Dietary restriction to optimize T cell immunity is an ancient survival strategy conserved in vertebrate evolution

Abstract

Recent advances highlight a key role of transient fasting in optimizing immunity of human and mouse. However, it remains unknown whether this strategy is independently acquired by mammals during evolution or instead represents gradually evolved functions common to vertebrates. Using a tilapia model, we report that T cells are the main executors of the response of the immune system to fasting and that dietary restriction bidirectionally modulates T cell immunity. Long-term fasting impaired T cell immunity by inducing intense autophagy, apoptosis, and aberrant inflammation. However, transient dietary restriction triggered moderate autophagy to optimize T cell response by maintaining homeostasis, alleviating inflammation and tissue damage, as well as enhancing T cell activation, proliferation and function. Furthermore, AMPK is the central hub linking fasting and autophagy-controlled T cell immunity in tilapia. Our findings demonstrate that dietary restriction to optimize immunity is an ancient strategy conserved in vertebrate evolution, providing novel perspectives for understanding the adaptive evolution of T cell response.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-023-04865-x.

Introduction

Food intake provides animals with essential nutrients and is of paramount importance to cell growth, development, and survival [1, 2]. Recent studies have reported that dietary composition, quantity, and frequency regulate immunity, thus highlighting an intimate relationship between diet and immune system [3–5]. Severely reduced dietary intake causes undernutrition, which not only affects normal physiological activities but also weakens the immune cell response and performance and destroys the mucosal barrier integrity, thus rendering the animals more vulnerable to pathogenic infection and autoimmune diseases [6, 7]. Remarkably, a constant dietary supply may also be detrimental to the immune system. Owing to industrialization, the dietary structure of present-day human is markedly different from that of their ancestors and early human, which allows for a continuous and sufficient supply of calories at the individual level. However, the incidence of obesity, hypoimmunity, and autoimmune diseases has increased worldwide with increasing industrialization [8, 9]. In accordance with this observation, recent studies have emphasized that transient reductions in dietary intake limit aberrant inflammatory responses but enhance protective immunity. For example, dietary restriction reprograms the metabolism of immune cells [10] and strengthens the pulmonary mucosal barrier, thus resisting Mycobacterium tuberculosis infection and reducing tissue damage [10, 11]; intermittent fasting induces hematopoietic stem cells to produce new immune cells [12] and triggers autophagy to reuse cell components [13, 14]. In an inflammatory bowel disease model, cycles of a very-low-calorie diet optimized the gut microbiota and in turn ameliorated immunopathology and promoted intestinal regeneration [15]. In addition, during bacterial infection, dietary restriction protects the brain by producing ketone bodies that limit ROS-induced neuronal damage [16]. Thus, these findings emphasize the profound effect of dietary restriction on the immune system and highlight how it may inform the design of novel approaches to prevent and treat diseases.

Although dietary restriction affects numerous aspects of the immune system, recent studies have reported that T cell immunity is the key facilitator of the fasting strategy. Dietary restriction enhances the functional capacity of memory T cells and results in improved protection against secondary bacterial infection [17]. A combination of the fasting strategy and PD-L1 mAb optimizes T cell-mediated immunotherapy against several types of tumor cells [3]. In addition, fasting inhibits excessive activation of T helper cells and prevents an aberrant inflammatory response by activating the FOXO4–FKBP5 axis [4]. In experimental autoimmune encephalomyelitis (EAE), caloric restriction alters the composition and metabolic function of the microbiota, which in turn increases the number of T_reg cells but decreases the pathogenic Th17 cells, thus ameliorating neuronal immunopathology [18]. Moreover, fasting also helps to reshape T cell metabolism and aging. For example, dietary restriction maintains the energy balance of T cells by promoting glucocorticoid production and inhibiting glycolytic pathways [10, 17]. In old mice, dietary restriction improved the maintenance of naive T cells and prevented the deterioration of the T cell receptor (TCR) repertoire and the decline of antigen presentation and T cell proliferation [19]. Therefore, these findings indicate that reshaping T cell immunity by restricting dietary intake represents an effective approach to manipulate immune defense.

Although dietary interventions optimize T cell immunity in mouse and human, it remains largely unknown whether this survival strategy was acquired by mammals independently during evolution or instead represents a gradually evolved function common to all jawed vertebrates. In fact, the diets of wild animals in nature consist exclusively of foraged foods and vary depending on seasonal fluctuations, making starvation or fasting a common phenomenon. Some non-mammals, especially ectothermic animals, such as snake, frog, and fish, have even evolved remarkable tolerance to starvation [19–21]. Moreover, they actively reduce or even stop dietary intake in response to pathogenic infection or to adapt to environmental changes [22]. However, the immunological mechanisms underlying such physiological behavior in these vertebrates remain unknown. For example, it remains to be studied whether the fasting strategy controls immunity, how the immune system adapts and responds to the dietary restriction, and what role T cells play in this process. Answers to these questions will help to elucidate the regulatory mechanism of T cell immunity in non-mammals and are expected to provide novel perspectives and evidence to understand the evolution of a survival strategy in which dietary restriction optimizes immunity.

Unlike mammals, bony fish do not consume additional energy to maintain body temperature and are able to survive long-term starvation [20]. Meanwhile, bony fish have evolved a well-developed adaptive immune system, which plays a central role in resisting pathogenic infection [23, 24]. Moreover, many detailed mechanisms underpinning T cell immunity of the bony fish, Nile tilapia, have been illustrated in our previous studies [24–31]. Owing to these physiological properties and research advances, bony fish represent an ideal model to investigate the relationship between the fasting strategy and T cell immunity in early vertebrates and its evolution. In the present study, using Nile tilapia (Oreochromis niloticus) as a model, we investigated the effects of dietary restriction on the immune response. We found that fasting triggered AMPK-dependent T cell autophagy, and fasting duration determined the intensity of autophagy, resulting in opposite effects on T cell immunity. Long-term (more than 7 days’) fasting induced intense autophagy, leading to T cell apoptosis, spleen atrophy, and aberrant inflammation. In contrast, short-term (2 or 3 days’) dietary restriction triggered moderate autophagy to promote immunity by maintaining T cell homeostasis; alleviating inflammation and tissue damage; and enhancing the T cell activation and proliferation potential and T cell function. Thus, we demonstrated a bidirectional regulation of T cell immunity in response to dietary intervention in this bony fish, which provides novel insights into the evolution of a survival strategy that optimizes immunity through dietary restriction.

Materials and methods

Ethics statement

All animal experiments were conducted in accordance with the guidelines for the Care and Use of Laboratory Animals of the Ministry of Science and Technology of China, and approved by the East China Normal University Experimental Animal Ethics Committee with an approve number of AR2021-247. All efforts were made to minimize the pain of animals.

Fish maintenance and study design

Nile tilapia larvae (3.5 ± 0.5 cm of body length) were obtained from an aquatic farm in Guangzhou, Guangdong Province, China, and maintained in a freshwater system with biofilters and continuous aeration at 28 °C at the Biological Station of East China Normal University. Fish were fed twice daily using commercial pellets. Only healthy fish with body length of 9–13 cm were used for the study. For the fasting study, long-term fasting group (fast for 7 days or more) was set first, followed by the short-term fasting group (fast for 2 or 3 days), so that the samples were collected at the same time at the end of fasting. The normal diet group was sampled 2 h after the end of feeding.

Sequence, structure, and phylogenic analysis

All the sequences were obtained from the NCBI GenBank. The amino sequences were submitted to ExPASy (http://web.expasy.org/protparam/) to compute the protein molecular weights (MW). The functional domains were predicted using the Simple Modular Architecture Research Tool (SMART, http://smart.embl-heidelberg.de/), and the domain organization was displayed by DOG 2.0 software. The phylogenetic tree was constructed using MEGA v7.0 software with neighbor-joining (NJ) method, and 1000 bootstraps replications were set for phylogenetic evaluation. The information of amino acid sequences used is listed in Table S1.

Aeromonas hydrophila infection

A. hydrophila was cultured in Luria Broth liquid medium at 37 °C. Bacteria were collected during the exponential phase and resuspended in PBS. One hundred microliters of A. hydrophila with a final concentration of 3.6 × 10⁶ CFU was injected intraperitoneally into per fish, and the control group was injected with the same volume of PBS. The injected fish were kept at 28 °C with aeration and daily feeding as before. For the A. hydrophila infection experiment after fasting, the fish with normal diet and 2 days of fasting were infected as above, after which the normal diet group maintained daily feeding, and the fasting group was fed every 2 days.

Leukocyte isolation

Spleen or head kidney leukocytes were isolated according to our previous report [24]. Briefly, the spleen and head kidney were harvested and ground in pre-cooled L-15 medium to obtain cell suspension and then filtered with a cell strainer and finally adjusted to 3 ml. The cell suspension was added onto 52/34% discontinuous Percoll (GE Healthcare) density gradient. After centrifuging at 500×g, at room temperature for 30 min with the lowest acceleration and deceleration, leukocytes at the interface of 52 and 34% Percoll were harvested and washed twice with L-15 medium.

Cell counting and calculation

The isolated spleen or head kidney leukocytes were resuspended in 1 ml of L-15 medium, and then, 20 μl cell suspension was added into another 480 μl of L-15 medium. The diluted cells were subsequently counted under a microscope using a hemocytometer, and the total number of isolated leukocytes was calculated. After determining the percentage of lymphoid cells among leukocytes by flow cytometry, the absolute number of lymphoid cells was calculated using the leukocyte number multiplied by the lymphoid cell percentage among leukocytes. And the absolute number of T or B cells was calculated using the lymphoid cell number multiplied by the T or B cell percentage among lymphoid cells.

Leukocyte stimulation

For T cell activation, spleen leukocytes were cultured at 28 °C in DMEM containing 10% FBS and 1% penicillin/streptomycin and then stimulated with 2 μg/mL PHA, or 2 μg/mL mouse anti-tilapia CD3ε mAb according to our previous report [29], for different times. For glucose-free stimulation, 2 × 10⁷ spleen leukocytes were cultured in 24-well plates in DMEM with or without glucose for 2, 4 or 6 h. The stimulated cells were harvested for the indicated assay.

RNA-seq and data analysis

Spleen leukocytes harvested from four fed tilapia individuals and four tilapia individuals that fasted for 3 or 7 days were applied for RNA-seq. Total RNA was extracted using Trizol reagent kit (Invitrogen, USA) according to the manufacturer’s protocol. RNA was examined by an Agilent 2100 Bioanalyzer (Agilent Technologies, USA), and then subjected to commercial RNA-seq analyses using Illumina HiSeq2500 in Gene Denovo Biotechnology Co. (Guangzhou, China). Raw data were filtered by fastp (version 0.18.0) to obtain high quality and reliability clean reads. Paired-end clean reads were mapped to the Nile tilapia genome (Ensembl release100) using HISAT2.2.4 [32]. The differential expression analysis was performed by DESeq2 software, and genes/transcripts with the parameter of false discovery rate (FDR) below 0.05 and absolute fold change ≥ 2 were considered significantly differentially expression genes/transcript. The differentially expressed genes were mapped to Gene Ontology (GO) terms in the Gene Ontology database and subjected to pathway enrichment analysis. The gene expression trend analysis was performed on all differential genes of grouped samples through the Short Time-series Expression Miner software [33]. For gene set enrichment analysis (GSEA), software GSEA and MSigDBto were used to identify whether a set of genes in specific GO terms/pathways shows significant differences in two groups [34]. For weighted gene co-expression network analysis (WGCNA), genes with expression > 1 FPKM in 90% of samples were selected from the three groups RNA seq datasets. The filtered data were used to generate co-expressing genes modules, and intramodular connectivity and module correlation degree (MM) of each gene were calculated by R package of WGCNA [35]. The data of RNA-seq used in this study are listed in Data file S1.

Inhibition or activation of AMPK and autophagy

For in vivo studies, Nile tilapia was intraperitoneally injected with or without 5 mg/kg AMPK inhibitor Compound C (MedChemExpress) or 10 mg/kg autophagy inhibitor 3-MA (MedChemExpress) or 100 mg/kg AMPK activator Metformin (MedChemExpress) every 2 days. For in vivo inhibition during bacterial challenge, the tilapia was injected with inhibitor one day before infection and was then pursued to inject as above. The same volume of PBS was injected as control. For in vitro studies, the leukocytes were cultured in DMEM (10% FBS, 1% penicillin/streptomycin) containing 10 μM Compound C or 300 μM 3-MA for indicated time points, and the same volume of inhibitor-solvent was used as control.

Examination of blood glucose

The peripheral blood was collected from caudal vessels of fed or fasted Nile tilapia and placed at room temperature for 30 min to precipitate the serum. After centrifugation at 4000 rpm for 5 min, the serum was collected to measure the blood glucose using commercially Glucose (GO) Assay Kit (Rsbio, China).

Detection of ATP and AMP concentration

The ATP or AMP concentration was measured by the commercial kit. Briefly, the freshly harvested spleen was weighed, smashed, and lysed with the provided lysis solution. After centrifuging at 12,000 rpm at 4 °C for 5 min, the supernatant was collected for subsequent determination. ATP and AMP contents were measured by using the ATP Assay Kit (Beyotime, China) and AMP Assay Kit (J&l Biological, China), respectively.

Quantitative real-time PCR (qPCR)

Total RNA was extracted from the freshly harvested or in vitro-stimulated spleen leukocytes using RNAiso Plus reagent (Takara), and cDNA was then synthesized using the NovoScript Plus All-in-one 1st Strand cDNA Synthesis SuperMix (Novoprotein) according to the manufacturer’s protocols. qPCR was performed using NovoStart SYBR qPCR SuperMix Plus (Novoprotein) on QuantStudio5 Q5 (Applied Biosystems). β-Actin was used as the internal reference, and the relative gene expression levels were analyzed with the 2^–ΔΔCt method. The information of primers used in this study is listed in Table S2.

Flow cytometry and cell sorting

Spleen or head kidney leukocytes were resuspended in FACS buffer (PBS containing 2% FBS) for FACS analysis. The FITC-conjugated mouse anti-tilapia CD3ε mAb (1 mg/mL) and biotinylated-conjugated mouse anti-tilapia IgM mAb (1 mg/mL) were generated in our previous study [29]. For CD3ε and IgM staining, leukocytes were stained with 1:400-diluted CD3ε mAb and 1:400-diluted IgM mAb on ice for 30 min and then washed twice with FACS buffer. The cells were then stained with Alexa Fluor 647-conjugated streptavidin on ice for another 30 min. For intracellular staining, leukocytes were first stained with mouse anti-tilapia CD3ε mAb as above and then fixed with BD Cytofix/Cytoperm Buffer on ice for 30 min and washed twice with BD Perm/Wash solution. The fixed cells were then stained with 1:300 diluted LC3A/B, cleaved caspase 3, cleaved caspase 8 or p-AMPKα antibody (Cell Signaling Technology), or p-ERK1/2 antibody (eBioscience) on ice for 30 min. All samples were collected using a BECKMAN CytoFLEX flow cytometer, and data were analyzed using FlowJo software. To sort CD3⁺ T cells, spleen leukocytes were stained with mouse anti-tilapia CD3ε mAb in DMEM containing 2% FBS as above, and the cells were resuspended in DMEM containing 10% FBS and sorted with a BD FACSAria II flow cytometer.

Death and apoptosis assay

Spleen leukocytes were stained with the FITC-conjugated mouse anti-tilapia CD3ε mAb as above. The cells were then resuspended in Annexin V binding buffer (0.01 M HEPES/NaOH, pH 7.4, 0.14 M NaCl, 2.5 mM CaCl₂) containing 1:400 diluted Alexa Fluor 647-conjugated Annexin V mAb (BioLegend) and stained at room temperature for 15 min. 1:400 diluted 7-AAD (Life Technologies) was added to the samples shortly before collection by flow cytometer.

Detection of autophagy

To measure autophagy level, spleen leukocytes were first stained with FITC-conjugated mouse anti-tilapia CD3ε mAb as above. Subsequently, the cells were incubated with 50 nM LysoTracker Red (Sigma, USA) on ice for 30 min and then washed twice with pre-chilled L-15 medium. The cells were collected and observed by fluorescence microscope (Leica, Germany), and subjected to flow cytometry assay.

BrdU incorporation

Nile tilapia was i.p. injected with 0.8 mg of BrdU (Sigma) in 150 μl of PBS 1 day before spleen leukocytes were isolated. The cells were stained with mouse anti-tilapia CD3ε mAb as above, and fixed with BD Cytofix/Cytoperm solution on ice for 30 min. After washed twice with BD perm/washing solution and permeabilized with BD Cytoperm Plus solution on ice for 10 min, cells were fixed with BD Cytofix/Cytoperm at room temperature for 5 min and then digested with 300 μg/mL DNase at 37 °C for 1 h. Subsequently, the cells were stained with 1:100 diluted FITC-conjugated anti-BrdU antibody (BD Biosciences) in perm/washing buffer at room temperature for 20 min. After twice washing, the cells were analyzed by flow cytometry.

Histological studies

Spleen and liver were fixed in 4% Paraformaldehyde Fix Solution (Beyotime) for 24 h. The fixed samples were then dewatered in gradient ethanol solutions, cleaned in xylene and then embedded in paraffin. Sections were prepared and stained according to hematoxylin and eosin (HE) staining procedure for histological observation. The stain sections were observed under Olympus BX53 light microscope, and AxioVision software was used to acquire and analyze the image. For immunohistochemistry (IHC) staining, paraffin sections of spleen and liver were incubated in 20 μg/mL proteinase K at 37 °C for 30 min for antigen retrieval. Then, the samples were washed with PBST (PBS with 0.05% Tween-20) for three times and blocked with 2% BSA at room temperature for 1 h. After being washed with PBST, the sections were stained at 37 °C for 1 h, using mouse anti-tilapia CD3ε or IgM mAb as primary antibody, and 1:200 diluted biotin-labeled goat anti-mouse IgG as secondary antibody, respectively. After another three times of wash, the BCIP/NBT Alkaline Phosphatase Color Development Kit (Beyotime) was used for the color development. For Sudan Black B staining, sections were incubated in Sudan Black staining solution (500 mg Sudan Black B in 20 ml acetone, 15 ml acetic acid and 85 ml H₂O) at 37 °C for 45 min, followed by a rapid wash with 70% ethanol for 10 s, and a final rinse with distilled water for 1 min.

Western blotting

The freshly isolated or stimulated leukocytes were lysed in lysis buffer (1% NP-40, 150 mM NaCl, 50 mM Tris, pH 7.4) with phosphatase and protease inhibitor. After centrifugation at 12,000 rpm at 4 °C for 10 min, the supernatants were subjected to SDS-PAGE assay and then transferred onto the nitrocellulose membrane. The blots were blocked with 4% nonfat powdered milk in PBST at room temperature for 1 h, and followed by the incubation with 1:1000 diluted primary antibodies for Cleaved Caspase-3 (#9664, clone 5A1E), phospho-AMPKα Thr¹⁷² (#2535, clone 40H9), phospho-LKB1 Ser⁴²⁸ (#3482, clone C67A3), AMPKα (#5831, clone D5A2), LKB1 (#3047, clone D60C5), SQSTM1/p62 (#5114), phospho-AKT Thr³⁰⁸ (#13,038, clone D25E6), phospho-ERK1/2 Thr²⁰²/Tyr²⁰⁴ (#4370, clone D13.14.4E), anti-β-actin (#4967) from Cell Signaling Technology, or phospho-NFκB P65 Ser²⁷⁶ (#AF5875), LC3B (#AF5225) from Beyotime at 4 °C overnight. The blots were washed three times with PBST and then incubated with 1:10,000 diluted Alexa Fluor 800 or 680-conjugated goat-anti rabbit or mouse IgG H&L (Thermo Fisher Scientific) as secondary antibody at room temperature for 1 h. All the blots were finally scanned by Odyssey CLx Image Studio.

Statistical analysis

Prism version 8.0 software (GraphPad Software Inc., San Diego, CA, USA) was used for statistical analysis. The statistical significance of results was analyzed using two-tailed, unpaired, Student’s t test. Animal survival data were analyzed by log-rank analysis. The significance was indicated by *, p < 0.05, **, p < 0.01, ***, p < 0.001.

Results

T cells may be a potential executor of the fasting strategy in tilapia

The strategy of dietary restriction is employed by mammals to optimize the anti-infection immune response. In the present study, tilapia infected with A. hydrophila showed markedly reduced dietary intake (Fig. 1A), indicating that this behavior also existed in this bony fish. To investigate the relationship between fasting and immunity, healthy tilapia was fasted for different durations. Prolonged fasting for more than 7 days markedly decreased the spleen size, but this was not observed in fish fasted for 3 days (Fig. 1B, C). Next, we investigated the effect of fasting on the lymphoid cells of tilapia. Since more than 90% of the spleen leukocytes we isolated belong to an obvious population (Fig. 1D, left panel), which has been identified as lymphoid cell population in our previous study [24], we could consider these spleen leukocytes as spleen lymphoid cells in our subsequent experiments. Upon long-term fasting, spleen lymphoid cells became smaller (Fig. 1D, right panel) and showed markedly higher apoptosis (Fig. 1E, F). However, slightly increased lymphoid cell proliferation was synchronously observed in the spleen of long-term-fasted tilapia (Fig. 1G, H). These findings thus indicate that dietary restriction may affect the biological properties of tilapia lymphoid cells. To further confirm this, we analyzed the transcriptome of spleen lymphoid cells isolated from tilapia with normal diet, short-term (3 days) fasting, and long-term (7 days) fasting. A set of differentially expressed genes (DEGs) were identified in fasted tilapia when compared with the fed group (Fig. 1I). Moreover, the GO enrichment analysis showed that short-term fasting triggered processes associated with amino acid metabolism and protein modification, whereas long-term fasting induced ribosome biosynthesis, cell activation and migration, and cytokine response (Supplementary Fig. S1A–C). These results were also confirmed by GSEA (Supplementary Fig. S1D–F). Furthermore, the heatmap analysis indicated that fasting altered the metabolic pattern of tilapia lymphoid cells, reducing glycolysis and enhancing fatty acid metabolism (Fig. 1J, K). Notably, we found that the DEGs between fed and fasted groups or between short-term-fasted and long-term-fasted groups were enriched in pathways associated with T cells and antigen recognition (Fig. 1L–N). Moreover, an analysis of 409 common DEGs identified by the pairwise comparison showed that T cell-related processes were indeed significantly enriched (Fig. 1O, P), indicating that dietary restriction had a marked effect on T cell immunity of tilapia. Overall, these findings suggest that dietary restriction affects lymphoid cell, including T cell processes in tilapia.

Fig. 1.

Tilapia T cells respond actively to fasting. A Feed intake (% body weight) of tilapia that infected with A. hydrophila or not for 1 or 3 days. B–H Tilapia individuals were fed or fasted for indicated days, and the spleen was harvested. B Scatter plot figure showing the spleen weight/body weight of tilapia fed or fasted for 3 or 7 days. C Image showing the spleen size of tilapia fed or fasted for 7 or 14 days. D FACS plot showing the purity of lymphoid cells in isolated spleen leukocytes, and overlaid histogram showing the comparation of FSC on gated lymphoid cell population in tilapia that was fed or fasted for 7 or 14 days. E, F Representative FACS plots (E) and scatter plot figure (F) showing the gating and statistic of 7-AAD and Annexin V staining on lymphoid cell population in tilapia fed or fasted for 3 or 7 days. G, H Tilapia was i.p. injected with BrdU 1 day before sacrifice, and representative FACS plots (G) and scatter plot figure (H) showing the BrdU⁺ cells on gated lymphoid cells in tilapia fed or fasted for 3 or 7 days. I–P RNA-seq analysis of spleen leukocytes isolated from tilapia fed or fasted for 3 or 7 days, n = 4. I Quantity and volcano plots of the DEGs, p value < 0.001. J, K Heatmaps showing the expression of key genes involving glycolysis (J) and lipid metabolism (K). L–N Pathway enrichment analysis of the DGEs basing on indicated comparation showing the top 10 pathways in immune system, q value < 0.05. O Venn diagram showing the 409 common DEGs among the three indicated comparations. P Pathway enrichment analysis of the 409 common DGEs showing the top 10 pathways in immune system, q value < 0.05. Experiments in A–H were repeated for three independent times. *: p < 0.05, **: p < 0.01, ***: p < 0.001, determined by a two-tailed Student’s t test

Long-term fasting impairs T cell homeostasis in tilapia by inducing auto-inflammation and apoptosis

Next, we investigated the effects of fasting on tilapia T cells. In the spleen, short-term fasting did not affect the number of lymphoid cells (Fig. 2A) and the frequency and absolute number of CD3⁺ T cells (Fig. 2B, C). In contrast, long-term fasting resulted in a reduced number of lymphoid cells and T cell frequency (Fig. 2A, B) and consequently severely decreased the absolute number of T cells (Fig. 2C). Although the absolute number of IgM⁺ B cells and CD3⁻IgM⁻ lymphoid cells also decreased after long-term fasting (Fig. 2D, E), the extent of this decrease was less than that in T cells (Fig. 2F), indicating that T cells were more susceptible to dietary restriction. Similar results were obtained in the head kidney; however, the effect of long-term fasting on lymphoid cells seemed to be weaker in the head kidney than in the spleen (Supplementary Fig. S2A–E). These findings thus suggest that long-term fasting affects lymphoid cell, especially T cell homeostasis in tilapia. In addition, we found that the decrease in T cell number in long-term-fasted tilapia was associated with the increase in programmed cell death (Fig. 2G, H), which was also confirmed by the high level of cleaved caspase 3 and caspase 8 in both lymphoid cells and gated T cells (Fig. 2I–K). Moreover, long-term fasting but not short-term fasting significantly induced the expression of the pro-inflammatory cytokines IFN-γ, IL-2, IL-1β, TNF-α, and IL-6 and the cytotoxic molecules perforin A and granzyme B in lymphoid cells or sorted T cells (Fig. 2L, M, Supplementary Fig. S2F, G). Additionally, marked infiltration or aggregation of inflammatory cells (i.e., T cells, B cells or neutrophils) was observed in the liver and spleen of long-term-fasted fish (Fig. 2N–Q), suggesting that long-term fasting induced a spontaneous inflammatory response in tilapia. Collectively, these results suggest that long-term but not short-term dietary restriction in tilapia disrupts lymphoid cell, especially T cell homeostasis by inducing inflammation and apoptosis.

Fig. 2.

Long-term fasting affects T cell homeostasis in tilapia. A–M Spleen leukocytes were isolated from tilapia fed or fasted for 3 or 7 days. A–E Lymphoid cell number (A), representative FACS plots (B), percentage and absolute number of CD3⁺ T cells (C), IgM⁺ B cells (D), or CD3⁻IgM⁻ lymphoid cells (E) were shown. F Bar figure showing the ratio of the absolute number of indicated cells in 7-days’ fasted tilapia to that in fed tilapia. G, H Representative FACS plots (G) and scatter plot figure (H) showing the gating and statistic of 7-AAD and Annexin V staining on gated T cell population. I Western blot assay showed the expression of cleaved caspase 3 and caspase 8. J, K Protein levels of cleaved caspase 3 or caspase 8 in gated T cell population. L, M T cells were sorted from tilapia that fed or fasted for 3 or 7 days, and relative mRNA levels of the indicated molecules were examined by qPCR, n = 4. N–Q Liver or spleen were collected from tilapia fed or fasted for 7 days. N, O H&E staining. Black arrow: inflammatory cell infiltration; red arrow: immune cell aggregation. P, Q Immunohistochemistry staining showing the T cells and IgM⁺ B cells in live and spleen (P, Q), and Sudan black B staining showing the neutrophils in spleen (Q). These experiments were repeated for three independent times. *: p < 0.05, **: p < 0.01, ***: p < 0.001, determined by a two-tailed Student’s t test

Short-term fasting enhances T cell immunity of tilapia

To directly assess the effects of dietary restriction on tilapia immunity, we infected the fed fish, short-term-fasted fish, and long-term-fasted fish with A. hydrophila. Notably, long-term-fasted fish were susceptible to bacterial infection, whereas short-term-fasted fish showed significantly improved survival (Fig. 3A). Further analysis of gene set profiles 4–7, which showed upregulation trends during fasting (Supplementary Fig. S3A, B), suggested that pathways regarding lymphoid cell/T cell activation and proliferation, nucleic acid synthesis, ribosome biosynthesis, transcription and translation, and mitosis were significantly enriched (Fig. 3B, Supplementary Fig. S3C–E). In addition, under short-term fasting, receptors or molecules crucial for cellular immunity, such as TCRα, CD3ε, NF-κB, Myc, JAK1, and Stat5, were upregulated in spleen lymphoid cells, whereas immunosuppressive molecules such as CTLA-4, IL-10, and IRF10 were downregulated (Fig. 3C), suggesting that short-term fasting enhanced the immunological potential of lymphoid cells, including T cells in tilapia. Consistent with these findings, short-term fasting led to an increase in T cell size (Fig. 3D, E); CD3ε, CD8α, and CD4-1 mRNA levels (Supplementary Fig. S3F); and CD3 protein expression (Fig. 3F). However, under long-term fasting, T cells were smaller in size, and the expression level of the receptors gradually dropped back (Fig. 3D–F, Supplementary Fig. S3F). Next, we stimulated spleen lymphoid cells with T cell mitogen PHA or mouse anti-tilapia CD3ε mAb to induce T cell activation. Upon T cell activation, the cells isolated from short-term-fasted tilapia expressed more IL-2 and its receptor CD122 (Fig. 3G, H). Moreover, the phosphorylation levels of AKT, NF-κB p65, and ERK1/2 were significantly higher in lymphoid cells (Fig. 3I, J) and T cells (Fig. 3K) from fasted tilapia, in both resting and stimulated conditions. These observations regarding cell size, gene expression, and signaling pathway activation collectively indicated that short-term fasting enhanced the activation of T cell in tilapia. Furthermore, upon T cell activation, the short-term fasting elevated mRNA levels of IFN-γ and IL-17A and their transcription factors T-bet and RORα in spleen lymphoid cells (Fig. 3L–O), however, but not GATA-3 (Fig. 3P, Q) Notably, the expression of cytotoxic gene granzyme B was also substantially upregulated (Fig. 3R, S). Taken together, our findings demonstrate that short-term dietary restriction may enhance the immunity in tilapia by potentially and partially boosting the T cell response.

Fig. 3.

Short-term fasting optimizes T cell response of tilapia. A Tilapia fed or fasted for 3 or 7 days were infected with A. hydrophila or not, and Kaplan–Meyer survival plot showing the survival percentage, n = 25. B Spleen leukocytes isolated from tilapia fed or fasted for 3 days were used for RNA-seq analysis, n = 4. Pathway enrichment analysis basing on the gene clusters 4–7 in Supplementary Fig. S3A that showing up-regulation trends in the transcriptome. C Heatmap showing the expression of key genes involving T cell immunity in the spleen leukocytes that fasted for 3 days compared with the fed ones. D Overlaid histogram showing the comparation of FSC in spleen T cells. E, F Scatter plot figures showing the mean fluorescence intensity of FSC (E) or CD3ε (F) in spleen T cells. G-S Spleen leukocytes from fed or 2-day fasted tilapia were stimulated with 2 μg/mL PHA or 2 μg/mL mouse anti-tilapia CD3ε mAb or not. G, H, L–S Relative mRNA levels of the indicated molecules were examined by qPCR at 0, 3, 6 h after stimulation, n = 4–6. I, J Western blot assay showed the phosphorylation levels of AKT, NF-κB P65, and ERK1/2 at 0, 1, 3 h after stimulation. K Flow cytometry showing the phosphorylation levels of ERK1/2 in gated T cells at 1 h after stimulation. These experiments were repeated for three independent times. *: p < 0.05, **: p < 0.01, ***: p < 0.001, determined by a two-tailed Student’s t test

Moderate autophagy initiated by short-term fasting is required for tilapia to enhance T cell immunity

The weighted gene co-expression network analysis (WGCNA) was used to identify the relevant gene clusters/modules. Five among the 23 co-expressed gene modules exhibited significant changes in terms of the correlation between fasting states/times (Supplementary Fig. S4A). Analysis of the module “salmon” with the most significant difference showed that the autophagy pathway was significantly enriched (Fig. 4A), indicating that it is a factor associated with the effect of fasting on immunity in tilapia. Next, an in-depth exploration against the tilapia genome sequence identified a set of components involved in autophagy signaling, including ULK1, SQSTM1/P62, Beclin-1, ATG5/7/12/14, and LC3B (Supplementary Fig. S4B). Notably, their functional domains and tertiary structure were highly conserved with the mammalian homologues (Supplementary Fig. S4C, D), suggesting that tilapia possesses an evolutionarily conserved autophagy pathway. Short-term dietary restriction for 3 days significantly elevated the mRNA expression of ULK1, Beclin-1, ATG5/7/12/14, and LC3B in both spleen lymphoid cells and sorted T cells (Fig. 4B, Supplementary Fig. S5A). Consistent with these observations, upregulated Beclin-1 and LC3B expression and increased P62 degradation were observed in lymphoid cells or CD3⁺ T cells (Fig. 4C, D). Notably, autophagy-lysosomal staining in lymphoid cells was stronger in fasted tilapia than in fed tilapia (Fig. 4E). Therefore, these findings collectively suggest that short-term fasting induces autophagy in tilapia lymphoid cells and T cells. However, prolonged dietary restriction markedly increased autophagy in both lymphoid cells and T cells, as evidenced by the sharply increased mRNA expression levels of autophagy-related genes (Fig. 4B, Supplementary Fig. S5A) and the more significant enrichment of autophagy pathway gene sets (Supplementary Fig. S5B–D). Considering that excessive autophagy triggers apoptosis [36], we suggest that intense autophagy triggered by long-term fasting is a potential factor contributing to lymphoid cells and T cells apoptosis in tilapia.

Fig. 4.

Fasting-induced autophagy is necessary for tilapia to enhance T cell immunity. A Spleen leukocyte isolated from tilapia fed or fasted for 3 days was used for RNA-seq analysis, n = 4. KEGG enrichment analysis of genes in the “salmon” module showing the significantly enriched pathways, p value < 0.001. B Spleen T cells were sorted from tilapia fed or fasted for 3 or 7 days, and relative mRNA levels of the indicated molecules in T cells were examined by qPCR, n = 4. C Western blot showing the protein and phosphorylation levels of indicated molecules in spleen leukocytes isolated from tilapia fed or fasted for 1–4 days. D Expression of LC3 protein in gated spleen T cell population from tilapia fed or fasted for 4 days. E Fluorescence microscope image of leukocytes stained with LysoTracker, microscopy setting: eyepiece (e.g. 10 ×) and objective lens (e.g. 40 ×). F–H Fasted tilapia was i.p. injected with or without 10 mg/kg 3-MA on days 1 and 2, and spleen leukocytes were isolated on day 3 of fasting. Representative FACS plots (F) and scatter plot figure (G, H) showing the gating and statistic of CD3 and Annexin V staining. I Spleen leukocytes cultured with or without glucose were treated or untreated 300 μM 3-MA for 12 h, and flow cytometry showing the 7-AAD and Annexin V staining on T cell population. J–L Spleen leukocytes from 2-days’ fasted tilapia were treated with 2 μg/mL PHA and 300 μM 3-MA or not. J Western blot assay showing the phosphorylation levels of AKT, NF-κB P65, ERK1/2 at 1 h after stimulation. K Flow cytometry showing the phosphorylation levels of ERK1/2 in gated T cells at 1 h after stimulation. L Relative mRNA levels of CD122 and IFN-γ were examined by qPCR at 3 h after stimulation, n = 4. M–O Spleen leukocytes of normally fed tilapia were stimulated with 2 μg/mL PHA, and treated with 300 μM 3-MA or not. M Western blot assay showing the phosphorylation levels of indicated proteins after 2 h after stimulation. N Flow cytometry showing the phosphorylation levels of ERK1/2 in gated T cells at 2 h after stimulation. O Relative mRNA levels of indicated genes were examined by qPCR at 3 h after stimulation, n = 4–6. P–U Tilapia individuals that infected with A. hydrophila on day 1 were i.p. injected with or without 10 mg/kg 3-MA on days 2, 3 and 4. P, Q Representative FACS plots (P) and scatter plot figures (Q) showing the gating and statistic of percentage and absolute number of T cells within the spleen lymphoid cells on 5 DPI. R Tilapia was i.p. injected with BrdU 1 day before killing, and flow cytometry showing the CD3 and BrdU staining on gated lymphoid cells on 5 DPI. S, T Spleen T cells were sorted from tilapia on 5 DPI, and relative mRNA levels of the indicated molecules were examined by qPCR, n = 4. U Kaplan–Meyer survival plot showing the survival percentage of 10 mg/kg 3-MA-treated or untreated fish after infection, n = 25. These experiments were repeated for three independent times. *: p < 0.05, **: p < 0.01, ***: p < 0.001, determined by a two-tailed Student’s t test

Next, we investigated the correlation between the enhanced anti-bacterial immunity and moderate autophagy after short-term fasting. The inhibitor 3-MA was used to inhibit autophagy (Supplementary Fig. S5E, F). Although short-term fasting did not affect T cell homeostasis (Fig. 2B, C), apoptosis of T cells significantly increased when autophagy was suppressed (Fig. 4F, G), leading to an obvious decrease of the T cell frequency among lymphoid cells (Fig. 4H). This result was evidently supported by the in vitro experiment, in which autophagy blockade rendered T cells more susceptible to apoptosis under glucose-free condition (Fig. 4I). Moreover, once spleen lymphoid cells were stimulated by PHA, the T cell activation events including phosphorylation of NF-κB, AKT, and ERK1/2 and upregulation of CD122 and IFN-γ, observed under short-term fasting was obviously attenuated by the autophagy blockade in lymphoid cells or CD3⁺ T cells (Fig. 4J–L). These findings indicate that short-term dietary restriction enhanced T cell immunity of tilapia by initiating moderate autophagy. To further confirm the regulatory effect of autophagy on T cell immunity, we treated normally fed tilapia with 3-MA. Blockade of autophagy signaling markedly impaired the T cell activation-induced phosphorylation of NF-κB, AKT, and ERK1/2 (Fig. 4M, N) and the upregulation of CD122 and IFN-γ (Fig. 4O), suggesting that autophagy is essential for the proper activation of tilapia T cells. The critical role of tilapia T cells in resisting bacterial infection has been revealed in our previous studies [24, 25, 29]. In the present study, we demonstrated that autophagy blockade impaired pathogen-induced T cell proliferation, as evidenced by the significant reduction of the frequency and absolute number of T cells (Supplementary Fig. S5G, Fig. 4P, Q) and disruption of BrdU incorporation in CD3⁺ T cells (Fig. 4R). Meanwhile, autophagy suppression decreased the absolute number of IgM⁺ B cells while increasing their frequency within lymphoid cells (Supplementary Fig. S5H). Furthermore, loss of autophagy reduced the pathogen-induced expression of pro-inflammatory cytokine IFN-γ, TNF-α, IL-2, and IL-1β, and cytotoxic gene Perforin A and Granzyme B in both spleen lymphoid cells and sorted T cells (Fig. 4S, T, Supplementary Fig. S5I, J), and eventually rendered the tilapia more vulnerable to the bacterial infection (Fig. 4U). Thus, our results indicate that upon tentative lack of diet, tilapia trigger moderate autophagy to enhance T cell immunity, which may contribute to the improved survival of the tilapia during bacterial infection.

AMPK signaling is critical for dietary restriction to induce autophagy and support T cell immunity in tilapia

Autophagy is tightly regulated by several pathways, and the bioenergy sensor AMPK plays a key role in this process [37]. However, whether and how AMPK signaling regulates T cell autophagy and governs T cell immunity in fish remains unknown. In the present study, we found that AMPK signaling was significantly enriched together with the autophagy pathway in the WGCNA co-expression module (Fig. 4A). Essential components of the AMPK pathway, including LKB1, AMPKα1, AMPKα2, and AMPKγ1, are evidently encoded in tilapia (Supplementary Fig. S6A). A comparative analysis regarding the functional domain, tertiary structure, and phylogeny revealed high conservation of these molecules between tilapia and mammals (Supplementary Fig. S6B–E). Upon dietary restriction, transcription of these components was significantly induced in both spleen lymphoid cells (Supplementary Fig. S7A, B) and sorted T cells (Fig. 5A, B); meanwhile, the LKB1 or AMPKα phosphorylation in lymphoid cells and T cells was also enhanced (Fig. 5C, D). The potential association between fasting and AMPK activation in immune cells was further indicated by GSEA (Supplementary Fig. S6F). Moreover, we found that energy restriction caused by both in vitro lymphoid cell culture under glucose-free condition (Fig. 5E, F) and in vivo administration with metformin (Fig. 5G, H), dramatically enhanced the phosphorylation of AMPKα in lymphoid cells and T cells, which was accompanied by the upregulation expression of LKB1 and AMPKα in lymphoid cells (Fig. 5I). The fasting-induced AMPK activation in tilapia lymphoid cells, especially T cells was associated with the reduced blood glucose levels (Fig. 5J). However, the abundance of ATP and AMP in spleen and the AMP/ATP ratio did not change significantly under fasting conditions (Fig. 5K), indicating that AMP or ATP is not indispensable for AMPK activation in tilapia lymphoid cells under energy restriction.

Fig. 5.

AMPK signaling is essential for fasting to induce autophagy and support T cell immunity in tilapia. A, B Spleen T cells were sorted from tilapia fed or fasted for 3 or 7 days, and relative mRNA levels of the indicated molecules were examined by qPCR, n = 4. C Western blot assay showing the protein and phosphorylation levels of LKB1 and AMPKα in spleen leukocytes isolated from tilapia fed or fasted for 1–5 days. D Phosphorylation of AMPKα in gated spleen T cell population from tilapia fed or fasted for 3 days. E Western blot showing the protein and phosphorylation levels of AMPKα in spleen leukocytes after 0, 2, 4, 6 h of incubation in glucose-free medium. F Phosphorylation of AMPKα in gated spleen T cells after 4 h in glucose-free medium. G–I Tilapia individuals were i.p. injected with or without 100 mg/kg Metformin on day 1 and 2, and the spleen leukocytes were isolated on day 3. G Western blot showing the protein and phosphorylation levels of AMPKα. H Flow cytometry showing the phosphorylation of AMPKα in gated T cells from tilapia treated with or without Metformin. I Relative mRNA levels of indicated molecules were examined by qPCR, n = 4–5. J, K Tilapia individuals were fed or fasted for 3 and 5 days. Bar figure showing the levels of blood glucose in serum, and ATP and AMP level in spleen, n = 4–5. L, M Spleen leukocytes cultured with or without glucose were treated with 10 μM Compound C or not. L Western blot showing the protein or phosphorylation levels of indicated molecules in spleen leukocytes at 6 h after stimulation. M Overlaid histogram showing the comparation of LysoTracker in T cells at 12 h after stimulation. N–P Spleen leukocytes of normally fed tilapia were treated with 2 μg/mL PHA and 10 μM Compound C or not. N Relative mRNA levels of CD122 and IFN-γ were examined by qPCR at 3 h after stimulation, n = 4–5. O Western blot assay showing the phosphorylation levels of AKT, NF-κB P65, ERK1/2 at 2 h after stimulation. P Flow cytometry showing the phosphorylation levels of ERK1/2 in gated T cells at 2 h after stimulation. Q–V Tilapia individuals that infected with A. hydrophila on day 1 were i.p. injected with or without 5 mg/kg Compound C on days 2, 3, 4, and spleen leukocytes were isolated on day 5. Q–T Representative FACS plots for CD3 and IgM staining (R), lymphoid cell number (Q), and percentage and absolute number of T cell (S) or IgM⁺ B cells (T) were shown. U, V Spleen T cells were sorted from tilapia on 5 DPI, and relative mRNA levels of the indicated molecules were examined by qPCR, n = 4. W Spleen leukocytes cultured with or without glucose were treated with 10 μM Compound C or not for 12 h, and flow cytometry showing the 7-AAD and Annexin V staining on T cell population. These experiments were repeated for three independent times. *: p < 0.05, **: p < 0.01, ***: p < 0.001, determined by a two-tailed Student’s t test

Then, we found that the blockade of AMPK signaling during energy restriction reduced the increased SQSTM1/P62 expression (Fig. 5L) and impaired the autophagy-lysosome activity in T cells (Fig. 5M), suggesting that AMPK signaling is crucial for low-energy-induced T cell autophagy in tilapia. Similar to the autophagy inhibition, the blockade of AMPK signaling affected T cell activation, as revealed by the impaired induction of CD122 and IFN-γ (Fig. 5N) and the reduced phosphorylation of NF-κB, AKT, and ERK1/2 (Fig. 5O, P) in spleen lymphoid cells and T cells upon PHA-induced T cell activation. Moreover, AMPK inhibition during A. hydrophila infection markedly reduced the expansion of T cells but not IgM⁺ B cells (Fig. 5Q–T), and impaired the elevation of Perforin A, Granzyme B, IFN-γ, TNF-α, IL-2 and IL-1β in both lymphoid cells and sorted T cells (Fig. 5U, V, Supplementary Fig. S7C, D). In addition, the lack of AMPK signaling exacerbated T cell apoptosis in lymphoid cells cultured in glucose-free conditions (Fig. 5W). Taken together, these findings support the notion that tilapia AMPK signaling, which is crucial for maintaining T cell activation, proliferation, and survival, serves as a central hub to induce autophagy and ensure proper T cell immunity under energy restriction.

Fasting-induced autophagy rescues T cell apoptosis and reduces inflammatory tissue damage during bacterial infection

Bacterial infection may also induce autophagy [38]. At 72 h after A. hydrophila infection, the expression of SQSTM1/P62 in spleen lymphoid cells was downregulated, whereas that of LC3B was upregulated (Fig. 6A). At the same time, the lysosome activity was markedly increased (Fig. 6B), suggesting that autophagy was induced by A. hydrophila infection. Since bacterial infection also reduced the dietary intake of tilapia (Fig. 1A), which could further trigger autophagy, the autophagy we observed during the bacterial infection may be caused by two different factors: bacterial infection and fasting. The precise function of autophagy occurring during bacterial infection was further investigated by infecting fed and short-term-fasted tilapia with A. hydrophila. At 24 h after infection, LC3B expression was slightly upregulated in spleen lymphoid cells of normally fed tilapia, but the AMPKα phosphorylation and the SQSTM1/P62 degradation were limited (Fig. 6C, left panel), indicating a weak autophagy response. However, in lymphoid cells of fasted tilapia, a robust autophagy response had been initiated (Fig. 6C, left panel). At 72 h after infection, AMPK signaling was evidently activated in the normally fed group, and the expression pattern of SQSTM1/P62 and LC3B also signified the occurrence of autophagy (Fig. 6C, middle panel); in contrast, in fasted tilapia, the autophagy response had already reached a higher level (Fig. 6C, middle panel). At 120 h after infection, autophagy in the lymphoid cells of fed tilapia reached a peak, whereas autophagy in the fasting group already returned to the normal levels (Fig. 6C, right panel). Importantly, not only at the lymphoid cell level, these observations were also echoed at the T cell level; in gated CD3⁺ T cells, both lysosomal activity (Fig. 6D, E) and LC3 expression (Fig. 6F, G) exhibited the same trends with those in lymphoid cells. Thus, these findings suggest that short-term fasting during bacterial infection triggers lymphoid cell and T cell autophagy at an earlier stage of infection and maintains autophagy within a controllable degree.

Fig. 6.

Fasting-induced AMPK-dependent autophagy rescues T cell apoptosis and inflammatory tissue damage during infection. A–K Fed or fasted tilapia was infected with A. hydrophila or not, and the spleen leukocytes was harvested. A, C Western blot assay showing the expression of the indicated molecules at 24, 72 or 120 h after infection. B Fluorescence microscope image of spleen leukocytes stained with LysoTracker at 72 h after infection, microscopy setting: eyepiece (e.g. 10 ×) and objective lens (e.g. 40 ×). D–G Overlaid histogram showing the comparation of LysoTracker (D) and LC3 (F) in gated T cells, and scatter plot figures showing their MFI (E, G), n = 4–6. H–K Representative FACS plots (H, I) and scatter plot figures (J, K) showing the gating and statistic of 7-AAD and Annexin V staining on T cell population at 24, 72 or 120 h after infection, n = 4–6. L Fed or fasted tilapia that infected with A. hydrophila were i.p. injected with or without 10 mg/kg 3-MA. Images showing the H&E staining of the liver or spleen at 48, 144 h after infection. Red arrow: inflammatory cell infiltration; black arrow: tissue vacuolization; blue arrow: immune cell aggregation. These experiments were repeated for three independent times. *: p < 0.05, **: p < 0.01, ***: p < 0.001, determined by a two-tailed Student’s t test

Next, we sought to determine whether this autophagy regulatory pattern triggered by short-term fasting at the early stage of infection is beneficial to the immune response of tilapia. At 72 or 120 h after A. hydrophila infection, tilapia T cells underwent significant apoptosis, whereas short-term fasting significantly reduced T cell apoptosis (Fig. 6H–K). However, inhibition of autophagy counteracted the rescue of T cell apoptosis achieved by fasting (Fig. 6I, K) or even aggravated cell death (Fig. 6H, J), suggesting that autophagy induced by short-term fasting plays an important role in maintaining T cell survival during bacterial infection. In addition, normally fed tilapia exhibited marked hepatic tissue damage and cellular vacuolization at 48 and 144 h post-infection, which was significantly rescued by short-term fasting (Fig. 6L, left panel). In fasted tilapia, the spleen showed higher immune cell infiltration after infection (Fig. 6L right panel, Supplementary Fig. S8). However, when autophagy was inhibited, fasting did not effectively rescue the hepatic tissue damage induced by bacterial infection and impaired the expansion of immune cells in the spleen (Fig. 6L), suggesting that autophagy induced by short-term fasting promotes cellular immune response and avoids inflammatory tissue damage during bacterial infection. Taken together, our findings suggest that in tilapia, short-term fasting during bacterial infection triggers the AMPK-dependent autophagy at an earlier stage of infection to maintain T cell survival, rescue inflammatory tissue damage, and promote cellular immune response, thus representing an evolutionarily ancient strategy to enhance anti-infection immunity via dietary restriction.

Discussion

Transient dietary restriction or fasting optimizes immunity in human and mouse and has the potential to inform the development of novel therapies to prevent and treat disease [7, 39]. However, the regulation of dietary restriction on the immune response of early vertebrates and underlying mechanisms remain unknown. In the present study, we found that long-term fasting severely impaired T cell immunity, whereas short-term dietary restriction enhanced the T cell response in tilapia, and such diverse immunological outcomes are directly associated with the intensity of autophagy induced by fasting. These findings highlight the key role of diet-modified lymphoid cell, especially T cell immunity in early vertebrates and provides a novel perspective for understanding the adaptive evolution of the immune system.

Short-term or intermittent fasting did not change the proportion of T cells in human peripheral blood, but refeeding significantly increased the frequency of activated T cells [4]. On the basis of this strategy, dietary interventions that promote the activation of CD8⁺Granzyme B⁺ and CD4⁺FoxP3⁻ T cells have been developed to optimize tumor therapy in mouse models [3, 39]. In the present study, we speculate that such a starvation strategy is an evolutionarily conserved principle, as similar results were observed in the early vertebrate Nile tilapia. Short-term fasting for 3 days did not affect T cell homeostasis but improved T cell activation and function. Although antigen signaling is considered to be essential for T cell activation, findings in both tilapia and human indicate that fasting could induce T cell activation, to some extent, in the absence of antigen stimulation. Our study reveals an evolutionarily conserved immunological strategy responding to energy restriction and provides novel insights into T cell activation.

Because mammals such as mouse and human have a limited tolerance to starvation, the effects of prolonged fasting on the immune system remain largely unknown. Bony fish survive for months without dietary intake, providing an ideal model for investigating this question. We found that a dietary restriction for more than 7 days induced spontaneous lymphoid cell proliferation and inflammatory response, and triggered severe spleen atrophy and T cell apoptosis and reduced the number of T cells, which consequently impaired the anti-infection immune response of fish. These findings demonstrate that although bony fish survive long-term food deprivation, their immune systems are consequently weakened. Therefore, we speculate that in bony fish, the immune system integrates dietary cues to modify the immune response on the basis of the nutritional status, and under conditions of severely limited resources, trade-offs and priorities emerge to sustain processes most critical to survival at the expense of immunity. Remarkably, we found that T cells in tilapia head kidney were considerably less susceptible to long-term fasting than those in the spleen. The head kidney of bony fish is similar to mammalian bone marrow and represents a site for the development of a variety of immune cells [40, 41]; therefore, we speculate that head kidney, similar to mammalian bone marrow, provides a “metabolic refuge” for T cells and prevents substantial T cell apoptosis caused by prolonged exposure to fasting-induced glucocorticoids [17]. Although further investigation is still needed to explore the underlying mechanisms, our findings provide a novel perspective for understanding the immune response under severe nutritional restriction.

Although emerging studies have indicated the essential roles of metabolic programs and their reprogramming in the determination of T cell function and fate [11, 42, 43], how metabolic switching under fasting conditions affects T cell immunity remains unclear. Similar to the observations reported in mammals [3, 44], immune cells of tilapia shifted their metabolic profile away from glycolysis and showed enhanced fatty acid oxidation. In addition, enhanced transcription levels of genes involved in fatty acid synthesis were also observed in tilapia, which may be associated with the T cell activation and proliferation triggered by fasting. In fact, a metabolic pathway closely associated with dietary restriction is autophagy, which recycles building blocks and energy for cellular renovation and homeostasis, allowing cells to adapt to stress [45, 46]. Recently, the significant role of autophagy in maintaining an optimal T cell response and preventing pathogenic infection and a wide variety of diseases has also been highlighted [47, 48]. For example, in mouse, the dysfunctional or defective autophagy resulting from inhibitor administration or gene knockout impaired the differentiation of Th1, Th17, and T_reg cells [49–51]; weakened the CD8⁺ effector T cell response during influenza infection [48]; and aggravated the apoptosis of CD4⁺ T cells and failed to establish CD8⁺ memory T cells [47, 52]. In the present study, the blockade of autophagy also markedly impaired T cell activation, proliferation, and effector function in resisting A. hydrophila infection in the bony fish tilapia; therefore, we speculated that using autophagy to ensure normal T cell function is a strategy conserved in vertebrate evolution. However, whether and how fasting regulates the T cell response through autophagy remains unclear. In tilapia, short-term fasting triggered T cell autophagy, which prevented T cell apoptosis and enhanced their activation, indicating the optimization of T cell immunity. These findings demonstrate that autophagy bestowed tilapia T cells with high adaptability and extreme resilience to thrive in conditions of fluctuating nutrient availability and infection. Remarkably, autophagy is a double-edged sword; long-term or excessively activated autophagy is harmful for cells and may result in cell death [36, 53]. We found that in tilapia, fasting for more than 7 days induced intense immune cell autophagy, accompanied by spontaneous inflammation and severe T cell apoptosis, which impaired the immune response. This finding explains why strategies that use dietary restriction and autophagy to boost immunity need to be kept within limits. However, from an energy cycle perspective, the impairment of immune cell autophagy under conditions of severe malnutrition may represent a survival strategy of bony fish to ensure nutrition supply at the expense of the immune system. Therefore, the present study elucidates the molecular mechanism of the bidirectional regulation of T cell immunity by fasting from a novel perspective and is expected to provide insights for disease prevention and therapy.

As a central regulator of metabolism and energy balance in many cells, AMPK can sense the upstream signaling from LKB1 or CaMKK2 to be activated [54, 55] and then manipulates the T cell function and differentiation [56, 57]. However, AMPK signaling exerts both positive and negative effects on T cell responses, as observed in different cell types, infection or tumor models, and inhibition or knockout approaches [57, 58]. For example, activating AMPK signaling with agonists impaired IFN-γ expression and effector T cell function [57]; this finding was corroborated by the observation that LKB1 deficiency led to an enhanced CD4⁺ Th1 response [59], but this effect was not observed in AMPKα1-deficient CD4⁺ T cells [57]. In contrast, other studies report positive effects of AMPK signaling on T cell immunity; both antibacterial and antitumor responses of CD8⁺ T cells require robust AMPK signaling [57, 60]. In tilapia, the blockade of AMPK using a specific inhibitor impaired T cell activation and expansion and compromised the inducible expression of cytotoxic genes during bacterial infection, indicating that AMPK plays an indispensable role in the primary T cell response. Mammalian AMPK triggers autophagy under energy restriction; a similar regulatory mechanism was also observed in tilapia T cells, representing an evolutionarily conserved mechanism to balance energy consumption and immune response. Moreover, we found that AMPK is an important regulator that triggers autophagy under fasting conditions and therefore acts as a central hub linking energy, autophagy, and T cell immunity in tilapia. Notably, although fasting decreased blood glucose levels, it did not elevate the AMP/ATP ratio as reported in mammals [61], suggesting that an AMP-independent mechanism of AMPK activation exists in tilapia. In addition, A. hydrophila infection also activated the AMPK pathway in tilapia, which may be a consequence of elevated Ca²⁺ influx downstream of TCR signaling [62] or the metabolic imbalance caused by nutrient usage by the bacteria. Although the antibodies regarding T cell activation signaling, autophagy signaling, and AMPK signaling in this study are commercial ones against human species, they actually recognize tilapia proteins specifically (Supplementary Fig. S9), therefore providing helpful tools for us to elucidate the exact regulatory mechanism in this teleost species. Notably, autophagy and AMPK inhibition could not exclude extrinsic interference caused by other cell lineages, and the tissue-specific knockout models is expected to more precisely elucidate the intrinsic regulation of autophagy and AMPK on T cell immunity in tilapia.

Although a tight association exists between autophagy and bacteria clearance [63], autophagy seems to play different roles in resisting the infection caused by intracellular and extracellular bacteria. The intracellular bacteria can be ubiquitinated after invading cells and then be degraded by an autophagic process named “xenophagy” [64]. In contrast, although extracellular bacteria cannot invade the host cells, they destroy the cell structure via secreting various virulence factors, while autophagy plays a key role in recycling the damaged cellular structures [65, 66]. The A. hydrophila used in present study is generally considered as an extracellular bacterium [67, 68], whose infection could trigger the autophagy in common carp [69]. Similarly, here we found that A. hydrophila was able to initiate autophagy in tilapia lymphoid cell and T cells. However, we still know very few about the regulatory mechanism by which tilapia T cells activate the autophagy during A. hydrophila infection, and further study is needed to elucidate it in the future.

Considering that both fasting and bacterial infection trigger AMPK-dependent autophagy in tilapia, we speculate that a strategy to balance energy and T cell immunity has evolved in early vertebrates (Fig. 7). Although the precise mechanism of activation of AMPK signaling under the dual factors of dietary restriction and bacterial infection has not been fully understood, our findings at least support that transient fasting enhances T cell immunity in tilapia through the AMPK-dependent autophagy. This is demonstrated not only by the finding that mild fasting helped tilapia to initiate autophagy earlier during bacterial infection and optimize T cell function but also by the observation that it maintained autophagy at a moderate level to alleviate T cell apoptosis, inflammatory response, and tissue damage (Fig. 7). However, a long-term fasting triggers intense autophagy of immune cells, accompanied by spontaneous inflammation and severe T cell apoptosis, representing another survival strategy of the tilapia to ensure nutrition supply at the expense of the immune system (Fig. 7). Our findings highlight the mechanism by which tilapia enhances the plasticity of the immune system by regulating its diet and have important implications for understanding the fasting strategy used upon infection or environmental changes.

Fig. 7.

Dietary restriction bidirectionally modulates T cell immunity of tilapia via AMPK-dependent autophagy

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

KML designed and performed experiments, analyzed data, and drafted the manuscript. XW acquired funding, conceived and administrated the project. KL, QZ and JZ performed experiments. DW kept the experimental animals. JY acquired funding, conceived the project, designed experiments, analyzed data, and drafted the manuscript.

Funding

This research was supported by grants from National Natural Science Foundation of China to Dr. Jialong Yang (no. 32022086) and Dr. Xiumei Wei (no. 31972822); and Natural Science Foundation of Shanghai to Dr. Xiumei Wei (no. 20ZR1417500).

Data availability

All data needed to evaluate the conclusions in the paper are presented in the paper or the Supplementary Materials.

Declarations

Conflict of interest

The authors declare no competing interests.

Ethics approval

All animal experiments were conducted in accordance with the guidelines for the Care and Use of Laboratory Animals of the Ministry of Science and Technology of China, and approved by the East China Normal University Experimental Animal Ethics Committee with an approve number of AR2021-247. All efforts were made to minimize the pain of animals.

Consent to participate

The study does not contain clinical studies or patient data.

Consent to publish

Not applicable.