IgA dysfunction induced by the early-lifetime disruption of gut microbiota aggravates diet– induced metabolic syndrome in mice


 BackgroundLow-dose antibiotic contamination in animal food is still a severe food safety problem worldwide. Penicillin is one of the main classes of antibiotics being detected in food. Previous studies have shown that transient exposure of low-dose penicillin (LDP) during early life resulted in metabolic syndrome (MetS) in mice. However, the underlying mechanism(s) and efficient approaches to counteracting this are largely unknown.MethodsWild-type (WT) or secretory IgA (SIgA)-deficient (Pigr-/-) C57BL/6 mice were exposed to LDP or not from several days before birth to 30 d of age. Five times of FMT or probiotics (a mixture of Lactobacillus bulgaricus and L. rhamnosus GG) treatments were applied to parts of these LDP-treated mice from 12 d to 28 d of life. Bacterial composition from different regions (mucosa and lumen) of the colon and ileum were analyzed through 16S rDNA sequencing. Intestinal IgA response was analyzed. Multiple parameters related to MetS were also determined. In addition, germ-free animals and in vitro tissue culture were also used to determine the correlations between LDP, gut microbiota (GM) and intestinal IgA response.ResultsLDP disturbed the intestinal bacterial composition, especially for ileal mucosa, the main inductive and effective sites of IgA response, in 30-d-old mice. The alteration of early GM resulted in a persistent inhibition of the intestinal IgA response, leading to a constant reduction of fecal and caecal SIgA levels throughout the 25-week experiment, which is early life-dependent, as transfer of LDP-GM to 30 d germ-free mice only resulted in a transient reduction in fecal SIgA. LDP-induced reduction in SIgA led to a decrease in IgA+ bacteria and a dysbiosis in the ileal mucosal samples of 25 week wild-type but not Pigr-/- mice. Moreover, LDP also resulted in increases in ileal bacterial encroachment and adipose inflammation, along with an enhancement of diet-induced MetS in an intestinal SIgA-dependent manner. Furthermore, several times of FMT or probiotic treatments during LDP treatment are efficient to fully (for FMT) or partially (for probiotics) counteract the LDP-effect on both GM and metabolism.ConclusionsEarly-life LDP-induced enhancement of diet-induced MetS is mediated by intestinal SIgA, which could be (partially) restored by FMT or probiotics treatment.


Results
LDP treatment affected the early microbiota Study 1 design ( Figure 1A): Nine-week-old wild-type (WT) female and male C57BL/6 mice were randomly paired (1:1) and co-housed for 4 d (one estrous cycle for house mouse). After co-housing, females was separated from males and fed solely and pregnancies and due dates were speculated according to the body weight changes. Several days before birth, pups were exposed to LDP (10 mg/L, ~1.5 mg/kg body weight) 3 through their mothers or drinking water until 30 d of age (LDP, F and P groups) or not (Ctr group).
At 12d, 16d, 20d, 24d and 28d of age, the pups were gavaged with sterilized pre-reduced PBS (Ctr and LDP groups), and fecal microbiota (F group, pooled fecal samples from Ctr mice with a concentration 0 .3 g feces/mL PBS) or probiotics (P group, a bacterial mixture of Lactobacillus bulgaricus and L. rhamnosus GG (LGG) at ~10 8 /mL for each bacterium). For 12 day and 16 day mice, 50 μL and 100 μL liquid were gavaged, respectively; for the rest, 150 μL was gavaged. Pups with a bodyweight near the average level were selected for experiments before the rst gavage. The mice in each group came from at least three dams and housed in at least 2 cages (n = 2-3 per cage) to avoid the cage-effect. Pups were separated from their mothers at 21 d of age and female and male pups were housed together. At 30 d of age (2 d after the last gavage), mice were killed and bacterial and tissue samples were collected for analysis.
As female and male pups were housed together throughout the 30 day experiment in Study 1, their GM composition was analyzed together here. Quantitative PCR (qPCR) using 16S universal or ITS1 primers 20 (Table S1) showed no signi cant differences in the bacterial counts or fungal loads of the intestinal samples among all groups, suggesting that the overall microbial loads were not in uenced by LDP (Figures S1A and S1B). Nonetheless, signi cant differences were found in the bacterial compositions of the ileal mucosal (p < 0.01, Adonis test) and colonic luminal (P < 0.001 Adonis test) samples of the LDP and Ctr mice, as shown by weighted UniFrac distance measurements ( Figures 1B and 1C). Notably, there was no signi cant difference between Ctr and F and P mice ( Figures 1B and 1C), suggesting that both FMT and probiotics treatments e ciently restored the overall bacterial composition of these two intestinal regions. Similar trends were also found in the ileal lumen and colonic mucosa as shown by the difference of PC1 in Figures S1F to S1I.
We then analyzed the in uence of LDP on OTUs based on Linear discriminant analysis Effect Size (LEfSe). LDP treatment mainly resulted in a reduction in Lactobacillus and Candidatus Arthromitus (segmented lamentous bacteria (SFB)) in all examined intestinal regions (Figures 1D and S1C-S1E). It is also worth noting that a more broad difference was observed between the ileal samples of the LDP and Ctr mice than that in colonic samples (Figures 1D and S1C-S1E).
Next, we evaluated the e ciency of FMT and probiotics on restoring these two bacterial taxa. Based on the 16S rDNA sequencing data, FMT e ciently restored the intestinal levels of both SFB and Lactobacillus, while the probiotics treatment, as expected, only restored Lactobacillus ( Figures 1E and  1F). Moreover, three days after every gavage (i.e., one day before the next gavage) and at the end of (30 d) the study 1, the fecal and (or) intestinal levels of SFB and two treating probiotics (L. bulgaricus and LGG) were also determined by qPCR using species-speci c primers (Table S1) and similar results were obtained as compared to the 16S rDNA sequencing data ( Figure S1J). These results indicate that at least during the treatment period (from 12 d to 30 d of age), the intestinal SFB and (or) Lactobacillus levels were consistently affected by LDP, FMT and probiotic treatment.
In conclusion, LDP disturbed the intestinal bacterial composition, including a signi cant reduction in SFB and Lactobacillus, which was fully (for FMT) or partially (for probiotic) restored by several FMT and probiotic treatments.

LDP treatment persistently dampened intestinal IgA response
Early-life high-dose antibiotic treatment has been shown to transiently inhibit the intestinal IgA response in mice [17][18] . As intestinal SIgA is closely related to the development of MetS 21 , we then sought to determine the in uence of early-life LDP on intestinal IgA response. There were no signi cant difference in the fecal SIgA among all 14 d and 21 d age of mice, suggesting that the passive SIgA received from their mothers via breast milk was not in uenced by LDP ( Figure S2A). Nonetheless, in both 30-d-old male and female pups, when mice were able to actively generate intestinal SIgA 6 , LDP tended to decrease the SIgA levels in ileal, caecal and colonic contents, which was (partially) restored by FMT and probiotic ( Figures S2B, S2C, S2E, and S2F). However, because the small number (n = 2) of female pups in Ctr group of Study 1, we were not able to obtain statistically signi cant comparison results.
To further verify as well as determine the long-term effects of LDP on intestinal IgA response, we performed mouse Study 2 ( Figure 2A). In this study, all the LDP, FMT, and probiotic treatments during early life were the same to Study 1, but the experimental period was extended to 25 weeks. And as a western diet (45% energy from lard) has been shown to accelerate the metabolic effects of LDP 3 , the diet of mice was changed to a western diet at 6 weeks of age. In addition, a high-dose penicillin [HDP, 1.5 g/L (according to the recommended dose of penicillin for children 22 ) in drinking water from 17 to 24 d of age] group was added to study the difference in intestinal IgA response between therapeutic and subtherapeutic antibiotics. At 30 d of age, parts of pups were killed for the analysis of serum and intestinal (S)IgA and IgA-related lymphocytes. At 5, 10, 16, and 25 weeks of age, feces were collected to analyze the SIgA levels ( Figure 2A).
Consistent to the previous studies [17][18] , a decrease in fecal SIgA upon transient HDP treatment was found in 5-and 10-week-old mice ( Figure 2B). This effect was disappeared, however, in 16-and 25-week-old mice ( Figure 2B). Nonetheless, transient LDP reduced fecal SIgA throughout the 25-week experiment, showing a more persistent in uence on SIgA than HDP (Figure 2A). Determination of caecal SIgA at 30 d and 25 weeks of age also obtained consistent results ( Figure 2C). The reduction in intestinal SIgA upon LDP treatment was fully restored by FMT but only partially by probiotics (Figure 2A and 2C). In addition to the intestinal SIgA, a moderate reduction in serum IgA was observed in LDP and HDP mice at 30 d of age as compared to that in Ctr, F and P mice, which, however, was disappeared at 25 weeks of age ( Figure   S2D). The inhibition of intestinal IgA response by LDP was GM-and early life-dependent As intestinal IgA response is greatly dependent on the colonization and composition of GM 25-26 , we then sought to determine the correlation between LDP-induced changes in GM and inhibition of intestinal IgA response.
We rstly examined the in vitro IgA-inducing capabilities of the antigens derived from the feces of 21-dold mice in Study 1 as previously described 7,27 . Brie y, antigens separated from the feces through centrifugation were normalized and co-cultured with ileum tissue samples (containing no visible PPs) obtained from 8-week-old SPF C57/BL6 mice to study the effects of these antigens on ileal IgA production. After 2 d of cultivation, the ileum tissue produced signi cantly more IgA when co-cultured with antigens derived from the feces of Ctr and F mice than from LDP mice ( Figure 3A). A similar trend was evident for the antigens from the feces of P mice compared to the LDP mice but did not reach signi cance (P = 0.094). Moreover, the mRNA expression of the Jchain and IL-6 of the ileum tissue was also higher, while tumor necrosis factor (ligand) superfamily, member 13 (also known as APRIL) and 13b (also known as BAFF) were not in uenced, when cultured with antigens derived from the feces of Ctr and F mice than those from the LDP mice ( Figure 3B), suggesting that Ctr and F antigens may increase IgA production through enhancing the survival and activity of IgA+ PCs 19 .
These results suggest that LDP inhibit the intestinal immune response to GM, RNA-sequencing of the ileum from 30-d-old mice in Study 1 con rmed these results. Several biological functions related to IgA production and immune response to microbes that were downregulated in the LDP mice compared to the Ctr, F, and P groups, including PI3K signaling in B lymphocytes, B cell receptor signaling, dendritic cell maturation, and Th1 and Th17 pathways ( Figure 3D).
To further determine the correlation between LDP-induced changes in GM and IgA production, we transferred fecal microbiota from 30-d-old mice in Study 1 to even-aged male germ free (GF) mice, and measured the fecal SIgA levels in recipients. Three weeks post-transfer, LDP recipients (LDPR) exhibited signi cantly lower fecal SIgA levels compared to that of Ctr recipients (CtrR) and F recipients (FR), as well as the P recipients (PR) ( Figure 3C). The difference in the fecal SIgA levels of the LDPR, CtrR, FR, and PR lasted 6 weeks after the transfer ( Figure 2C). However, contrary to the donors, the fecal SIgA differences disappeared 12 weeks after the transfer ( Figure 3C), suggesting that the time factor (early life) also play an important role in the prolonged in uence of LDP-induced GM alteration on IgA responses. Moreover, no signi cant difference in serum IgA 12 weeks after transfer was observed among all recipients ( Figure  3C).
Together, these results indicated that LDP-induced changes in GM mediated the inhibition of intestinal IgA response and that early life is critical for the long-lasting effects of LDP on intestinal IgA response.
LDP induced persistent changes in the ileal microbiota in a SIgAdependent manner As intestinal SIgA is a master controller of the GM, we then sought to determine the effects of LDPinduced decrease in SIgA on GM. To this end, we performed Study 3 ( Figure 4A) using SIgA-de cient (Pigr -/-) mice generated by Pigr +/females and males. The LDP, FMT and probiotic treatments were the same to Study 1 but the experimental period was extended to 25 weeks. All mice were transferred to a western diet (45% energy from lard) at 6 weeks of age and lasted to the end of the experiment to accelerate the effects of LDP 3 . In addition, due to the insu cient number of male Pigr -/pups, only female Pigr -/pups were selected for the experiment at 10 days of age.
We rstly analyzed the fecal, serum, and caecal (S)IgA levels. Before weaning, no signi cant difference in fecal SIgA was observed between WT (Pigr +/+ and Pigr +/-) and Pigr -/pups while higher fecal SIgA levels were found in WT pups after weaning ( Figure S4A). In 25-week-old mice, there was no signi cant difference in serum IgA among all groups in both WT and Pigr -/mice ( Figure S4B). However, a moderate (P = .065) increase in serum IgA was found in Pigr -/as compared to WT females ( Figure S4D). In consistent to the results found in Study 1, LDP decreased the caecal SIgA levels in 25-week-old WT mice, which was (partially) restored by FMT and probiotics ( Figure S4C). In addition, no signi cant difference in serum and fecal (S)IgA was found between the WT mice of two genotypes (Pigr +/+ and Pigr +/-) ( Figure  S4E).
We then analyzed the in uence of LDP on GM in Study 3. No signi cant difference in fecal bacterial composition was observed between 21-d-old Pigr -/mice and their WT littermates ( Figure S4G), suggesting that the starting GM was not in uenced by the genotypes. At 25 weeks of age, WT LDP females showed a signi cantly different bacterial composition in the ileal but not colonic mucosa compared to WT Ctr females ( Figures 4B and S4H, Adonis test), which was in line with previous study showing a more closely interaction between SIgA and ileal microbiota 28 . Similar difference was found for WT males but did not reach signi cance (P = 0.082) ( Figure 4C, Adonis test). Moreover, FMT (moderately) restored this LDP-induced changes in ileal mucosal microbiota in WT mice ( Figures 4B and   4C). In contrast to WT mice, no signi cant difference in the ileal bacterial composition were found among all groups in Pigr -/mice ( Figure 4D, Adonis test), suggesting that intestinal SIgA may mediate this longterm effects of LDP on ileal microbiota.
To further verify this assumption, we performed mouse Study 4 using B cell-de cient (μMT) mice ( Figure  4A). Brie y, pan B cells (including PCs) isolated from the spleens, PPs, mesenteric lymph nodes and ileal and colonic lamina propria of 30 day LDP-treated (LDP, F, and P groups) or -free (Ctr) WT SPF mice were puri ed and adoptively transferred to even-aged μMT mice. After transfer, mice were housed 2-3/cages (2 cages for each group) to avoid the cage-effect. Twelve weeks after the transfer, recipients were killed for GM analysis. B cells were successfully colonized in the recipients, as considerable SIgA was detected in the caecal contents except for recipients received PBS treatment ( Figure S4F). As expected, we found a moderate (P = 0.143, Adonis test) difference in the ileal microbiota between Ctr and LDP recipients while a signi cant difference was found between LDP and F recipients ( Figure 4E, Adonis test). In addition, PBS recipients showed a signi cant difference in the ileal microbiota compared to all other B cell-received recipients ( Figure 4E, Adonis test).
Collectively, we demonstrated that transient LDP during early life can induce persistent changes in the ileal microbiota in an intestinal SIgA-dependent manner, which could be (partially) restored by FMT and probiotics.

LDP-induced reduction of SIgA increased bacterial encroachment and adipose in ammation
Intestinal SIgA can bind to the microbial antigens, preventing bacterial encroachment and adhesion to intestinal epithelial cells (IECs) as well as the translocation of bacteria and their metabolites, such as lipopolysaccharides (LPS) and agellin, therefore protecting hosts from in ammation 19 . We then sought to investigate the in uences of LDP-induced reduction in intestinal SIgA on bacterial encroachment and in ammation in 25-week-old mice from Study 3.
As expected, confocal microscopy, using mucus-preserving Carnoy xation 29 , showed that WT LDP females exhibited reduced distance between the bacteria and IECs, and showed exaggerated bacterial translocation into the in the distal ileum as compared to Ctr females, which was restored in FMT and P females ( Figures 5A and 5B). An enhanced bacterial translocation was observed in Pigr -/mice than WT mice while no signi cant difference was observed among all Pigr -/mice ( Figures S5C and S5D). In consistent to this, a decrease in IgA+ bacteria in ileum was observed in LDP females as compared to that in Ctr, F, and P females ( Figures 5C and 5D). This compromised control of bacteria by SIgA was accompanied with an increase in serum LPS in WT but not Pigr -/-LDP females as compared to that in Ctr, F, and P females with identical genotypes ( Figure 5E), while no signi cant difference in serum TNF-α and IL-6 was observed among all groups ( Figures S5M and S5N).
We then determined the in uence of LDP on adipose in ammation, a key factor involved in the development of insulin resistance 30 . A signi cant increase in macrophage in ltration in visceral adipose tissue, an indicator of adipose in ammation, was found in WT LDP females compared to that in WT Ctr females, as determined by immunostaining of F4/80 ( Figures 5F and 5G). Gene expression analysis using qPCR also demonstrated increases in the mRNA expression of TNF-α and IL-6 in the visceral adipose tissue of WT LDP females than that of WT Ctr, F, and P females ( Figures S6E and S6F). This LDP-induced adipose in ammation was disappeared in Pigr -/females ( Figure S6).
With respect to WT females, similar in uences of LDP and FMT treatments on bacterial encroachment and adipose in ammation were observed in WT males, while a moderate in uence of probiotics was found ( Figures S5 and S6). Collectively, transient exposure to LDP during early life caused increases in ileal bacterial encroachment and adipose in ammation in a SIgA-dependent manner, which could be (partially) counteracted by FMT and probiotic treatments.

Dysbiosis and bacterial encroachment enhanced the development of in ammation and MetS
Previous studies have shown that LDP can enhance the development of MetS, however, the mechanism(s) and the therapies were poorly determined 3,16 . Considering an essential role of intestinal SIgA in controlling the GM and MetS 21 , and the above results showing that intestinal SIgA mediated the LDP-induced increases in in ammation, a key factor contributing to MetS, we speculated that LDPinduced decreases in intestinal SIgA may also mediated the enhancement of MetS. To verify these assumptions, we measured the parameters related to MetS, including body weight (BW), fat mass (rate), serum hormones and glucose, and insulin sensitivity, in 25-week-old WT and Pigr -/mice in Study 3.
LDP enhanced the development of western diet-induced MetS in WT females, including increases in BW, fat and adipocyte tissues masses, adipocytes size, hepatic triglyceride, and serum glucose, insulin, and leptin levels ( Figures 6A-6F, S7C and S7D). In consistent to the elevated fast serum glucose and insulin, glucose and insulin tolerance tests also showed impaired glucose tolerance and insulin sensitivity in LDP females as compared to that in Ctr females ( Figures S7G and S7H). In addition, LDP decreased the serum levels of peptide YY (PYY), an GM-related anti-obesity peptide secreted by L cells in the gastrointestinal tract 31 ; while increased serum insulin-like growth factor-1 (IGF-1), a hormone secreted by the liver and intestine that can promote growth and preadipocyte proliferation and is also closely linked to obesity and the GM 32-33 ( Figure 6G). Similar effects of LDP on the metabolism were also observed in WT males (Figures 6 and S7). Moreover, FMT and probiotic treatments counteracted the LDP-effect on the development of MetS to different degrees: a more profound in uence was found upon FMT treatment compared to probiotics (Figures 6 and S7).
Lastly, in contrast to the WT mice, no signi cant difference in these parameters was found among all groups in their female Pigr -/littermates, demonstrating an indispensible role of intestinal SIgA in mediating LDP-induced effects on the development of MetS ( Figure S8).

Discussion
In this study, we showed that the LDP treatment during early life disturbed GM composition, resulting in a persistent inhibition of the intestinal IgA response and decrease in intestinal SIgA, which led to a disruption in the ileal bacterial composition, increases in bacterial encroachment and adipose in ammation, and an enhancement of diet-induced MetS. Several FMT or probiotic administrations during the LDP treatment period fully (FMT) or partially (probiotics) restored the intestinal IgA levels and prevented LDP-induced MetS.
The intestinal mucosal GM is more stable and resistant to disturbance factors than the luminal GM 34 . In line to this, we found that transient LDP exposure disturbed both the mucosal and luminal GM in 30-d-old WT mice while only the ileal mucosal GM was still altered by LDP in 25-week-old WT mice. This persistently in uence on ileal microbiota by LDP is mediated by the intestinal SIgA as the ileal microbiota was not affected by LDP in their even-aged SIgA-de cient (Pigr -/-) littermates. The uneven distribution of SIgA-producing ASCs and SIgA in the intestine may also contribute to the different in uences of LDP on GM between the ileum and colon as well between lumen and mucosa: (1) The distal small intestine (including the PPs) is the main inductive and effective site of IgA+ B cells, which is also the main place where the interaction between SIgA and GM taking place 28, 35 ; and (2) the intestinal mucosa has a higher SIgA level than the lumen 19 .
The intestinal IgA+ B cell repertoire can be greatly shaped by the routine of GM colonization in GF mice 36 . Although this has not been veri ed in the newborn mice, considering their similarities (both rstly experience a GM-induced intestinal IgA+ B cells differentiation and maturation), similar mechanisms may be also the same with newborns. This in uence on IgA repertoire can last for a long time, which may be partially attributed to the long-living property of IgA+ B cells [37][38] . In line to this, we found that transient changes in GM (as observed in Pigr -/mice) during early life led to a persistent inhibition of intestinal IgA response. In addition, our results also demonstrated that IgA+ B cells differentiation in newborn mice is more sensitive to GM alteration than that in GF mice, since the intestinal SIgA reduction continued until the end of the experiment (25 weeks) in the WT SPF mice (whose GM was disturbed since birth) but disappeared in conventionalized GF mice (received GM at 30 d old) 12 weeks after the GM transfer.
It has been previously demonstrated that transient early-life exposure of high-dose antibiotics induced a similar inhibition of intestinal IgA response and decrease in intestinal and serum (S)IgA to LDP [17][18] . However, in contrast to the persistent (at least for 25 weeks) effect of LDP on intestinal IgA response as observed in our study, high-dose antibiotics only show a transient (up to 10 weeks) effect on the fecal SIgA, which was also veri ed by our results ( Figure 2B). Through determination of cells related to intestinal IgA response using ow cytometry, we found a decrease in the total and RORγ+ Tregs within the SLP in mice of both 30 d and 25 weeks of age upon LDP treatment ( Figure S3). This decrease in RORγ+ Tregs, which has been previously shown to be inversely correlated with the intestinal SIgA 24 , may therefore contribute to the restoration of intestinal SIgA levels in HDP mice. In addition, the treating time of HDP may be also important. In this study, we treated the mice with HDP around weaning (from 17 d to 24 d of age). This is a critical time when a so-called "weaning action" has a fundamental in uence on the later-life intestinal immunity 23 . Elimination of the GM by high-dose antibiotics during this time can disturb this process and persistently impairing the intestinal immunity, including a reduction in Tregs 23 , as observed in our study ( Figure S3). Therefore, more works are warranted to further determine the time factor of high-dose antibiotics on the intestinal IgA response and its contribution to the observed difference in the intestinal IgA response between HDP and LDP. In addition, FMT and probiotics are both interventions with certain effectiveness to improve LDP-induced MetS. Although FMT is more e cient than probiotics, nding proper donors, as well as the acceptability and safety (especially for infants), are signi cant issues that need to be addressed. Probiotics are safer and more acceptable than FMT but exhibit lower e ciency. Additionally, the initiation and duration times, as well as the dosage level and frequency, are factors that should be considered. Moreover, future studies should focus on the differences between FMT and probiotics (such as the e ciency of SFB restoration in mice) and seek e cient and safe ways to treat Abx-induced MetS.

Animals
Wild-type speci c pathogen free animals Wild-type (WT) C56BL/6J mice purchased at 8 weeks of age (Vital River Laboratory Animal Technology. Co., Ltd., China) were randomly paired (1:1) after adapting for one week and fed a standard diet (#12450B, Research Diets). After cohousing for 4 days (one estrous cycle for house mouse), females were separated from the males and fed solely. During and after cohousing, pregnancies and due dates were monitored and calculated according to the body weight changes.
Pups (and their mothers) were randomly assigned to four groups: (1) low-dose penicillin (LDP)-free control group (Ctr) that received no antibiotics, (2) LDP-treated control group (LDP) that received antibiotics only, (3) FMT-treated group (F) that received LDP and fecal microbiota transplantation treatments, and (4) probiotics-treated group (P) that received LDP and probiotic cocktail treatments. Pups were separated from their mothers at 21 days. Pups in every group were from at least three dams and born within 24 hours. Pups in the Ctr and F groups were born within 24 hours. For the LDP, F and P groups, dams received antibiotics at a dose of 10 mg/L to deliver approximately 1.5 mg per kg body weight about one week prior to birth and were continuously maintained on penicillin 3 . Pups in the three groups were exposed to penicillin until 30 days of age either through their mother or through drinking water. After the LDP treatment, pups were divided by sexes and kept 2~3/cages. Pups in the F and P groups were gavaged with fecal microbiota from Ctr and probiotics, respectively at 12, 16, 20, 24 and 28 days of age. For the FMT group, feces were collected from pups in the Ctr group, pooled, and immediately placed in prereduced anaerobically sterilized PBS (0.3 g feces/mL PBS), homogenized under anaerobic conditions, settled with gravity for 2 minutes, and then the supernatant was transferred to pups in the F group. For the P group, pups were gavaged with prereduced anaerobically sterilized PBS containing Lactobacillus bulgaricus and Lactobacillus rhamnosus GG at ~10 8 /mL for each bacterium. For the Ctr and LDP groups, pups were gavaged with prereduced anaerobically sterilized PBS on the same days. For 12 day and 16 day mice, 50 μL and 100 μL liquid were gavaged, respectively; for the rest, 150 μL was gavaged. All pups were fed on normal diet (#D12450B, Research Diets) but transferred to a western diet (45% energy from lard, #D12451, Research Diets) at 6 weeks of age and allowed ad libitum access to food and water.

Germ-free animals
For the microbiota transfer to germ-free (GF) animals experiment, feces were collected from 30-day-old donors from each group and immediately placed in prereduced anaerobically sterilized PBS, homogenized under anaerobic conditions, settled with gravity for 2 minutes, and then the supernatant was transferred to even-aged GF C57BL/6 mice (n = 8 for each group, n (females) = n (males) = 4). After transfer, the conventionalized GF mice were housed in standard SPF conditions, and food and water were provided ad libitum.

SIgA-de cient mice
Pigr +/males and females, which can produce normal SIgA and therefore have similar GM to WT mice 6 , were used to generate SIgA-de cient pups. Speci cally, 8-week-old Pigr +/females and males were mated as above speci ed, genotypes of pups were identi ed at 10 days of age and only Pigr -/female and WT mice pups were chosen for the following experiment as speci ed above.

B cell-de cient (μMT) mice
Separation and transfer of pan B cells to B cell-de cient (μMT) mice was conducted as previously described with some modi cation 39 . Pan B cells (includes plasma cells) from the spleens, Peyer's patches, mesenteric lymph nodes and ileal and colonic lamina propria of 30 day LDP-treated or -free WT SPF mice were puri ed using negative selection (>90% purity, EasySep; StemCell Technologies) and injected intraperitoneally (~5.5 × 10 6 cells) in even-aged μMT mice.

Animal management and sampling
Mice were housed in standard speci c pathogen-free (SPF) conditions (12/12-hour light-dark cycle, humidity at 50 ± 15%, temperature of 22 ± 2°C), and food and water were provided ad libitum. The food used in this study was sterilized using radiation (25.0 kGy). Food intake was recorded every week. Body weight was recorded weekly. At the end of the experimental period, the mice were fasted for 12 hours, and plasma was collected by eyeball extirpation. The lumen contents of the distal ileum and proximal colon were collected by washing the lumen with sterilized PBS, and the mucosal samples were collected by scraping the intestinal wall with sterilized glass slides. Then, the samples were stored at -80°C for microbial analysis. The contents of the cecum were collected and stored at -80°C for the analysis of secreted IgA (SIgA). The weights of the liver, inguinal white adipose tissue (iWAT), epididymal white adipose tissue (eWAT) and mesenteric white adipose tissue (mWAT) were measured. Tissues were preserved at -80°C for gene expression analysis (for ileum, all fat and mesentery were removed, and Peyer's patches were excised from the ileum), and the liver, iWAT, colon and ileum were xed using 4% paraformaldehyde and used for hematoxylin-eosin (H&E) staining, immuno uorescence and immunohistochemistry analysis.
The guidelines of the institute regarding the care and use of laboratory animals were followed. This study was approved by the Animal Experiment Committee of the College of Food Science and Nutritional Engineering at China Agricultural University.

Ileum tissue culture
Mouse ileum was obtained and cultured as previously described with some modi cation 27  from the previous step were ltered twice through a 100-and 40-μmnylon cell strainer to obtain a singlecell suspension. PPs were mechanically disrupted in the harvest medium and the homogenate was ltered through a 70 μm cell strainer to obtain a single-cell suspension. Single-cells were stimulated with 50 ng/mL phorbolmyristate acetate and 1 μM Ionomycin in the presence of 5 μg/mL Brefeldin A.

Quantitative Real-time PCR (qPCR) Analysis
Total RNA was extracted using TRIzol TM reagent (Invitrogen) according to the manufacturer's instructions. Reverse transcription of the total RNA (2.5 μg) was performed with a high-capacity cDNA reverse transcription kit (Promega Biotech Co., Ltd). qPCR was run in triplicate for each sample and analyzed in a LightCycler 480 real-time PCR system (Roche). Data were normalized to the internal control β-actin and analyzed using the △△CT method. The expression of genes in iWAT, the liver, ileum and colon, as well as the bacterial and fungal load, were determined through qPCR (the related genes and primers used are shown in Table S1).
Quanti cation of the bacterial and fungal loads through qPCR was conducted as previously described 41 . Brie y, the total bacterial DNA was isolated from the samples with a QIAamp DNA Stool Mini Kit (Qiagen) following the manufacturer's instructions. For the isolation of fungal DNA, samples were suspended in 50 mM Tris buffer (pH 7.5) supplemented with 1 mM EDTA, 0.2% b-mercaptoethanol and 1000 units/ml of lyticase (Sigma), incubated at 37°C for 30 min to disrupt fungal cells as described 42 , prior to processing through the QIAamp DNA Stool Mini Kit (Qiagen). The DNA was then subjected to qPCR using a QuantiFast SYBR Green PCR kit (Bio-Rad) with speci c primers (Table S1).

Determination of Body Composition through MRI
MRI experiments were performed on 30-day-old and 25-week-old mice. The body composition was determined using MesoQMR instrument (Testniumag, Shanghai, China) with a 60 mm receiver and 0.5 ± 0.08 T magnetic eld strength. To obtain high resolution scanned MRI images, MRI measurements were performed on a 7.0 T Varian MRI instrument (Varian Medical Systems, Palo Alto, CA, USA) using a 40 mm volume and receiver coil at the Institute of Laboratory Animal Sciences, Chinese Academy of Medical Sciences. Prior to the experiments, the mice were initially anesthetized with 2% iso urane in a dedicated chamber. During the course of MRI, anesthesia levels were reduced to 1.5-1% in a combination of medical air and medical oxygen. The mice were positioned in the prone position, and respiratory-gated image acquisition was performed. MRI images of the mice were analyzed by Argus software.

Plasma Biochemical Parameters
The plasma biochemical parameters, including, glucose, cholesterol and triglyceride (TG) levels, were determined by a 3100 Clinical Analyzer (Hitachi High-Technologies Corporation, Japan).

Quanti cation of Serum Hormones and SIgA
Serum leptin, insulin, PYY, IGF-1, and IgA, as well as intestinal/fecal SIgA levels were determined using ELISA kits (Sangon Biotech, Shanghai, China) according to the manufacturer's recommendations.
Separation of antigens and determination of IgA to the speci c antigens were conducted as described with some modi cation 7,40 . Brie y, to prepare antigens from cecal contents of mice carrying GM composed of "pure or none Firmicutes", cecal contents of mice from each group were normalized by bacterial loads (determined by qPCR as previously described 20

Immuno uorescence and immunohistochemistry
Tissue sections for immunohistochemical testing were prepared on poly-L-lysine-pretreated coverslips.
Immunohistochemical staining was performed according to a standard protocol using antibodies against F4/80 at a 1:200 dilution. The samples were incubated overnight in a humidi ed chamber at 4°C. Secondary antibodies for immunohistochemical staining were purchased from Invitrogen. All images were acquired on an Olympus BX51 system and processed using ImageJ software, version 1.8.0. Immunostaining for IgA was conducted using FITC-conjugated anti-IgA antibody at a 1:500 dilution. The samples were incubated overnight at 4°C. Observations and analyses were performed with a Zeiss LSM 700 confocal microscope.

Localization of bacteria by uorescent in situ hybridization
The localization of bacteria by uorescent in situ hybridization (FISH) was conducted as previously described with some modi cations 29 . Brie y, distal ileum (second cm from the caecum) containing fecal material was placed in methanol-Carnoy's xative solution (60% methanol, 30% chloroform, 10% glacial acetic acid) for a minimum of 3 h at room temperature. The hybridization step was performed at 50 ℃ overnight with an modi ed EUB338 probe 43 (EUB338-II, 5'-GCAGCCACCCGTAGGTGT-3', with a 5' Texas Red label) diluted to a nal concentration of 10 mg/ml in hybridization buffer (20 mM Tris-HCl, pH7.4, 0.9 M NaCl, 0.1% SDS, 20% formamide). After washing, Mucin 2 primary antibody was diluted to 1:500 in block solution and applied overnight at 4 ℃. After washing, block solution containing anti-rabbit FITCconjugated secondary antibody diluted to 1:1000 was applied to the section for 2 h. After washing, slides were mounted using Prolong anti-fade mounting media (Life Technologies). Observations were performed with a Zeiss LSM 700 confocal microscope with software Zen 2011 version 7.1. This software was used to determine the distance between bacteria and the epithelial cell monolayer.

Histology
Tissues xed in 4% paraformaldehyde were cut into 5 μm sections after being embedded in para n.
Multiple sections were prepared and stained with hematoxylin and eosin (H&E) for general morphological observation.

GM Analysis
The microbial community of fecal, mucosal and lumen samples of colon and ileum were analyzed through the sequence of 16S rRNA gene V4 region. Brie y, Total genome DNA from samples was extracted using CTAB/SDS method. DNA concentration and purity was monitored on 1% agarose gels. According to the concentration, DNA was diluted to 1 ng/μL using sterile water. 16S rRNA gene V4 region were ampli ed used speci c primer for V4 region (515F-806R) with the barcode. All PCR reactions were carried out in 30 μL reactions with 15 μL of Phusion® High-Fidelity PCR Master Mix (New England Biolabs); 0.2 μM of forward and reverse primers, and about 10 ng template DNA. Thermal cycling consisted of initial denaturation at 98°C for 1 min, followed by 30 cycles of denaturation at 98°C for 10 s, annealing at 50°C for 30 s, and elongation at 72°C for 30 s. Finally 72°C for 5 min. Mix same volume of 1×loading buffer (contained SYB green) with PCR products and operate electrophoresis on 2% agarose gel for detection. PCR products were mixed in equidensity ratios. Then, mixture PCR products were puri ed with GeneJETTM Gel Extraction Kit (Thermo Scienti c). Sequencing libraries were generated using Ion Plus Fragment Library Kit 48 rxns (Thermo Scienti c) following manufacturer's recommendations. The library quality was assessed on the Qubit@2.0 Fluorometer (Thermo Scienti c).
At last, the library was sequenced on an Ion S5TM XL platform and 400 bp/600 bp single-end reads were generated.
Single-end reads was assigned to samples based on their unique barcode and truncated by cutting off the barcode and primer sequence. Quality ltering on the raw reads was performed under speci c ltering conditions to obtain the high-quality clean reads according to the Cutadapt (V1.9.1, http://cutadapt.readthedocs.io/en/stable/) quality controlled process. The reads were compared with the reference database (Silva database, https://www.arb-silva.de/) using UCHIME algorithm (UCHIME Algorithm, http://www.drive5.com/usearch/manual/uchime_algo.html) to detect chimera sequences, and then the chimera sequences were removed 44 . Then the Clean Reads nally obtained.
Alpha diversity is applied in analyzing complexity of species diversity for a sample through 6 indices, including Observed-species, Chao1, Shannon, Simpson, ACE, Good-coverage. All this indices in our samples were calculated with QIIME (V1.7.0) and displayed with R software (V2. 15.3). Beta diversity analysis was used to evaluate differences of samples in species complexity, Beta diversity on both weighted and unweighted unifrac was calculated by QIIME software (V1.7.0). Principal Coordinate Analysis (PCoA) was performed to get principal coordinates and visualize from complex, multidimensional data. A distance matrix of weighted or unweighted unifrac among samples obtained before was transformed to a new set of orthogonal axes, by which the maximum variation factor is demonstrated by rst principal coordinate, and the second maximum one by the second principal coordinate, and so on. PCoA analysis was displayed by WGCNA package, stat packages and ggplot2 package in R software (V2.15.3).

RNA-sequencing
Total RNA of the liver and ileum were extracted using TRIzol TM reagent (Invitrogen) according to the manufacturer's instructions. RNA degradation and contamination was monitored on 1% agarose gels; RNA purity was checked using the NanoPhotometer® spectrophotometer (IMPLEN, CA, USA); RNA concentration was measured using Qubit® RNA Assay Kit in Qubit®2.0 Flurometer (Life Technologies, CA, USA); RNA integrity was assessed using the RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies, CA, USA). Then a total amount of 3 μg RNA per sample was used as input material for the RNA sample preparations. Sequencing libraries were generated using NEBNext® UltraTM RNA Library Prep Kit for Illumina® (NEB, USA) following manufacturer's recommendations and index codes were added to attribute sequences to each sample. Brie y, mRNA was puri ed from total RNA using poly-T oligo-attached magnetic beads. Fragmentation was carried out using divalent cations under elevated temperature in NEBNext First Strand Synthesis Reaction Buffer (5X). First strand cDNA was synthesized using random hexamer primer and M-MuLV Reverse Transcriptase (RNase H -). Second strand cDNA synthesis was subsequently performed using DNA Polymerase I and RNase H. Remaining overhangs were converted into blunt ends via exonuclease/polymerase activities. After adenylation of 3' ends of DNA fragments, NEBNext Adaptor with hairpin loop structure were ligated to prepare for hybridization. In order to select cDNA fragments of preferentially 250~300 bp in length, the library fragments were puri ed with AMPure XP system (Beckman Coulter, Beverly, USA). Then 3 μL USER Enzyme (NEB, USA) was used with size-selected, adaptor-ligated cDNA at 37°C for 15 min followed by 5 min at 95°C before PCR. Then PCR was performed with Phusion High-Fidelity DNA polymerase, Universal PCR primers and Index (X) Primer. At last, PCR products were puri ed (AMPure XP system) and library quality was assessed on the Agilent Bioanalyzer 2100 system. The clustering of the index-coded samples was performed on a cBot Cluster Generation System using TruSeq PE Cluster Kit v3-cBot-HS (Illumia) according to the manufacturer's instructions. After cluster generation, the library preparations were sequenced on an Illumina Hiseq platform and 125 bp/150 bp paired-end reads were generated.
Raw data (raw reads) of fastq format were rstly processed through in-house perl scripts. In this step, clean data (clean reads) were obtained by removing reads containing adapter, reads containing ploy-N and low quality reads from raw data. At the same time, Q20, Q30 and GC content the clean data were calculated. All the downstream analyses were based on the clean data with high quality. Reference genome and gene model annotation les were downloaded from genome website directly. Index of the reference genome was built using Hisat2 (V2.0.5) and paired-end clean reads were aligned to the reference genome using Hisat2. We selected Hisat2 as the mapping tool for that Hisat2 can generate a database of splice junctions based on the gene model annotation le and thus a better mapping result than other non-splice mapping tools. featureCounts (V1.5.0-p3) was used to count the reads numbers mapped to each gene. And then FPKM of each gene was calculated based on the length of the gene and reads count mapped to this gene. FPKM, expected number of Fragments Per Kilobase of transcript sequence per Millions base pairs sequenced, considers the effect of sequencing depth and gene length for the reads count at the same time, and is currently the most commonly used method for estimating gene expression levels. Differential expression analysis of two conditions/groups (two biological replicates per condition) was performed using the DESeq2 R package (V1.16.1).

Statistical Analysis
All data reported in this paper are expressed as the means ± SEMs. The data were evaluated by one-way ANOVA, Wilcox, or Mann-Whitney U tests. All statistics were analyzed by SPSS software, and all analyses were performed with GraphPad Prism 7.

Supplementary Information
Additional le 1: Table S1. Primers used in this study. Table S2. Key resources table.
Additional le 2: Figure S1. Effects of the LDP treatment on microniota. Figure S2. The relative abundances of bacteria at phylum and genus level. Figure S3. The long-term effects of LDP on intestinal microbiota. Figure S4. Changes in the metabolism of mice. Figure S5. Changes of the serum hormones and adipose tissue.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
The RNA-seq and 16S rRNA gene sequencing data supporting this research are available on NCBI with accession number PRJNA577425. All other data that support the ndings of this study are available from the corresponding author upon reasonable request.

Competing interests
The authors declare no competing interests.