (Fig. 1A). As shown in Fig. 1B and 1C, wt-wt and ko-wt mice did not differ before CUMS; however, after 4 weeks of CUMS treatment, mice without LF intake during the suckling period (ko-wtM and ko-wtF) showed a lower sucrose preference in the SPT. In the TST (Fig. 1D), there was no significant difference between the wt-wtM and ko-wtM groups during the first week of CUMS. At week 4, the immobility time of the wt-wtM and ko-wtM groups was higher than that of week 1, and the immobility time of the ko-wtM group was longer than that of the wt-wtM group. Although a similar trend was observed in the female group (Fig. 1E), no significant difference was observed between the wt-wtF and ko-wtF groups at weeks 1 and 4. In the FST (Fig. 1F), there was no difference between wt-wtM and ko-wtM mice at week 1; at week 4, the immobility time of the wt-wtM and ko-wtM groups increased compared with that of week 1, and the immobility time of the ko-wtM mice was longer than that of the wt-wtM mice. Figure 1G shows that the immobility time in the FST of the ko-wt female mice (ko-wtF) was significantly longer than that of the wt-wtF group at week 1, and increased compared with that of the wt-wtF mice at week 4, but the difference was not significant. In the OFT, the distance in the central area for the ko-wtM group was significantly smaller than that for the wt-wtM group, with no difference between the ko-wtF and wt-wtF groups (Fig. 1H). Male mice with LF feeding deficiency during lactation showed more severe anxiety-like behaviours in the OFT compared to male mice that drank normal milk. Figure 1H also shows that the distance in the central area for the wt-wtF group was significantly lower than that for the wt-wtM group, indicating that female mice exhibited more severe anxiety-like behaviours compared to male mice.
Serum indicators associated with depression were measured. BDNF is a well-known growth factor in the brain, and the level of BDNF in the ko-wt group was significantly lower than that in the wt-wt group, for both male and female mice (ko-wtM vs. wt-wtM; ko-wtF vs. wt-wtF). The levels of BDNF in female mice were significantly lower than those in male mice (wt-wtF vs. wt-wtM; ko-wtF vs. ko-wtM) (Fig. 1I). Corticosterone (CORT) and adrenocorticotropic hormone (ACTH) are hormones related to the hypothalamic–pituitary–adrenal axis and are associated with depression. Our results revealed a significantly increased production of CORT and ACTH in the ko-wt mice compared to that in the wt-wt mice (ko-wtM vs. wt-wtM; ko-wtF vs. wt-wtF) (Fig. 1J, 1K); the wt-wtF group showed significantly higher CORT levels than the wt-wtM group. The levels of pro-inflammatory cytokines, such as tumour necrosis factor (TNF)-α (Fig. 1L) and interleukin (IL)-1β (Fig. 1M), in the ko-wt group were significantly higher than those in the wt-wt group (ko-wtM vs. wt-wtM; ko-wtF vs. wt-wtF), and IL-1β levels in the wt-wtF group were significantly higher than those in the wt-wtM group. Additionally, LPS production was significantly increased in the ko-wt mice serum (ko-wtM vs. wt-wtM; ko-wtF vs. wt-wtF) (Fig. 1N), which can induce inflammation. Furthermore, the LPS levels in the ko-wtF group were significantly higher than those in the ko-wtM group.
CUMS mice with lactation LF feeding deficiency showed intestinal, brain and microbial flora disorders
Dysfunction of the microbiota–gut–brain axis is thought to be the main pathological basis of depression. We compared the histological changes in the colon, inflammation of the colon and hippocampus and composition of intestinal microorganisms in depressed mice with LF feeding deficiency during lactation with those in depressed mice that consumed normal mouse milk during lactation. As shown in Fig. 2A, the crypt depth of ko-wtM mice was significantly lower than that of wt-wtM mice, there was no significant difference in colonic crypt depth between the wt-wtF and ko-wtF groups. Damage to the intestinal epithelial barrier increased the translocation of this luminal LPS to systemic circulation, and expression of the zonula occludens gene (zo-1) in ko-wt mice was significantly lower than that in wt-wt mice (ko-wtM vs wt-wtM, analysis of variance, P < 0.05; ko-wtF vs wt-wtF, t test, P < 0.05). Occludin expression in ko-wtM mice was lower than that in wt-wtM mice but the difference was not significant (Fig. 2B). Occludin expression in wt-wtF was significantly lower than that in wt-wtM. The lactating LF feeding deficient mice showed more severe damage to their intestinal barrier integrity after CUMS, which may explain the increased LPS concentration in the serum of the ko-wtF and ko-wtM groups. Moreover, the intestinal barrier function was more severely damaged in female mice than in male mice.
As shown in Fig. 2C, the colons of the ko-wtM mice showed more severe inflammatory infiltration than that of wt-wtM mice; however, among females, inflammatory infiltration was more severe in wt-wtF mice than that in ko-wtF mice. We further examined the activation of the LPS-TLR4 signaling pathway in the colon and hippocampus. In the colon (Fig. 2D) of male mice, the expression of TLR-4, Myd88, CD14, nuclear factor (NF)-κB, TNF-α and IL-1β was significantly higher in ko-wtM group than in the wt-wtM group and expression of P38MAPK was higher in the ko-wtM group than in wt-wtM group, but the difference was not significant (P = 0.07). JNK expression in the wt-wtM group was significantly higher than that in the ko-wtM group; other genes showed no significant differences between these two groups. In female mice, the expression of TLR-4, CD14, p38MAPK, JNK, IL-1β and TNF-α was significantly lower in ko-wtF group than in the wt-wtF group. In the hippocampus (Fig. 2E), the expression of JNK was significantly higher in the ko-wtM group than in the wt-wtM group; the expression of Myd88, CD14, NF-κB, and p38MAPK tended to be higher in the ko-wtM group than in the wt-wtM group, but the difference was not significant. The expression of TLR-4, LBP, Myd88, NFκB, IL-1β, and TNF-α was significantly higher in ko-wtF mice than in wt-wtF mice. The expression of Iba-1 (a marker of activation for microglia) was significantly higher in ko-wtF mice than in wt-wtF mice.
Thus, after 4 weeks of CUMS, the colon of LF-free male mice showed more severe inflammation which occurred through activation of the TLR4-Myd88-NF-κB signaling pathway. However, female mice that consumed normal mouse milk showed more severe intestinal inflammation via activation of the MAPK and JNK signals in the TLR4 signaling pathway. The difference in colonic inflammation between female and male mice may be related to different levels of oestrogen [24]. Based on the significantly different levels of BDNF in the serum, we measured the expression of genes related to the BDNF signalling pathway in the hippocampus of depressed mice. None of the genes differed between the ko-wtM and wt-wtM group (Fig. 2F); however, in the female groups, the expression of GRB2, CaMK, and CREB in the ko-wtF group was significantly lower than that in the wt-wtF group, and the expression of BDNF in the ko-wtF mice was lower than that in the wt-wtF group but not significantly (P = 0.09).
Next, we examined the composition of the gut flora. Alpha diversity represents the richness and diversity of the microbial community. The Shannon formula was used to estimate diversity, and the richness of the microbial community was estimated using the Chao index. Mice without LF intake during the suckling period (ko-wtM and ko-wtF) showed a lower Shannon index of the intestinal microbiota but the difference was not significant (Fig. 2G). The Chao index was significantly lower in the ko-wtF than in the wt-wtF mice, but there was no significant difference between the ko-wtM and wt-wtM groups (Fig. 2G). Taken together, LF-free feeding during the suckling period decreased the microbial community richness compared with LF feeding after CUMS. Principal coordinate analysis (PCoA) revealed major differences between the ko-wtM and wt-wtM mice as well as differences in the microbial composition between the ko-wtF and wt-wtF mice (Fig. 2H, P = 0.097). There was a significant difference in the pattern of intestinal flora between the ko-wt-depressed and wt-wt-depressed mice. The predominant phyla (Fig. 2I) were Firmicutes, Bacteroidetes, Desulfobacterota, Patescibacteria, and Actinobacteria in the faecal samples of the four groups, and there was no difference between ko-wtM and wt-wtM mice. In contrast, the abundances of Desulfobacterota, Actinobacteria, and Chloroflexi significantly increased in the ko-wtF group (Fig. 2J). At the genus level, we used linear discriminant analysis effect size (LEfSe) to identify bacteria whose relative abundance significantly differed between ko-wt and wt-wt mice. Figure 2K shows that there was a greater abundance of Eubacterium_xylanophilum_group, and a lower abundance of Lactobacillus and SCFA-producing bacteria (Alistipes and Dubosiella) in the ko-wtM group than in the wt-wtM group. SCFAs exert anti-inflammatory functions and are beneficial to the composition of the gut microbiota and intestinal barrier [25]. Figure 2K shows that Lactobacillus, Desulfovibrio, Arenimonas, Enterorhabdus, and Bifidobacterium were significantly enriched in the ko-wtF group, and Ileibacterium was significantly enriched in the wt-wtF group.
To further explore the correlation between the gut microbiota (significantly different microbiota at the genus level) and depression-related behavioural indices (OFT, SPT, TST, and FST), depression-related hormones (ACTH and CORT), serum inflammatory factors (IL-1β and TNF-α), and BDNF, a heatmap of Spearman’s correlation analysis was generated. As shown in Fig. 2L, Desulfovibrio was significantly negatively correlated with SPT and positively correlated with FST (P = 0.06); Bifidobacterium was negatively correlated with TST (P = 0.05); and Alistipes was significantly positively correlated with OFT. Furthermore, Lactobacillus and Enterorhabdus were positively correlated with TST, ACTH, CORT, IL-1β, TNF-α, and LPS, and negatively correlated with OFT, SPT, and BDNF. Blautia showed a positive correlation with ACTH, FST, TST, CORT, IL-1β, TNF-α, LPS, and depression.
Intestinal growth retardation and intestinal microbiological disorder in 18-day-old suckling mice with LF feeding deficiency during lactation
Our experiments showed that LF intake during lactation significantly affected depression in adult mice, which was induced by differences in body development caused by LF intake during lactation. Therefore, we used 18-day-old mice with almost no feed intake for analysis. The body weight of 18-day-old suckling mice without LF feeding during lactation did not differ from that of control mice (Fig. 3A) but the intestinal index and density of the small intestine and colon decreased significantly (Fig. 3B, 3C). Crypt depth did not differ significantly between the two groups (Fig. 3D). The number of colonic goblet cells in the ko-wt group was significantly lower than that in the control group (Fig. 3E).
To further explore the effect of LF feeding deficiency on intestinal development during lactation, we performed genome-wide transcriptional profiling of the small intestine of the two groups of mice for RNA sequencing. PCoA showed that the transcriptomes of ko-wt and wt-wt mice were clearly separated (Fig. 3F). Transcriptome sequencing showed that 143 genes were upregulated and 72 genes downregulated in the small intestine of the ko-wt mice compared to those of the wt-wt mice (fold-change > 1, P adjust < 0.05, Fig. 3G). These DEGs were subjected to pathway analysis using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database, which showed that upregulated mRNAs were mainly involved in infectious, immune, and neurodegenerative diseases, and were enriched in the nervous, immune, and endocrine systems (Fig. 3H). We explored immune-related genes in the small intestine using an RNA-seq transcriptomic profiling approach, as shown in Fig. 3I; genes encoding α-defensin, such as Defa23, Defa36, Defa3, Defa17, Defa38, and Defa39, were significantly decreased in the ko-wt group. The RNA-Seq results showed no difference in the expression of the tight junction proteins zo-1 and occludin between the two groups. However, the expression of the gene encoding the actin protein (Gm12715) was significantly decreased (Fig. 3J); this gene is related to the assembly of tight junctions.
To determine the effect of LF deficiency during lactation on intestinal microorganisms, we analysed the gut microbiota composition by 16S rRNA gene amplicon sequencing of the caecal contents of 18-day-old mice. Alpha diversity measures (Shannon and Chao diversity indices), calculated for each group, showed that the Chao diversity index of the ko-wt group did not differ from that of the wt-wt group; however, the Shannon index was significantly higher than that of the control group (Fig. 3K). These results indicate that the intestinal microbial alpha diversity was increased in 18-day-old mice not fed LF. PCoA of the beta-diversity comparison revealed significant separation between the two groups of microbial communities (Fig. 3L). At the phylum level, Firmicutes, Bacteroidetes, and Desulfobacterota accounted for the majority of the bacteria (Fig. 3M), with no differences between the two groups (Fig. 3N). At the genus level, the LEfSe results revealed a greater abundance of Muribaculum, Oscillibacter, Bilophila, Colidextribacter, Harryflintia, Bifidobacterium, Odoribacter, UCG-009, Lachnospiraceae_UCG-006, Eubacterium_xylanophilum_group, Ruminococcus, and some unclassified microbiotas; there was a lower abundance of Blautia, Desulfovibrio, Parabacteroides, Romboutsia, Eubacterium_brachy_group, and Turicibacter in the ko-wt group (Fig. 3O).
LF feeding deficiency during lactation induced hippocampal growth retardation
Considering that depression is a mental illness, it is necessary to explore the effects of LF deficiency during lactation on the hippocampus. Figure 4A shows that there was no significant difference in the brain index between ko-wt and wt-wt 18-day-old mice. The CA1 region of the hippocampus is involved in cognitive processes, learning, and memory[26]. The hippocampal dentate gyrus (DG) is a key target region for both antidepressant effects and stress-related processes [27]. In the wt-wt group, pyramidal cells in the CA1 region were arranged neatly in dense layers and had regular and clear structures. No significant difference was observed between the ko-wt and wt-wt groups in cell morphology in the DG region (Fig. 4B).
RNA-seq of the hippocampi of 18-day-old mice was performed. PCoA showed that the microbiota profiles of the two groups were clearly distinct (Fig. 4C). Transcriptome sequencing results showed that 199 genes were upregulated and 297 genes downregulated in the hippocampi of the ko-wt mice compared to those of the wt-wt mice (fold-change > 1, P < 0.05, Fig. 4D). The KEGG pathway annotation results showed that the downregulated mRNAs mainly participated in signalling molecules and interactions, and signal transduction, and were enriched in the immune system (Fig. 4E). Figure 4F shows that the expression of genes involved in innate immunity, such as Lrg1, Gbp4 [28], Gbp10 [28], and Stx11 [29], was decreased. Genes involved in adaptive immunity, such as Ccl21a [30], Cd247 [31], Cd28 [32], and IL25 [33], showed increased expression. The gene expression of Pigr [34], Ltf, Lcn2 [35], S100a9 [36], and Ccl28 [37], which are involved in both innate and adaptive immune responses, decreased. Some genes involved in the development of the nervous system in the ko-wt group were significantly different from those of the wt-wt group. The gene expression of BDNF, Cdh3, Ccn5, and Folr1 decreased, and that of Prdm12 and Pcdha8 increased (Fig. 4G). These differences may lead to the delayed development of the nervous system. Figure 4H shows that the expression of genes related to neuronal signal transduction (Trpv4, Gabra6, Kcne2, and Kcnj13) decreased in the ko-wt group.
To explore whether the influence of LF-deficient feeding during lactation on the hippocampus extends into adulthood, we assayed differential gene expression in the hippocampus of adult mice. Figure 4I shows that there was no significant difference between KO-WT mice (mKO-WT, fKO-WT) and WT-WT mice (mWT-WT, fWT-WT) in immune-related differential expression genes (DEGs), but gene expression in female mice was significantly higher than that in male mice. Figure 4J shows that the expression of BDNF, Ccn5, and Cdn3 in KO-WT mice (mKO-WT, fKO-WT) did not differ from that in WT-WT mice (mWT-WT, fWT-WT). Similar to in 18-day-old mice, the expression of Prdm12 and Pcdha8 in adult female KO-WT mice (fKO-WT) was significantly higher than that in WT-WT mice (fWT-WT); however, this difference was not observed in adult male mice. Figure 4K shows that the expression of Trpv4, Gabra6, Kcne2, and Kcnj13 in KO-WT mice (mKO-WT, fKO-WT) did not differ from that in WT-WT mice (mWT-WT, fWT-WT), and the expression of Gabra6, Kcne2, and Kcnj13 was significantly higher in female mice than in male mice. When the mice reached adulthood, the difference induced by LF feeding deficiency in the hippocampus was decreased, but some genetic differences related to the nervous system development process persisted.
Lactation LF feeding-deficient adult mice still showed some differences in intestinal and intestinal microorganisms compared with the control group
To explore whether the effects of LF feeding deficiency during lactation on intestinal and intestinal microorganisms continue into adulthood, mice were fed normally until adulthood after weaning.
Intestinal development was evaluated using four groups (male: mWT-WT, mKO-WT; female: fWT-WT, fKO-WT) of mice drinking normal water which was different from that of 18-day-old mice. There were no significant differences in the body weight (Fig. 5A), small intestinal index (Fig. 5B), small intestinal density (Fig. 5C) and colon length (Fig. 5D) between WT-WT and KO-WT mice either male or female. The maltase/lactase values of the ileum in the mKO-WT and fKO-WT mice were significantly lower than those in the mWT-WT and fWT-WT mice, respectively (Fig. 5E, 5F, respectively). As shown in Fig. 5G, there was no difference between KO-WT and WT-WT mice in the crypt depth (mKO-WT vs. mWT-WT; fKO-WT vs. fWT-WT). Figure 5H shows that the expression of zo-1 and occludin in KO-WT mice was lower than that in WT-WT (mKO-WT vs. mWT-WT; fKO-WT vs. fWT-WT), but not significantly. Occludin expression was significantly lower in female mice than in male mice (fWT-WT vs. mWT-WT; fKO-WT vs. mKO-WT). Thus, the influence of LF feeding deficiency on the intestines of normal adult mice was weakened; however, female mice that consumed LF-free milk during lactation may have a greater risk of the intestinal barrier dysfunction in adulthood. In addition, the ileal maturity of LF-deficient mice was lower than that of LF-drinking mice, and we have previously shown that the maturity of the three segments of the small intestine of 18-day old mice drinking LF-free milk was lower than that of normal mice [17].
To examine the effects of different feeding methods during lactation on the composition of intestinal microflora in healthy adult mice (male: mWT-WT, mKO-WT; female: fWT-WT, fKO-WT), the cecal content was investigated using 16S rRNA gene sequencing. Alpha-diversity analysis revealed no difference in the Shannon and Chao indices between WT-WT and KO-WT mice (mKO-WT vs mWT-WT; fKO-WT vs fWT-WT, Fig. 5I). PCoA of the beta-diversity comparison revealed separation between the mKO-WT and mWT-WT groups, but the difference was not significant (P = 0.053). PCoA showed no separation among fKO-WT and fWT-WT groups on OUT level (Fig. 5J). These results suggest that there was no difference in alpha and beta diversity between lactating LF-deficient adult mice and normal mice. The phylum level was dominated by Firmicutes, Bacteroidetes, Desulfobacterota, Actinobacteria, and Patescibacteria (Fig. 5K). Patescibacteria was significantly higher in the mWT-WT group, and there was no difference between the fKO-WT and fWT-WT groups at the phylum level (Fig. 5L). At the genus level, the LEfSe results demonstrated a greater abundance of Lactobacillus, Lysinibacillus, Ileibacterium, Bifidobacterium, and Lachnospiraceae_UCG-001, whereas there was a lower abundance of Kurthia, Enterorhabdus, Candidatus, Saccharimonas, Eubacteriumxylanophilum_group, Monoglobus, and ASF356 in the mKO-WT group. Dubosiella, Lactococcus, Sporosarcina, Pseudogracilibacillus, and Eubacteriumbrachygroup were significantly higher in the fKO-WT group; Staphylococcus and Eubacterium_nodatum_group were significantly higher in the fWT-WT group (Fig. 5M).
Lactation LF feeding-deficient adult mice showed severe intestinal injury and intestinal microbial disorder based on DSS-induced colitis
Following LF deficiency during lactation, the intestinal development and composition of intestinal microorganisms in 18-day-old mice significantly differed from those of normal mice; however, these differences were covered and became not obvious after 9 weeks of normal feeding under physiological conditions. To examine whether the two types of mice exhibited the same pathological reaction, a DSS model was established in adult mice.
In the DSS colitis model mice, the body weight of mDKO-WT mice was lower than that of mDWT-WT mice at d 1–5, but this difference disappeared at d 6–7 (body weight decreased to 73.4% and 73.1% in mDWT-WT and mDKO-WT mice, respectively), possibly because of the longer treatment time with DSS (Fig. 6A). In female mice, the body weights of fDKO-WT mice were consistently lower than those of fDWT-WT mice from d 3 to 7, but the difference was not significant, and the body weights of fDWT-WT and fDKO-WT mice decreased to 77.3% and 74.6%, respectively (Fig. 6B). The disease activity index (DAI) score is an effective indicator of colon inflammation and damage. The DAI score of the mDKO-WT group was higher than that of the mDWT-WT group but not significantly (Fig. 6C). The DAI score of the fDKO-WT group was also higher than that of the fDWT-WT group, and the difference was significant at d 6 and 7 (Fig. 6D). The survival rate of the mDKO-WT group was lower than that of the mDWT-WT group, and there were no deaths in the two female mouse groups (Fig. 6E). There was no difference in colon length between the two groups of male mice but the colon length of fDKO-WT mice was significantly shorter than that of fDWT-WT mice (Fig. 6F). Hematoxylin and eosin (H.E) staining of the colonic tissue sections showed that DSS-treated colons had great histological damage (cellular infiltration, goblet cell depletion, and damage to the crypt architecture), but the difference between KO-WT and WT-WT mice was unclear. The crypt depth of fDKO-WT mice was lower than that of fDWT-WT mice, but the difference was not significant (Fig. 6G). Figures 6H–K shows that the expression of IL-1β and IL-10 in the colon in the fDKO-WT group was significantly higher than that in the fDWT-WT group. Thus, LF-deficient feeding of mice during lactation led to more severe DSS-induced weight loss (female), an increased DAI score (male and female), decreased survival (male mice), and colon shortening (female mice).
Ulcerative colitis is typically associated with dysbiosis of the gut microbiota. Alpha-diversity analysis showed that the Shannon index of the fDKO-WT group was significantly lower than that of the fDWT-WT group. The Chao index at the OTU level of fDKO-WT mice was significantly lower than that of fDWT-WT mice (Fig. 6L). These results suggest that alpha diversity was lower in LF-deficient female mice compared to those with normal milk; no difference was observed in male mice after acute DSS injury. PCoA revealed significant separation between the mDKO-WT and mDWT-WT groups at the OTU level, and the microflora of female LF-deficient DSS mice (fDKO-WT) differed significantly from that of the fDWT-WT group (Fig. 6M). At the phylum level, Bacteroidetes, Firmicutes, Proteobacteria, Campilobacterota, Desulfobacterota, Actinobacteriota, and Deferribacterota accounted for the majority of the bacteria (Fig. 6N). The relative abundance of Firmicutes was significantly higher in the mDKO-WT group, and of Proteobacteria was significantly higher in the mDWT-WT group. In the fDKO-WT group, the relative abundance of Campilobacterota was significantly higher, whereas that of Desulfobacterota and Cyanobacteria was significantly lower (Fig. 6O). At the genus level, the relative abundance of Lachnospiraceae_NK4A136 group, Ruminococcus_torques_group, Odoribacter, Romboutsia, Marvinbryantia, and Family_XIII_AD3011_group was higher in the mDKO-WT group than in the mDWT-WT group (Fig. 6P). As shown in Fig. S2B, the abundance of Ruminococcus_torques _group, Odoribacter, Romboutsia, and Family_XIII_AD3011_group was higher in the mDKO-WT group than in the mKO-WT group, indicating that mDKO-WT mice experienced a more substantial increase in these genera. The relative abundances of Leucobacter and Parvibacter were lower in the mDKO-WT group than in the mDWT-WT group (Fig. 6P). In female mice, the relative abundances of pathogenic bacteria such as Escherichia-Shigella, Enterococcus, Clostridium_sensu_stricto_1, Erysipelatoclostridium, Colidextribacter, and Akkermansia were higher in the fDKO-WT group than in the fDWT-WT group (Fig. 6P). The relative abundances of beneficial genera, such as Alloprevotella, Desulfovibrio, and many unclassified bacteria on genus level were lower in the fDKO-WT group than in the fDWT-WT group (Fig. 6P). Similarly, the abundance of Escherichia-Shigella, Enterococcus, Erysipelatoclostridium, Colidextribacter and Akkermansia were increased and Desulfovibrio was decreased in the DSS model (Fig. S2 D). fDKO-WT mice showed a more significant increase in Escherichia-Shigella, Enterococcus, Erysipelatoclostridium, Colidextribacter and Akkermansia and a serious decline in Desulfovibrio. LF feeding-deficient mice clearly showed more severe microbial disorders after DSS processing.