The generation and characterization of Ercc1^+/Δ7 and Ercc1^−/+ mice has been previously described (25). Ercc1^−/Δ7 mice were obtained by crossing Ercc1^+/Δ7 with Ercc1^−/+ mice of pure C57Bl6/J and FVB backgrounds to yield Ercc1^−/Δ7 with an F1 C57Bl6J/FVB hybrid background. Genotyping was performed as described previously (33). Wild-type littermates (C57Bl6J/FVB F1) were used as controls. Four-month-old and 18-month-old C57Bl6/J mice were purchased from Harlan (Horst, The Netherlands; only used in Figure 1).

Animals were housed in individual ventilated cages under SPF conditions. Experiments were performed in accordance with the Principles of Laboratory Animal Care and with Dutch legislation. This study was carried out in accordance with the recommendations of the Dutch Ethical Committee of Wageningen that approved the work. Blood was taken from mice being sacrificed, and serum was frozen in −80°C for later use. After mice (n = 4–6) were sacrificed, feces from colon was collected and snap-frozen. Distal ileum and proximal colon sections were isolated and fixed in Carnoy or snap-frozen in liquid nitrogen. BM, thymus, spleen, mesenteric lymph nodes (MLN), and Peyer’s patches (PPs) were isolated.

L. plantarum WCFS1, L. casei BL23, and B. breve DSM20213 were grown on MRS medium (Merck, Darmstadt, Germany) until stationary phase, frozen in glycerol, and stored in −80°C until use. Upon use, bacteria were thawed and 10× diluted in NaHCO3/PBS buffer. Around 2 × 10⁸ CFU in 200 μL were administered to mice by gavage, three times per week. Treatment of mice started at 6 weeks of age until 1 day before sacrifice at 16 weeks or until death.

Carnoy-fixed proximal colon sections were embedded in paraffin. Paraffin sections (5 μm) were attached to poly-l-lysine-coated glass slides (Thermo Scientific, Germany). After overnight incubation at 37°C, slides were de-waxed and rehydrated. Sections were stained with hematoxylin and eosin (H&E) and PAS/Alcian blue. Mucus layer thickness was measured using ImageJ software (NIH, MD, USA), as previously published (34). For detection of bacteria, tissue sections were used for fluorescence in situ hybridization (FISH), as previously published (8).

Microbiota composition in colonic content was analyzed by Mouse Intestinal Tract Chip (MITChip), as described previously (35). The data were normalized and analyzed using a set of R-based scripts in combination with a custom-designed relational database, which operates under the MySQL database management system. For the microbial profiling, the Robust Probabilistic Averaging signal intensities of 2667 specific probes for the 94 genus-level bacterial groups detected on the MITChip were used (36). Diversity calculations were performed using a microbiome R-script package (https://github.com/microbiome). Multivariate statistics, redundancy analysis (RDA), and principal response curves were performed in Canoco 5.0 and visualized in triplots or a principal response curves plot (37).

Total RNA was isolated from proximal colon (n = 3–6 per group) using the RNeasy kit (Qiagen) with a DNase digestion step according to the manufacturer’s protocol. Transcriptome analysis on individual samples was performed as previously described (8). The gene expression datasets were deposited in NCBI’s Gene Expression Omnibus (GEO) and are accesible through GEO Series accession number: GSE87368.

Single-cell suspensions of BM were obtained by crushing femurs, tibias, iliac crests, and sternum with mortar and pestle. BM cells were then filtered on a 40-μm cell strainer. A proportion of the BM cells was frozen for later use in in vitro cultures. Spleen, MLN, PP, thymus, and peritoneal cavity single-cell suspensions were obtained by gently pushing cells through a 40-μm cell strainer with a syringe. All cells were stained for extracellular markers and dead cells were identified with fixable live/dead stain (Ebioscience, San Diego, CA, USA), after which intracellular staining was enabled by fixing and permeabilizing cells with Fix/Perm buffer (Ebioscience) according to manufacturer’s instructions. Antibodies used for flow cytometric measurements are listed in Supplementary Table 1 in Data Sheet 1. All flow cytometric measurements were performed on a Canto II flow cytometer (BD Biosciences, Erembodegem, Belgium). FlowJo vX.07 software (Tree Star) was used for data analysis. Gating of all presented immune cell populations was based on single live cells.

Splenic cells were cultured at 10⁶ cells/mL for 4 days in the absence or presence of 5 μg/mL concanavalin A (ConA). Proliferation was measured by Ki-67 (Ebioscience). Supernatants were stored at −20°C. After thawing, levels of IL-2, IL-4, IL-6, IL-10, IL-17A, IFN-γ, and TNF were measured with the Cytometric Bead Array Th1/Th2/Th17 Kit (BD Biosciences), according to manufacturer’s instructions. Samples were acquired on a Canto II flow cytometer. Data were analyzed using FCAP Array version 3.0 (BD Biosciences) software.

Levels of IgM, IgG1, IgG2a, IgG2b, IgG3, IgE, and IgA were analyzed in serum using ProcartaPlex Mouse Antibody Isotyping Panel kit on the Luminex platform (Affymetrix, Santa Clara, CA, USA) according to the manufacturer’s instructions. Data were acquired on a BioPlex 200 (Bio-Rad, Hercules, CA, USA) and analyzed with BioPlex software (version 5.0, Bio-Rad).

Primary and secondary T cell-dependent (TD) immune responses against TNP-KLH were measured 7 days after primary i.p. immunization and 7 days after i.p. booster immunization. The primary immunization was performed at 8 weeks of age (TNP-KLH in alum), booster doses were injected at 12 weeks of age (TNP-KLH in PBS). Total and TNP-specific Ig subclasses were determined by sandwich ELISA as previously described (38).

Values are expressed as mean + SEM. Normal distribution of the data was confirmed using the Kolmogorov–Smirnov test. Statistical comparisons were performed using the two-sided Student’s t-test. Where non-Gaussian distribution was demonstrated, we applied the non-parametric Mann–Whitney U test. Where no equal variances were observed, we applied the two-sided Student’s t-test with Welch’s correction. Statistical comparisons for lifespan data were performed using the log-rank (Mantel–Cox) test. Statistical comparisons for serum immunoglobulins were performed using two-way ANOVA, with subsequent Bonferroni posttests. Values of p < 0.05 were considered to be statistically significant. Values between p > 0.05 and p < 0.10 were considered as a trend.

To assess the mucus barrier in normal and accelerated aging, we compared the proximal colon of 4-month-old (young) with 18-month-old (aged) C57Bl/6 mice, and of 6-week-old (young) with 16-week-old (aged) Ercc1^−/Δ7 mice. We observed that in aged C57Bl/6 and Ercc1^−/Δ7 mice, a thinner mucus layer was present, compared with young C57Bl/6 and Ercc1^−/Δ7 mice (Figure 1A). With ImageJ, we measured the thickness of the mucus layer. In young C57Bl/6 and Ercc1^−/Δ7 mice, a mucus layer of ~20 μm was present, whereas in normal and accelerated aged mice, a significantly thinner mucus layer of less than 10 μm was observed (p < 0.001; Figure 1B).

To determine the effects of the three selected bacterial strains in the young intestine, we analyzed proximal colon tissues of WT mice that were treated with L. plantarum WCFS1, L. casei BL23, or B. breve DSM20213 for 10 weeks. No change in tissue integrity (H&E) or mucus layer (PAS/Alcian Blue) was observed in the colon after supplementation with bacterial strains (Figure 1C).

Because the mucus layer declines with age, we questioned whether bacterial supplementation of Ercc1^−/Δ7 mice prevents the decline in mucus barrier. Colon tissue of 10-week treated Ercc1^−/Δ7 mice was checked for tissue integrity and mucus layer thickness. In contrast to our findings in WT mice, bacterial supplementation had significant effects on tissue integrity and the mucus layer. In Ercc1^−/Δ7 mice supplemented with L. plantarum, the colon showed a thicker mucus layer than their controls (Figure 1D). L. plantarum supplementation completely prevented age-related decline in the mucus layer compared with controls (p < 0.001; Figure 1E), resulting in a mucus thickness comparable to young WT mice. Spatial compartmentalization of bacteria in the colon was improved after L. plantarum supplementation (Figure 1F), as demonstrated by FISH analyses. On the contrary, Ercc1^−/Δ7 mice supplemented with L. casei or B. breve showed loss of tissue integrity (Figure 1D). No difference in mucus thickness was observed after supplementation with L. casei (Figure 1E). B. breve supplementation resulted in a deteriorated mucus layer and a loss in mucus thickness (p < 0.001; Figure 1E). B. breve supplementation also resulted in less spatial compartmentalization of bacteria in the colon of Ercc1^−/Δ7 mice (Figure 1F).

Collectively, these data show that L. plantarum supplementation improves the mucus layer in the aged (but not young) colon. In addition, supplementation with L. casei or B. breve exacerbates the age-related decline of mucus barrier in the colon.

As we introduced bacteria by bacterial supplementations into the intestinal microbial community, we investigate whether changes in the microbiota composition were underlying the observed changes in the mucus barrier of Ercc1^−/Δ7 mice. Microbiota composition was determined by performing 16S rRNA gene microbiota profiles of colonic content. The bacterial supplementations did not significantly alter microbial diversity nor richness (data not shown).

Redundancy analysis showed that 10.1% of the total variability of the gut microbiota can be related to the bacterial supplementations (Figure 2). No statistical significance was established. The first ordination axis explained 4.9% of the variability and separated Ercc1^−/Δ7 mice supplemented with either of the three bacterial strains from the control Ercc1^−/Δ7 mice. The second ordination axis explained 3.6% of the variability but did not result in a separation between groups. The third ordination axis explained an additional 1.6% of the variability (data not shown).

To assess whether significant changes in the microbial genus-like bacterial groups existed after different bacterial supplementations in Ercc1^−/Δ7 mice, we performed the Wilcoxon test. Subdoligranulum was higher in mice supplemented with L. casei (p < 0.05), whereas it tended to be higher in mice supplemented with B. breve (p = 0.05), as compared with control mice (Figure 3). Akkermansia muciniphila tended to be less present (p = 0.06) in mice supplemented with L. plantarum compared with control mice. Eubacterium plexicaudatum and a close relative to Anaerostipes caccae tended to be higher (p = 0.06) in Ercc1^−/Δ7 mice supplemented with L. casei.

These data demonstrate some differences in microbial species between control-treated Ercc1^−/Δ7 mice and Ercc1^−/Δ7 mice treated with bacterial supplementations.

To understand the mechanisms by which bacterial supplementation changes the mucus barrier, we performed transcriptome analysis on the proximal colon of Ercc1^−/Δ7 mice. Gene expression microarrays on total proximal colon samples from Ercc1^−/Δ7 mice treated with bacterial supplementations or control treatment revealed relatively low numbers of differentially expressed genes: 84 by L. plantarum, 238 by L. casei, and 384 by B. breve. Only a few genes were overlapping between two or three different bacterial supplementations, whereas most of the differentially expressed genes were distinctly regulated by one of the treatments (Figure 4).

Several growth- and immune-related genes were differentially expressed after bacterial supplementation. Apolipoprotein (APO) A-1, APOA-4, suppressor of cytokine signaling (SOCS) 3, and toll-like receptor (TLR) 4 were upregulated more than 1.2-fold after L. plantarum supplementation (Data Sheet 2 in the Supplementary Material). Several immunoglobulin variable genes and TLR13 were upregulated after administration of L. casei, whereas defensin 40β was 1.3-fold downregulated. Defensin 24α, amphiregulin, and keratinocyte growth factor 7 (FGF7) were upregulated more than 1.4-fold after administration of B. breve, while TLR6, TLR7, and CCL3 (MIP-1α) were more than 1.2-fold downregulated (Data Sheet 2 in the Supplementary Material). Remarkably, we found no significant up- or downregulation in any mucin.

Because we found relatively low numbers of differentially expressed genes, we applied a gene set enrichment analysis (GSEA) (39) to gain insight into the regulated pathways by bacterial supplementations. Upstream regulators that can explain the observed changes in gene expression were identified using Ingenuity Upstream Regulator Analysis.

Gene set enrichment analysis revealed that L. plantarum supplementation significantly enhanced several processes involved in growth and cell cycle, and immunity (Supplementary Table 2 in Data Sheet 1), such as “Type II Interferon Signaling,” “VIP pathway,” and “IL8/CXCR1 pathway.” Interestingly, in the top-10 of upregulated pathways, three pathways were linked to DNA repair: the “Fanconi Pathway,” “ATRBRCA pathway,” and the “Fanconi anemia pathway.” Several growth factors were activated after L. plantarum supplementation: leptin, epidermal growth factor (EGF), platelet-derived growth factor (PDGF) BB, early growth response protein (EGR) 1, and insulin-like growth factor (IGF) 1 (Table 1). Inflammatory cytokines (IFN-γ, IL-1β, IL-4, TNF) and CD40L (CD154) were activated in colon of mice supplemented with L. plantarum, compared with colon of mice supplemented with control treatment.

Lactobacillus casei supplementation enhanced several processes involved in growth and cell cycle, like “Mitotic G1-G1 S Phases,” “DNA replication,” “Synthesis of DNA,” and “G1 S transition” (Supplementary Table 2 in Data Sheet 1). In addition, the “NOD-like receptor signaling pathway” was enhanced after L. casei supplementation, as well as and the “Unfolded protein response” (UPR), indicated endoplasmatic reticulum (ER) stress. Upstream regulators resistin-like β (RTNLB; activated) and GATA3 (inhibited) were regulated in the colon of mice supplemented with L. casei (Table 1).

Several metabolic pathways were enhanced in colon of B. breve-supplemented mice (Supplementary Table 2 in Data Sheet 1). Of note, “Protein folding” was upregulated. In contrast to L. plantarum and L. casei supplementation, B. breve supplementation significantly inhibited several processes involved in immunity, such as “IL2 STAT5 pathway,” “Immunoregulatory interactions between lymphoid/non-lymphoid cells,” “Type II Interferon signaling,” “IL4 2 pathway,” and “IL6 7 pathway” (Supplementary Table 3 in Data Sheet 1). B. breve supplementation activated EGF and inhibited fibroblast growth factor (FGF) 2. IgG, GATA2, and IL-2-inducible T cell kinase (ITK) were inhibited in colon of mice supplemented with B. breve. In line with GSEA, IFN-γ was inhibited as well after B. breve supplementation.

These data indicate that immune pathways in the colon are enhanced by L. plantarum and L. casei, but are inhibited by B. breve supplementation.

Based on the regulation of immune genes by bacterial supplementations, we tested whether the distribution of immune cells was altered in mucosal immune organs of Ercc1^−/Δ7 mice.

First, we evaluated changes in distribution of immune cells in PPs and MLN. B cell frequencies were reduced in PP and MLN (p < 0.05) after L. casei supplementation in Ercc1^−/Δ7 mice (Figure 5A). By contrast, frequencies of T cells were increased in PP (p < 0.01) and MLN (p < 0.05; Figure 5B). The frequencies of regulatory T (Treg) cells in MLN were increased after L. plantarum and L. casei supplementation (p < 0.05; Figures 5C,D). No changes in distribution of B and T cells were observed upon bacterial supplementation in WT mice, except for a tendency to decreased Treg cells after L. casei supplementation (p = 0.09; Supplementary Figure 1 in Data Sheet 1).

Next, we assessed distribution of immune cells in the spleen. We noted that the relative spleen weight increased after L. casei supplementation in Ercc1^−/Δ7 mice (Supplementary Figure 2A in Data Sheet 1). Absolute numbers of spleen cells were not affected by bacterial supplementations (data not shown). Splenic B cell frequencies tended to be decreased after L. casei supplementation (p = 0.06; Figure 6A), but no changes in T cell frequencies were observed (Figure 6B). Treg cell frequencies in the spleen were increased after L. casei supplementation in Ercc1^−/Δ7 mice (p < 0.05; Figure 6C).

Increased frequencies of CD11b⁺Ly6G⁻CD68⁺Ly6C^hi monocytes (p < 0.01; Figures 6D,E) and a tendency to increased frequencies of CD11b⁺CD68^intLy6C^intLy6G⁺ neutrophils were observed after L. casei supplementation (p = 0.07; Figure 6F). In addition, the proportions of CD3⁺CD4⁺RORγt⁺ Th17 cells (Supplementary Figure 2B in Data Sheet 1) were increased after L. casei supplementation (p < 0.05; Figure 6G). A 4-day culture of splenocytes stimulated with concanavalin A (ConA), also showed increased IL-17A production (p < 0.01; Figure 6H) and a decreased T cell proliferation in splenocytes derived from L. casei-treated mice (p < 0.01; Figure 6I). None of these changes were observed in Ercc1^−/Δ7 mice treated with L. plantarum or B. breve, and in WT mice treated with each of the bacterial supplementations (Supplementary Figure 3 in Data Sheet 1). These data suggest that L. casei, in contrast to L. plantarum and B. breve, raises several inflammatory markers in Ercc1^−/Δ7 mice.

We subsequently investigated the development of B cells and myeloid cells in BM and of T cells in thymus of Ercc1^−/Δ7 mice, as the observed changes in cell distribution in PP, MLN, and spleen might be explained by an altered migration or production. Absolute numbers in the BM were unchanged after bacterial supplementation (data not shown). In the BM, we observed significantly higher Lin⁻CD117^hiCD11c⁻CD135⁻CD16/32⁺ granulocyte–monocyte precursor (GMP), CD11b⁺Ly6G⁺ neutrophil, and Ly6C^hiCD31⁻ monocyte frequencies after L. casei supplementation (Figures 7A–C). Frequencies of total CD19⁺CD45R⁺ B-lineage cells were decreased after L. plantarum (p < 0.05) and L. casei supplementation (p < 0.001), but not after B. breve supplementation (Figure 7D). We observed a reduction in all B-lineage subsets, except pro-B cells, after L. casei and L. plantarum supplementation (data not shown). In thymus, only L. casei supplementation caused changes in cell distribution, with significantly reduced CD3⁻CD4⁺CD8⁺ double-positive (DP) cell numbers (Figures 7E–H).

In WT mice, we found no effect of bacterial supplementations on distribution of immune cells in BM or thymus, except for B-lineage cells after L. plantarum and L. casei supplementation (Supplementary Figure 4 in Data Sheet 1).

The accelerated aging of Ercc1^−/Δ7 mice enabled us to assess the potential life-extending properties of the bacterial strains. No significant change in lifespan of Ercc1^−/Δ7 mice was observed after lifelong supplementation with L. plantarum or L. casei (Supplementary Figure 5 in Data Sheet 1).

Because L. casei supplementation led to decreased B cell proportions in several immune organs of Ercc1^−/Δ7 mice, we tested whether serum antibody titers in Ercc1^−/Δ7 mice were altered. Total IgG1 and IgG2b (but not IgG2a, IgG3, IgE, and IgA) titers were significantly increased after L. casei supplementation (Figure 8). L. plantarum and B. breve supplementation did not significantly change titers of any Ig subclass.

To test whether changes in immune cell distribution also impact immune competence, we analyzed the B cell response of Ercc1^−/Δ7 mice to the T cell-dependent antigen TNP-KLH. Specific anti-TNP-KLH Ig titers of the three tested isotype classes (IgM, IgG1, IgG2a) after primary and booster immunization were consistently higher after L. plantarum and L. casei supplementation (Figure 9). In particular, IgG1 titers after booster immunization increased in both L. plantarum- and L. casei-supplemented mice compared with control-treated mice (p < 0.001).

From these findings, we conclude that L. plantarum and L. casei enhance T cell-dependent B cell responses in Ercc1^−/Δ7 mice.