Graphical abstract

   graphic file with name fx1.jpg
   [51]Open in a new tab

Highlights

     * •
       A single-cell transcriptomic immune atlas for EAU and CR in lymph
       nodes
     * •
       CR alleviated EAU symptoms and regulated Teff/Treg balance
     * •
       CR influenced the balance of CD4^+ T cells by inhibiting the
       Pi3k/Akt/c-Myc pathway
     __________________________________________________________________

   Immune response; Transcriptomics; Diet

Introduction

   Uveitis is an autoimmune disease characterized by inflammation within
   the eye, affecting the central nervous system
   (CNS).[52]^1^,[53]^2^,[54]^3 It is a significant cause of blindness,
   accounting for 10–25% of global vision loss, and severe visual
   impairment or blindness occurs in up to 35% of patients.[55]^4^,[56]^5
   Currently, corticosteroid therapy is the primary treatment approach,
   which comes with potential long-term side effects. Therefore, gaining a
   comprehensive understanding of uveitis pathogenesis is essential for
   the development of effective treatment strategies. Experimental
   autoimmune uveitis (EAU) serves as an extensively accepted animal model
   to explore potential pathogenic mechanisms underlying uveitis.[57]^6 T
   lymphocytes exert a pivotal effect in regulating adaptive immune
   responses, and their impaired tolerance can lead to the polarization of
   T cells (TCs) into hyperactive and pathogenic inflammatory phenotypes,
   contributing to T cell-mediated autoimmune diseases such as
   uveitis.[58]^1^,[59]^3 However, the precise mechanisms responsible for
   T cell polarization into inflammatory phenotypes remain poorly
   understood. A deeper understanding of intrinsic and extrinsic factors
   influencing T cell polarization will aid in identifying novel targets
   for uveitis treatment.

   Caloric restriction (CR) is a dietary approach that decreases calorie
   intake but does not cause malnutrition. It has been demonstrated to
   modulate nutrient-sensing mechanisms, resulting in improved metabolic
   profiles, reduced oxidative stress, enhanced stress resistance,
   suppressed inflammation, and alleviation of metabolic
   disorders.[60]^7^,[61]^8 Therefore, it may be a critical strategy for
   targeting the metabolic system to achieve immune modulation.
   Nonetheless, the signaling pathways and networks that regulate
   inflammatory induction across tissues and the impact of CR on EAU are
   largely unknown and require further investigation.

   Lymph nodes play a crucial role in autoimmune
   processes[62]^9^,[63]^10^,[64]^11 as they serve as sites where distinct
   immune cells including T cells (TCs), B cells (BCs), as well as
   dendritic cells (DCs) interact. Lymph nodes facilitate antigen
   presentation, immune cell activation, and the initiation of autoimmune
   responses.[65]^12 Among them, cervical-draining lymph nodes (CDLNs) are
   particularly important for efficient drainage of macromolecules as well
   as immune cells from the CNS.[66]^13 Consequently, CDLNs may act as
   production sites for autoreactive TCs as well as BCs targeting CNS
   autoantigens. In this study, our objective was to employ single-cell
   RNA sequencing (scRNA-seq) to map the immune cell landscape of CDLNs in
   mice fed ad libitum (AL) or subjected to CR, with or without EAU.

Results

CR induced complicated and extensive changes within the immune profile of
CDLNs

   The influence of CR on the immune profile of CDLNs was investigated by
   generating scRNA-seq data from CDLNs of normal mice with CR (NR) and
   normal mice with AL (NA) ([67]Figures 1A and [68]S1A). Weight changes
   in these groups at different time points are shown in [69]Figure 1B.
   The general effect of CR on the immune profile of CDLNs in normal mice
   was illustrated. Cell clustering revealed seven immune cell lineages:
   TCs, BCs, macrophages (Macro), neutrophils (Neu), conventional DC
   (cDC), plasmacytoid DC (pDC), as well as natural killer cells (NK)
   based on classical markers ([70]Figures 1C and [71]S1B–S1D). To find
   out the global gene signatures related to CR in CDLN cells, DEGs were
   identified for all cell lineages in CDLNs from normal mice. Gene
   Ontology (GO) analysis of these genes was conducted
   ([72]Figures 1D–1F). DEG analysis revealed downregulation of Fos, Jun,
   and Junb in NR mice ([73]Figure 1D). These genes encode components of
   activator protein 1 (AP-1), which is implicated in various inflammatory
   pathologies, such as rheumatoid arthritis and
   psoriasis.[74]^14^,[75]^15^,[76]^16 Gapdh, the main enzyme involved in
   glycolysis, was also downregulated in NR mice. Conversely, Il27ra, a
   suppressive regulator of effector Th1 and Th17 cells via its role as
   the receptor for IL-27, was upregulated in these mice.[77]^17 Ddit4, an
   mTORC1 inhibitor and a regulator of endoplasmic reticulum
   stress,[78]^18 was more highly expressed in NR mice than in NA mice
   ([79]Figure 1D). Ppp2cb, Ppp2ca, and Ppp2r5a were upregulated in NR
   mice ([80]Figure 1D), encoding protein phosphatase 2A (PP2A) known to
   inhibit phosphatidylinositol 3-kinase (PI3K)/AKT, dephosphorylate
   forkhead transcription factor O class 1 (FOXO1), and promote Treg
   functions.[81]^19 Enrichment analysis revealed that downregulated DEGs
   among NR mice were related to the biosynthetic processes of the
   cytoplasm and the activation of the immune system ([82]Figure 1E).
   Upregulated genes in NR mice were enriched in phosphatase and tensin
   homolog (Pten) regulation, antigen processing and presentation, and
   oxidative phosphorylation ([83]Figure 1F). Pten, a tumor suppressor,
   acts as a negative regulator of PI3K signaling.[84]^20 Collectively,
   these DEGs and GO analyses indicated a decreased immune effector
   response, particularly in autoimmunity, and a reduced rate of CDLN
   glycolysis in NR mice.

Figure 1.

   [85]Figure 1
   [86]Open in a new tab

   Study design and CR induces complicated and extensive changes in the
   immune profile of CDLNs

   (A) Schematic of the experimental design for single-cell RNA
   sequencing. CDLNs were harvested from normal (N) mice with ad libitum
   (A) or caloric restriction (R) diet. Samples were processed via
   scRNA-seq by using the 10x Genomics platform.

   (B) Line chart showing weights of NR and NA mice at different time
   points.

   (C) UMAP plot showing clusters of immune cell subsets.

   (D) Volcano plot showing upregulated and downregulated DEGs of all
   immune cell types in the NR/NA comparison group. Red and blue dots
   indicate upregulated and downregulated DEGs in NR group compared to NA
   group, respectively.

   (E and F) Representative GO terms and KEGG pathways enriched in
   downregulated (E) or upregulated (F) DEGs of total immune cells in the
   NR/NA comparison group.

   (G and H) Representative GO terms and KEGG pathways enriched in
   downregulated (G) or upregulated (H) DEGs of immune cell subsets in the
   NR/NA comparison group.

   (I) Heatmap showing average inflammation pathway scores of all immune
   cell types in NR and NA group.

   (J) Heatmap showing average metabolic pathway scores of all immune cell
   types in NR and NA group.

   Further investigation into cell type-specific transcriptome alterations
   induced by CR revealed that downregulated DEGs of TCs in NR mice were
   enriched in the activation as well as the proliferation of TCs,
   Th17-cell differentiation, and IL-17 signaling pathways
   ([87]Figure 1G). Meanwhile, negative regulation markers of
   phosphorylation and PTEN regulation were upregulated ([88]Figure 1H).
   In BCs, the mitogen-activated protein kinase (MAPK) signaling pathway
   and the regulation of BC differentiation and activation pathways were
   downregulated ([89]Figure 1G). Within myeloid cells, cDC as well as pDC
   from NR mice expressed fewer genes enriched in positive regulation of
   myeloid leukocyte differentiation and cellular responses to stress
   ([90]Figure 1G), while these cells highly expressed genes enriched in
   transforming growth factor (TGF)-beta signaling and negative regulation
   of PI3K/AKT pathway ([91]Figure 1H). CR-downregulated biological
   pathways related to the tumor necrosis factor(TNF) signaling pathway,
   positive regulation of immunity mediated by NK cells, and response to
   IL-1 were observed in NK cells ([92]Figure 1G), whereas upregulated
   pathways included PI5P, PP2A, and IER3, which regulate PI3K/AKT
   signaling and positive regulation of dephosphorylation pathways
   ([93]Figure 1H). Additionally, proinflammatory cytokine production
   scores were compared between the two groups. TNF-α, IFN-γ, and IL1-β
   production scores were significantly decreased in TCs, NK cells, cDC,
   and macrophages in NR mice compared to NA mice ([94]Figures 1I and
   [95]S1D–S1F; [96]Table S1). Consistently, the immunoglobulin production
   score in BCs from NR mice was significantly decreased as well
   ([97]Figure S1G).

   To further investigate the metabolic changes induced by CR, glycolysis
   or gluconeogenesis, the TCA cycle, and oxidative phosphorylation
   processes were examined in all cells of the two groups. The results
   showed that glycolysis or gluconeogenesis was suppressed by CR in all
   cells, particularly in BCs and TCs ([98]Figures 1J and [99]S1H).

CR induced functional changes of TC and BC compartments in CDLNs

   CR induces functional changes in the TC and BC compartments in CDLNs.
   Lymph nodes primarily consist of T and BCs, which play crucial roles in
   adaptive immunity. Therefore, our focus was on the alterations induced
   by CR in both of these cell types.

   We categorized eight subtypes of TCs from NR and NA mice based on
   classical markers, such as naive CD4^+ TCs (NCD4), naive CD8^+ TCs
   (NCD8), cytotoxic TCs (CTL), T helper 17 cells (Th17), regulatory TCs
   (Treg), T follicular helper cells (Tfh), T helper 1 cells (Th1), and
   proliferative TCs (ProT) ([100]Figures 2A, [101]S2A, and S2B).

Figure 2.

   [102]Figure 2
   [103]Open in a new tab

   CR induces functional changes of T cell and B cell compartments in
   CDLNs

   (A) UMAP plot showing clusters of T cell subsets from NA and NR mice.

   (B and C) Representative GO terms and KEGG pathways enriched in
   downregulated (B) or upregulated (C) DEGs of T cell subsets in the
   NR/NA comparison group.

   (D) Heatmap showing expression of Il17a, Il1r1, and Il23r in Th 17
   cells from NA and NR group.

   (E) Heatmap showing average glycolysis pathway scores of T cell
   subtypes in NR and NA group.

   (F) Boxplots showing TNF-α and IFN-γ production score in Th1 and Th17
   cells from NA and NR group. Data are represented as mean ± SD.
   Significance was determined using Wilcoxon rank-sum test. ∗p < 0.05,
   ∗∗p < 0.01, ∗∗∗∗p < 0.0001.

   (G and H) Proportions of Treg cells from lymph nodes (LN) (G) or spleen
   (SP) (H) of NA and NR group were measured by flow cytometry. Each group
   contains six mice. Data are represented as mean ± SD. Significance was
   determined using unpaired two-tailed Student’s t test. ∗∗p < 0.01,
   ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

   The GO analysis demonstrated that downregulated DEGs of Th17 from NR
   mice were enriched during proinflammatory cytokine production,
   specifically regulation of TNF production and IFN-γ production, IL-17
   signaling pathway, as well as Th17 cell differentiation
   ([104]Figure 2B). Moreover, the expression of Il17a, Il1r1, and Il23r
   was decreased in Th17, while TNF-α and IFN-γ production scores were
   lower in Th1 and Th17 cells from NR mice compared to those from NA mice
   ([105]Figures 2D, [106]S2C, and 2F), suggesting suppressed function of
   Th17 cells in NR mice. NCD4 and NCD8 cells exhibited downregulated
   biological pathways related to MAPK activation, TC activation, and
   differentiation induced by CR ([107]Figure 2B). Tregs from NR mice
   highly expressed genes enriched in MAPK activation ([108]Figure 2B),
   while positive regulation of TC differentiation and activation, TC
   proliferation, and the T cell receptor (TCR) signaling pathway were
   upregulated ([109]Figure 2C). Additionally, other TC subtypes in NR
   mice showed upregulated genes related to pathways such as PTEN
   regulation, regulation of dephosphorylation and phosphorylation,
   negative regulation of the PI3K/AKT network, and FOXO signaling pathway
   ([110]Figure 2C).

   Flow cytometry confirmed that CR treatment increased CD4^+ Foxp3+ TCs
   in NR mice ([111]Figures 2G and 2H), but led to no change in the
   proportion of CD4^+ IL17A + T as well as CD4^+ IFN-γ+ TCs
   ([112]Figures S2F–S2I), which was observed in both lymph nodes and
   spleens. Furthermore, we investigated metabolic alterations in
   CR-induced TCs. The glycolysis pathway score in NCD4, Tfh, Th17, and
   Treg cells from NR mice was significantly lower compared to that in NA
   mice ([113]Figures 2E and [114]S2D). Moreover, inflammatory genes such
   as Fos, Junb, and Pim1, as well as glycolysis-related genes like Gapdh,
   were expressed less in NR mice. Conversely, Ppp2ca, Ppp2cb, Ppp2r5a,
   and Pik3ip1 were more expressed in these mice relative to their
   counterparts ([115]Figure S2E). To further elucidate the role of
   glycolysis in the anti-inflammatory effects of CR, we assessed the
   expression levels of key glycolytic molecules and enzymes. Our findings
   indicated that CR reduced the expression levels of HK2, PKM2, and LDHA
   in Th17 cells ([116]Figures S2K–S2M), but led to no change in the
   expression levels of c-Myc in Th17 cells ([117]Figure S2J).

   These findings indicate that TCs in CR-induced CDLNs exhibited
   anti-inflammatory and regulatory phenotypes while displaying decreased
   autoimmune ability. Thus, CR may induce various immune alterations in
   different TC subsets and suppress glycolysis in these cells. We also
   observed impaired Th17 cell function in CR-induced CDLNs.

   Regarding BC compartments, we identified three subtypes of BCs from NR
   and NA mice based on classical markers: naive BCs (NBC), germinal
   center BCs (GC), and plasma cells (PC) ([118]Figures S3A–S3D). The GO
   analysis of BCs demonstrated that the upregulated DEGs were enriched in
   PTEN pathways and regulation of dephosphorylation in NBC, as well as
   PI3K/AKT signaling in GBC ([119]Figure S3E). CR downregulated
   regulation of BC proliferation and activation, MAPK signaling pathway,
   regulation of BC receptor signaling pathway, and regulation of
   translation in BCs, particularly in NBC ([120]Figure S3F). We then
   examined the metabolic changes among BCs and found that the glycolysis
   pathway score induced by CR was significantly lower in GC and NBC
   compared to AL ([121]Figure S3G). Similarly to TCs, the changes in gene
   expression in BCs from NR mice showed similar tendencies, including
   Ppp2r5a, Ppp2ca, and Gapdh ([122]Figures S3H and S3I). Thus, CR may
   suppress the primary functions of BCs.

CR mitigated EAU symptoms and altered immune cell response to the challenge
of EAU

   To understand this phenomenon and further investigate the relationship
   between CR and EAU, we conducted an analysis of changes in autoimmune
   responses during EAU in mice subjected to CR ([123]Figure 3A). We
   assessed the weights of EAU mice with CR (EAU-CR) and with AL (EAU-AL)
   at different time points and documented the severity of EAU symptoms
   using a published clinical grading scale.[124]^21 Evaluation of the
   eyes using a fundus camera revealed multiple chorioretinal lesions
   and/or infiltration as manifestations of EAU ([125]Figure 3C). Upon
   histological examination of sectioned eyeballs stained with H&E, we
   observed inflammatory infiltration together with retinal folding within
   the EAU lesions ([126]Figure 3D). EAU-CR mice showed milder symptoms of
   EAU, characterized by lower clinical and pathological scores, in
   contrast with EAU-AL mice ([127]Figures 3C and 3D).

Figure 3.

   [128]Figure 3
   [129]Open in a new tab

   Study design and CR mitigates EAU symptoms and alters immune cell
   response to EAU challenge

   (A) Schematic of the experimental design for single-cell RNA
   sequencing. CDLNs were harvested from normal (N) mice and EAU.

   (E) mice with ad libitum (AL) or caloric restriction (CR) diet. Samples
   were processed via scRNA-seq by using the 10x Genomics platform.

   (B) Line chart showing weights of EAU-CR and EAU-AL mice at different
   time points.

   (C) Representative fundus images and clinical scores of eyes from the
   EAU-AL and EAU-CR group after immunization at day 14. Each group
   contains six mice. Data are represented as mean ± SD. Significance was
   determined using unpaired two-tailed Student’s t test. ∗∗∗∗p < 0.0001.

   (D) Representative histopathological images (hematoxylin and eosin
   staining) and pathological scores of eyes from EAU-AL group and EAU-CR
   group after immunization at day 14. Each group contains six mice. Data
   are expressed as mean ± SD. Significance was determined using unpaired
   two-tailed Student’s t test. ∗∗∗∗p < 0.0001. Scale bars, 20 μm.

   (E) Volcano plot showing upregulated and downregulated DEGs of all
   immune cell types in the EAU-CR/EAU-AL comparison group. Red and blue
   dots indicate upregulated and downregulated DEGs in NR group compared
   to NA group, respectively.

   (F and G) Representative GO terms and KEGG pathways enriched in
   downregulated (F) or upregulated (G) DEGs of total immune cells in the
   EAU-CR/EAU-AL comparison group.

   (H) Boxplots showing average glycolysis pathway scores of BC and TC in
   NR, NA, EAU-CR, and EAU-AL group. Data are represented as mean ± SD.
   Significance was determined using Wilcoxon rank-sum test. ∗p < 0.05,
   ∗∗∗∗p < 0.0001.

   We performed single-cell analysis on the CDLNs of EAU-CR and EAU-AL
   mice, as depicted in [130]Figure 3A. Using classical markers, we
   identified the major immune cell clusters, as shown in and
   [131]Figures S4A, [132]S1A and S1B. Our initial analysis focused on
   identifying DEGs between EAU-CR and EAU-AL mice. We observed similar
   changes in DEGs as those observed in NR mice compared to NA mice, such
   as Gapdh, Ppp2ca, Pim1, and Pik3ip1 ([133]Figures 3E and [134]S4B). GO
   analysis revealed downregulation of positive regulation of immune
   response, cell activation, and translation pathways ([135]Figure 3F),
   while upregulated pathways included PTEN regulation, negative
   regulation of the PI3K/AKT network, and negative regulation of
   phosphorylation ([136]Figure 3G). Additionally, CR attenuated the
   EAU-induced increase in glycolysis pathway scores in both T and BCs
   ([137]Figure 3H).

   We proceeded to investigate the influence of CR on the response of TC
   subsets to the EAU challenge. We identified eight TC subsets in EAU-CR
   and EAU-AL mice, similar to those found in normal mice
   ([138]Figure S4C). To gain insights into the functional alterations
   induced by CR in response to EAU, we conducted GO analysis on these TC
   subsets ([139]Figures S4D and S4E). Our findings revealed
   downregulation of MAPK signaling, NF-κB signaling, as well as PI3K/AKT
   signaling pathways, while the PTEN regulation pathway was upregulated
   in the T cell subsets of EAU-CR mice ([140]Figures S4D and S4E).
   Additionally, we observed downregulated pathways associated with
   inflammation (annotated as positive regulation of IFN-γ production,
   positive regulation of cytokine production, along with IL-17 signaling
   pathway) and immune responses (annotated as TC differentiation and
   activation, and positive regulation of the immune response) in EAU-CR
   mice ([141]Figure S4E). These results suggest that CR induces a
   suppressive immune process and mitigates the severity of inflammation
   in response to the challenge of EAU. Changes in gene expression were
   further explored, specifically in Th17 cells among the four groups.
   Differential expression analysis of Th17 cells between EAU-CR and
   EAU-AL mice revealed genes affected by CR, including Ppp2cb, Pik3ip1,
   Pim1, and Gapdh ([142]Figures 4A and 4B). Moreover, the expression of
   Il17a, Il23r, and Il1r1, which were increased during the EAU challenge,
   was reduced in Th17 cells by CR ([143]Figures 4A and 4B, and 4C),
   indicating a restrained function of Th17 cells induced by CR. Notably,
   CR also decreased the glycolysis pathway score in Th17 cells, both in
   normal and EAU mice, which exhibited enhancement during the EAU
   response ([144]Figure 4D). Furthermore, we assessed the expression
   levels of transcription factors in the four groups. Transcription
   factors associated with Th17 differentiation and function showed
   reduced expression in Th17 cells induced by CR, such as Jun, Runx1,
   Rora, and Rorc ([145]Figures 4E and [146]S4F). Conversely, Treg cells
   exhibited enhanced expression of Foxo1 and Stat3 under CR conditions
   ([147]Figure 4E). Therefore, it appears that CR weakens the role of
   Th17 cells and promotes immunosuppression during EAU development.

Figure 4.

   [148]Figure 4
   [149]Open in a new tab

   CR-induced alterations in T cell subsets and cell-cell communication

   (A) Volcano plot showing upregulated and downregulated DEGs of Th17
   cells in the EAU-CR/EAU-AL comparison group. Red and blue dots indicate
   upregulated and downregulated DEGs in NR group compared to NA group,
   respectively.

   (B) Heatmap showing gene expression in Th17 cells of NA, NR, EAU-AL,
   and EAU-CR group.

   (C) Violin plots showing the expression of Il17a, Csf2, Il23r, and
   Il1r1 in Th17 cells of NA, NR, EAU-AL, and EAU-CR group.

   (D) Boxplot showing average glycolysis pathway scores of Th17 cells in
   NR, NA, EAU-CR, and EAU-AL group. Data are represented as mean ± SD.
   Significance was determined using Wilcoxon rank-sum test. ∗p < 0.05,
   ∗∗p < 0.01.

   (E) Heatmap showing single-cell regulon scores inferred by SCENIC in
   T cell subtypes of NA, NR, EAU-AL, and EAU-CR group (g, genes;
   extended, SCENIC-annotated additional genes).

   (F) Heatmap showing the number of possible interactions between immune
   cells analyzed in NA, NR, EAU-AL and EAU-CR groups.

   (G) The interaction of Th17 cells with myeloid cells and B cells in NA,
   NR, EAU-AL and EAU-CR mice.

   We also classified three subtypes of BCs in EAU-CR and EAU-AL mice
   based on classical markers: NBC, GC, and PC ([150]Figure S5A). GO
   analysis conducted on BCs revealed upregulated pathways enriched in
   PTEN pathways, negative regulation of dephosphorylation, and negative
   regulation of immune system processes, particularly in GBCs
   ([151]Figure S5B). Downregulated pathways included BC proliferation,
   differentiation, activation, and immune effector processes
   ([152]Figure S5C). Furthermore, the glycolysis pathway score, which may
   be enhanced in EAU mice, was reduced in these BC subtypes under CR
   conditions, especially in NBCs ([153]Figure S5D).

   In summary, our findings elucidated the alterations in TC and BC
   subsets in the CDLNs in response to the challenge of EAU, in normal
   mice as well as EAU mice. CR resulted in a mitigated inflammatory
   status, a suppressive immune response process, and a restrained role of
   Th17 cells during the EAU challenge. Moreover, the reduction in the
   glycolysis process induced by CR suggested decreased activation of
   immune cells, potentially contributing to the amelioration of EAU
   symptoms.

CR alleviated aberrant intercellular communication in EAU

   To understand the intercellular communication among immune cells in the
   four groups, the cell-cell communication network was explored utilizing
   CellphoneDB2 analysis ([154]Figures 4F, 4G and [155]S6A). We found that
   the enhanced interactions observed in response to EAU were rescued by
   CR, specifically the interactions involving Th17 cells and other immune
   cells ([156]Figures 4F and 4G). Notably, the interaction between CSF2
   and its receptors, which is known to be associated with CNS
   autoimmunity, occurred between Th17 cells and myeloid cells after EAU
   development and was suppressed by CR ([157]Figure 4G). Additionally,
   the interactions involving CD244 and CD48, IL-15R and IL-15, as well as
   IGF1 and IGF1R between Th17 cells and myeloid cells or macrophages were
   reduced under CR conditions ([158]Figure 4G). These findings suggest
   that the activation and function of Th17 cells were suppressed by CR.

   Further analysis revealed that the reduction in interaction pairs
   observed during EAU development, such as the interaction between TGFB1
   and its receptors, was enhanced by CR ([159]Figure S6A), indicating a
   higher immunosuppressive status induced by CR in EAU mice. Interactions
   associated with proinflammatory cytokines, such as IL1R2, CCL8, and
   CCR5, were enhanced after EAU development ([160]Figure S6A).
   Interestingly, the interaction involving CCL7 and its receptors between
   macrophages and other cells, which can amplify the inflammatory
   process, was decreased in EAU-CR mice compared to EAU-AL mice
   ([161]Figure S6A). These results further demonstrate that CR attenuates
   the inflammatory process.

CR suppressed inflammatory status and reduces Th17 pathogenicity via
PI3K/AKT/c-Myc axis

   The analysis of DEGs and GO revealed that CR induces the attenuation of
   inflammation and immune processes. We utilized flow cytometry to
   validate these responses and changes. Our findings demonstrated that CR
   decreased the population of Th1 (IFN-γ+) cells as well as Th17
   (IL-17A+) cells, while increasing the number of Treg (Foxp3+) cells
   during the development of EAU. Moreover, CR resulted in decreased
   expression of inflammatory and pathogenic factors. These observations
   confirmed the suppressive inflammatory status and immune process
   induced by CR ([162]Figures 5A–5J).

Figure 5.

   [163]Figure 5
   [164]Open in a new tab

   CR weakened Th17 pathogenicity

   (A–E) Proportions of Th1 cells (A), Th17 cells (B and C), regulatory
   T cells (Treg) (D) and GM-CSF+ Th17 cells (E) from LN of EAU-AL and
   EAU-CR group were measured by flow cytometry. Each group contains six
   mice. Data are represented as mean ± SD. Significance was determined
   using unpaired two-tailed Student’s t test. ∗∗p < 0.01, ∗∗∗p < 0.001.

   (F–J) Proportions of Th1 cells (F), Th17 cells (G and H), regulatory
   T cells (Treg) (I) and GM-CSF+ Th17 cells (J) from spleen (SP) of
   EAU-AL and EAU-CR group were measured by flow cytometry. Each group
   contains six mice. Data are represented as mean ± SD. Significance was
   determined using unpaired two-tailed Student’s t test. ∗∗p < 0.01,
   ∗∗∗p < 0.001.

   To identify the impact of CR on Th17 cell’s function, which was
   essential for the pathogenicity of AU,[165]^22 we analyzed the DEGs
   specifically in Th17 cells among the four groups. The expression levels
   of Il17a, Csf2, as well as Il23r were lower among Th17 cells from CR
   mice compared to those from AL mice following EAU development
   ([166]Figures 4A–4C). Additionally, flow cytometry analysis showed
   reduced production of granulocyte-macrophage colony stimulating factor
   (GM-CSF), predominantly secreted by Th17 cells along with a
   characteristic feature distinguishing pathogenic Th17 cells from
   non-pathogenic ones,[167]^23 in the lymph nodes and spleens of EAU mice
   subjected to CR ([168]Figures 5E and 5J). Collectively, these results
   indicate impaired Th17 pathogenicity resulting from CR during EAU.

   To further confirm these findings within the IRBP[1-20]-specific
   response, CDLN cells were collected from EAU-CR and EAU-AL mice,
   cultured with IRBP[1-20] as a stimulator, and we observed a decrease in
   the number of Th1, Th17, and in the expression of GM-CSF, due to CR
   ([169]Figures 6A–6C, and 6E). Additionally, the proportion of Tregs was
   elevated among EAU-CR mice ([170]Figure 6D). Furthermore, through
   adoptive transfer experiments, we found that CD4^+ TCs obtained from
   CDLNs of EAU-AL mice successfully induced EAU, while those from EAU-CR
   mice failed among normal mice ([171]Figures 6F and 6G). We obtained
   similar results using an alternative method to induce
   IRBP[1-20]-specific T cells[172]^24 ([173]Figures S7B–S7G).

Figure 6.

   [174]Figure 6
   [175]Open in a new tab

   CR weakened IRBP-specific Teff differentiation

   (A–E) CDLNs cells from EAU-AL and EAU-CR group cultured with or without
   IRBP[1-20] for 72 h. Proportions of Th1 cells (A), Th17 cells (B and
   C), regulatory T cells (Treg) (D) and GM-CSF+ Th17 cells (E) of EAU-AL
   and EAU-CR group were measured by flow cytometry. Data are represented
   as mean ± SD from six independent experiments. Significance was
   determined using unpaired two-tailed Student’s t test. ∗∗p < 0.01,
   ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

   (F) The representative fundus images after induction by CD4^+ T cells
   from EAU-AL (EAU-AL-- > AT) or EAU-CR (EAU-CR-- > AT) groups after
   immunization at day 14.

   (G) Clinical scores of EAU-AL-- > AT and EAU-CR-- > AT groups (n = 6).
   Clinical scores significantly decreased (∗∗∗∗p < 0.0001) in
   EAU-CR-- > AT group compared to EAU-AL-- > AT group at day 14. Data are
   shown as mean ± SD from three independent experiments. Data were
   analyzed using unpaired student t tests.

   (H) The representative HE staining images and pathological scores after
   induction by CD4^+ T cells from EAU-AL (EAU-AL-- > AT) or EAU-CR
   (EAU-->CR-AT) groups after immunization at day 14. Scale bars, 20 μm.
   Each group contains six mice. Significance was determined using
   unpaired Student’s t test. ∗∗∗p < 0.001.

   To investigate whether the specific mechanism by which CR effectively
   inhibits EAU-related inflammation is related to glycolysis, we further
   conducted an analysis of single-cell data. Our findings revealed that
   the expression levels of glycolysis-related genes, such as Pik3r1,
   Akt1, Myc, and Hif1α, as well as key glycolytic enzymes, were
   significantly elevated in the presence of EAU and were notably reduced
   following CR treatment ([176]Figure S6B). This trend was confirmed
   through flow cytometry analysis ([177]Figures 7A–7D). In summary, these
   results collectively indicate that CR effectively restrains
   inflammatory and immune processes, as well as weakens the pathogenicity
   of Th17 cells, via PI3K/AKT/c-Myc/glycolysis axis.

Figure 7.

   [178]Figure 7
   [179]Open in a new tab

   CR repressed PI3K/AKT/c-Myc/glycolysis pathway

   (A–F) The proportions of CD4^+ p-PI3K+ cells (A), CD4^+p-AKT+ cells
   (B), CD4+IL-17A+c-Myc+ T cells (C), CD4+IL-17A + HK2+ T cells (D),
   CD4+IL-17A + PKM2+ T cells (E), and CD4+IL-17A+LDHA+ T cells (F) were
   measured by flow cytometry. Each group contains six mice. Data are
   represented as mean ± SD. Significance was determined using unpaired
   two-tailed Student’s t test. ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Discussion

   In our study, we have elucidated the transcriptomic shifts within CDLNs
   of mice subjected to CR and their subsequent response to EAU. Our
   findings underscore the modulatory role of CR in immune cell dynamics,
   particularly in dampening the pathogenicity of Th17 cells, which is
   pivotal in mitigating EAU symptoms. CR was found to influence the
   PI3K/AKT/c-Myc signaling axis and the glycolytic pathway, thereby
   altering Th17/Treg cell differentiation and restoring T cell
   homeostasis.

   Cellular metabolism, including aerobic glycolysis, fatty acid
   oxidation, and oxidative phosphorylation, is crucial for T cell fate
   determination.[180]^25 Recent studies indicate that enhanced glycolysis
   in immune cells is essential for effective processes, such as cell
   activation, proliferation, and effector responses.[181]^26 Nutrient
   metabolism significantly influences the differentiation and function of
   various immune cells. Dietary interventions like calorie restriction,
   time-restricted eating, and fasting can modulate immune cell
   function.[182]^8 Previous research has shown that CR can reduce the
   expression of inflammatory cytokines associated with EAU.[183]^27 Our
   study offers a detailed analysis of the mechanisms by which CR
   alleviates EAU through a comprehensive immune cell map. We found that
   CR regulates Th17 and Treg cell differentiation by affecting the
   glycolytic pathway via the PI3K/AKT/c-Myc/glycolysis axis and restores
   T cell homeostasis by reshaping the immune microenvironment.

   The activation of effector TCs (Teff cells) heavily relies on the
   PI3K/AKT/mTOR signaling pathway. Conversely, the low-energy state
   induced by CR suppresses the PI3K/AKT/mTOR axis while reciprocally
   activating adenosine monophosphate-activated protein kinase
   (AMPK).[184]^28^,[185]^29^,[186]^30 The hypoglycemia induced by CR
   reduces anabolic hormones and inhibits insulin-dependent anabolic
   metabolism by impeding the PI3K and MAPK signaling pathways, thereby
   preventing mTOR activation. Inactivated mTOR triggers autophagy,
   contributing to the suppression of inflammation by downregulating
   IFN/proinflammatory cytokine secretion, as well as inhibiting
   inflammasome activity.[187]^31 Besides, inactivation of mTOR hampers
   hypoxia-inducible factor 1 (HIF-1)-dependent activation of genes in
   relation with inflammation, the proinflammatory effects of reactive
   oxygen species (ROS), as well as NF-κB activation.[188]^32^,[189]^33
   The inhibition of the PI3K signaling pathway entails critical metabolic
   consequences, including increased lipolysis and ketogenesis, together
   with a shift in substrate utilization for energy production from
   glucose to fatty acids and ketone bodies.[190]^34 Our study employed
   single-cell metabolic analysis, which revealed reduced glycolytic
   metabolic pathway activity during CR. We observed downregulation in the
   expression of glycolysis-related genes such as Gapdh and Pim1, along
   with increased expression of inhibitors of the PI3K/AKT/mTOR/glycolysis
   pathway, including Pten, Pik3ip1, and PP2A-related genes, in TCs and
   BCs.

   Although previous studies have extensively investigated alterations in
   the immune system during CR using scRNA sequencing in various rat
   organs,[191]^35 a comprehensive immune cell atlas of CR lymph nodes,
   which play a crucial role in immune responses, was lacking prior to our
   study. Furthermore, detailed investigations of cellular changes in CR
   CDLNs and their association with autoimmune diseases were also lacking.
   Therefore, our study fills these knowledge gaps and provides valuable
   insights into the single-cell expression landscape of EAU and CR.

   Consistent with previous research highlighting the anti-inflammatory
   effects of CR in age-related inflammation and other eye
   diseases,[192]^36^,[193]^37^,[194]^38^,[195]^39 our study demonstrated
   that CR can attenuate the inflammatory response of immune cells within
   CDLNs, particularly during autoimmune challenges. Specifically, TCs
   exhibited reduced proliferation, activation, IL-17 signaling, along
   with differentiation toward Th17 cells. BCs showed decreased production
   of immunoglobulins, and other innate immune cells displayed lower
   levels of inflammation-associated pathways. GO analysis revealed
   impaired cell activation among T cell subsets as well as compromised
   antigen presentation and cell division among BC subsets. Within the
   various immune cell types, Th1, Th17, and Treg cells are recognized as
   key contributors to either uveitis or other autoimmune
   diseases.[196]^22^,[197]^40^,[198]^41 In the present study, CR
   suppressed the proliferation of Th1/Th17 cells while promoting the
   expansion of Treg cells. Notably, the pathogenicity of Th17 cells was
   diminished, as evidenced by reduced secretion of GM-CSF in CR mice.

Limitations of the study

   Our findings offered novel ideas into the intricate interaction of CR
   and autoimmunity, serving as valuable knowledge for researchers
   interested in this field of study. However, there are some limitations
   in this study. Firstly, the absence of human AU samples, such as the
   correlation between body mass index (BMI) and AU scores or other
   disease activity indicators, limits the direct clinical applicability
   and transformative potential of our findings. Secondly, while a
   decrease in glucose metabolism was noted, the study did not delve into
   other metabolic profiles, including amino acids and lipids, in serum or
   cell lysates for both human and murine samples. Additionally, the
   translation of these findings into potential treatments for human
   patients with AU, such as clinical trials for CR treatment, remains to
   be explored.

Resource availability

Lead contact

   Further information and requests for resources and reagents should be
   directed to and will be fulfilled by the lead contact, Wenru Su
   (suwr3@mail.sysu.edu.cn).

Materials availability

   This study did not generate new unique reagents.

Data and code availability

     * •
       Single-cell RNA-seq data have been deposited at Genome Sequence
       Archive (GSA) and are publicly available as of the date of
       publication. Accession numbers: CRA016398.
     * •
       This paper does not report original code.
     * •
       Any additional information required to reanalyze the data reported
       in this paper is available from the [199]lead contact upon request.

Acknowledgments