Zeaxanthin augments CD8⁺ effector T cell function and immunotherapy efficacy

Summary

The detailed mechanisms underlying the regulatory significance of dietary components in modulating anti-tumor immunity remain largely unknown. Here, we apply a co-culture-based screening approach using a blood nutrient compound library and identify zeaxanthin (ZEA), a dietary carotenoid pigment found in many fruits and vegetables and known for its role in eye health, as an immunomodulator that enhances the cytotoxicity of CD8⁺ T cells against tumor cells. Oral supplementation with ZEA, but not its structural isomer lutein (LUT), enhances anti-tumor immunity in vivo. Integrated multi-omics mechanistic studies reveal that ZEA promotes T cell receptor (TCR) stimulation on the CD8⁺ T cell surface, leading to improved intracellular TCR signaling for effector T cell function. Hence, ZEA treatment augments the efficacy of anti-PD1 immune checkpoint inhibitor in vivo and the cytotoxicity of human TCR gene-engineered CD8⁺ T cells in vitro. Our findings uncover a previously unknown immunoregulatory function of ZEA, which has translational potential as a dietary element in bolstering immunotherapy.

Graphical abstract

Highlights

• •Plant-produced nutrient zeaxanthin enhances the cytotoxicity of CD8⁺ T cells
• •Orally supplemented zeaxanthin augments anti-tumor immunity
• •Zeaxanthin promotes T cell receptor stimulation on CD8⁺ T cell surface
• •Zeaxanthin treatment improves immunotherapy efficacy

Mechanistic insights into how dietary components regulate anti-tumor immunity remain limited. Zhang et al. identify zeaxanthin—a carotenoid in many fruits and vegetables—as an immunomodulator that enhances CD8⁺ T cell function, highlighting its translational potential in supporting immunotherapy.

Introduction

The evolution of diet has played an essential role in shaping human physiology and pathology, including the development of the immune system and its responses to environmental stimuli.^1,2,3 However, despite extensive epidemiological studies, the mechanistic understanding of how different diets impact the human immune system is very limited due to the intricate nature of the immune system, whole-organism metabolism, and the vast diversity of dietary components available. Emerging studies on the diet-immune axis have uncovered diet-derived nutrients that, besides performing their metabolic functions to provide energy and biosynthesis building blocks, serve immunoregulatory functions to influence diverse immune cell populations.⁴ We thus developed an approach to focus on the bioactivity of individual nutrients, which allows us to conduct mechanistic studies to decipher the links between diet and immunity, despite the vast diversity of foods and dietary origins. We assembled a library of “blood nutrients” that are commercially available, including dietary supplements, inorganics, organic metabolites, peptides, and lipids, and screened for circulating blood nutrients that affect CD8⁺ T cell function. We uncovered that dietary nutrient trans-vaccenic acid (TVA) selectively promotes effector CD8⁺ T cell function and anti-tumor immunity in vivo, through TVA-mediated antagonism of an immunomodulatory G-protein-coupled receptor, GPR43, on the cell surface of CD8⁺ T cells.⁵

Here, using this blood nutrient compound library, we applied a co-culture-based screening approach to identify circulating nutrients that function as immunomodulators to affect the cytotoxicity of mouse Pmel-1 CD8⁺ T cells against mouse B16F10 melanoma tumor cells. We uncovered zeaxanthin (ZEA), a dietary nutrient found in many fruits and vegetables, as an unexpected immunomodulator, which enhances CD8⁺ effector T cell function and consequently improves anti-tumor immunity in vivo.

ZEA is a non-provitamin A carotenoid synthesized only by plants and microorganisms, which can be found in grains such as corn and corn products, as well as many green leafy vegetables, and is commonly used as a feed additive and colorant for birds, swine, and fish.^6,7,8 While more than 700 carotenoids are found in nature, only 40–50 have nutritional values and about 20 of them reach detectable levels in human circulation and tissues. ZEA and its structural isomer lutein (LUT) are among the most prevalent circulating carotenoids in the human body, but they can only be incorporated into the human circulation through consumption from diet and dietary supplements.^9,10 The chemical structures of ZEA and LUT differ only in the position of a double bond in one cyclic ring, resulting that ZEA has two β-ionone rings while LUT has a β-ionone ring and an ε-ionone ring.¹¹ In addition, among the three stereoisomeric forms of ZEA, (3R,3′R)-zeaxanthin is the most common natural form found in many fruits and vegetables and is the predominant isomer present in the human retina. Therefore, (3R,3′R)-zeaxanthin was used in our study. Both ZEA and LUT are found to concentrate at the macular region of the human retina, with ZEA being the dominant component in the central macula and LUT distributed more broadly throughout the retina.¹² Because of their blue-light-filtering and antioxidant roles, ZEA and LUT are commonly consumed as dietary supplements to protect the retina and eye tissues for vision health. Previous studies have revealed ZEA’s role in photoprotection, energy transfer, and antioxidation.^12,13 However, the distinct immunomodulatory functions of ZEA and LUT in the context of anti-tumor immunity were previously unknown.

Here, we show that ZEA interacts with the TCR complex and promotes stimulation on the cell surface of CD8⁺ T cells, leading to enhanced TCR complex formation, activation of intracellular TCR signaling to augment CD8⁺ effector T cell function, and consequent anti-tumor immunity.

Results

Carotenoid nutrient ZEA enhances cytotoxicity of CD8⁺ T cells and anti-tumor immunity

Using our blood nutrient library⁵ including dietary supplements, lipids, peptides, organic metabolites, and inorganics (Figure S1A), we performed a co-culture screen to identify circulating nutrients that influence the cytotoxicity of mouse Pmel-1 CD8⁺ T cells stimulated by anti-CD3/CD28 antibodies against co-cultured B16F10-luc2 mouse melanoma cells (Figure 1A). The screen results (Figure 1B; Table S1) show that ZEA, but not its structural isomer LUT, effectively enhanced Pmel-1 mediated cell death of B16F10-luc2 cells. In addition, treatment with ZEA, but not LUT, augmented the ability of OT-I CD8⁺ T cell-mediated cytotoxicity to induce cell death against co-cultured B16-OVA cells (Figure 1C). Notably, ZEA exhibited no significant direct cytotoxicity toward either tumor cell line at the concentration used in the co-culture assays (Figure S1B). Moreover, pre-treatment of cancer cells with ZEA did not intrinsically sensitize them to killing mediated by CD8⁺ T cells (Figure S1C).

Figure 1.

Oral ZEA enhances CD8⁺ T cell-mediated anti-tumor immunity

(A) Schematic depicting experimental design for effector T cell cytotoxicity screen. This schematic was generated using BioRender.com.

(B) Chemical structures of ZEA and LUT (upper) and volcano plot showing results from the library screen with ZEA and LUT highlighted (lower).

(C) Effects of 5 μM ZEA or LUT on B16-OVA cell death when treating B16-OVA cells alone and treating OT-I primary murine T cell and B16-OVA cell co-culture (n = 3 technical replicates).

(D) Schematic depicting experimental design for in vivo tumor-bearing mouse model. This schematic was generated using BioRender.com.

(E) Effect of orally administered ZEA (500 mg/kg b.w.) on the growth of subcutaneous B16F10 melanoma in C57BL/6 mice (n = 11 mice).

(F) Effect of orally administered ZEA (500 mg/kg b.w.) on the growth of subcutaneous MC38 colon tumors in C57BL/6 mice (n = 10 mice).

(G) Schematic depicting experimental design for CD8⁺ (or CD4⁺) T cell depletion experiment. This schematic was generated using BioRender.com.

(H) Effect of orally administered ZEA (500 mg/kg b.w.) on B16F10 tumor growth in C57BL/6 mice treated with isotype control (left) or CD8⁺ T cell-depleting antibody (right) (n = 8 mice).

b.w., body weight. Data are mean ± SEM (E, F, and H) or mean ± SD (C). p values are calculated using two-way ANOVA (E, F, and H) or one-way ANOVA with Dunn’s multiple comparisons test (C) (ns, not significant; ∗0.01 < p < 0.05; ∗∗0.001 < p < 0.01; ∗∗∗p < 0.001). Also see Figure S1 and Table S1.

We next found that in vivo tumor growth of poorly immunogenic B16F10 cells (Figure 1D) was significantly attenuated in syngeneic mice receiving oral gavage of ZEA (Figures 1E and S1D), but not in mice receiving LUT (Figure S1E), compared to B16F10 syngeneic mice receiving oral gavage of control vehicle. We detected increased levels of ZEA and LUT in the tumor interstitial fluid (TIF) of tumors in the syngeneic mice (Figure S1F), whereas oral administration of neither ZEA nor LUT affected body weights of syngeneic tumor-bearing mice (Figure S1G). Further absolute quantification showed that oral gavage of ZEA significantly increased its levels in both mouse plasma and TIF, with overall higher concentrations observed in the TIF (Figure S1H). Similar results were obtained using mouse colon cancer MC38 cells in syngeneic mice receiving oral gavage of ZEA or vehicle control (Figures 1F and S1I). Moreover, we found that depletion of CD8⁺ T cells by anti-CD8 antibody in B16F10 syngeneic mice (Figures 1G and S1J) resulted in abolishment of ZEA-mediated reduction of tumor growth (Figure 1H). In contrast, depletion of CD4⁺ T cells by anti-CD4 antibody had minimal effects on ZEA-mediated tumor growth in B16F10 syngeneic mice (Figures S1K and S1L). These data together suggest that ZEA reprograms CD8⁺ T cells and consequently enhances anti-tumor immunity.

Oral ZEA improves recruitment and activation of tumor-infiltrating CD8⁺ T cells

Flow cytometry analyses revealed that oral ZEA supplementation to B16F10 syngeneic mice significantly increased both the number and proportion of CD8⁺ T cells within the CD45⁺ tumor-infiltrating leukocyte (TIL) population (Figure 2A). In contrast, tumor-infiltrating populations of CD4⁺ T cells, B cells, dendritic cells, M1 and M2 macrophages, neutrophils, and natural killer cells were not altered in B16F10 syngeneic mice by oral ZEA supplementation, compared to mice receiving vehicle control (Figures 2B and S1O). Further analysis of tumor-infiltrating CD8⁺ T cells with additional representative markers revealed that oral ZEA supplementation promoted CD8⁺ T cell function with increased expression levels of the activation marker CD69; the co-stimulatory receptor ICOS; and cytokines including tumor necrosis factor (TNF)-α, interferon (IFN)γ, and interleukin (IL)-2 (Figures 2C, S1N, and S1P). Oral ZEA supplementation increased the proportion of Ki-67⁺ cells but had no effect on the frequency of cleaved caspase-3⁺ cells in tumor-infiltrating CD8⁺ T cells (Figure S1P). While ZEA did not significantly alter the proportions of stem-like or effector-like exhausted T cells within the tumor microenvironment, it reduced the population of terminally exhausted CD8⁺ T cells (Figure S1Q). In addition, ZEA supplementation significantly increased the proportion of tumor-reactive CD8⁺ T cells within the TILs of MC38-OVA tumor-bearing mice (Figure S1R). Oral ZEA also resulted in increased CD4⁺ T helper 1 (Th1) but reduced CD4⁺CD25⁺Foxp3⁺ Treg cell populations in B16F10 tumors, while it had minimal effects on other tumor-infiltrating CD4⁺ T cell populations including Th2, Th17, Th9, and Th22 (Figure 2D). In contrast, oral ZEA supplementation did not affect the cell populations of diverse immune cells in spleens or draining lymph nodes (dLNs) in B16F10 syngeneic mice (Figures S1S and S1T). Lastly, oral ZEA supplementation did not significantly alter the diversity and composition of gut microbiota (Figure S2). Given that Th1 cells are critical in cellular immune responses against intracellular viruses, including activation of cytotoxic T cells,¹⁴ and that Treg cells suppress CD8⁺ T cells,¹⁵ these results together suggest that oral ZEA enhances recruitment and activation of CD8⁺ effector T cells to tumors for improved anti-tumor immunity.

Figure 2.

ZEA reprograms CD8⁺ T cells in vivo and in vitro

(A) Schematic depicting experimental setup for harvesting the spleen, tumor, and draining lymph node (dLN) to examine zeaxanthin’s effect on leukocytes. This schematic was generated using BioRender.com.

(B) Effect of orally administered ZEA (500 mg/kg b.w.) on the composition of tumor-infiltrating leukocytes (n ≥ 10 mice).

(C) Effects of orally administered ZEA (500 mg/kg b.w.) on tumor-infiltrating CD8⁺ T cell activation, co-stimulation, and pro-inflammatory cytokine TNF-α production (n ≥ 7 mice).

(D) Effects of orally administered ZEA (500 mg/kg b.w.) on tumor-infiltrating CD4⁺ T cell composition (n ≥ 7 mice).

(E) Schematic depicting experimental setup for in vitro primary murine CD8⁺ T cells treatment. This schematic was generated using BioRender.com.

(F) Effects of 5 μM ZEA or LUT treatment for 24 h on αCD3/CD28-stimulated murine CD8⁺ T cell proliferation, activation, and pro-inflammatory cytokine levels (n = 3 technical replicates).

(G) Chemical structures of ZEA and 6 natural structural analogs of ZEA.

(H) Effect of 24-h treatment with 5 μM ZEA or ZEA derivatives on primary murine CD8⁺ T cell TNF-α production (n = 3 technical replicates).

Data are mean ± SD. p values are calculated using Student’s two-sided unpaired t test (B, C, and D) or one-way ANOVA with Dunn’s multiple comparisons test (F and H) (ns, not significant; ∗0.01 < p < 0.05; ∗∗0.001 < p < 0.01). Also see Figure S1.

Consistent with these findings, we next found that ZEA but not LUT treatment had direct effects on and effectively promoted mouse primary CD8⁺ T cell function (Figure 2E) with increased expression levels of proliferation marker Ki-67; activation marker CD69; and cytokines including TNF-α, IFNγ, and IL-2 (Figure 2F). Given that ZEA and LUT are structurally similar but have distinct effects on CD8⁺ T cells, we next explored the structure-activity relationship (SAR) by testing several natural compounds that share structural similarity with ZEA, all of which are available in dietary sources (Figure 2G). Notably, only ZEA and D5 (fucoxanthin) significantly promoted CD8⁺ T cell activation, despite the fact that all compounds have long chains of conjugated double bonds (Figure 2H). This result suggests that the antioxidant properties of conjugated double bonds in these compounds are not sufficient to mediate enhanced activation of CD8⁺ T cells. Instead, the structurally defined symmetrical conjugated polyene capped with six-membered rings is crucial for ZEA to enhance CD8⁺ T cell activation. In addition, the different six-membered ring structures at both ends of the double bonds account for the differential bioactivity of ZEA compared to D1, D2, D4, and D5 for enhanced CD8⁺ T cell activation, while D3 with an open-chain structure and D6 with a short-chain monomer lacked bioactivity (Figure 2H). These data together suggest that the distinct structure and conformation of ZEA are vital for its bioactivity to enhance CD8⁺ T cell activation and function.

ZEA engages the T cell receptor complex and augments CD8⁺ T cell surface TCR stimulation

To determine the mechanisms underlying ZEA’s bioactivity in enhancing CD8⁺ T cell function, we performed kethoxal-assisted single-stranded DNA sequencing (KAS-seq) (40 min–2 h)¹⁶ to investigate the initial influences of ZEA treatment on mouse primary CD8⁺ T cells by capturing global transcription dynamics and enhancer activity. Functional Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment of the top-ranking altered genes from the genome-scale KAS-seq results (Figure 3A; Table S2) revealed the T cell receptor (TCR) signaling pathway among the top-enriched ontologies in ZEA-treated CD8⁺ T cells.

Figure 3.

ZEA augments CD8⁺ T cell surface TCR stimulation

(A) KEGG pathway enrichment for upregulated genes generated from KAS-seq analysis of αCD3/CD28-stimulated murine CD8⁺ T cells treated with 5 μM ZEA compared to cells treated with DMSO control (n = 3 biological replicates).

(B) Schematic depicting experimental setup for flow cytometry and confocal microscopy detection of αTCRβ, αCD3ε, and αCD28 antibody binding in murine CD8⁺ T cells with or without 5 μM ZEA or LUT treatment with different stimulation strategies. This schematic was generated using BioRender.com.

(C) Flow cytometry analysis showing the effects of 5 μM ZEA or LUT on αTCRβ, αCD3ε, and αCD28 antibody binding in murine CD8⁺ T cells at 5, 15, and 30 min after treatment and stimulation with αCD3ε and αCD28 antibodies (n = 3 technical replicates).

(D) Representative confocal microscopy images showing the effects of 5 μM ZEA on αTCRβ, αCD3ε, and αCD28 antibody binding in murine CD8⁺ T cells at 5, 15, and 30 min after treatment and stimulation with αCD3ε and αCD28 antibodies (n = 3 technical replicates). Scale bars represent 5 μm.

(E) Quantification of fluorescence intensities from confocal microscopy images shown in (D). TCRβ, CD3, and CD28 antibody signals were measured at 5, 15, and 30 min after treatment and stimulation with CD3 and CD28 antibodies (n = 3 technical replicates).

(F) The chemical structure of the ZEA-based photo-affinity labeling (PAL) probe and schematic workflow for validating its binding to the TCR complex.

(G) Pull-down of the TCR complex using the biotinylated ZEA probe. Mouse primary CD8⁺ T cells were incubated with DMSO and 50 μM ZEA probe or pre-treated with 250 μM ZEA for 1 h followed by 50 μM ZEA probe. After UV-induced crosslinking, samples were subjected to click chemistry with a biotin tag, followed by streptavidin pull-down and immunoblotting.

Data are mean ± SD. p values are calculated using two-way ANOVA (ns, not significant; ∗∗∗p < 0.001). Also see Figure S3 and Table S2.

TCR signaling in CD8⁺ T cells is triggered when the cell surface TCR binds to a major histocompatibility complex (MHC) class I-bound peptide antigen. This interaction triggers the assembly of the TCR signaling complex and activates downstream intracellular signaling cascades to turn on CD8⁺ effector T cell function including cytokine production and cytotoxicity.¹⁷ Given that polar carotenoids including ZEA and LUT can insert into plasma membranes and alter the membrane fluidity,^18,19 we examined the effects of ZEA or LUT treatment on TCR complex formation on the cell surface of mouse primary CD8⁺ T cells upon stimulation with anti-CD3/CD28 antibodies or on the cells with overnight pre-stimulation (Figure 3B). Flow cytometry analysis revealed that ZEA treatment for 5, 15, and 30 min effectively enhanced TCR complex formation on mouse primary CD8⁺ T cells assessed by increased levels of TCRβ, CD3ε, and CD28 on CD8⁺ T cell surface upon stimulation (Figure 3C; left two panels). In contrast, LUT treatment had minimal effects on TCR complex formation on mouse primary CD8⁺ T cell surface upon stimulation (Figure 3C; right two panels). Moreover, we performed confocal microscopy to visualize cell surface TCR complex formation on mouse primary CD8⁺ T cells upon stimulation with anti-CD3/CD28 antibodies. ZEA treatment for 5, 15, or 30 min effectively promoted TCR complex formation with increased detection of cell surface TCRβ, CD3ε, and CD28 on mouse primary CD8⁺ T cells (Figures 3D and 3E). Similar results were obtained from flow cytometry (Figure S3A) and confocal microscopy (Figure S3B) analyses using mouse primary CD8⁺ T cells with overnight pre-stimulation by anti-CD3/CD28 antibodies prior to treatment with ZEA or LUT. These findings prompted us to investigate whether ZEA binds to the TCR complex. To this end, we synthesized a ZEA photo-affinity labeling (PAL) probe. The ZEA PAL probe was incubated with intact mouse primary CD8⁺ T cells. Upon UV crosslinking, followed by click chemistry, we performed either in-gel fluorescence (via TAMRA labeling) or streptavidin pull-down (via biotin tagging) to assess probe-protein interactions (Figure 3F). TAMRA fluorescence revealed strong probe binding in wild-type CD8⁺ T cells, but significantly reduced signal in TCRα knockout cells, suggesting TCR-dependent binding (Figure S3C). Pull-down assays confirmed that the probe enriched TCRα, TCRβ, and CD3ζ subunits, and these interactions were specifically outcompeted by excess unmodified ZEA (Figure 3G). Functionally, ZEA-induced activation was completely abrogated in TCRα-deficient CD8⁺ T cells (Figure S3D). These results together suggest that ZEA engages the TCR complex and facilitates or stabilizes its formation at the cell membrane, thereby enhancing TCR-mediated signaling and CD8⁺ T cell activation.

ZEA-mediated CD8⁺ T cell activation involves intracellular Ca²⁺ signaling and nuclear factor κB pathway

We next sought to explore the downstream intracellular TCR signaling cascades that are crucial for ZEA-enhanced CD8⁺ T cell activation and function using mouse primary CD8⁺ T cells (Figure 4A). We designed integrated, temporal mechanistic studies, including (1) TCR-specific phospho-antibody array (5 min) for changes of proximal TCR signaling cascades following ZEA treatment, (2) conventional phospho-antibody array (40 min–6 h) for downstream cellular signaling changes, and (3) RNA sequencing (RNA-seq) (24 h) for whole-transcriptome analysis. TCR complex assembly upon antigen recognition activates tyrosine kinase Lck, leading to consequent phosphorylation and activation of downstream signaling molecules including ZAP-70 and phospholipase Cγ (PLCγ). This initiates activation of transcription factors such as nuclear factor κB (NF-κB) and NFAT, as well as a signaling cascade involving Ca²⁺ release from endoplasmic reticulum for CD8⁺ T cell functions including cytokine production and cytotoxicity.^20,21,22,23 Indeed, TCR phospho-antibody array analysis on mouse primary CD8⁺ T cells treated with ZEA revealed increased phosphorylation levels of ZAP-70, Lck, PLC, LAT, and NF-κB and reduced phosphorylation levels of NFAT as activation and nuclear localization of NFAT require dephosphorylation by Ca²⁺-dependent phosphatase calcineurin²⁴ (Figure 4B; Table S3). In contrast, LUT treatment had minimal effects on phosphorylation levels of these key proximal and downstream signaling molecules of the TCR complex in CD8⁺ T cells (Figure S4A). The results of the conventional cell phospho-antibody array (Figures S4B and S4C; Table S4) also revealed increased phosphorylation and activation of proteins in the JAK-STAT and ERK pathways, which are known to be activated by TNF in CD8⁺ T cells^25,26 and in mouse primary CD8⁺ T cells treated with ZEA, but not in cells treated with LUT.

Figure 4.

ZEA enhances TCR signaling cascades and effectiveness of T cell-based immunotherapies

(A) Schematic depicting temporal, mechanistic studies performed with murine CD8⁺ T cells treated with and without 5 μM ZEA or LUT.

(B) Volcano plot showing the effects of 5 min ZEA treatment on TCR-related protein phosphorylation (left). A list of changed protein phosphorylation in TCR signaling pathways (right).

(C) Chord diagram showing KEGG pathway enrichment of differentially expressed genes between ZEA treatment and DMSO in RNA-seq (n = 3 biological replicates).

(D) GSEA analysis of the gene set of IFNα-stimulated CD8⁺ T cell upregulated genes after 24 h ZEA or LUT treatment (n = 3 biological replicates).

(E) Effect of LCK inhibitor PP2 (200 nM) treatment on ZEA-dependent murine CD8⁺ T cell effector function assessed by TNF-α levels (n = 3 technical replicates).

(F) Effect of PLCγ inhibitor 3-nitrocoumarin (500 nM) treatment on ZEA-dependent murine CD8⁺ T cell effector function assessed by TNF-α levels (n = 3 technical replicates).

(G) Effect of 5 μM ZEA or LUT treatment on calcium indicator Fluo-4 levels in murine CD8⁺ T cells. Cells were stimulated with soluble anti-CD3 (1 μg/mL) and anti-CD28 (0.5 μg/mL) in the presence of 5 μM ZEA, 5 μM LUT, or DMSO control for 30 min, followed by flow cytometry analysis (n = 3 technical replicates).

(H) Effect of calcineurin inhibitor cyclosporin A (7 nM) treatment on ZEA-dependent murine CD8⁺ T cell effector function assessed by TNF-α levels (n = 3 technical replicates).

(I) Effect of NFAT inhibitor (1 μM) treatment on ZEA-dependent murine CD8⁺ T cell effector function assessed by TNF-α levels (n = 3 technical replicates).

(J) Effect of NF-κB inhibitor INK4 (2 μM) treatment on ZEA-dependent murine CD8⁺ T cell effector function assessed by TNF-α levels (n = 3 technical replicates).

(K) Effect of TVA (10 μM) and ZEA (5 μM) combined treatment on murine CD8⁺ T cell effector function assessed by TNF-α levels (n = 3 technical replicates). CI, combination index.

(L) Effect of αPD-1 antibody on B16F10 tumor growth in C57BL/6 mice orally supplemented with (CI = 0.886) or without ZEA (n ≥ 9 mice). CI, combination index

(M) Effect of aPD-1 antibody on MC38 tumor growth in C57BL/6 mice orally supplemented with (CI = 0.885) or without ZEA (n ≥ 7 mice). CI, combination index.

(N) Effect of 5 μM ZEA or LUT on human CD8⁺ T cell effector function assessed by IFNγ, IL-2, and TNF-α levels in human CD8⁺ T cells isolated from peripheral blood mononuclear cells of healthy donors (n = 10 biological replicates).

(O) Schematic depicting experimental design for human TCR-T transduction and ex vivo co-culturing with tumor cells to assess TCR-T cell-mediated cytotoxicity. This schematic was generated using BioRender.com.

(P) Flow cytometry analysis showing the effects of 5 μM ZEA or LUT on human 19305DP-TCR-T cell-mediated cytotoxicity against A375 melanoma cells and A375 melanoma cells alone. Human CD8⁺ cells were derived from a healthy donor (n = 3 technical replicates).

(Q) Flow cytometry analysis showing the effects of 5 μM ZEA or LUT on A375 melanoma cells alone, human untransduced (mock) CD8⁺ cells against A375 melanoma cells, T cell human 19305DP-TCR-T cell-mediated cytotoxicity against A375 melanoma cells, and human AL-TCR-T cell-mediated cytotoxicity against A375 cells. Human CD8⁺ cells were derived from a second healthy donor (n = 3 technical replicates).

Data are mean ± SEM (L and M) or mean ± SD (E–J, K, and M–Q). p values are calculated using two-way ANOVA (E, F, H–M, and Q), one-way ANOVA with Dunn’s multiple comparisons test (G and P), or Student’s two-sided paired t test (N) (ns, not significant; ∗0.01 < p < 0.05; ∗∗0.001 < p < 0.01; ∗∗∗p < 0.001). Also see Figure S4 and Tables S3, S4, S5, S6, and S7.

Consistent with these findings, principal-component analysis of RNA-seq results revealed that control mouse primary CD8⁺ T cells and cells treated with LUT can be grouped together and are separated from mouse primary CD8⁺ T cells treated with ZEA (Figure S4D), suggesting that ZEA selectively alters CD8⁺ T cells while LUT has minimal effects. Moreover, functional KEGG pathway analysis²⁷ and global gene set enrichment analysis (GSEA)²⁸ of RNA-seq results revealed that ZEA treatment enhanced the expression of the genes enriched in the TCR signaling pathway (Figure 4C; Tables S5, S6, and S7). Notably, ZEA treatment upregulated the expression of genes enriched in interferon-stimulated effector CD8⁺ T cells more effectively than LUT did (Figure 4D), correlating with enhanced CD8⁺ T cell function following ZEA treatment.

We next sought to determine which downstream pathway(s) of the TCR complex is required for ZEA function. ZEA but not LUT treatment significantly increased phosphorylation levels of TCR downstream signaling molecules including CD3ζ, Lck, ZAP-70, LAT, and S6 kinase that is a downstream effector of the PI3K-AKT pathway (Figures S4E–S4I). Consistent with these findings, Lck or PLCγ inhibitors abolished ZEA-dependent enhancement of CD8⁺ T cell function (Figures 4E and 4F). PLCγ is a downstream signaling effector of TCR for rapid release of Ca²⁺ from the endoplasmic reticulum lumen to cytosol.²⁹ Indeed, we found that ZEA but not LUT treatment effectively increased intracellular Ca²⁺ level in CD8⁺ T cells, assessed by increased Ca²⁺-bound indicator Fluo-4 (Figures 4G and S4J). Consistent with this finding, treatment with specific inhibitors targeting Ca²⁺-dependent phosphatase calcineurin and its downstream substrate transcription factor NFAT obliterated ZEA-enhanced CD8⁺ T cell function, assessed by TNF-α production (Figures 4H and 4I). Moreover, we found that treatment with NF-κB inhibitor (Figure 4J), but not inhibitors targeting ERK or JAK that are activated by TNF (Figures S4K–S4N), also abolished ZEA-enhanced CD8⁺ T cell function assessed by TNF-α production. Taken together, our results reveal that TCR signaling cascades involving the Lck-PLC-Ca²⁺-NFAT and NF-κB pathways are crucial for ZEA-enhanced CD8⁺ T cell function.

ZEA augments T cell-based therapies

Our studies show that animal-derived TVA⁵ and plant-derived ZEA enhance CD8⁺ effector T cells and anti-tumor immunity through different molecular and signaling mechanisms. Consistent with these findings, combined treatments with different dosages of TVA and ZEA resulted in synergistic effects to enhance primary mouse CD8⁺ T cell function in vitro (Figures 4K, S4N, and S4O). We found that oral ZEA supplementation combined with anti-PD-1 antibody, a representative form of immune checkpoint inhibitor therapy,³⁰ synergistically suppressed tumor growth in both B16F10 and MC38 tumor models (Figures 4L and 4M). In addition, ZEA treatment was significantly more potent than LUT treatment in enhancing human CD8⁺ T cell activation and function assessed by the production of cytokines, including IFNγ, IL-2, and TNF-α (Figure 4N). Furthermore, ZEA treatment improved in vitro cytotoxicity of human TCR-engineered T (TCR-T) cells derived from primary CD8⁺ T cells of healthy donors (Figure 4O). For example, ZEA treatment was significantly more potent than LUT in enhancing cytotoxicity of A∗02-restricted NY-ESO-1 tumor antigen-specific 19305DP³¹ TCR-T cells derived from a healthy donor against co-cultured A∗02⁺NY-ESO-1⁺ A375 human melanoma cells (Figure 4P; upper), while both ZEA and LUT have minimal direct cytotoxic effects on A375 cells (Figure 4P; lower). ZEA but not LUT also effectively promoted cytotoxicity of both 19305DP and A∗02-restricted NY-ESO-1-specific AL-TCR-T³¹ cells derived from another healthy donor against co-cultured A375 cells (Figure 4Q). Moreover, consistent with our findings using mouse primary CD8⁺ T cells, treatment with PLCγ inhibitor abolished ZEA-enhanced cytotoxicity of AL-TCR-T cells on A375 cells (Figure S4P). Additionally, we included a third donor and co-cultured TCR-T cells with A∗02⁺NY-ESO-1⁺ A375, A∗02⁺NY-ESO-1⁺ U266 (multiple myeloma), and A∗02⁺NY-ESO^-U87 (glioblastoma) cell lines under ZEA treatment to demonstrate its anti-tumor effects on human T cells across different tumor types. ZEA enhanced the cytotoxic activity of TCR-T cells against all three tumor cell lines (Figure S4Q), while exhibiting minimal direct cytotoxicity toward the tumor cells themselves at the concentration used in the co-culture treatment (Figure S4R). These findings together align with the concept that ZEA supplementation potentially improves clinical responsiveness to T cell-based immunotherapies.

Discussion

Hereby our findings reveal a previously unknown immunomodulatory function of ZEA, a carotenoid antioxidant known for its role in eye health. Dietary ZEA can be found in many fruits and green leafy vegetables such as kale, spinach, broccoli, peas, lettuce, durum wheat, and corn, as well as egg yolks, because common feeds for chickens contain grains such as corn and vegetables enriched with ZEA.³² It is widely recognized that plant-based or vegan diets can potentially benefit the human immune system.³³ However, the underlying mechanistic link between plant-derived nutrients and human immunity remains largely unknown. Our strategy focusing on individual diet-derived nutrients despite the vast diversity of food and diet origins makes mechanistic studies feasible to explore nutritional influences on human health and the immune system. Using this approach, we successfully identified dietary TVA derived from animal-based diets⁵ and hereby ZEA as a plant-produced nutrient, both of which have direct yet mechanistically distinct effects on CD8⁺ effector T cells and anti-tumor immunity. Therefore, our approach has broad implications to uncover unprecedented roles of nutrients in human health and pathologic responses.

Our findings support the high translational potential of ZEA, as a natural food component, to serve as a dietary supplement that enhances clinical outcomes of immunotherapies such as ICIs and TCR-T cell therapy. By investigating the physiological influences of diverse diet-derived nutrients on the immune system, our studies contribute to a comprehensive understanding of how dietary nutrients shape human biology, which is crucially informative for dietary choices that may be beneficial for maintaining human health, reducing disease risk, and improving response to therapies. Our studies on TVA from animal-based diets⁵ and ZEA from plant-based foods demonstrate that the scientific and mechanistic understanding of how nutrients from different diets influence human health are still limited. The finding that TVA and ZEA have mechanistically distinct effects on CD8⁺ effector T cells and anti-tumor immunity suggests that a balanced diet may provide complementary benefits to human health.

TCR signaling plays a vital role in activation and differentiation of CD8⁺ T cells, while stronger signaling strength of TCR generally promotes cytotoxic activity.³⁴ Our findings suggest that ZEA enhances cytotoxicity of CD8⁺ T cells through multiple intracellular TCR signaling cascades, which may act in concert to achieve fulfillment of effective cytotoxicity. Our results demonstrate a crucial coordination between the Lck-PLC-Ca²⁺-NFAT axis and the NF-κB pathway to mediate the pro-cytotoxic effects of ZEA on CD8⁺ T cells. Targeting either of them results in abolishment of ZEA-enhanced CD8⁺ T cell function, suggesting that the Lck-PLC-Ca²⁺-NFAT axis and the NF-κB pathway are indispensable to each other. In contrast, the ERK and JAK-STAT pathways are likely downstream effectors of TNF that is produced by CD8⁺ T cells when TCR signaling is activated upon antigen recognition; thus, treatment with specific inhibitors of ERK or JAK did not abolish ZEA-enhanced CD8⁺ T cell function assessed by TNF-α production. These findings also support a crucial role of TCR signaling cascades involving Ca²⁺ signaling and the NF-κB pathway in promoting the cytotoxic activity of CD8⁺ T cells.

Limitations of the study

ZEA and LUT as polar carotenoids can insert into plasma membranes and alter the membrane fluidity.^18,19 Future studies are warranted to elucidate the potential effects of ZEA on plasma membrane properties that may contribute to the alteration of TCR clustering on the cell surface of CD8⁺ T cells, leading to enhanced formation of TCR complex. We noted that, although tumor-infiltrating CD8⁺ T cells were the only immune cell population showing an increased proportion among the cell types analyzed following oral ZEA supplementation, we cannot completely rule out the possibility that ZEA may also modulate the function of other immune cells, which could in turn indirectly influence CD8⁺ T cell responses. Another limitation of our study is that, although CD4⁺ T cell depletion did not affect the anti-tumor efficacy of ZEA in vivo, the differential effects of ZEA on CD8⁺ versus CD4⁺ T cells remain to be fully elucidated. This discrepancy may arise from differences in their TCR co-receptors—CD8 and CD4—which interact with distinct MHC molecules and influence TCR signaling strength. It will also be crucial to determine the structural basis that distinguishes the differential bioactivities of ZEA and LUT to promote cytotoxicity of CD8⁺ T cells, despite their structural similarity. Our studies on several natural homologs of ZEA revealed that the antioxidant properties of conjugated double bonds in their long chains are not sufficient for enhancing cytotoxicity of CD8⁺ T cells. The specific structure of ZEA with symmetrical conjugated polyene and two specially arranged six-membered rings at both ends is vital for its bioactivity to promote CD8 T cell activation. Note that D5 (fucoxanthin) has more potent effects than ZEA to enhance TNF-α production by CD8⁺ T cells. Fucoxanthin is also a plant-produced carotenoid nutrient found in marine products such as seaweed, which has anti-cancer effects.^35,36 Future studies are warranted to further evaluate fucoxanthin as a natural, dietary nutrient for its translational potential to enhance CD8⁺ T cell function and related immunity in vivo. Future detailed SAR studies on ZEA and fucoxanthin will provide insights into the design and development of derivatives with improved potency for CD8⁺ T cell activation and for optimizing efficacy of T cell-based therapies as a dietary supplement.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Jing Chen (jingchen@uchicago.edu).

Materials availability

All reagents generated in this study are available from the lead contact with a completed materials transfer agreement.

Data and code availability

• •16S amplicon sequencing data have been deposited in the NCBI Sequence Reach Archive (SRA) : PRJNA1189274 and are publicly available as of the date of publication. The KAS-seq data have been deposited in the GEO : GSE282686 and are publicly available as of the date of publication. The RNA-seq data have been deposited in the GEO : GSE282703 and are publicly available as of the date of publication.
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

This work was supported in part by NIH grants including CA140515, CA174786, CA276568 (J.C.), HG006827 (C.H.), and Predoctoral T32 5T32CA009594 (F.Q.Z.); Ludwig Center at the University of Chicago (C.H.); Sigal Fellowship in Immuno-oncology (H.F.); and Harborview Foundation Gift Fund (C.H. and J.C.). C.H. is an Investigator of the Howard Hughes Medical Institute.

Author contributions

Conception and design: F.Q.Z., J.L., H.F., and J.C. Development of methodology: F.Q.Z., J.L., R.Z., J.T., H.S., and J.H. Acquisition of data: F.Q.Z., J.L., H.F., R.Z., J.T., Z.X., H.S., K.X., F.C., C.Y., Y.L., and H.M. Analysis and interpretation of data (e.g., statistical analysis, biostatistics, and computational analysis): F.Q.Z., J.L., H.F., M.O.R., R.L., L.D., and R.S. Providing essential materials: T.T., J.M., A.T., H.R.Y., S.S., M.W.D., K.O., Y.-Y.H., and B.M.B. Writing, review, and/or revision of the manuscript: F.Q.Z., J.L., H.F., and J.C. Study supervision: H.F., C.H., and J.C. All authors approved the final manuscript.

Declaration of interests

J.C. has patents pending on ZEA. C.H. is a scientific founder, a member of the scientific advisory board, and an equity holder of Aferna Bio, Inc. and Ellis Bio Inc.; a scientific cofounder and equity holder of Accent Therapeutics, Inc.; and a member of the scientific advisory board of Rona Therapeutics.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rat anti-IgG2b isotype	BioXCell	Cat#BE0090; Clone#LTF-2; RRID:AB_1107780.
Mouse anti-CD8α	BioXCell	Cat#BE0061; Clone#2.43; RRID:AB_1125541.
Mouse anti-PD-1(CD279)	BioXCell	Cat#BE0146; Clone# RMP1-14; RRID:AB_10949053.
Mouse anti-CD4	BioXCell	Cat#BE0288; Clone# GK1.5; RRID: AB_2687811
Mouse FITC anti-CD3ε Antibody	Biolegend	Cat#100305; Clone#145-2C11; RRID:AB_312670.
Mouse PE anti-CD28 Antibody	Biolegend	Cat#102105; Clone#37.51; RRID: AB_312870.
Mouse APC anti-TCRβ Antibody	Biolegend	Cat#109211; Clone#H57-597; RRID: AB_313434.
Human FITC anti-TNF-α Antibody	Biolegend	Cat#502906; Clone#MAb11; RRID: AB_315258.
Human APC anti-IL-2 Antibody	Biolegend	Cat#310605; Clone#W19046A; RRID:.AB_3083299.
Human PE anti- IFN-γ Antibody	Biolegend	Cat#383405; Clone#W19227C; RRID: AB_2924584.
Purified anti-human CD3 Antibody	Biolegend	Cat#317302; Clone# OKT3; RRID:AB_571927.
Purified anti-human CD28 Antibody	Biolegend	Cat#302902; Clone#CD28.2; RRID:AB_314304.
Purified anti-mouse CD3 Antibody	Biolegend	Cat#100202; Clone#17A2; RRID:AB_312659.
Purified anti-mouse CD28 Antibody	Biolegend	Cat#102102; Clone#; 37.51; RRID:AB_312867.
Human/mouse/rat FITC anti-CD278 (ICOS) Antibody	Biolegend	Cat#313505; Clone#C398.4A; RRID:AB_416329.
Mouse PE/Cyanine5 anti-CD69 Antibody	Biolegend	Cat#104509; Clone#H1.2F3; RRID:AB_313112.
Mouse PE/Cyanine5 anti-CD4 Antibody	Biolegend	Cat#100409; Clone#GK1.5; RRID:AB_312694.
Mouse Brilliant Violet 421™ anti-IL-2 Antibody	Biolegend	Cat#503825; Clone#JES6-5H4; RRID:AB_10895901.
Mouse PE anti-IL-2 Antibody	Biolegend	Cat#503807; Clone#JES6-5H4; RRID:AB_315301.
Mouse APC anti-IL-2 Antibody	Biolegend	Cat#503810; Clone#JES6-5H4; RRID:AB_315304.
Mouse APC anti-CD45.2 Antibody	Biolegend	Cat#109813; Clone#104; RRID:AB_389210.
Mouse APC anti-IFN-γ Antibody	Biolegend	Cat#505810; Clone#XMG1.2; RRID:AB_315404.
Mouse PE anti- IFN-γ Antibody	Biolegend	Cat#505808; Clone#XMG1.2; RRID:AB_315402.
Mouse PerCP/Cyanine5.5 anti-TNF-α Antibody	Biolegend	Cat#506321; Clone#MP6-XT22; RRID:AB_961435.
Mouse FITC anti-TNF-α Antibody	Biolegend	Cat#506304; Clone#MP6-XT22 RRID:AB_315425.
Mouse Brilliant Violet 711™ anti-CD8a Antibody	Biolegend	Cat#100747; Clone#53–6.7; RRID:AB_11219594.
Mouse Brilliant Violet 421™ anti-FOXP3 Antibody	Biolegend	Cat#126419; Clone#MF-14; RRID:AB_2565933.
Mouse APC anti-CD3 Antibody	Biolegend	Cat#100235; Clone#17A2; RRID:AB_2561455.
PE anti-RPS6 Phospho (Ser235/Ser236)	Biolegend	Cat#608603; Clone#A17020B; RRID:AB_2750251.
Mouse FITC anti-F4/80 Recombinant Antibody	Biolegend	Cat#157309; Clone#QA17A29; RRID:AB_2876535.
Mouse APC anti-Ly-6G (Gr1) Antibody	Biolegend	Cat#127613; Clone#1A8; RRID:AB_1877163.
Mouse/human APC anti-CD11b Antibody	Biolegend	Cat#101211; Clone#M1/70; RRID: AB_312794.
Mouse PerCP anti-CD11c Antibody	Biolegend	Cat#117325; Clone#N418; RRID: AB_893236.
Alexa Fluor® 647 anti-mouse CD16 Antibody	Biolegend	Cat#158021; Clone#S17014E; RRID: AB_2904300.
PE/Cyanine5 anti-mouse CD28 Antibody	Biolegend	Cat#102108; Clone#37.51; RRID: AB_312873.
Mouse PE/Cyanine7 anti-CD14 Antibody	Biolegend	Cat#123315; Clone#Sa14-2; RRID:AB_10641133.
Mouse/human PE anti-Ki-67 Antibody	Biolegend	Cat#151210; Clone#11F6; RRID:AB_2716008.
Mouse/human PE anti-Phospho-CD247 (CD3zeta) (Tyr142)	Invitrogen	Cat#12-2478-42; Clone#3ZBR4S; RRID:AB_2744700.
PE anti-Lck Phospho (Tyr394)	Biolegend	Cat#933103; Clone#A18002D; RRID:AB_2820203.
PE anti-ZAP70 Phospho (Tyr292) Antibody	Biolegend	Cat#693803; Clone#A16038B; RRID:AB_2686946.
Alexa Fluor® 647 anti-LAT Phospho (Tyr171) Antibody	Biolegend	Cat#946603; Clone#A20005D; RRID:AB_2936765.
PE Mouse anti-S6 (pS235/pS236)	BD Biosciences	Cat# 560433; Clone#N7-548; RRID:AB_2827879.
FITC anti-mouse CD3e Antibody	BD Biosciences	Cat#553062; Clone#145-2C11; RRID:AB_394595.
Mouse Alexa Fluor® 647 anti-IL-9 Antibody	Biolegend	Cat#514107; Clone#RM9A4; RRID:AB_2126637.
Mouse BV510 anti-IFN-γ Antibody	Biolegend	Cat#505841; Clone#XMG1.2; RRID:AB_2562187.
Mouse BV605 anti-IL-4 Antibody	Biolegend	Cat#504125; Clone#11B11; RRID:AB_2562101.
Mouse PE anti-Ahr Antibody	Biolegend	Cat#162703; Clone#W16012C; RRID:AB_2892335.
Mouse BV421 anti-RORγt Antibody	BD Biosciences	Cat#562894; Clone#Q31-378; RRID:AB_2687545.
Mouse APC anti-FOXP3 Antibody	Proteintech	Cat#APC-65089; Clone#3G3; RRID:AB_2882984.
BUV661 Rat Anti-Mouse Ly-6G	BD Biosciences	Cat#741587; Clone#1A8; RRID:AB_2871000
BUV737 Rat Anti-Mouse I-A/I-E	BD Biosciences	Cat#748708; Clone# 2G9; RRID:AB_2873112
BV480 Rat Anti-Mouse CD103	BD Biosciences	Cat#566201; Clone#M290; RRID: AB_2739520
BV480 Rat Anti-Mouse CD25	BD Biosciences	Cat#566120; Clone#PC61; RRID:AB_2739522
BV510 anti-mouse CD19	Biolegend	Cat#152423; Clone# 1D3; RRID:AB_3106189
BV570 anti-mouse/human CD11b	Biolegend	Cat#101233; Clone# M1/70; RRID:AB_10896949
BV711 anti-mouse CD3	Biolegend	Cat#100241; Clone# 17A2; RRID:AB_2563945
BV421 anti-mouse Ly-6C	Biolegend	Cat#128032; Clone# HK1.4; RRID:AB_2562178
FITC anti-mouse CD8a	Biolegend	Cat#100706; Clone# 53–6.7; RRID:AB_312745
CD45 Monoclonal Antibody (30-F11), Alexa Fluor™ 532	Thermofisher	Cat#58-0451-82; Clone#30-F11; RRID:AB_11218871
F4/80 Monoclonal Antibody (BM8), PerCP-Cyanine5.5	Thermofisher	Cat#45-4801-82; Clone#BM8; RRID:AB_914345
Alexa Fluor® 700 anti-mouse CD206	Biolegend	Cat#141734; Clone# C068C2; RRID:AB_2629637
APC/Cyanine7 anti-mouse NK-1.1	Biolegend	Cat#156510; Clone# S17016D; RRID:AB_2876527
Alexa Fluor® 647 anti-mouse CD11c	Biolegend	Cat#117312; Clone# N418; RRID:AB_389328
Mouse BUV563 anti-CD69 Antibody	BD Biosciences	Cat#741234; Clone#H1.2F3; RRID:AB_2870786.
TCF1/TCF7 (E6O1K) Rabbit mAb (Pacific Blue Conjugate)	Cell Signaling Technology	Cat#62283; Clone#E6O1K
BV750 Mouse Anti-Mouse Ly-108	BD Biosciences	Cat#747169; Clone#13G3 (RUO); RRID:AB_2871904.
Alexa Fluor® 647 anti-mouse CX3CR1 Antibody	Biolegend	Cat#149004; Clone#SA011F11; RRID: AB_2564273
CD101 Monoclonal Antibody (Moushi101), PE-Cyanine7, eBioscience	Invitrogen	Cat#25-1011-82; Clone#Moushi101; RRID:AB_2573378
FITC Rabbit Anti-Active Caspase-3	BD Biosciences	Cat#570334; Clone#C92-605.rMAb; RRID:AB_3685671
PE anti-mouse/human Ki-67	Biolegend	Cat#151210; Clone# 11F6; RRID:AB_2716008
Brilliant Violet 605™ anti-mouse CD223 (LAG-3)	Biolegend	Cat#125257; Clone# C9B7W; RRID:AB_3106205
Spark NIR™ 685 anti-mouse CD4	Biolegend	Cat#100476; Clone#GK1.5; RRID:AB_2819770
Mouse BV650 anti-CXCR5 Antibody	Biolegend	Cat#145517; Clone#L138D7; RRID:AB_2562453.
TRBC1/TCRβ constant region 1 (E6Z3S) Rabbit mAb	Cell Signaling Technology	Cat#79485; Clone#E6Z3S; RRID:AB_3675279
TCRα (H28-710) HRP Antibody	Santa Cruz Biotechnology	Cat#sc-101410; Clone#H28-710; RRID:AB_1130051
CD3-ζ Antibody (3F67) m-IgG1 BP-HRP Bundle (sc-541392)	Santa Cruz Biotechnology	Cat#sc-541392; Clone#3F67; RRID:AB_1120424
GAPDH (14C10) Rabbit mAb #2118	Cell Signaling Technology	Cat#2118; Clone#14C10; RRID:AB_561053
Biological samples
Whole blood	Michael Drazer Lab	N/A
Human Peripheral Blood Mononuclear Cells (PBMCs)	Zen-Bio Inc	Cat#SER-PBMC-200-F
Human TCR-T cells	Kunle Odunsi Lab	N/A
Chemicals, peptides, and recombinant proteins
Human IL-2 IS	Miltenyi Biotec	Cat#130-097-743
Mouse IL-2 IS	Miltenyi Biotec	Cat#130-120-662
Mouse IL-7	Thermo Fisher Scientific	Cat#50813299
Mouse IL-15	Thermo Fisher Scientific	Cat#50813273
PE-conjugated H-2Kb/OVA (SIINFEKL) MHC class I tetramers	Creative Biolabs	Cat# MHC-LC649
Fixable viability dye 780	Invitrogen	Cat#65-0865-14
DNase I	Millipore	Cat#EN0521
N-methylpyrrolidone	Sigma-Aldrich	Cat#PHR1352 CAS:872-50-4
Zeaxanthin	Sigma-Aldrich	Cat#14681 CAS:144-68-3
Zeaxanthin	aablocks	Cat# AA00BWRX CAS:144-68-3
Zeaxanthin	Wuxi AppTech	Cat#144-68-3 CAS:144-68-3
Lutein	Wuxi AppTech	Cat# LN01318215 CAS:127-40-2
Lutein	Sigma-Aldrich	Cat#204-840-0 CAS:127-40-2
Lutein	Angene Chemicals	Cat#AG000WIY(AGN-PC-0PIJ3S) CAS:127-40-2
beta carotene	Sigma-Aldrich	Cat#C4582 CAS:7235-40-7
(±)-α-Carotene	Sigma-Aldrich	Cat#74283 CAS:432-70-2
Lycopene	Sigma-Aldrich	Cat#75051 CAS:502-65-8
Astaxanthin	Sigma-Aldrich	Cat#SML0982 CAS:7542-45-2
Fucoxanthin	Sigma-Aldrich	Cat#16337 CAS:3351-86-8
Retinoic acid	Sigma-Aldrich	Cat#R2625 CAS:302-79-4
Ficoll® Paque Plus	Sigma-Aldrich	Cat#GE17-1440-03
TWEEN-80	Sigma-Aldrich	Cat#P1754
Lck inhibitor PP2	Sigma-Aldrich	Cat#529576
PLCγ inhibitor 3-Nitrocoumarin	MedChemExpress	Cat#HY-111919
Calcineurin inhibitor Cyclosporine	Sigma-Aldrich	Cat#1158504
NFAT Inhibitor	Cayman Chemical	Cat#13855-1
NF-κB inhibitor INK4	MedChemExpress	Cat#HY-13453
JAK2 inhibitor AZD1480	MedChemExpress	Cat#HY-10193
ERK inhibitor U0126	MedChemExpress	Cat#HY-12031A
Collagenase/Hyaluronidase	Stemcell Technologies Inc	Cat#07912
Ammonium Chloride Solution	Stemcell Technologies Inc	Cat#07800
Critical commercial assays
Annexin V apoptosis detection kit I	BD Biosciences	Cat#556547
EasySep™ Mouse T cell Isolation Kit	Stemcell Technologies Inc	Cat#19851
EasySep™ Mouse Naive CD8+ T cell Isolation Kit	Stemcell Technologies Inc	Cat#19858
EasySep™ Mouse Naive CD4+ T cell Isolation Kit	Stemcell Technologies Inc	Cat#19765
EasySep™ Mouse CD8+ T cell Isolation Kit	Stemcell Technologies Inc	Cat#19853
EasySep™ Mouse CD4+ T cell Isolation Kit	Stemcell Technologies Inc	Cat#19852
EasySep™ Human T cell Isolation Kit	Stemcell Technologies Inc	Cat#17951
EasySep™ Human CD8+ T cell Isolation Kit	Stemcell Technologies Inc	Cat#17953
Bright-Glo Luciferase Assay System	Promega	Cat# E2620
CaspaseGlo 3/7	Promega	Cat#G8090
Cell Activation Cocktail (with Brefeldin A)	Biolegend	Cat#423304
eBioscience 7-AAD Viability Staining Solution	Thermo Fisher Scientific	Cat#00-6993-50
eBioscience IC Fixation Buffer	Thermo Fisher Scientific	Cat#00-8222-49
eBioscience Intracellular Fixation & Permeabilization Buffer Set	Thermo Fisher Scientific	Cat#G762684
FoxP3/Transcription Factor Staining Buffer Set	Thermo Fisher Scientific	Cat#00-5523-00
Phosflow Perm Buffer III	Biolegend	Cat# 558050
Phosflow Lyse/Fix buffer	BD Biosciences	Cat#558049
Fluo-4 Direct™ Calcium Assay Kit	Fisher Scientific Company	Cat#F10471
Purelink RNA Mini Kit	Fisher Scientific Company	Cat#12183018A
Pierce™ Rapid Gold BCA Protein Assay Kit	Thermo Fisher Scientific	Cat#A53225
QIAprep Spin Miniprep Kit	Qiagen	Cat#27106
RBC lysis buffer	Invitrogen	Cat#00-4300-54
T cell receptor Phospho Antibody Array	Full Moon Biosystems	Cat#PTC188
Antibody Array Assay Kit	Full Moon Biosystems	Cat#KAS02
Proteome Profiler Phospho-Kinase Array Kit	R&D Systems	Cat#ARY003C
CellTiter-Glo® 2.0 Cell Viability Assay Kit	Promega	Cat# G9241
Zombie NIR™ Fixable Viability Kit	Biolegend	423106
Deposited data
RNA-seq data: effect of Zeaxanthin treatment on gene expression of mouse CD8+ T cells	This paper	GEO: GSE282703
KAS-seq data	This paper	GEO: GSE282686
Microbiome 16S rRNA sequencing data	This paper	NCBI SRA: PRJNA1189274
Experimental models: Cell lines
Mouse: B16F10 cells	ATCC	Cat#CRL-6475; RRID:CVCL_0159
Mouse: MC38 cells	Kerafast	Cat#ENH204-FP; RRID:CVCL_B288
Mouse: B16F10-luc2 cells	ATCC	Cat#CRL-6475-LUC2
Mouse: B16F10-OVA cells	Hongbo Chi Lab	N/A
Experimental models: Organisms/strains
Mouse: C57BL/6J	The Jackson Laboratory	JAX:000664; RRID:IMSR_JAX:000664
Mouse: Pmel-1 (B6.Cg-Thy1a/CyTg(TcraTcrb)8Rest/J)	The Jackson Laboratory	JAX:005023; RRID:IMSR_JAX:005023
Mouse: OT-I (C57BL/6-Tg(TcraTcrb)1100Mjb/J)	The Jackson Laboratory	JAX:003831; RRID:IMSR_JAX:003831
Software and algorithms
ImageJ	Schneider et al.37	https://imagej.net/ij/download.html
GraphPad Prism 9	GraphPad Software	RRID:SCR_002798 https://www.graphpad.com/
FlowJo v10.8.1	BD Biosciences	https://www.bdbiosciences.com/en-us/products/software/flowjo-v10-software
FragPipe	Nesvilab	https://fragpipe.nesvilab.org/
STAR	Dobin et al.38	https://github.com/alexdobin/STAR/releases
featureCounts	Liao et al.39	https://subread.sourceforge.net/
DESeq2	MI Love et al.40	https://bioconductor.org/packages/release/bioc/html/DESeq2.html
clusterProfiler	Wu et al.41	https://bioconductor.org/packages/release/bioc/html/clusterProfiler.html
Origin (Pro), Version 2021	OriginLab Corporation	https://www.originlab.com/demodownload.aspx
GSEA 4.3.3	Subramanian et al.30	https://www.gsea-msigdb.org/gsea/downloads.jsp
KAS-Analyzer	Lyu et al.42	https://github.com/Ruitulyu/KAS-Analyzer.
Leica Application Suite X	Leica Microsystems	RRID:SCR_013673 https://www.leica-microsystems.com/products/microscope-software/details/product/leica-las-x-ls/

Experimental model and study participant details

Mice

Animal experiments were conducted and designed according to protocols approved by the Institutional Animal Care and Use Committee of The University of Chicago. C57BL/6 mice (6–8 weeks), Pmel-1 mice (6–8 weeks), TCRα knockout (B6.129S2-Tcratm1Mom/J) (The Jackson Laboratory, JAX:002116; RRID:IMSR_JAX:002116) mice (6–8 weeks), and OT-I mice (6–8 weeks) were obtained from The Jackson Laboratory. Mice were housed in a temperature (23°C–25°C) and humidity-controlled colony room, maintained on a 12-h light/dark cycle (08:00 to 20:00 light on) with standard food and water provided ad libitum and environmental enrichments in a pathogen-free facility. Please refer to key resources table for more details.

Primary cells

Mouse naive CD8⁺ T cells were isolated from the spleen of C57BL/6 mice (6–8 weeks), TCRα knockout (B6.129S2-Tcratm1Mom/J) (The Jackson Laboratory, JAX:002116; RRID:IMSR_JAX:002116) mice (6–8 weeks), Pmel-1 mice (6–8 weeks), or OT-I mice (6–8 weeks) by magnetic bead purification using EasySep Mouse naive CD8⁺ T cell Isolation Kit (Stemcell Technologies) according to the manufacturer’s instructions. Naive CD8⁺ T cells were activated in vitro for 18 h with plate-bound anti-mouse CD3 (1μg/mL; Biolegend) and soluble anti-mouse CD28 (0.5 μg/mL; Biolegend) antibodies in complete Click’s media (Fujifilm Irvine Scientific) at 37°C and 5% CO₂ incubator for further experiments.

Human naive CD8⁺ T cells were isolated by magnetic bead purification using EasySep Human naive CD8⁺ T cell Isolation Kit (Stemcell Technologies) from peripheral blood mononuclear cells (PBMCs) obtained via Ficoll gradient centrifugation (Ficoll-Paque PLUS, GE Healthcare) from 10 anonymized healthy donor blood samples (with no identifiable information). Human naive CD8⁺ T cells were activated in vitro for 18 h with plate-bound anti-human CD3 (1μg/mL; Biolegend) and soluble anti-human CD28 (0.5 μg/mL; Biolegend) antibodies in Click’s media (Fujifilm Irvine Scientific) at 37°C and 5% CO₂ incubator for further experiments. Cells from each donor were then treated with ZEA, LUT, or vehicle control for analysis.

Human CD8⁺ untransduced mock T cells, 19305DP-TCR-T cells, and AL-TCR-T cells were provided by Dr. Kunle Odunsi. These cells were derived from three anonymized donors with no personal identifiers, and all human samples were fully de-identified prior to use. Mock and transduced TCR-T cells were cultured in RPMI 1640 medium at 37°C and 5% CO₂ incubator, and split into vehicle, ZEA, and LUT treatment groups, enabling paired comparisons within each donor-derived batch.

Cell lines

All cell lines were authenticated by genomic short tandem repeat (STR) profiling at the University of Chicago Integrated Genomics Core (EIGC) upon purchase and at least annually as appropriate. Cell lines were not tested for mycoplasma contamination.

Mouse melanoma cell line B16F10-luc2, mouse melanoma cell line B16F10, human multiple myeloma cell line U266, and human glioblastoma cell line U-87 MG were purchased from ATCC. Mouse melanoma cell line B16-OVA was provided by Dr. Hongbo Chi. Mouse colorectal adenocarcinoma cell line MC38-OVA was provided by Dr. Ralph R. Weichselbaum. Mouse colorectal adenocarcinoma cell line MC38 was purchased from Kerafast. B16F10 and MC38 cells were cultured in Dulbecco Modified Eagle Medium (DMEM). B16F10-luc2 and B16-OVA were cultured in the Click’s media (Fujifilm Irvine Scientific) prior to co-culture experiments to assess murine T cell-mediated cytotoxicity. Human melanoma cell line A375 was provided by Dr. Kunle Odunsi. Human melanoma A375 cells were cultured in RPMI 1640 medium prior to co-culture experiments to assess TCR-T cell-mediated cytotoxicity. All the cells were cultured at 37°C and 5% CO₂. Cell line experiments were conducted and designed according to protocols approved by the Institutional Biosafety Committee (IBC) of the University of Chicago.

Method details

Effector T cell cytotoxicity screen

T cells were isolated from Pmel-1 mouse by magnetic bead purification using EasySep Mouse T cell Isolation Kit (Stemcell Technologies) and seeded at a density of 5×10⁶ cells/well in a 6-well plate pre-coated with anti-mouse CD3 (Biolegend) and anti-mouse CD28 (Biolegend) for overnight stimulation. 5×10⁵ stimulated Pmel-1 cells were then co-cultured with 1 × 10⁵ B16F10-luc2 cells (E:T = 5:1) and treated with each of the compounds in the blood nutrient library for 24 h. Bright-Glo Luciferase Assay System (Promega) was used to measure luminescence level as a readout for B16F10-luc2 cell death. Luminescence was measured on the BioTek Synergy H4 Hybrid Reader. Please refer to key resources table for more details.

Co-culture cytotoxicity experiments

To determine mouse OT-I T cell-mediated tumor cytotoxicity, OT-I T cells were isolated using EasySep Mouse T cell Isolation Kit (Stemcell Technologies) and seeded at a density of 5×10⁶ cells/well in 6-well plate (pre-coated with anti-CD3 and anti-CD28) for overnight stimulation. B16-OVA cells were stained with CellTracker Deep Red dye (Invitrogen) according to manufacturer’s instruction prior to co-culturing with OT-I T cells. 5×10⁵ stimulated OT-I cells were then co-cultured with pre-stained 1 × 10⁵ B16F10-OVA cells (E:T = 5:1) and treated with DMSO control, 5μM ZEA (Sigma-Aldrich) or 5μM LUT (Sigma-Aldrich) for 24 h. B16-OVA cells alone were treated under the same conditions. All cells were collected, stained with PI, and analyzed by flow cytometry.

To determine human TCR-T cell-mediated tumor cytotoxicity, AL-TCR-T cells were generated by retrovirally transduce human primary CD8⁺ T cells with TCR genes from an A∗02-restricted NY-ESO-1-specific CD8⁺ single-positive T cell clone³³ and 19305DP-TCR-T cells were generated in parallel with TCR genes from an A∗02-restricted NY-ESO-1-specific CD4⁺CD8⁺ double-positive T cell clone (19305DP T cells).³³ Human A∗02⁺NY-ESO-1⁺ A375 melanoma cells, A∗02⁺NY-ESO-1⁺ U266 multiple myeloma cells, or A∗02⁺NY-ESO-1^- U-87 MG glioblastoma cells were stained with CellTrace Violet dye (Invitrogen) according to manufacturer’s instruction prior to co-culturing with TCR-T cells. Mock and transduced T cells were co-cultured at a density of 3x10⁵ cells/well (E:T = 3:1) with 1x10⁵ human A∗02⁺NY-ESO-1⁺ A375 melanoma cells, A∗02⁺NY-ESO-1⁺ U266 multiple myeloma cells, or A∗02⁺NY-ESO-1^- U-87 MG glioblastoma cells and treated with or without 5μM ZEA (Sigma-Aldrich) or LUT (Sigma-Aldrich) for 24 h. Tumor cells alone were treated under the same conditions. All cells were collected, stained with Annexin V-APC/FITC and propidium iodide (PI), and analyzed by flow cytometry. Please refer to key resources table for more details.

Syngeneic tumor models in C57BL/6 mice

For C57BL/6 mice tumor models, female C57BL/6 mice (8 weeks) were anesthetized with isoflurane, shaved at the injection site, and then injected subcutaneously in the abdominal flank with 2x10⁵ B16F10 or MC38 cells for tumor development. Daily oral gavage of ZEA (Wuxi AppTech), LUT (Angene Chemicals), or vehicle control (PBS+1% Tween-80) was administered once tumors reached a size of 100 mm³. Mice’s body weight and tumor size were measured every other day. Tumor sizes were measured using a caliper, and tumor volumes were calculated using the following formula: tumor volume = (Length x Width²)/2. Mice were euthanized at humane endpoints or on days 15–16 for tissue collection.

Antibody-mediated T cell depletion models

For antibody-mediated T cell depletion models, female C57BL/6 mice (8 weeks) were injected subcutaneously in the abdominal flank with 2x10⁵ B16F10 cells and then injected intraperitoneally with six doses of depleting antibodies (anti-CD8a, BioXCell) or isotype control (rat IgG2b isotype control, BioXCell) on day 1 (200 μg), 2 (200 μg), 4 (200 μg), 8 (200 μg), 12 (200 μg) and 16 (200 μg) relative to tumor injection (day 0). Depletion efficiency was assessed by staining for CD8 (Biolegend) or CD4 (Biolegend) and flow cytometry analysis. Tumor sizes were measured using a caliper and tumor volumes were calculated using the following formula: tumor volume = (Length x Width²)/2. Mice were euthanized at humane endpoints.

Anti-PD-1 treatment mouse model

For the anti-PD-1 treatment mouse model, female C57BL/6 mice (8 weeks) were injected subcutaneously in the abdominal flank with 2x10⁵ B16F10 cells or 1x10⁶ MC38 cells for tumor development. Daily oral gavage of ZEA or vehicle control (PBS+1% Tween 80) was given to tumor-bearing mice after tumors became palpable. On days 17, 19, and 21, 200 μg anti-PD-1 (BioXCell) or IgG control (BioXCell) was injected intraperitoneally per mouse. Tumor sizes were measured using a caliper and tumor volumes were calculated using the following formula: tumor volume = (Length x Width²)/2. Mice were euthanized at humane endpoints. Please refer to key resources table for more details.

CD45⁺ tumor-infiltrating leukocyte isolation

Tumor tissues were dissected from euthanized tumor-bearing mice, minced into small pieces (≤2 mm) using a scalpel in a dish, and then transferred to a 14 mL round-bottom tube containing 5mL tumor digestion medium (500 μL Collagenase/Hyaluronidase Solution,750 μL1 mg/mL DNase I Solution, and 3.75 mL RPMI 1640 Medium). After incubation at 37°C for 25 min on a shaking platform, the digested tumor tissues were transferred into a 70 μm mesh nylon strainer on a 50 mL conical tube, pushed through the strainer using the rubber end of a syringe plunger, and rinsed with the recommended medium. After centrifugation at 300×g for 10 min at room temperature with the brake on low, the resulting cell pellets were added 10 mL of ammonium chloride solution for incubation at room temperature for 5 min, followed by centrifugation at 300×g for 10 min at room temperature with the brake on low. The resulting cell pellets were re-suspended at 1-10 ×10⁶ cells/mL in PBS and then subjected to tumor-infiltrating leukocyte isolation via Ficoll gradient centrifugation (Ficoll-Paque PLUS, GE Healthcare).

Mouse spleen lymphocyte isolation

Mouse spleens were disrupted with a syringe plunger in a 70 μm strainer, filtered to a 50 mL tube, washed with PBS, and centrifuged at 300×g for 10 min. The resulting cell pellets were re-suspended with 2 mL red cell lysis buffer (Invitrogen), incubated at room temperature for 10 min, and centrifuged at 300×g for 5 min after adding 15 mL PBS. The resulting cell pellets were designated splenocytes and used for the following experiments. For cytokine detection, T cells were stimulated with a cell activation cocktail (Biolegend) for 4h, and then intracellular staining for cytoplasmic protein was performed, followed by flow cytometry analysis.

Mouse draining lymph node lymphocyte isolation

Draining lymph nodes were disrupted with a syringe plunger in a 70 μm strainer filtered to a 50 mL tube, washed with PBS, and centrifuged at 300×g for 10 min. The resulting cell pellets were designated dLN lymphocytes and used for the following experiments.

Flow cytometry analysis

Murine primary cells isolated from tumor, spleen, draining lymph node were stained with fluorescent antibodies and analyzed by flow cytometry.

For experiments with live/dead criteria, cells were first stained with Fixable Viability Dyes (FVD) (Thermo Fisher Scientific) according to the manufacturer’s instructions. Subsequent surface marker staining was performed in Cell Staining Buffer (Thermo Fisher Scientific).

Intracellular staining for flow panels containing nuclear proteins was performed using the eBioscience FoxP3/Transcription Factor Staining Buffer Set (Thermo Fisher Scientific) according to the manufacturer’s instructions. For intracellular staining of cytoplasmic proteins, eBioscience Intracellular Fixation & Permeabilization Buffer Set (Thermo Fisher Scientific) was used according to the manufacturer’s instructions.

For phospho-antibody staining, cells were incubated with FVD (cell viability dye) for 15 min at room temperature in a tube, re-suspended with 200 μL pre-warmed 1× BD Phosflow Lyse/Fix buffer directly into the tube, and incubated at 37°C for 10–15 min, followed by centrifugation at 300g for 5 min. The resulting cell pellets were washed once with FACS buffer, permeabilized with 200 μL of BD Phosflow Perm Buffer III for 45 min on ice, and centrifuged at 300g for 5 min. The cell pellets were washed again with FACS buffer, centrifuged at 300g for 5 min, and incubated with antibodies in FACS buffer for 45 min-1 hour at room temperature.

The gating strategy for CD45⁺ TIL subset analysis is shown in Figure S1L. For CD4 subtypes analysis, CD3⁺ cells were gated within the CD45⁺ population, and CD4⁺ and CD8⁺ cells were subsequently identified. For CD45⁺CD4⁺ cells, further analysis was performed to identify subtypes: Treg cells were identified as CD25⁺Foxp3⁺ within the CD45⁺CD4⁺ population. In the Foxp3^- (Treg⁻) population, Th1 cells were gated as IFNγ⁺, Th2 cells as IL4⁺, Th17 cells as RORγt⁺, Th9 cells as IL9⁺, and Th22 cells as AhR⁺.

Tumor-specific activated CD8⁺ T cells were identified using PE-conjugated H-2K^b/OVA (SIINFEKL) MHC class I tetramers (Cat# MHC-LC649, Creative Biolabs). Tumor-infiltrating lymphocytes (TILs) were isolated from B16-OVA and MC38-OVA tumors, and cells double-positive for CD8 and tetramer were quantified by flow cytometry and defined as tumor-specific activated CD8⁺ T cells.

To analyze exhaustion phenotypes, CD45⁺CD3⁺CD8⁺ T cells were first gated, and PD-1⁺ cells were identified within this population. PD-1⁺CD8⁺ T cells were then further classified into the following subsets based on established markers:Stem-like exhausted T cells: TCF1⁺ Ly108⁺, effector-like exhausted T cells: TCF1^- Ly108^- CX3CR1⁺ CD101^-, terminally exhausted T cells: TCF1^- Ly108^- CX3CR1^- CD101⁺.

Data were collected with LSR-Fortessa 4–15, Attune NxT 4–14, Aurora Red/Blue, or Agilent NovoCyte Quanteon flow cytometer and analyzed using FlowJo v10.8.1. Please refer to key resources table for more details.

Cell culture treatment

For all in vitro ZEA (Sigma-Aldrich) or LUT(Sigma-Aldrich) treatments, ZEA or LUT were added to the cell culture in the dark, and the cell culture plates were covered with aluminum foil to avoid light throughout the duration of incubation.

Murine naive CD8⁺ T cells were isolated from C57BL/6 mice, stimulated, and subjected to the following treatments. For in vitro inhibitor treatment, stimulated murine CD8⁺ T cells were pre-treated with LCK inhibitor PP2 (200 nM), PLCγ inhibitor 3-Nitrocoumarin (500 nM), calcineurin inhibitor cyclosporin A (7 nM), NFAT inhibitor VIVIT (1 μM), NF-κB inhibitor INH14 (2 μM), JAK2 inhibitor AZD1480 (500 nM), or ERK inhibitor U0126 (100μM) for 1–2 h, followed by treatment with 5 μM ZEA (Sigma-Aldrich), 5 μM LUT (Sigma-Aldrich), or DMSO control for 24 h. Cells were then stained with FVD and treated with Cell Activation Cocktail (with Brefeldin A) (Biolegend) for 4–6 h before intracellular staining to detect pro-inflammatory cytokine levels.

For in vitro derivative treatment, stimulated murine CD8⁺ T cells were treated with 5 μM ZEA (Sigma-Aldrich), or 5 μM of each of the derivatives D1-6 for 24 h before intracellular staining to detect pro-inflammatory cytokine levels.

Cell viability assay

Cancer cell lines B16, MC38, and A375 were seeded in 96-well plates at a density of 3,000 cells per well, while U266 and U87 cells were seeded at 6,000 cells per well. After 24 h, cells were treated with serial dilutions of ZEA, LUT, or Doxorubicin. Following 72 h of incubation, cell viability was assessed using the CellTiter-Glo 2.0 Cell Viability Assay Kit (Promega, Cat# G9241) according to the manufacturer’s instructions.

Intracellular calcium detection with Fluo4 calcium indicator

Murine naive CD8⁺ T cells were isolated from C57BL/6 mice and loaded with Fluo-4 using either the Fluo-4 Direct Calcium Assay Kit (ThermoFisher Scientific; used for 30-min endpoint analysis) or the Fluo-4 NW Calcium Assay Kit (ThermoFisher Scientific, F36206; used for time-course kinetic analysis), following the manufacturer’s instructions. For stimulation, loaded cells were treated with soluble anti-CD3 (1 μg/mL; clone 145-2C11) and anti-CD28 (0.5 μg/mL; clone 37.51) antibodies in the presence of 5 μM ZEA (Sigma-Aldrich), 5 μM LUT (Sigma-Aldrich), or DMSO control. For kinetic measurements, real-time fluorescence was recorded by flow cytometry immediately after stimulation. For endpoint measurements, cells were incubated for 30 min before analysis by flow cytometry. Please refer to key resources table for more details.

TCR complex analysis with flow cytometry and confocal microscopy

Murine naive CD8⁺ T cells were isolated from C57BL/6 mice. For short-term stimulation conditions, murine naive CD8⁺ T cells were stimulated with anti-CD3ε-FITC (1μg/ml; Clone:145-2c11) and anti-CD28 -PE (0.5μg/ml; Clone: 37.51) antibodies, and simultaneously treated with 5μM ZEA (Sigma-Aldrich), 5 μM LUT (Sigma-Aldrich), or DMSO control for 5 min, 15 min, or 30 min. For overnight pre-stimulation conditions, murine naive CD8⁺ T cells were stimulated with anti-CD3ε-FITC (1μg/ml; Clone:145-2c11) and anti-CD28 -PE (0.5μg/ml; Clone: 37.51) antibodies and incubated overnight in Click’s medium at 37°C and 5% CO₂ before centrifugation and treating anti-CD3ε-FITC (1μg/ml; Clone:145-2c11) and anti-CD28 -PE (0.5μg/ml; Clone: 37.51) antibodies again simultaneously with 5μM ZEA (Sigma-Aldrich), 5μM LUT (Sigma-Aldrich), or DMSO control for 5 min, 15 min, or 30 min.

For confocal microscopy analysis, cells were collected and stained with anti-TCRβ-APC antibody, fixed, and stained with DAPI (1μg/ml), and subject to analysis using Leica SP8 Confocal Microscope. For flow cytometry analysis, cells were collected and stained with FVD anti-CD8-BV421 and anti-TCRβ-APC antibodies, followed by flow cytometry analysis. Please refer to key resources table for more details.

Kethoxal-assisted single-stranded DNA sequencing (KAS-seq) and data analysis

Primary murine CD8⁺ T cells were isolated from the spleens of 8-week-old female C57BL/6 mice using the EasySep Mouse Naive CD8⁺ T cell Isolation Kit (STEMCELL Technologies, 19858). A total of 10,000,000 primary Naive CD8⁺ T cells were activated overnight with αCD3/CD28 antibodies, followed by treatment with 5 μM ZEA for 40 min or 2 h. DMSO-treated cells served as controls. Each condition included three biological replicates. Samples for KAS-seq were processed following the protocol developed by Tong Wu et al. with minor modifications.⁴³ Cells were lysed using a lysis buffer containing 0.1% NP40, 0.1% Tween 20, 0.01% Digitonin, 10 mM Tris-HCl (pH 7.4), 10 mM NaCl, and 3 mM MgCl₂, and incubated on ice for 3 min. The lysed cell pellets were resuspended in 100 μL of 10 mM N3-Kethoxial in PBS and incubated at 37°C with shaking at 500 rpm for 10 min. Cells were then collected for genomic DNA (gDNA) purification using the PureLink Genomic DNA Mini Kit (Invitrogen). Approximately 1 μg of purified gDNA in a 100 μL reaction volume was incubated with 1 mM DBCO-PEG4-biotin (Sigma) in PBS at 37°C for 1.5 h with gentle shaking to achieve biotinylation. Biotinylated gDNA was further purified using the DNA Clean & Concentrator-5 kit (Zymo Research) and fragmented to approximately 250 bp by sonication. Biotinylated DNA fragments were captured using 10 μL of Dynabeads MyOne Streptavidin C1 (Thermo Fisher) following the manufacturer’s instructions. DNA was eluted from the beads by heating in 15 μL of H₂O at 95°C for 10 min. Library preparation was performed using the xGen Methyl-Seq DNA Library Prep Kit (IDT), and the final libraries were sequenced on the Illumina NovaSeq 6000 platform.

For bioinformatics analysis, we used the Trim Galore package to trim low-quality sequences, adapter sequences, and primer sequences from paired-end KAS-seq raw FASTQ files. Reads shorter than 30 bp were discarded. The remaining reads were aligned to the reference genome (mm10) using Bowtie2 (v2.3.3.1)⁴⁴ with the alignment carried out under default parameters. The SAM files were then converted and sorted into BAM files using SAMtools sort (v1.9).³⁷ Duplicate reads were removed with the assistance of Picard MarkDuplicates (v1.141), and the uniquely mapped reads were extended to 150 bp to match the average length of DNA fragments, using the awk command. The extended data, in BED file format, were converted to BedGraph files using bedtools genomecov, and subsequently transformed into BigWig files through bedGraphToBigWig. Shell scripts for KAS-seq data mapping and quality control can be accessed at GitHub (https://github.com/Ruitulyu/KAS-Analyzer). For peak calling, MACS2 was utilized to identify broad KAS-seq peaks, and 2kb bins overlapping with KAS-seq peaks were used to detect differentially generated time-course KAS-seq signals. Promoters were defined as regions spanning 1kb upstream and 1kb downstream of the transcription start site (TSS), while gene bodies were defined from 1kb downstream of the TSS to the transcription end site (TES). Differential analysis of time-course KAS-seq data at promoter and gene body regions was performed using DESeq2 software.³⁸ KEGG pathway enrichment analysis was performed using the DAVID platform.

T cell receptor (TCR) phospho-antibody array

Murine naive CD8⁺ T cells were isolated, stimulated with soluble anti-CD3e and anti-CD28, then immediately treated with DMSO control, 5 μM ZEA or 5μM LUT for 5 min followed by TCR-related phosphoprotein profiling using T cell receptor Phospho Antibody Array kit (Full Moon BioSystem) according to the manufacturer’s instructions. Please refer to key resources table for more details.

Conventional phospho-antibody array

Murine naive CD8⁺ T cells were isolated, stimulated overnight, and then treated with DMSO control, 5 μM ZEA or 5μM LUT for 40 min, 2 h, or 6 h, followed by phosphoprotein profiling using Proteome Profiler Human Phospho-Kinase Array Kit (R&D Systems, ARY003C) according to the manufacturer’s instructions. Analysis was performed using Fiji for relative pixel density comparison. Please refer to key resources table for more details.

16S rRNA sequencing

Feces were collected from C57BL/6 mice (6–8 weeks) orally administered with ZEA (500 mg/kg b.w.) or vehicle control (PBS +1% Tween 80) daily for 3 weeks (10 samples per group), and then subjected to microbiome 16S sequencing by Zymo Research (Irvine, CA). In brief, The ZymoBIOMICS-96 MagBead DNA Kit (Zymo Research) was used to extract DNA using an automated platform. Bacterial 16S ribosomal RNA gene targeted sequencing was performed using the Quick-16S NGS Library Prep Kit (Zymo Research). The bacterial 16S primers amplified the V3-V4 region of the 16S rRNA gene. The sequencing library was prepared using an innovative library preparation process in which PCR reactions were performed in real-time PCR machines to control cycles and therefore limit PCR chimera formation. The final PCR products were quantified with qPCR fluorescence readings and pooled together based on equal molarity. The final pooled library was cleaned with the Select-a-Size DNA Clean & Concentrator (Zymo Research), then quantified with TapeStation (Agilent Technologies) and Qubit (Thermo Fisher Scientific). The ZymoBIOMICS Microbial Community Standard (Zymo Research) was used as a positive control for each DNA extraction, if performed. The final library was sequenced on Illumina MiSeq with a v3 reagent kit (600 cycles). The sequencing was performed with 10% PhiX spike-in. For Bioinformatics Analysis, unique amplicon sequences variants were inferred from raw reads using the DADA2 pipeline.³⁹ Potential sequencing errors and chimeric sequences were also removed with the Dada2 pipeline. Chimeric sequences were also removed with the DADA2 pipeline. Taxonomy assignment was performed using Uclust from Qiime v.1.9.1 with the Zymo Research Database, a 16S database that is internally designed and curated, as reference. Composition visualization, alpha-diversity, and beta-diversity analyses were performed with Qiime v.1.9.1.⁴⁰

RNA-seq and data analysis

Spleens were harvested from 8-week-old female C57BL/6 mice, and Naive CD8⁺ T cells were isolated using the EasySep Mouse Naive CD8⁺ T cell Isolation Kit (STEMCELL Technologies, 19858). For each sample, a total of 6,000,000 primary Naive CD8⁺ T cells were activated overnight with αCD3/CD28 antibodies, followed by treatment with 5 μM ZEA, 5 μM LUT, or DMSO for 24 h (n = 3). After treatment, cells were collected and washed once with PBS, followed by RNA extraction using RNeasy Mini Kit according to the manufacturer’s instructions. Library preparation and sequencing were performed by Novogene Co, Ltd (Sacramento, California). RNA samples with RIN>7.5 were used for Illumina Next Generation Sequencing. Alignment was performed by STAR software⁴¹ with the reference genome being mm10. Post-alignment quantification was performed using featureCounts software.⁴² Differential gene analysis was conducted using DESeq2,³⁸ with multiple testing corrections applied using FDR. Enrichment analysis was carried out using clusterProfiler 4.0,⁴⁵ and gene set enrichment analysis was performed using GSEA 4.3.2.²⁸ Please refer to key resources table and Tables S5, S6, and S7 for more details.

Click-reaction pulldown

Mouse CD8⁺ T cells (3 × 10⁷) were incubated with 50 μM photo-affinity ZEA probe (Wuxi AppTech) at 37°C for 1 h in the dark, followed by 365 nm UV irradiation for 8 min on ice. Cells were lysed in 200 μL lysis buffer containing protease inhibitors at 4°C for 15 min with rotation. Protein concentration was measured by BCA assay and adjusted to 1.5 mg/mL. Click chemistry was performed with 100 μM azide–PEG3–biotin (Sigma-Aldrich, 762024), 100 μM TBTA (Sigma-Aldrich, 678937), 1 mM CuSO₄ (Sigma-Aldrich, 451657), and 1 mM TCEP (Sigma-Aldrich, 75259) at room temperature for 1 h in the dark. Proteins were precipitated by adding 4 volumes of cold acetone, incubated at −20°C for 3 h, and pelleted by centrifugation at 17,000 × g for 15 min at 4°C. Pellets were resuspended in 0.2% SDS in PBS (final volume 400 μL). 40 μL of Pre -washed Dynabeads MyOne Streptavidin C1 beads (Invitrogen, 65001) were added and incubated overnight at 4°C with rotation. Beads were washed three times with 0.1% SDS in PBS, eluted in 2× SDS sample buffer, and boiled for western blotting.

TAMRA click labeling and in-gel fluorescence

Mouse CD8⁺ T cells (1 × 10⁷) were incubated with 50 μM photo-affinity ZEA probe (Wuxi AppTech) at 37°C for 45 min in the dark, followed by UV irradiation at 365 nm for 8 min on ice. Cells were lysed in 200 μL lysis buffer supplemented with protease inhibitors and rotated at 4°C for 15 min. Protein concentration was determined by BCA assay and adjusted to 1 mg/mL. Click chemistry was performed by incubating the lysates with 200 μM TAMRA-azide (Vectorlabs, CCT-1245), 500 μM TBTA, 2 mM CuSO₄, and 2 mM TCEP at room temperature for 1 h in the dark. After labeling, samples were mixed with 5× SDS loading buffer and protected from light during SDS-PAGE electrophoresis. Gels were washed in distaining buffer (50% methanol, 10% acetic acid, 40% ddH₂O) for 1 h to remove nonspecific dye binding. TAMRA-labeled proteins were visualized using the Krypton channel of the ChemiDoc Imaging System.

Zeaxanthin and lutein detection by liquid chromatography-mass spectrometry (LC-MS)

To 5 μL of tumor interstitial fluid (TIF) and plasma sample, 225 μL of ice-cold methanol containing 0.1mg/mL butylated hydroxytoluene and 750 μL of methyl-tert-butyl ether were added followed by mixing for 15 min at 500 rpm and 15°C. The phase separation was induced by adding 225 μL of ice-cold water and a vortex for 15 min at 15°C and 500 rpm then samples were spun down at 18,000 g at 15°C for 15 min. The upper 650 μL of organic layer was dried using the Genevac EZ 2.4 elite evaporator. The dried samples were re-suspended in 50μL of isopropanol/acetonitrile, vortexed for 15 min at 15°C and 2000 rpm, and centrifuge at 18,000 g at 15°C for 15 min before collecting the supernatant in amber-color LC-MS vials. Zeaxanthin stock solution was prepared in 0.1mg/m (MTBE) and serial dilution external calibration curve (24.41nM–781.250nM) was prepared in methanol (0.1mg/mL BHT). Equal volume of calibration standard, TIF and plasma sample was extracted using the previously described procedure. The exposure to lights were minimized during the extraction process. Zeaxanthin and Lutein were separated on the BEH C18 column (Waters Corporation, 2.1 × 150 mm, 2.5 μm) connected to a Vanquish Horizon UHPLC system and IQ-X tribrid mass spectrometer. The column temperature, injection volume, and flow rate were 45°C, 5 μL, and 0.4 mL/min, respectively. The mobile phase A (MPA) was 60/40 acetonitrile/water, 10mM ammonium formate+0.1% formic acid and MPB was 89.1/9.9/0.99 isopropanol/acetonitrile/water, 10 mM ammonium formate +0.1% formic acid. The chromatographic gradient was 0 min: 10%B, 2.00 min: 10% B, 9 min: 100%B, 10.50 min: 100%B, 11.00 min: 10%B, 14.00 min: 10%B. The MS parameters were as follows: spray voltage: 3600 V for positive ionization, sheath gas: 40, auxiliary gas: 10, sweep gas: 1, ion transfer tube temperature: 300°C, vaporizer temperature: 350°C, RF lens(%): 60. The isobaric compounds, zeaxanthin, and lutein were isolated by quadrupole and targeted MS/MS fragmentation was performed with the normalized HCD energy(%): 20, maximum injection time of 54 milliseconds and normalized AGC target (%): 50 followed by detection in orbitrap analyzer at 30K resolution. The data acquisition was performed using the Thermo Scientific Xcalibur and data analyses using the Tracefinder 5.1. Zeaxanthin and lutein are chromatographically separated and have distinct fragmentation patterns: Zeaxanthin 568.4264 m/z → 476.3642m/z (Quantifier) and lutein 568.4266 m/z → 338.2601m/z. The retention time and fragmentation patterns from the TIF and plasma samples were matched with the respective reference standards.

Quantification and statistical analysis

All statistical analyses were performed using GraphPad Prism 8. Before statistical testing, datasets were evaluated for normality. Data following a normal distribution were analyzed using two-tailed Student’s t-tests or one-way ANOVA for comparisons of two conditions or more than two conditions, respectively. p values were corrected for multiple comparisons using Dunnett’s test (when multiple experimental groups were compared to a single control group) or Tukey’s test (when multiple conditions were compared within a single experiment). two-way ANOVA was applied to analyze experiments involving two-factor variables.Data are expressed as mean ± s.e.m or mean ± s.d as illustrated on the figure legend. Statistical significance was defined as p < 0.05. Significance levels are indicated as follows:(ns, not significant; ∗0.01 < p < 0.05; ∗∗0.001 < p < 0.01; ∗∗∗p < 0.001). Sample size (n) and the definition of n for each experiment are provided in the corresponding figure legends.

Published: September 1, 2025