Rapamycin reduces lymphocytes in LNs and spleen. To investigate the immunological effects of rapamycin within secondary lymphoid organs, C57BL/6 mice received rapamycin (5 mg/kg/day i.p.) (43) and were characterized after 3 (early), 7 (intermediate), and 30 (late) days of treatment. We assessed cell populations in peripheral LNs (pLNs), mesenteric LNs (mLNs), and spleen, focusing on CD4+ and CD8+ T lymphocytes and Tregs via flow cytometry. Reductions in pLN weight and overall cell counts were observed at all 3 time points, consistent with known rapamycin inhibitory effects on cell proliferation and metabolism (Supplemental Figure 1, A–D; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.186505DS1). Rapamycin reduced CD4+ and CD8+ T lymphocytes and Foxp3+ Tregs at all time points in the pLNs (Figure 1, A and D), demonstrating substantial and sustained immunomodulatory effects. In the mLNs, rapamycin reduced CD4+ and CD8+ T lymphocytes and Foxp3+ Tregs after 7 and 30 days but not on day 3 (Figure 1, B and D). In the spleen, decreases in CD4+ T cells and Foxp3+ Tregs were noted on days 7 and 30, with decreased CD8+ T cells only on day 7 (Figure 1, C and D). Rapamycin decreased the Treg/non-Treg ratio in the spleen but not in LNs (Supplemental Figure 1, E–G), suggesting an uneven effect on various T cell populations. Overall, these data demonstrate that rapamycin markedly reduced lymphocytes and Treg cell counts in LNs and spleen, and this is a manifestation of its immunosuppressive effect. Notably, rapamycin showed more pronounced effects in mLNs and spleen at later time points, while earlier changes were primarily observed in pLNs. This suggests that rapamycin immunomodulatory effects are both time dependent and site specific.

Rapamycin elicits significant changes in immune cell populations in LNs and spleen. (A–C) Flow cytometry for the total number (CD45+ cells), CD4+ T cells, CD8+ T cells, and Foxp3+ Tregs (Foxp3+CD4+) in pLN (A), mLN (B), and spleen (C) after 3, 7, and 30 days of rapamycin treatment. (D) Heatmap depicts changes in cell numbers relative to the control for pLN, mLN, and spleen after rapamycin treatment versus no drug control; red represents “increased,” white represents “unchanged,” and blue represents “decreased.” There were 5 mice/group. One-way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Time-dependent effect of rapamycin on LN architecture and Treg distribution. The architecture and distribution of specific immune cells within the LN microenvironments, including the CR and HEVs, are key in fostering immune tolerance and suppression (37). We conducted quantitative IHC of these LN domains to characterize rapamycin spatiotemporal effects. Rapamycin rapidly induced early proinflammatory changes by decreasing the La4/La5 in pLN HEVs by day 3. This extended to pLN CR by day 7 and diminished to undetectable levels by day 30 (Figure 2, A and E, and Supplemental Figure 2, A and B), indicating an early proinflammatory effect. In mLNs, rapamycin increased La4/La5 on day 30, with no changes on days 3 or 7 (Figure 2, B and F, and Supplemental Figure 2C), indicating a late protolerogenic effect.

Effects of rapamycin on LN cell content, cell distribution, and structure. (A and B) IHC of CR and HEV La4/La5 on days 3, 7, and 30 of pLN (A) and mLN (B). (C and D) IHC of CR and HEV Foxp3+ Tregs on days 3, 7, and 30 of pLN (C) and mLN (D). (E and F) Heatmaps depict changes in expression with rapamycin relative to control in pLN (E) and mLN (F). There were 1–3 LNs/mouse, 3 mice/group, 2–3 sections/LN group, and 7–30 fields/slide. Each dot in the graph represents 1 field. One-way ANOVA. *P < 0.05; **P < 0.01, ***P < 0.001, ****P < 0.0001.

Using IHC, in pLNs, rapamycin increased Foxp3+ Tregs by day 30, especially in the CR, without alterations at earlier time points (Figure 2, C and E, and Supplemental Figure 2D). Flow cytometry analysis of total pLN Treg percentages corroborated these findings, demonstrating stable levels through days 3 and 7, followed by an increase at day 30 (Supplemental Figure 1H), indicating a late protolerogenic effect. In mLNs, rapamycin reduced Foxp3+ Tregs on day 3 but not at later time points (Figure 2, D and F, and Supplemental Figure 2E). Analysis of total mLN Foxp3+ Treg percentages revealed an increase by day 30 (Supplemental Figure 1I), suggesting a temporal transition in rapamycin effects from early localized proinflammatory to later protolerogenic states. Overall, both pLNs and mLNs exhibited early proinflammatory states, characterized by decreased La4/La5 or reduced Tregs followed by a late protolerogenic environment, marked by either increased La4/La5 or elevated Tregs. Notably, the data indicate that day 7 marked a transition from proinflammatory to protolerogenic regulation by rapamycin.

Rapamycin increases laminin α5 and decreases La4/La5 in LNSCs. Given the effect of rapamycin on LN architecture via La4/La5, we next sought to determine whether LNSC expression of laminin α4, laminin α5, or both directly mediated these effects after 7 days treatment, given that this time point marked the critical transition from proinflammation to protolerance. Flow cytometry was used to quantify laminin α4 and α5 levels in live CD45– LNSCs, including FRCs, BECs, and LECs. In pLNs, rapamycin upregulated laminin α5 in all groups with little influence on laminin α4, leading to decreased La4/La5 (Figure 3, A–G). In mLNs, treatment with rapamycin similarly decreased the La4/La5 in LNSCs by upregulating laminin α5 expression (Figure 3, H–M). Taken together, this cell type–specific analysis provided detailed insight into how rapamycin modulates proinflammatory responses by selectively increasing laminin α5 expression in LNSCs, hence decreasing La4/La5. The flow cytometry analysis revealed a subset-specific increase in laminin α5 expression within LNSCs (Figure 3), with no notable changes in laminin α4 levels. By contrast, IHC provided a broader architectural analysis of La4/La5 across all cell types and in specific regions (Figure 2). This broader approach, which does not differentiate between cell types, reflects contributions from both preexisting laminin expression along with laminin expression from nonstromal cells, diluting the changes in laminin α5 observed in stromal subsets.

Rapamycin increases laminin α5 and decreases the La4/La5 in LNSCs. (A) Flow gating of CD45– cells for FRCs (CD31–gp38+), BECs (CD31+gp38–), and LECs (CD31+gp38+) for laminin α4 (La4) and laminin α5 (La5). (B–M) Mean fluorescence intensity (MFI) and flow plots show La4, La5, and La4/La5 ratios in: pLN (B–G), FRCs (B and C), LECs (D and E), BECs (F and G); mLN (H–M); FRCs (H and I); LECs (J and K); and BECs (L and M). There were 3 mice/group. Two-tailed t test. *P < 0.05, **P < 0.01, ****P < 0.0001.

Laminin α5 is responsible for rapamycin-induced changes in LN architecture. Given the effect of rapamycin on pLNs La4/La5 on day 7 and the important role of FRCs in laminin expression, we further employed 2 laminin-KO strains to assess if laminin α4 or α5 expression was mediated by FRCs. FRC–laminin α4–KO (FRC-La4-KO) mice (Pdgfrb-Cre+/– × La4fl/fl) have the laminin α4 gene deleted in FRCs. At baseline, these mice showed decreased laminin α4 levels in both pLNs and mLNs compared with WT mice (Figure 4, A and D). Administration of rapamycin did not affect the expressions of laminin α4, laminin α5, and the La4/La5 in pLNs of the FRC-La4-KO mice (Figure 4, A–C, and Supplemental Figure 3A). In mLNs, rapamycin increased laminin α5 around HEV without affecting laminin α4 in FRC-La4-KO, leading to a reduced La4/La5 (Figure 4, E and F, and Supplemental Figure 3B). These results indicate that, even in the absence of FRC laminin α4, rapamycin can still upregulate laminin α5 expression, suggesting that FRC-derived laminin α4 is not required to mediate the LN architectural change under the influence of rapamycin.

Rapamycin regulates laminin α5 in FRC-La4-KO and FRC-La5-KO mice. IHC showing (A and H) laminin α4; (B and G) laminin α5; (C and I) La4/La5 in pLN; (D and K) laminin α4; (E and J) laminin α5; (F and L) La4/La5 in mLN for WT, untreated, and rapamycin-treated FRC-La4-KO mice (A–F) and FRC-La5-KO mice (G–L) on day 3. There were 1–3 LNs/mouse, 3–5 mice/group, 2–3 sections/LN group, and 7–30 fields/slide. Each dot in the graph represents one field. One-way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

FRC-La5-KO mice (Pdgfrb-Cre+/– × La5fl/fl) have the laminin α5 gene deleted in FRCs. At baseline, these mice showed decreased laminin α5 in both pLN and mLN compared with WT mice (Figure 4, G and J). Administration of rapamycin did not affect the expression of laminin α4 or α5 in the pLN, resulting in an unchanged La4/La5 in FRC-La5-KO mice (Figure 4, G–I, and Supplemental Figure 3A). In the mLN, there was a differential increase in both laminin α4 and α5 (Figure 4, J and K), leading to a decrease in the La4/La5 (Figure 4L and Supplemental Figure 3B). Since the laminin α5 gene is deleted in FRCs but not in other cells, the increased laminin α5 levels after rapamycin treatment suggest upregulation of laminin α5 expression by non-FRC cell types, such as LECs and BECs. Flow cytometry results (Figure 3) support this conclusion. Overall, the results show that rapamycin differentially regulates laminin α4 and α5, altering LN architecture. This effect is not exclusively mediated by FRC-derived laminin, as LECs and BECs also play crucial roles in these rapamycin-induced changes.

Time-dependent effect of rapamycin on gut microbiota composition and metabolic capacity. We next investigated rapamycin effects on the gut microbiome because of its critical role in interfacing with the immune system. Whole-community metagenomic sequencing of intraluminal fecal contents was performed at a sequencing depth of 41.1 ± 14.5 (mean ± SD) million reads after quality control steps per sample (Supplemental Table 2A). Taxonomic composition was estimated using the comprehensive mouse microbiota genome catalog (44) (Supplemental Table 2B). After rapamycin treatment, no significant differences in gut microbiota diversity were observed at days 3 and 7 (Figure 5A). However, by day 30, there was a decrease in community diversity (P < 0.05) and a shift in microbial composition and structure, characterized by increased Bacteroides to Firmicutes relative abundance ratio (B/F ratio) (Supplemental Figure 4A). The B/F ratio on day 3 was 0.29 ± 0.26 (mean ± SD), increased to 0.78 ± 0.53 on day 7, and reached 1.19 ± 1.13 by day 30, indicating a sustained effect of rapamycin on the structure and composition of the gut microbiota. Compared with the control group, no significant changes in taxonomic groups were observed on day 3. In Bacteroides, Duncaniella showed the most pronounced increase at day 7 and Muribaculum was most increased at day 30, in response to rapamycin treatment (Supplemental Figure 4B). In Firmicutes, Lachnospiraceae, Lactobacillales, and Christensenellales showed the greatest decrease under rapamycin treatment at 30 days (Supplemental Figure 4C). Using differential abundance analyses, sporadic alterations were noted on day 7 including Bacteroidale (i.e., Duncaniella sp., Muricubaculum intestinale) and Firmicutes (i.e., Eubacterium) (Supplemental Figure 4D and Supplemental Table 2C). By day 30, there was an increase in Bacteroidales (including Muribaculaceae, Bacteroides, Actinobacteria), and a decrease in Firmicutes (such as Lachnospiraceae, Oscillibacter, Lawsonibacter, Eubacterium) (Supplemental Figure 4E). These taxonomic groups drive the observed changes in composition and structure of the gut microbiota at different times and between groups (Figure 5B). These effects, starting at day 7 and most pronounced by day 30, highlight a sustained and potent influence of rapamycin on the microbiome and correspond to the same time frame for changes in LN immune architecture and content.

Rapamycin alters gut microbiome and metabolic potentials. (A–C) Gut microbiome characterization of rapamycin treated and no-treatment control for within-community diversity by total number of taxa and Shannon diversity index (A); community β-diversity with PCA (B); and gut microbiome functional pathway abundance in copies per million (CPM) difference between control and rapamycin groups (C). The height of the stacked bar represents CPM of associated MetaCyc pathways contributed by different taxa in control (green) or rapamycin (red). There were 3–5 mice/group and 1 stool sample collected/mouse/time point. (D) Intestinal IHC of Foxp3+ Tregs after 3 days, 7 days, and 30 days of rapamycin treatment. There were 3 mice/group as a block, 1 piece of intestine, 2–3 sections/block on a slide, and 7–30 fields/slide. Each dot in the graph represents 1 field of view from the slide. One-way ANOVA. *P < 0.05, **P < 0.01. (E) Hierarchical clustering heatmap of metabolites of rapamycin and control groups. Specimens were collected after rapamycin treatment for 7 days, compared with no drug control. Top 50 features shown. Color bar indicates the scaled z score of each feature.

Functional pathway characterization revealed an increase in nucleotide biosynthesis and a reduction in glycolysis and multiple sugar degradation metabolic pathways after 30 days of rapamycin treatment (Figure 5C) but not 3 or 7 days. Species-resolved functional pathway analysis revealed that the Muribaculaceae family (Bacteroides) was enriched after long-term rapamycin treatment and harbored pathways involved in nucleotide biosynthesis (Supplemental Figure 5A). However, B. thetaiotaomicron and a variety of Clostridiales taxa (Firmicutes) that were relatively depleted after long-term rapamycin treatment harbored functional pathways in amino acid biosynthesis (i.e., ornithine), branched and aromatic amino acid biosynthesis (horismate pathway), glycolysis, and energy processing (Supplemental Figure 5B). These findings revealed a distinct shift in the gut microbiome composition and function following rapamycin treatment, emphasizing rapamycin effects beyond immunosuppression and a potential mechanism for its diverse therapeutic effect.

Rapamycin temporally shifts intestinal immune responses. To elucidate the reciprocal interactions between the gut microbiome and host under rapamycin treatment, we analyzed the intestinal transcriptome. Days 7 and 30 were chosen as they corresponded to the major gut microbiota alterations following rapamycin treatment. Differentially expressed genes (DEGs) were identified by comparing the rapamycin group to the no-treatment control. A total of 69 and 234 upregulated DEGs were observed at days 7 and 30, respectively, and a total of 168 and 84 downregulated DEGs were observed at days 7 and 30, respectively (Supplemental Table 3, A and B). At day 7, 70.9% of DEGs were downregulated, while at day 30, 73.6% of DEGs were upregulated. Analysis of differential expression patterns, visualized using an UpSet plot, revealed distinct temporal responses to rapamycin treatment with minimal overlap between day 7 and day 30 (Supplemental Figure 6A). Of all DEGs, only 32 genes (16 upregulated and 16 downregulated) showed consistent modulation at both time points. Many changes were time point specific: 159 genes were uniquely regulated at day 7 (45 upregulated and 114 downregulated), while 121 genes were uniquely modulated at day 30 (82 upregulated and 39 downregulated). The day 7 response primarily reflected suppression of inflammatory and innate immune pathways as well as chromatin remodeling (Supplemental Table 3C). Downregulated genes included multiple immunoglobulin families (Ighv, Igkv), immune defense genes (GTPases Igtp, Iigp1, Irgm1, Irgm2), antimicrobial peptides (Reg3b, Reg3g, Defa family), and inflammatory mediators (Ccl8, Ccl24, Cxcl9). Upregulated genes included histone family members (H2bc, H4c), metabolic regulators (Scd2, Tmprss15), and immune modulators (Nos2, Ubd). By day 30, the response shifted toward metabolic reprogramming and selective immune modulation. Downregulated genes were involved in glucose metabolism (G6pc, Pck1), lipid metabolism (Srebf1, Cyp4a10), and stress response. Upregulated genes included B cell–related genes (Cd19, Cd79a, Blk, Ms4a1), MHC class II pathway components (H2-Aa, Ciita), metabolic modulators (Cyp2c55, Slc10a2), and defense peptides (Defa family). The small set of consistently regulated genes across both time points maintained aspects of B cell regulation and antimicrobial defense. Consistently downregulated genes included immunoglobulin family members (Ighv) and metabolic regulators, while consistently upregulated genes included B cell–related genes, antimicrobial peptides (Defa family), and metabolic regulators. This temporal pattern suggests that rapamycin induces distinct phases of intestinal adaptation, transitioning from broad immunosuppression at day 7 to more targeted metabolic and immune regulatory programs by day 30.

On the other hand, intestinal Foxp3+ Treg expression was strongest at days 3 and 7 but was attenuated by day 30 (Figure 5D and Supplemental Figure 7). The enriched upregulated immune pathways at both days 7 and 30 include B cell regulation, activation, proliferation, antigen binding, and immunoglobulin-mediated immune responses (Supplemental Figure 6B). On day 7, there was unique enrichment in cellular responses to interferon-λ, -α, and -β, as well as cytokine-mediated signaling pathways. At day 30, unique enrichment was observed in MHC class II protein complex binding, antigen processing and presentation, mucosal immune responses, and tissue-specific immune responses. Other immune pathways demonstrated a substantially stronger effect in most functional categories by day 30, such as immunoglobulin receptor binding, production, and circulation; positive regulation of lymphocyte activation; and phagocytosis (Supplemental Figure 6C). Overall, the transcriptional changes were substantial, marked by the number of DEGs and enriched pathways. These results collectively indicate a temporal shift from suppression to activation in both intestinal gene expression and the immune environment following rapamycin treatment.

Rapamycin reprograms amino acid metabolism in gut lumen. Given the alterations in the composition and functional makeup of the gut microbiome, we next assessed whether this translated to functional changes in metabolism through the gut luminal metabolome. Intraluminal stool was assessed using capillary electrophoresis–mass spectrometry (CE/MS) (45–47). The 7-day time point was used as it represents the transitional phase between early and late alloimmune responses in both LNs and the intestine. A no-treatment group served as control to provide a baseline for comparison. Luminal metabolites (n = 264) were exhaustively annotated by PubChem (48), Kyoto Encyclopedia of Genes and Genomes (KEGG) (49), and Human Metabolome Database (HMDB) (50) (Supplemental Table 4A). According to the KEGG BRITE hierarchical classification system, the most prevalent class of luminal metabolites was from amino acid metabolism, comprising 42.7% of all annotated metabolites (Supplemental Table 4B). These metabolites belonged to pathways of arginine and proline, histidine, tyrosine, and tryptophan metabolism. Other prevalent classes included carbohydrates (11.5%), cofactors and vitamins (10.4%), nucleotides (10.4%), lipids (6.3%), other amino acids (6.3%), and xenobiotic metabolism (5.2%). Distinct gut metabolic profiles were observed after rapamycin treatment, with amino acids such as Asn, Phe, Arg, and Leu and metabolic derivatives differentially abundant in rapamycin treatment group (Figure 5E and Supplemental Figure 8, A and B). These results suggest that rapamycin treatment either increased amino acid biosynthesis and/or reduced catabolism.

Rapamycin induces a rapid proinflammatory response and a gut microbiome shift during allogeneic stimulation. We next employed a mouse model with allogeneic stimulation (Allo) to characterize the effect of rapamycin on transplant-related alloimmune responses. Mice were injected with fully Allo (1 × 107 cells intravenously) followed by rapamycin treatment for 3 days. Compared with no-treatment control, Allo alone induced a proinflammatory shift by decreasing the La4/La5 in pLNs and mLNs (Figure 6, C and F, and Figure 7, E and F), consistent with our previous findings (51). When Allo was combined with rapamycin (Rapa+Allo), there was an increase in both laminin α4 and α5 compared with Allo alone. However, the increase in laminin α5 exceeded that of laminin α4, resulting in decreased La4/La5 in both pLNs and mLNs (Figure 6, A–F, and I). This pattern aligns with findings from WT and KO mice treated with rapamycin without allostimulation, as both models demonstrated early proinflammatory effect on day 3 (Figure 2 and Figure 4). Compared with no-treatment controls, Allo alone decreased Tregs in the pLN without affecting Tregs in mLN (Figure 6, G–I). After 3 days of Rapa+Allo, compared with Allo alone, there was no change in Treg distribution in pLNs, but Tregs were decreased in the mLNs (Figure 6, G–I), indicating a proinflammatory state. These data demonstrate that rapamycin fosters an early proinflammatory LN environment in the context of alloantigen-induced immune responses through altering La4/La5 and Treg distribution.

Rapamycin triggers a rapid inflammatory immune response in mice immunized with allogeneic splenocytes. (A–C) IHC for pLN laminin α4 (A), laminin α5 (B), and La4/La5 (C). (D–F) IHC for mLN laminin α4 (D), laminin α5 (E), and La4/La5 (F). (G and H) CR and HEV Foxp3+ Tregs in pLN (G) and mLN (H) in no-treatment group (control), Allo only (Allo), and Allo plus rapamycin (Rapa+Allo). (I) Heatmap changes in marker expression comparing Allo plus rapamycin to Allo only in pLN and mLN. There were 1–3 LNs/mouse, 3 mice/group, 2–3 sections/LN group, and 7–30 fields/slide. Each dot in the graph represents 1 field. One-way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Allo, allogeneic splenocytes.

Rapamycin-induced protolerogenic modulation persists in tissue-specific patterns after allogeneic stimulation. Four groups include: untreated B6 mice (control), mice receiving 30 days of rapamycin treatment (Rapa), mice receiving allogeneic stimulation alone with 1 × 107 BALB/c splenocytes i.v. (Allo), or mice receiving 30 days of rapamycin pretreatment followed by allogeneic stimulation (Rapa+Allo). (A–D) Flow cytometry for percentages of CD4+ T cells (A), CD8+ T cells (B), B cells (B220+) (C), and Foxp3+ Tregs (Foxp3+CD4+) (D) in pLN, mLN, and spleen. (E–H) IHC for La4/La5 in the CR and around HEV in pLN (E) and mLN (F), and the distribution of Tregs in the CR and HEV in pLN (G) and mLN (H). There were 1–3 LNs/mouse, 3 mice/group, 2–3 sections/LN group, and 7–30 fields/slide. Each dot in the graph represents 1 field. (I) Heatmap of marker expression changes comparing Rapa+Allo to Allo alone in pLN, mLN, and spleen. Red indicates an increase, white indicates no change, blue indicates a decrease, and “X” denotes no data. One-way ANOVA: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Rapa+Allo for 3 days led to significant changes in the gut microbiome, characterized by increased microbial diversity and altered community composition and structure. In contrast, Allo alone showed no notable differences compared with the untreated control (Figure 8, A and B, and Supplemental Table 5). While the total number of microbial taxa remained unchanged, the Shannon diversity index increased in the Rapa+Allo group (P < 0.05). This suggests that rapamycin, in the context of allostimulation, promotes a more even distribution of microbial species without altering the overall number of distinct taxa. A marked shift in microbiome composition was observed, accompanied by this increase in microbial diversity. The B/F ratio decreased substantially in Rapa+Allo group (0.12 ± 0.10) compared with untreated controls (0.45 ± 0.22) or Allo alone (1.16 ± 0.67), indicating a substantial restructuring of the microbial community (Figure 8B and Supplemental Figure 9A). Allo alone increased the relative abundance of potentially proinflammatory Muribaculaceae (i.e., Duncaniella and Paramuribaculum) within the Bacteroides phylum. However, the combination of Rapa+Allo led to a higher abundance of Firmicutes, including Lachnospiraceae, Butyricicoccaceae, and CAG-274 (Figure 8, C and D, and Supplemental Figure 9, B–D). These results demonstrated the rapid, phylogenetic-aware effect of rapamycin under allostimulation where multiple taxa within the same phylogenetic group swiftly shifted in a unified direction. While similar to observations in naive mice under rapamycin treatment, the specific taxonomic groups affected differed. Furthermore, allostimulation increased intestinal Foxp3+ Tregs, an effect further enhanced by rapamycin treatment (Figure 8E) mirroring changes seen in naive mice. Collectively, these findings highlight rapamycin’s context-dependent influence on the intestinal microenvironment, affecting both intestinal Treg populations and gut microbiome during allostimulation.

Effects of rapamycin and alloimmunity on gut microbiome and intestinal Foxp3+ Treg. (A–C) Gut microbiome characterization of no-treatment control, allostimulation, and rapamycin combined with allostimulation in diversity index using observed number of taxa and Shannon diversity index (A); taxonomic composition on phylum level (B); and cladogram of differentially abundant taxonomic groups using LDA effect size (LEfSe) (C). Each filled circle represents 1 biomarker. The diameter of a circle is proportional to the phylotype relative abundance scaled at log10. (D) Principal component analysis (PCA) plot with taxa loadings labeled. Length of each taxa loading vector indicates its contribution to each PCA axis shown. The univariable group distribution appears above the plot. (E) IHC of Foxp3 Tregs in intestine assessed after 3 days of rapamycin treatment in allostimulated mice. There were 3 mice/group. O-way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001. Allo, allogeneic splenocytes.

A tissue-specific persistence of rapamycin-induced tolerogenic effects during allostimulation. Rapamycin treatment demonstrated its most profound immunomodulatory effects by day 30, establishing a protolerogenic environment characterized by increased Tregs in the pLN and mLN (Supplemental Figure 1, H and I, and Figure 7D), as well as an increased La4/La5 in the mLN (Figure 7, F and I). To evaluate whether this protolerogenic state persists under allostimulation, we conducted experiments comparing 4 groups: untreated control mice, mice receiving 30 days of rapamycin alone (Rapa), mice receiving only allogeneic stimulation with 1 × 107 BALB/c splenocytes i.v. (Allo), and mice receiving 30 days of rapamycin pretreatment followed by allogeneic stimulation (Rapa+Allo). The durability of rapamycin effects showed distinct tissue-specific patterns. In pLNs, the Rapa+Allo group reduced CD4+ T cell percentages compared with Allo alone (Figure 7, A and I), while other immune parameters remain largely unchanged (Figure 7, B–E, and I). IHC showed reduced Treg distribution in the Rapa+Allo group (Figure 7, G and I), suggesting that allostimulation partially overcame the rapamycin protolerogenic effect in pLN. This is supported by the observation that Allo alone led to a proinflammatory shift by decreasing the La4/La5 and Treg distribution in pLNs (Figure 6, C and I, and Figure 7, E and G). Together, these findings indicate a partially weakened protolerogenic environment due to allostimulation.

In mLNs, the Rapa+Allo group maintained stronger immunoregulatory features. Rapa+Allo treatment reduced the CD4+ T cell percentage compared with Allo only (Figure 7, A and I), with no other changes (Figure 7, B, C, F, and I). Notably, the Rapa+Allo group exhibited an increase in both the percentage and distribution of Tregs (Figure 7, D, H, and I). This preservation of Treg populations and positioning indicates that the 30-day rapamycin pretreatment sustained protolerogenic regulation in mLNs despite allostimulation. The spleen demonstrated the most robust maintenance of a rapamycin-induced protolerogenic state. Rapa+Allo reduced CD4+ and CD8+ T cell percentages compared with Allo alone (Figure 7, A, B, and I), maintained B cells percentages (Figure 7, C and I), and increased Treg percentages (Figure 7, D and I). Together, these findings reveal that rapamycin established tissue-specific patterns of sustained immune regulation, with the strongest maintenance of protolerogenic features in mLNs and spleen, while pLNs show more susceptibility to allostimulation.