Abstract

   The gut microbiota is increasingly recognized as a promising
   therapeutic target in cancer treatment. However, the specific
   mechanisms by which gut bacteria and their metabolites exert
   therapeutic effects in melanoma remain poorly understood. In this
   study, it is unexpectedly demonstrated that prophylactic
   supplementation with Limosilactobacillus reuteri exhibits significant
   tumor‐suppressive properties, primarily mediated by its secreted
   metabolite, reuterin. This metabolite induces trained immunity through
   macrophage metabolic reprogramming, thereby enhancing antitumor immune
   responses. Mechanistically, this process involves stabilizing HIF‐1α
   via the AHR‐ROS signaling pathway, enhancing glycerophospholipid
   metabolism, and elevating arachidonic acid levels, thereby amplifying
   the trained immunity response. Similar to reuterin, arachidonic acid
   also induces trained immunity and facilitates macrophage‐mediated tumor
   cell killing. To enhance its therapeutic efficacy, reuterin is
   encapsulated in covalent organic frameworks (COFs). COF‐Reuterin
   demonstrates superior effects in tumor‐associated macrophages (TAMs),
   remodulating intratumor bacteria and directly facilitating tumor cell
   killing. Notably, COF‐Reuterin demonstrates superior therapeutic
   efficacy compared to cisplatin. Furthermore, COF‐Reuterin reprogrammed
   TAMs from an M2 to an M1 phenotype, increasing CD8^+ T cell
   infiltration and decreasing myeloid‐derived suppressor cells (MDSCs),
   reshaping the immunosuppressive tumor microenvironment. These findings
   highlight the potential of probiotics and their metabolites in the
   metabolic reprogramming of TAMs, offering a promising cancer
   therapeutic approach.

   Keywords: intratumor bacteria, Limosilactobacillus reuteri, melanoma,
   reuterin, trained immunity, tumor‐associated macrophages
     __________________________________________________________________

   Covalent organic framework (COF)‐reuterin induces trained immunity in
   tumor‐associated macrophages, promoting the generation of nitric oxide
   and reactive oxygen species, thereby enhancing the antitumor immune
   response. Additionally, COF‐Reuterin directly kills tumor cells while
   simultaneously eradicating pathogenic intratumoural bacteria. Thus,
   COF‐Reuterin enhances both innate and adaptive antitumor immunity,
   inhibiting tumor growth and remodeling the tumor microenvironment.

   graphic file with name ADVS-12-e04784-g001.jpg

1. Introduction

   Melanoma, the deadliest form of skin cancer, originates from
   melanocytes and is driven by complex molecular mechanisms. Its global
   incidence continues to rise, placing increasing strain on healthcare
   systems. In certain regions, the incidence of melanoma increases by ≈3%
   annually, and it accounts for nearly 90% of skin cancer‐related
   deaths.^[ [40]^1 ^] In 2023, the United States reported 97610 new cases
   of melanoma, resulting in an estimated 7990 deaths.^[ [41]^2 ^] Despite
   advances in targeted therapies and immunotherapy, the 5‐year survival
   rate for advanced melanoma has improved from less than 10% to ≈30%,^[
   [42]^3 ^] but challenges such as resistance to these therapies and the
   variable efficacy of immunotherapy remain. These ongoing challenges
   underscore the complexity of melanoma treatment and emphasize the
   urgent need for innovative approaches. Personalized cancer therapies
   that enhance the immune response to melanoma and address current
   treatment limitations are becoming increasingly critical.

   Trained immunity refers to a process in which innate immune cells
   undergo metabolic and epigenetic reprogramming following an initial
   stimulus, resulting in enhanced nonspecific immune responses upon
   re‐exposure to either homologous or heterologous challenges.^[ [43]^4
   ^] Inducing trained immunity in innate immune cells presents a
   promising strategy for cancer management, primarily by augmenting
   antitumor immune responses.^[ [44]^5 ^] Research has shown that
   inducing trained immunity in hematopoietic progenitor cells can
   effectively inhibit lung cancer growth.^[ [45]^6 ^] Furthermore,
   activating trained immunity in alveolar macrophages, which reside in
   mucosal tissues, has been demonstrated to suppress melanoma lung
   metastasis in mice,^[ [46]^7 ^] highlighting the role of trained
   immunity in lung macrophages in combating metastatic tumor cells.
   Emerging evidence further suggests that tumor cells can reactivate
   these trained immune cells,^[ [47]^8 ^] emphasizing their potential in
   cancer immunotherapy. However, limited research exists on whether
   trained immunity can be effectively induced in tumor‐associated
   macrophages (TAMs) within the tumor microenvironment (TME) to stimulate
   antitumor immune responses. Current cancer therapies primarily focus on
   promoting the polarization of M1‐TAMs or reprogramming M2‐TAMs into an
   M1‐like antitumoral phenotype.^[ [48]^9 ^] This underscores the urgent
   need for innovative strategies that harness and regulate trained
   immunity in TAMs to enhance antitumor responses, paving the way for
   novel therapeutic approaches.

   Gut microbiota can influence cancer progression and the outcomes of
   various cancer therapies, including chemotherapy, immunotherapy, and
   radiotherapy. As such, strategies aimed at modulating gut microbiota
   hold promise for enhancing cancer prevention and treatment.^[ [49]^10
   ^] Growing understanding has led to an increasing demand among cancer
   patients actively seeking probiotics to improve health and potentially
   support cancer treatment.^[ [50]^11 ^] Specific bacterial species have
   been linked to both tumor promotion and suppression in colorectal
   cancer. For instance, Fusobacterium nucleatum, Bacteroides fragilis,
   and Streptococcus gallolyticus have been implicated in promoting
   tumorigenesis, whereas probiotics such as Clostridium butyricum,
   Streptococcus thermophilus, Lacticaseibacillus paracasei, and
   Ruminococcus gnavus have demonstrated antitumor effects.^[ [51]^12 ^]
   However, the precise mechanisms through which probiotics and their
   metabolites influence cancer progression, particularly by modulating
   innate immune memory, remain poorly understood. This critical knowledge
   gap restricts the full potential of probiotics as a viable therapeutic
   strategy in cancer immunotherapy.

   Here, we demonstrate that the metabolite reuterin, produced by
   Limosilactobacillus reuteri, induces trained immunity in TAMs, leading
   to potent antimelanoma effects. Specifically, pretreatment with
   reuterin triggers trained immunity in macrophages, enhancing macrophage
   cytotoxicity, nitric oxide (NO) production, and reactive oxygen species
   (ROS) generation, thereby significantly inhibiting tumor growth. The
   antitumor effects of reuterin‐induced trained immunity are attributed
   to metabolic reprogramming in macrophages, driving TAMs to acquire an
   antitumor phenotype. To maximize the therapeutic efficacy of reuterin,
   efficient drug delivery is critical. Recent studies have underscored
   the potential of covalent organic frameworks (COFs) as a versatile drug
   delivery system for cancer therapy.^[ [52]^13 ^] COFs, with their high
   surface area, tunable porosity, and functionalized surface chemistry,
   serve as effective carriers for encapsulating and targeting therapeutic
   agents.^[ [53]^14 ^] Consequently, reuterin was incorporated into the
   COFs, which not only facilitated the induction of trained immunity in
   TAMs but also modulated the intratumor bacteria and enhanced tumor cell
   killing. This strategy significantly suppressed tumor progression,
   demonstrating the potential of reuterin‐loaded COFs as a promising
   immunotherapeutic approach for melanoma.

2. Results

2.1. Prophylactic Supplementation with L. reuteri Inhibitis Melanoma Growth

   The gut microbiota has emerged as a pivotal regulator of host immunity,
   influencing both cancer progression and immune surveillance.^[ [54]^15
   ^] Among the array of beneficial microbes, Limosilactobacillus spp.
   have demonstrated considerable promise in modulating host immune
   responses.^[ [55]^16 ^] Here, we assessed whether prophylactic
   supplementation with Limosilactobacillus spp., followed by a washout
   period, could prime the host immune system before tumor inoculation to
   enhance antitumor immunity (Figure [56]1a). Based on phylogenetic
   analysis of the Limosilactobacillus genus, several representative
   strains were selected for subsequent experimental evaluation to explore
   their roles in immune activation and antitumor immunity (Figure [57]S1,
   Supporting Information). Our data demonstrated that preventive
   supplementation with Limosilactobacillus reuteri (L. re) DSM20016 and
   SKB1241 significantly inhibited tumor growth compared to the control
   group. In contrast, supplementation with Limosilactobacillus caviae,
   Limosilactobacillus pontis, or Limosilactobacillus fermentum failed to
   suppress tumor growth (Figure [58]1b,e). Tumor weight measurements and
   representative tumor images further corroborated these findings
   (Figure [59]1c,d), suggesting that the antitumor effects of
   Limosilactobacillus spp. are species‐specific. We hypothesize that
   distinct probiotic strains, shaped by divergent phylogenetic lineages,
   may exert antitumor effects through differential mechanisms.

Figure 1.

   Figure 1
   [60]Open in a new tab

   Prophylactic supplementation with L. reuteri elicits antitumor immune
   responses. a) Experimental timeline for evaluating the antitumor
   effects of Limosilactobacillus spp. Mice were gavaged daily with 200 µL
   of a bacterial suspension containing 2 × 10^8 CFU of different
   Limosilactobacillus strains for 7 consecutive days. After a 7‐day rest
   period, B16‐F10 tumor cells were injected into the mice to evaluate the
   subsequent antitumor responses. b,e) Average tumor volume (b) and
   individual tumor volumes (e) of melanoma‐bearing mice (n = 5). c) Tumor
   tissue weight (n = 5). d) Representative images of tumor tissues (n =
   5). f) Experimental design for assessing tumor‐associated macrophage
   responses after gavage with PBS (control), L. reuteri, or L. caviae.
   Mice were gavaged daily with 200 µL of a bacterial suspension
   containing 2 × 10^8 CFU of L. reuteri or L. caviae for 7 consecutive
   days. After a 7‐day rest period, B16‐F10 tumor cells were injected into
   the mice, and immune responses of tumor‐associated macrophages were
   subsequently analyzed. g) Macrophages were cocultured with B16‐F10
   cells for 48 h, followed by tumor cell survival assessment via the
   CCK‐8 assay (n = 3). h,i) TNF‐α secretion h) and ROS levels i) in
   macrophages cultured in medium or B16‐F10 tumor‐conditioned medium
   (TCM; supernatant collected from B16‐F10 tumor cell culture) (n = 3).
   Data represent means ± SD from 2–3 independent experiments. ns, not
   significant; **p < 0.01, ***p < 0.001; NO, nitric oxide; TCM,
   tumor‐conditioned medium.

   To elucidate the mechanisms underlying the potent tumor‐suppressive
   effects of L. reuteri, we focused on tumor‐associated macrophages
   (TAMs), a key immune cell population within the tumor microenvironment
   (TME) that regulates both tumor progression and immune responses.^[
   [61]^17 ^] Since macrophage activity is a crucial determinant of
   antitumor immunity,^[ [62]^18 ^] we hypothesized that L. reuteri could
   enhance macrophage‐mediated tumor cytotoxicity. To test this
   hypothesis, mice were pretreated with PBS, L. reuteri, or L. caviae as
   a control (Figure [63]1f). Our results demonstrated that L. reuteri
   pretreatment significantly increased macrophage cytotoxicity against
   tumor cells (Figure [64]1g). Nitric oxide (NO), a critical effector
   molecule, mediates tumor cell cytotoxicity, enhances macrophage
   activation, and modulates immune responses.^[ [65]^19 ^]
   Mechanistically, conditioned media from B16‐F10 melanoma cells
   stimulated macrophages to produce elevated levels of NO and reactive
   oxygen species (Figure [66]1h,i). These findings suggest that L.
   reuteri primes macrophages prior to their recruitment into the TME,
   augmenting their immune response and promoting antimelanoma immunity.

2.2. Reuterin Recapitulates the Prophylactic Effect of L. reuteri against
Melanoma

   To investigate the underlying mechanisms through which L. reuteri
   enhances macrophage‐mediated antimelanoma immunity, we examined whether
   its effects are driven by its cellular components or its metabolites.
   We compared the effects of oral administration of heat‐killed L.
   reuteri and its cell‐free supernatant (CFS), which contains bioactive
   molecules secreted by the bacteria (Figure [67]2a). Notably, the
   administration of L. reuteri CFS led to a significant reduction in
   tumor growth (Figure [68]2b), as evidenced by decreased tumor weight
   (Figure [69]2c). This finding suggests that the tumor‐suppressive
   effects of L. reuteri may be attributed to its secreted metabolites
   rather than the bacterial cells themselves. Among these metabolites,
   reuterin—an antimicrobial compound produced by most L. reuteri
   strains—has been identified as a key bioactive molecule.^[ [70]^20 ^]
   Given its immunomodulatory properties, we hypothesized that reuterin
   plays a pivotal role in mediating the observed antitumor effects. Our
   analysis showed that among Limosilactobacillus spp., only L. reuteri
   strains possess the ability to convert glycerol into reuterin (Figure
   [71]S2, Supporting Information). This observation raised the question
   of whether reuterin specifically contributes to the observed
   tumor‐suppressive effects or if other bioactive compounds might also be
   involved. To address this, we selected gamma‐aminobutyric acid
   (GABA)—another metabolite produced by certain Limosilactobacillus
   species—as a control.^[ [72]^21 ^] Comparative analysis showed that
   reuterin significantly inhibited melanoma growth, while GABA had no
   detectable effect (Figure [73]2d,e). Further analysis revealed that
   reuterin supplementation altered the polarization profile of
   tumor‐associated macrophages. Specifically, flow cytometry revealed a
   significant increase in CD86^+ (M1‐like) macrophages and a concurrent
   decrease in CD206^+ (M2‐like) macrophages, leading to an elevated M1/M2
   ratio (Figure [74]2f). These results imply that reuterin may, at least
   in part, mediate its antitumor effects through macrophage‐driven
   inflammatory remodeling of the tumor microenvironment.

Figure 2.

   Figure 2
   [75]Open in a new tab

   Reuterin recapitulated the prophylactic effect of L. reuteri. a)
   Experimental timeline for evaluating the antitumor effects of L.
   reuteri cell‐free supernatant (CFS) and heat‐killed L. reuteri (HK L.
   reuteri). Each group was administered a daily oral gavage of either
   200 µL of L. reuteri CFS, 200 µL of HK L. reuteri suspension (2 × 10^8
   CFU), or 200 µL of MRS medium (control). b) Average tumor volume in
   melanoma‐bearing mice (n = 5). c) Tumor tissue weight with
   representative images (n = 5). d,e) Tumor growth curves d) and tumor
   weights with representative images e) of melanoma‐bearing mice treated
   with PBS, GABA, or reuterin (n = 5). f) Flow cytometry analysis of the
   proportions of CD86^+ (M1 macrophages) and CD206^+ (M2 macrophages)
   within CD45^+F4/80^+ cells isolated from tumor tissues of B16‐F10
   tumor‐bearing mice pretreated with PBS, reuterin, GABA. The M1/M2
   macrophage ratio in tumor tissues was analyzed (n = 3). g,h) Tumor
   growth curves g) and tumor weights with representative images h) of
   melanoma‐bearing mice treated with 5% glycerol, L. reuteri, or L.
   reuteri combined with 5% glycerol (n = 5). i) Schematic of the
   experimental setup for evaluating macrophage responses following
   intraperitoneal injection of PBS (control) or reuterin (10 µg).
   Macrophages were subsequently isolated from the peritoneal cavity for
   further analysis. j) Macrophages were cocultured with B16‐F10 cells for
   48 h, and tumor cell viability was determined (n = 3). k,l) TNF‐α k)
   and NO l) production by macrophages cultured for 24 h in medium or TCM
   (n = 3). Data represent means ± SD from 2–3 independent experiments.
   ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001; NO, Nitric
   oxide; GABA, γ‐aminobutyric acid; TCM, Tumor‐conditioned medium.

   To assess whether glycerol metabolism by L. reuteri directly
   contributes to tumor suppression, we administered L. reuteri in a 5%
   glycerol solution. This combination led to significantly greater tumor
   inhibition compared to the administration of either glycerol or L.
   reuteri alone (Figure [76]2g,h). These findings support the hypothesis
   that reuterin, a key metabolite derived from glycerol,^[ [77]^22 ^] is
   crucial in limiting tumor progression. To further elucidate the role of
   reuterin in antitumor immunity, we examined its effects on macrophage
   activity (Figure [78]2i). Our results showed that reuterin pretreatment
   significantly enhanced macrophage‐mediated tumor cell killing
   (Figure [79]2j), accompanied by increased NO and TNF‐α production upon
   stimulation with TCM (Figure [80]2k,l). These findings indicate that
   reuterin primes macrophages and enhances their response to
   tumor‐derived signals. Beyond its direct effects on macrophages, we
   explored whether reuterin modulates antitumor immunity by altering the
   gut microbiota. Microbial analysis conducted prior to tumor inoculation
   revealed that reuterin supplementation increased overall diversity,
   with an enrichment of Firmicutes and a reduction in Bacteroidetes, as
   well as an increase in Lactobacillus and Limosilactobacillus species
   (Figure [81]S3, Supporting Information). These findings suggest that
   reuterin enhances immune function by both activating macrophages and
   modulating the gut microbiota, thereby supporting antitumor immunity.

2.3. Reuterin Imprints Innate Memory‐Like Responses in Macrophages to Promote
Antimelanoma Activity

   Traditionally, the innate and adaptive immune systems are
   differentiated by their specificity and memory capacity. However,
   recent studies have shown that the innate immune system is capable of
   acquiring memory‐like properties after transient stimulation, resulting
   in an enhanced response to subsequent challenges.^[ [82]^23 ^] Research
   further demonstrates that prevaccination with bacteria‐derived outer
   membrane vesicles can enhance tumor vaccination through the induction
   of trained immunity.^[ [83]^24 ^] Given the pivotal role of macrophages
   in tumor progression and immune modulation, we hypothesized that
   pre‐exposure to reuterin could induce trained immunity in these cells,
   thereby enhancing their ability to recognize and eliminate tumor cells
   more effectively, and ultimately strengthening the antitumor immune
   response. To test this hypothesis, we first evaluated the cytotoxicity
   of reuterin (Figure [84]S4, Supporting Information), followed by an
   analysis of its effect on trained immunity activity using our
   previously established macrophage model (Figure [85]3a). Macrophages
   pretreated with reuterin demonstrated enhanced production of TNF‐α and
   IL‐6 following secondary stimulation with LPS and Pam3CSK4
   (Figures [86]3b,c), similar to the response observed with heat‐killed
   C. albicans (HK C. a), a known trained immunity inducer.^[ [87]^25 ^]
   To further explore the functional effects of reuterin on macrophage
   activation, we evaluated NO levels at various reuterin concentrations
   following subsequent LPS stimulation (Figure [88]3d). This effect was
   also confirmed in bone marrow‐derived macrophages, demonstrating the
   universality of reuterin‐induced trained immunity in macrophages
   (Figure [89]S5, Supporting Information). Considering glycolysis as a
   key metabolic driver of trained immunity, we next examined its role in
   reuterin‐induced trained immunity.^[ [90]^26 ^] However, pretreatment
   with the glycolysis inhibitor 2‐deoxy‐d‐glucose (2‐DG) failed to impair
   trained immunity, suggesting that glycolysis is not essential for this
   process (Figure [91]3e). Instead, the histone methyltransferase
   inhibitor 5′‐deoxy‐5′‐(methylthio)adenosine (MTA) significantly
   attenuated reuterin‐induced trained immunity (Figure [92]3f),
   indicating that reuterin drives trained immunity through alternative
   epigenetic mechanisms independent of glycolysis. Further supporting its
   functional relevance, macrophages pre‐exposed to reuterin displayed
   enhanced responsiveness to tumor‐derived signals. Tumor‐conditioned
   medium stimulation increased TNF‐α, NO, and ROS levels
   (Figures [93]3g‐i). In co‐culture experiments, reuterin‐trained
   macrophages exhibited significantly greater tumor cell cytotoxicity
   compared to untrained controls (Figures [94]3j). Collectively, these
   findings demonstrate that reuterin induces trained immunity, thereby
   enhancing macrophage‐mediated antitumor responses and strengthening
   host defense against melanoma.

Figure 3.

   Figure 3
   [95]Open in a new tab

   Reuterin triggers trained immunity responses in macrophages to enhance
   their antimelanoma activity. a) Experimental timeline for assessing
   macrophage‐trained immunity in vitro. b,c) TNF‐α and IL‐6 secretion in
   macrophages treated with reuterin (10 × 10^−6 m) or PBS, followed by
   100 ng mL^−1 LPS or 1 µg mL^−1 Pam3C stimulation after a 5‐day resting
   period (n = 3). d) NO production in macrophages treated with reuterin
   (10 nM=10 × 10^−9 m –50 × 10^−6 m) or PBS, followed by 100 ng mL^−1 LPS
   stimulation after a 5‐day resting period (n = 3). e,f) Macrophages were
   preincubated for 24 h with the glycolysis inhibitor 2‐deoxy‐D‐glucose
   (2‐DG, 10 × 10^−3 m) (e) or the histone methyltransferase inhibitor
   5′‐deoxy‐5′‐(methylthio)adenosine (MTA, 10 × 10^−6 m) (f), followed by
   training with reuterin for 24 h and a 5‐day resting period. TNF‐α
   levels in the macrophage supernatants were measured after LPS
   stimulation (n = 3). g‐i) TNF‐α g), NO h), and ROS i) production levels
   in macrophages cultured for 24 h in PBS or TCM (n = 3). j) Macrophages
   trained with PBS or reuterin were cocultured with B16‐F10 cells for 48
   h, and tumor cell viability was determined (n = 3). Data represent
   means ± SD from 3 independent experiments. ns, not significant; *p <
   0.05, **p < 0.01, ***p < 0.001; NO, Nitric oxide; TCM,
   Tumor‐conditioned medium.

2.4. Reuterin‐Induced Macrophage Trained Immunity via the AHR‐ROS‐HIF‐1α Axis

   To explore the molecular mechanisms underlying reuterin‐induced trained
   immunity, we investigated the role of the aryl hydrocarbon receptor
   (AhR) in macrophages (Figure [96]4a). Reuterin treatment on day 4
   significantly upregulated AhR expression, an effect that was completely
   reversed by the AhR antagonist stemregenin‐1 (Figure [97]4b).^[ [98]^27
   ^] Notably, blocking AhR signaling during the training phase suppressed
   the enhanced production of TNF‐α and NO upon LPS restimulation
   (Figures [99]4c, Figure [100]S6a, Supporting Information). More
   importantly, AhR inhibition abolished the increased cytotoxicity of
   trained macrophages against B16‐F10 tumor cells, establishing AhR as a
   key driver of reuterin‐induced trained immunity (Figure [101]4d). Given
   the established role of Akt/mTOR signaling in trained immunity via
   HIF‐1α induction,^[ [102]^28 ^] we examined whether this pathway
   contributes to reuterin‐induced macrophage training. However,
   reuterin‐treated macrophages showed no changes in p‐AKT or p‐mTOR
   levels but exhibited significant upregulation of HIF‐1α protein
   (Figure [103]4e). Inhibition of HIF‐1α with ascorbate attenuated the
   trained immunity phenotype, reducing TNF‐α and NO production (Figure
   4f; Figure [104]S6b, Supporting Information) and impairing tumoricidal
   activity against B16‐F10 cells (Figure [105]4g). Interestingly, AhR
   inhibition did not affect HIF‐1α mRNA levels (Figure [106]4h),
   suggesting post‐transcriptional regulation of HIF‐1α. To further
   investigate the molecular basis of HIF‐1α stabilization in trained
   macrophages, we focused on ROS, which have been implicated in HIF‐1α
   stabilization, particularly under normoxic conditions.^[ [107]^29 ^]
   Recent studies have shown that ROS produced by reuterin can disrupt
   Clostridioides difficile metabolism and pathogenicity.^[ [108]^30 ^]
   Quantitative analysis revealed that reuterin treatment induced a
   dose‐dependent increase in ROS levels on day 1 (Figure [109]4i).
   Pharmacological inhibition of AhR significantly suppressed ROS
   production (Figure [110]4j). Scavenging ROS with N‐acetylcysteine (NAC)
   notably reduced TNF‐α levels in reuterin‐trained macrophages
   (Figure [111]4k), which also decreased HIF‐1α expression
   (Figure [112]4l) and impaired the tumoricidal activity of these
   macrophages (Figure [113]4m). These results support a causal role for
   ROS in HIF‐1α activation, emphasizing that the AhR‐ROS‐HIF‐1α axis is
   crucial for reuterin‐induced trained immunity, driving enhanced
   macrophage‐mediated antitumor activity.

Figure 4.

   Figure 4
   [114]Open in a new tab

   Reuterin enhances macrophage cytotoxicity against tumor cells by
   inducing trained immunity via the AHR‐ROS‐HIF‐1α axis. a) Experimental
   timeline for assessing the molecular mechanisms of reuterin‐induced
   trained immunity in macrophages. b) Western blot analysis of AHR
   expression in macrophages after reuterin training and a 3‐day resting
   period. c) TNF‐α levels in macrophages pretreated with the AHR
   inhibitor stemregenin‐1 (10 × 10^−6 m), trained with reuterin, rested
   for 5 days, and stimulated with LPS (100 ng mL^−1) for 24 h (n = 3). d)
   Tumor cell viability after co‐culture of B16‐F10 cells with macrophages
   pretreated with stemregenin‐1, trained with reuterin, and rested for 5
   days (n = 3, 10:1 macrophage‐to‐tumor cell ratio). e) Western blot
   analysis of p‐AKT, p‐mTOR, and HIF‐1α in macrophages 3 days
   post‐reuterin training. f) TNF‐α levels in macrophages pretreated with
   the HIF‐1α inhibitor ascorbate (25 × 10^−6 m), trained with reuterin,
   rested for 5 days, and stimulated with LPS (n = 3). g) Tumor cell
   viability after co‐culture of B16‐F10 cells with macrophages pretreated
   with ascorbate, trained with reuterin, and rested for 5 days (n = 3,
   10:1 ratio). h) Relative mRNA expression of HIF‐1α in macrophages
   pretreated with the AHR inhibitor SR‐1 (25 × 10^−6 m), trained with
   reuterin, rested for 1 day (n = 3). i) ROS levels in macrophages
   treated with reuterin for 24 h (n = 3). j) ROS levels in macrophages
   pretreated with stemregenin‐1 and trained with reuterin (n = 3). k)
   TNF‐α levels in macrophages pretreated with the ROS inhibitor NAC (1 ×
   10^−3 m), trained with reuterin, rested for 5 days, and stimulated with
   LPS (n = 3). l) Western blot analysis of HIF‐1α expression in
   macrophages pretreated with NAC, trained with reuterin, and rested for
   3 days. m) Tumor cell viability after co‐culture of B16‐F10 cells with
   macrophages pretreated with NAC, trained with reuterin, and rested for
   5 days (n = 3, 10:1 ratio). Data represent means ± SD from 3
   independent experiments. ns, not significant; *p < 0.05, **p < 0.01,
   ***p < 0.001.

2.5. Glycerophospholipid Metabolism Drives Reuterin‐Induced Trained Immunity

   The AHR‐ROS pathway regulates macrophage activity and drives HIF‐1α
   activation, thereby modulating cellular metabolism. To explore the
   metabolic mechanisms underlying reuterin‐induced trained immunity, we
   performed nontargeted LC‐MS/MS metabolomics profiling. Principal
   component analysis (PCA) of 91 differentially regulated metabolites
   (VIP > 1, P < 0.05) revealed distinct metabolic shifts in
   reuterin‐trained macrophages (Figure [115]5a,b). Lipids and lipid‐like
   molecules dominated the metabolic profile, accounting for 40.6% of
   total metabolites (Figure [116]5c). KEGG pathway enrichment analysis
   identified significant enrichment in glycerophospholipid metabolism,
   choline metabolism, oxidative phosphorylation, and niacin and
   nicotinamide metabolism, with glycerophospholipid and choline
   metabolism being the most enriched pathways (Figure [117]5d). Gene Set
   Enrichment Analysis (GSEA) further corroborated these findings (Figure
   [118]S7a,b, Supporting Information). Reuterin treatment reduced key
   metabolites in choline metabolism, including choline, phosphocholine,
   and CDP‐choline (Figure [119]S7c–e, Supporting Information). This
   reduction was attributed to the upregulation of mRNA expression of the
   metabolic enzymes Chka, Chkb, Pcyt1b, and Chpt1, which mediate choline
   degradation (Figure [120]5e). HIF‐1α was found to directly bind the
   Chka gene promoter,^[ [121]^31 ^] enhancing its transcription, further
   supporting the central role of HIF‐1α in reuterin‐induced trained
   immunity. Metabolism analysis also revealed that phosphatidylcholine
   (PC) accumulated due to altered choline metabolism (Figure [122]S7f,
   Supporting Information), while PC‐derived phosphatidylethanolamine (PE)
   and phosphatidylserine (PS) levels were reduced (Figure [123]S7g,
   Supporting Information). Notably, the accumulation of PC provided a
   substrate for phospholipase A2 (PLA2)‐mediated hydrolysis, leading to
   increased levels of arachidonic acid and lysophosphatidylcholine (LPC),
   while linoleic acid remained unchanged (Figure [124]5f, Figure
   [125]S7h, Supporting Information). Arachidonic acid metabolites have
   been shown to modulate gene expression and epigenetic modifications,
   promoting trained immunity in immune cells.^[ [126]^32 ^] Similarly,
   our previous metabolomics study also demonstrated an increase in
   arachidonic acid levels in trained immunity induced by heat‐killed
   Candida albicans.^[ [127]^25b ^] To determine whether arachidonic acid
   contributes to reuterin‐induced trained immunity, we supplemented
   macrophages with arachidonic acid during reuterin treatment
   (Figure [128]5g). As expected, arachidonic acid supplementation
   enhanced trained immunity, acting synergistically with reuterin to
   further amplify trained immunity phenotype. Moreover, activation of
   PLA2 by Melittin^[ [129]^33 ^] further enhanced trained immunity
   (Figures [130]5h,i), whereas inhibition of PLA2 by 1‐naphthylacetic
   acid (NAA) partially attenuated reuterin‐induced trained immunity,^[
   [131]^34 ^] as reflected by a reduction, but not a complete loss, in
   the tumor cell‐killing capacity, TNF‐α, and NO production
   (Figure [132]5j,k, Figure [133]S7i, Supporting Information). These
   findings highlight that reuterin enhances glycerophospholipid
   metabolism to promote arachidonic acid biosynthesis, thereby amplifying
   trained immunity responses (Figure [134]5k).

Figure 5.

   Figure 5
   [135]Open in a new tab

   Glycerophospholipid metabolism mediates reuterin‐induced trained
   immunity. Macrophages were primed with PBS (control) or reuterin for 24
   h, followed by a 5‐day resting period. Untargeted metabolomics was
   performed to assess metabolic changes. a) PLS‐DA plot of metabolomic
   profiles from PBS‐ and reuterin‐treated macrophages. b) Volcano plot of
   differential metabolites. c) Classification of differential
   metabolites. d) KEGG enrichment of differential metabolites (VIP > 1, P
   < 0.05, |log[2]FC| ≥ 1.2). e) Relative mRNA expression levels of
   glycerophospholipid metabolism genes in PBS‐ or reuterin‐treated
   macrophages (n = 3). f) Linoleic acid and arachidonic acid levels
   identified in metabolomic profiles. g) Proposed model of trained
   immunity induced by arachidonic acid and melittin. h, i) NO production
   (h) and TNF‐α levels i) in reuterin‐trained macrophages supplemented
   with arachidonic acid (10 × 10^−6 m) and melittin (250 ng mL^−1) after
   LPS restimulation (n = 3). j) TNF‐α levels in macrophages pretreated
   with the PLA2 inhibitor 1‐naphthylacetic acid (NAA; 50 × 10^−6 m),
   trained with reuterin, rested for 5 days, and stimulated with LPS (n =
   3). k) Tumor cell viability (n = 3, macrophage:tumor cell ratio = 10:1)
   following coculture with reuterin‐trained macrophages pretreated with
   PLA2 inhibitor NAA (50 × 10^−6 m). l) Schematic diagram of
   glycerophospholipid metabolism. Data represent means ± SD from 3
   independent experiments. ns, not significant; *p < 0.05, **p < 0.01,
   ***p < 0.001.

2.6. Covalent Organic Framework Based Reuterin Induces Trained Immunity in
Tumor‐Associated Macrophages

   To enhance reuterin‐mediated trained immunity and improve antitumor
   efficacy, we developed a Covalent Organic Framework (COF) delivery
   system. The COF was synthesized via a Schiff base reaction between
   1,3,5‐benzotri‐carbaldehyde (BTCA) and 1,3,5‐tris(4‐aminophenyl)benzene
   (TAPB) (Figure [136]6a). Scanning electron microscopy (SEM) revealed
   monodispersed spherical nanoparticles, with a slight increase in size
   following reuterin loading (Figure [137]6b). Dynamic light scattering
   (DLS) measurements showed hydrodynamic diameters of 198.2 nm (COF) and
   204.8 nm (COF‐Reuterin), with polydispersity index (PdI) values of
   0.054 and 0.071, respectively (Figure [138]6c). Zeta potential analysis
   indicated values of ‐34.1 mV (COF) and ‐36.4 mV (COF‐Reuterin),
   suggesting good colloidal stability (Figure [139]6d). Fourier transform
   infrared spectroscopy (FT‐IR) confirmed successful COF formation and
   reuterin encapsulation (Figure [140]6e). In vitro degradation
   experiments showed that COF‐Reuterin nanoparticles underwent
   substantial structural breakdown under simulated acidic tumor
   microenvironment conditions (pH 5.0), with near‐complete disintegration
   observed at 48 h, as confirmed by scanning electron microscopy (Figure
   [141]S8, Supporting Information). Biocompatibility assessment revealed
   low cytotoxicity of COF in macrophages and RAW264.7 cells, compared to
   Alum (Figure [142]S9, Supporting Information). In addition, hemolysis
   assays indicated that COF‐Reuterin nanoparticles did not induce red
   blood cell damage, further supporting their therapeutic safety (Figure
   [143]S10, Supporting Information). Furthermore, in vivo toxicity
   assessments conducted post‐administration showed no significant changes
   in key hepatic and renal biomarkers (ALT, BUN, ALP, CREA, UA),
   highlighting the nanoparticles’ favorable long‐term safety profile
   (Figure [144]S11, Supporting Information). To assess whether
   COF‐Reuterin induces trained immunity, we first examined its impact on
   macrophage activation. Treatment with COF‐Reuterin notably increased
   the levels of TNF‐α and NO in response to LPS stimulation
   (Figure [145]6f,g). Additionally, COF‐Reuterin‐trained macrophages
   displayed enhanced phagocytic activity against Staphylococcus aureus
   and improved tumoricidal capacity against B16‐F10 cells
   (Figure [146]6h,i). TAMs are the predominant immune cell population
   within the TME. To determine whether COF‐Reuterin induces trained
   immunity in TAMs and boosts antitumor activity, we administered various
   formulations to mice bearing orthotopic B16‐F10 tumors
   (Figure [147]6j). TAMs were isolated to assess trained immunity, with
   COF‐Reuterin inducing significantly higher levels of TNF‐α, NO, and ROS
   production compared to PBS or free reuterin (Figure [148]6k‐m). These
   results demonstrate that COF‐Reuterin effectively induces trained
   immunity in TAMs, thereby enhancing antitumor responses.

Figure 6.

   Figure 6
   [149]Open in a new tab

   Covalent Organic Framework‐loaded reuterin triggers trained immunity in
   TAMs. a) Schematic representation of COF‐Reuterin synthesis. b) SEM
   images of COF and COF‐Reuterin (scale bars: 100 nm). c) Hydrodynamic
   size distribution of COF and COF‐reuterin (n = 3). d) Zeta potential of
   COF and COF‐Reuterin. e) FTIR spectra confirming COF‐reuterin
   synthesis. f,g) Macrophages were primed in vitro with COF‐Reuterin or
   control and stimulated with LPS. NO f) and TNF‐α g) levels were
   measured (n = 3). h) Macrophages were cocultured with Staphylococcus
   aureus for 30 min, and phagocytic activity was assessed by CFU
   quantification (n = 5). i) Macrophages were cocultured with B16‐F10
   cells (macrophage: tumor cell ratio 10:1) for 48 h, and tumor cell
   viability was measured (n = 3). j) Schematic diagram illustrating the
   in vitro trained immunity analysis for TAMs. k,l) TAMs were trained and
   stimulated with LPS. NO (k) and TNF‐α (l) levels were then measured (n
   = 3). Data represent means ± SD from 2–3 independent experiments. *p <
   0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. COF, covalent organic
   framework; TAMs, tumor‐associated macrophages.

2.7. COF‐Reuterin Drives Antigen‐Specific Cellular and Humoral Immune
Response

   To evaluate the potential of COF‐Reuterin as a cancer vaccine adjuvant,
   we employed ovalbumin (OVA) as a tumor‐associated model antigen. Based
   on its ability to suppress LPS‐induced TNF‐α and NO production in
   macrophages (Figure [150]S12, Supporting Information), we hypothesized
   that incorporating reuterin into a vaccine platform could enhance
   immunogenicity while mitigating excessive inflammatory responses
   post‐vaccination. To test this hypothesis, we formulated a
   COF‐Reuterin‐based vaccine and assessed adaptive immune responses in
   C57BL/6J mice (Figure [151]7a). COF‐Reuterin + OVA vaccination
   significantly increased OVA‐specific IgG titers and enhanced the
   affinity of OVA‐specific antibodies, as measured by ELISA compared to
   alum‐adjuvanted vaccines (Figure [152]7b,c). IgG subclass analysis
   revealed distinct immune polarization: while alum predominantly induced
   a Th2‐biased response, free reuterin promoted Th1 polarization.
   COF‐Reuterin effectively balanced Th1/Th2 responses, elevating both
   IgG1 and IgG2c, with a more pronounced increase in IgG2c compared to
   alum (Figure [153]7d,e). Post‐vaccination cytokine analysis showed no
   significant elevation in systemic TNF‐α levels (Figure [154]7f). Safety
   assessments revealed no organ abnormalities or significant weight
   changes, supporting the biocompatibility of COF‐Reuterin (Figure
   [155]S13, Supporting Information). Flow cytometry analysis demonstrated
   that COF‐Reuterin +OVA vaccination increased CD4⁺ and CD8⁺ T cell
   proportions compared to OVA alone or OVA + alum (Figure [156]7h,i).
   Further, OVA re‐stimulation revealed robust antigen‐specific lymphocyte
   proliferation in the COF‐Reuterin group (Figure [157]7g), indicating
   that COF‐Reuterin effectively activates an immune response specific to
   the OVA antigen. To assess its protective efficacy, mice were immunized
   weekly with saline, Reuterin‐OVA, COF‐OVA, or COF‐Reuterin‐OVA for
   three consecutive weeks (Figure [158]7j), followed by a B16‐OVA tumor
   challenge. COF‐Reuterin‐OVA exhibited superior antitumor efficacy,
   significantly reducing tumor burden, as evidenced by decreased tumor
   volume and weight (Figure [159]7k–m, Figure [160]S14, Supporting
   Information). Notably, COF‐loaded reuterin treatment significantly
   increased the proportion of CD86^+ (M1‐like) macrophages while
   decreasing CD206^+ (M2‐like) macrophages within the tumor
   microenvironment, indicating a shift toward a proinflammatory
   macrophage phenotype (Figure [161]S15a–c, Supporting Information). To
   evaluate the post‐training immune response induced by COF‐Reuterin,
   macrophages were isolated from tumor tissues of immunized mice from all
   groups and restimulated with LPS. Macrophages from the COF‐Reuterin
   group showed enhanced production of the proinflammatory cytokine TNF‐α
   upon restimulation, consistent with features of trained immunity
   (Figure [162]S15d, Supporting Information). These findings establish
   COF‐Reuterin‐OVA as an effective prophylactic vaccine candidate,
   eliciting potent antitumor immunity.

Figure 7.

   Figure 7
   [163]Open in a new tab

   COF‐Reuterin amplifies the antigen‐specific immune response to OVA. a)
   Schematic representation of the immunization schedule. Mice were
   subcutaneously immunized with OVA, OVA + Reuterin, OVA + Alum, OVA +
   COF, OVA + COF‐Reuterin at week 0, followed by a booster dose at week
   2. b) Serum levels of anti‐OVA IgG antibodies were quantified by ELISA
   at week 7 (n = 5). c) IgG avidity index was measured at week 4 using
   NaSCN‐based ELISA (n = 5). d, e) Serum levels of anti‐OVA IgG1 d), and
   IgG2c e) isotype antibodies were quantified by ELISA at week 7 (n = 5).
   f) Serum TNF‐α levels were measured 4 h after vaccination as an early
   indicator of inflammatory response (n = 3). g) At week 7
   post‐immunization, splenocytes were stimulated with OVA (100 ng mL^−1)
   for 72 h, and cell proliferation was measured (n = 3). h, i) Flow
   cytometry analysis for the detection of CD3^+ CD4^+ T cells h) and
   CD3^+ CD8^+ T cells in splenocytes isolated at week 7 from immunized
   mice. Quantitative analysis results are shown on the right (n = 3). j)
   Schematic schedule for the prophylactic experiments in subcutaneous
   B16‐OVA tumor‐bearing mice. k) Tumor growth curves of B16‐OVA
   tumor‐bearing mice (n = 5). l) Tumor tissue weight measurement. m)
   Tumor tissue representative images (n = 5). Data represent means ± SD
   from 2–3 independent experiments. ns, not significant; *p < 0.05, **p <
   0.01, ***p < 0.001, ****p < 0.0001.

2.8. COF‐Reuterin Exhibits Potent Anticancer Activity and Remodels the Tumor
Microenvironment

   The therapeutic efficacy of COF‐Reuterin was assessed in B16‐OVA
   tumor‐bearing mice (Figure [164]8a). Consistently, COF‐Reuterin‐OVA
   exhibited the most potent antitumor activity, achieving 85.93% tumor
   inhibition compared to the PBS group (Figure [165]8b–d and Figure
   [166]S16, Supporting Information), highlighting its potential as a
   therapeutic cancer vaccine. Notably, COF‐Reuterin‐OVA exhibited a
   superior therapeutic effect compared to cisplatin, as reuterin directly
   induced tumor cell cytotoxicity more efficiently, leading to a more
   significant reduction in tumor growth (Figure [167]S17, Supporting
   Information). No significant body weight loss was observed during
   treatment (Figure [168]S18, Supporting Information), further supporting
   the excellent biocompatibility of COF‐Reuterin‐OVA. To further
   investigate its immunomodulatory effects, we analyzed
   tumor‐infiltrating immune cells post‐vaccination. Flow cytometry
   revealed a significant increase in F4/80⁺CD86⁺ M1 macrophages, along
   with a concurrent reduction in F4/80⁺CD206⁺ M2 macrophages
   (Figure [169]8e). This shift toward an elevated M1/M2 macrophage ratio
   is indicative of a more pro‐inflammatory, antitumor microenvironment
   (Figure [170]S19, Supporting Information). Consistently,
   COF‐Reuterin‐OVA‐treated tumors displayed enhanced T‐cell infiltration,
   with a marked increase in CD3⁺ and CD8⁺ cytotoxic T cells compared to
   cisplatin (Figure [171]8f). Moreover, COF‐Reuterin significantly
   reduced CD11b⁺Gr‐1⁺ myeloid‐derived suppressor cells (MDSCs),
   alleviating immunosuppression and improving immune surveillance
   (Figure [172]8g). Biosafety is a fundamental consideration in vaccine
   development, requiring thorough evaluation prior to in vivo
   application. Histopathological analysis using H&E staining revealed no
   pathological abnormalities in major organs, further supporting their
   sustained biocompatibility (Figure [173]S20, Supporting Information).

Figure 8.

   Figure 8
   [174]Open in a new tab

   COF‐Reuterin enhances antitumor efficacy and remodels the tumor
   microenvironment. a) Schematic representation of the experimental
   timeline for anticancer studies in subcutaneous B16‐OVA tumor‐bearing
   mice. b) Tumor growth curves of B16‐OVA tumor‐bearing mice treated with
   different formulations (n = 5). c) Tumor tissue weight quantification
   (n = 5). d) Tumor tissue representative images (n = 5). e) Proportion
   of F4/80^+CD86^+ in CD45^+ cells (M1 macrophages) and F4/80^+CD206^+ in
   CD45^+ cells (M2 macrophages) in the tumor tissues of B16‐OVA
   tumor‐bearing mice (n = 3). f) Proportion of CD3^+ and CD3^+CD8^+ among
   total tumor‐infiltrating cells (n = 3). g) Proportion of CD11b^+ and
   Gr‐1^+ myeloid‐derived suppressor cells (MDSCs) among total
   tumor‐infiltrating cells (n = 3). h) Representative images of bacterial
   cultures from tumor tissues grown on tryptic soy broth (TSB) plates. i)
   Bubble plot showing the relative abundance of bacterial species in
   tumor tissues of B16‐OVA tumor‐bearing mice treated with saline or
   COF‐OVA‐Reuterin. Bubble size represents bacterial abundance, with key
   species highlighted in red significantly reduced in the
   COF‐Reuterin‐OVA group. j) Immunohistochemical analysis of melanoma
   tissues demonstrated the expression of Ki67, VEGF, and MMP9 following
   treatment (scale bar, 50 µm). k) Quantitative analysis of Ki67, VEGF‐A,
   and MMP9 expression was performed by measuring the percentage of
   positive staining in tumor tissues (n = 3). Data represent means ± SD
   from 2–3 independent experiments. ns, not significant; *p < 0.05, **p <
   0.01, ***p < 0.001.

   Recognizing that intratumor dysbiosis contributes to immune tolerance
   and immunotherapy resistance,^[ [175]^35 ^] we next assessed
   COF‐Reuterin's impact on the tumor microbiota. Colony‐forming unit
   (CFU) assays revealed a fivefold reduction in total bacterial load
   following treatment, with 16S rRNA sequencing confirming significant
   depletion of Rodentibacter pneumotropicus, Staphylococcus aureus, and
   Escherichia coli within the TME (Figure [176]8h,i). Notably, in vitro
   assays demonstrated that COF‐Reuterin‐OVA effectively inhibited the
   growth of Staphylococcus aureus, and Escherichia coli, suggesting a
   direct antimicrobial effect, which may contribute to intratumor
   bacterial remodulating (Figure [177]S21, Supporting Information).^[
   [178]^36 ^] Immunohistochemical analysis revealed that COF‐Reuterin‐OVA
   treatment significantly reduced the expression of Ki67 (a proliferation
   marker), VEGF‐A (vascular endothelial growth factor A), and MMP9
   (matrix metalloproteinase 9) in melanoma tissues, indicating its
   effective inhibition of tumor cell proliferation, angiogenesis, and
   invasiveness (Figure [179]8j,k). To further validate the
   generalizability of our findings beyond melanoma, we evaluated tumor
   growth kinetics and tumor weight in an orthotopic 4T1 breast tumor
   model (Figure [180]S22a, Supporting Information). Notably, intravenous
   administration of antibiotics alone was insufficient to inhibit tumor
   growth, whereas the combination of cisplatin and antibiotics
   demonstrated a significantly enhanced antitumor effect (Figure
   [181]S22b‐d, Supporting Information). Consistent with our melanoma
   findings, treatment with COF‐Reuterin‐TAs significantly inhibited tumor
   progression in this breast cancer model. These findings indicate that
   COF‐Reuterin exerts its therapeutic effects through multiple
   mechanisms, including macrophage polarization, T‐cell activation,
   suppression of immunosuppressive cells, tumor microbiota modulation,
   and direct tumor cell cytotoxicity. Collectively, these
   immunomodulatory and antimicrobial properties contribute to enhancing
   antitumor immunity and improving therapeutic outcomes.

3. Discussion

   Over the past two decades, significant breakthroughs have been achieved
   in the field of cancer immunotherapy. Immune checkpoint blockade (ICB)
   therapy and chimeric antigen receptor T (CAR‐T) cell therapy, as
   advanced immunotherapeutic approaches, have demonstrated remarkable
   therapeutic efficacy across a variety of tumor types. However, 30–70%
   of patients fail to respond to immune checkpoint inhibitors, and the
   performance of CAR‐T cell therapy against malignant solid tumors has
   been less than optimal.^[ [182]^37 ^] Therefore, there is an urgent
   need for innovative strategies that can convert immunologically “cold”
   tumors into “hot” tumors with immune reactivity. Recent research has
   introduced the concept of “immune mobilization strategies,” which aim
   to induce trained immunity, thereby enhancing the efficiency of the
   innate immune system over time.^[ [183]^24 ^] This elevated state
   enables the immune system to mount a stronger response upon subsequent
   immune stimulation. Inducing trained immunity in TAMs to activate and
   amplify antitumor immune responses represents an innovative strategy
   with the potential to overcome the immunosuppressive TME, offering new
   breakthroughs in cancer immunotherapy.

   The relationship between cancer and the gut microbiota has garnered
   increasing attention in recent years, revealing complex interactions
   that can influence tumorigenesis and treatment outcomes.^[ [184]^38 ^]
   However, our understanding of how microbial metabolites affect the
   initiation and progression of melanoma remains limited. Recent studies
   have highlighted the dual role of certain bacterial populations in
   either promoting or inhibiting cancer growth.^[ [185]^39 ^] Microbial
   byproducts, such as butyrate and other short‐chain fatty acids, have
   been shown to exert diverse effects on cancer cells. While some studies
   suggest that high concentrations of butyrate can induce cell death in
   cancer cells,^[ [186]^40 ^] other research indicates that
   carbohydrate‐derived metabolites, like butyrate, can promote excessive
   proliferation of colonic epithelial cells, increasing the risk of
   colorectal tumors.^[ [187]^41 ^] In the present study, we identified a
   unique antimelanoma effect of Limosilactobacillus reuteri, a species
   within the Limosilactobacillus genus, and demonstrated that its
   metabolite reuterin plays a critical role in inducing trained immunity,
   which is essential for the antimelanoma response.

   Trained immunity is characterized by the epigenetic and metabolic
   reprogramming of innate immune cells in response to specific stimuli.
   This rewiring not only confers broad, nonspecific protection against a
   variety of diseases but also enhances the immune response to
   established immunotherapies and vaccines. While much of the current
   research on trained immunity has focused on biologically derived
   molecules like β‐glucan and BCG vaccines, there is growing evidence
   that other molecular inducers, such as uremic toxin indoxyl sulfate,
   also play a role in modulating immune responses. Specifically, indoxyl
   sulfate has been shown to induce trained immunity via the AhR‐dependent
   arachidonic acid pathway.^[ [188]^42 ^] Interestingly, while many other
   molecules, such as hippuric acid, indole 3‐acetic acid, and kynurenic
   acid, are AhR ligands, they did not show trained immunity responses
   like indoxyl sulfate.^[ [189]^42 ^] This suggests that specific
   molecular characteristics may determine the induction of trained
   immunity. Additionally, indoxyl sulfate promotes ROS production through
   the aryl hydrocarbon receptor‐NADPH oxidase pathway.^[ [190]^43 ^]
   Previous studies have identified the Akt/mTOR/HIF‐1α signaling pathway
   as a key mediator in the glycolytic induction of trained immunity.
   However, our results demonstrate that in reuterin‐induced trained
   immunity, the phosphorylation levels of AKT and mTOR remain largely
   unchanged. In contrast, we observed a notable increase in HIF‐1α
   expression. Interestingly, we also observed that reuterin‐induced
   training in macrophages results in a substantial accumulation of ROS.
   ROS have been shown to act as alternative inducers of HIF‐1α,
   independent of hypoxia, promoting HIF‐1α stabilization and facilitating
   the metabolic reprogramming essential for trained immunity.
   Furthermore, studies have demonstrated that ROS generated during
   trauma‐induced injury and tumor progression play a crucial role in
   stabilizing HIF‐1α responses,^[ [191]^44 ^] highlighting their
   importance in immune reprogramming.

   Glycolysis, glutamine metabolism, and cholesterol synthesis pathways
   are known to be essential for β‐glucan‐induced trained immunity.^[
   [192]^45 ^] However, no significant alterations were observed in these
   pathways in our metabolomics data. In contrast, glycerophospholipid
   metabolism and choline metabolism in cancer cells were significantly
   activated, leading to increased choline consumption and
   phosphatidylcholine accumulation, which correlated with elevated HIF‐1α
   levels. Phosphatidylcholine is hydrolyzed by phospholipase A2 to
   produce both arachidonic acid and lysophosphatidylcholine.^[ [193]^46
   ^] Previous studies have shown that the glycerophospholipid‐linoleic
   acid metabolic pathway plays a key role in BCG vaccine‐induced trained
   immunity.^[ [194]^47 ^] Our findings indicate that phosphatidylcholine
   accumulation plays a critical role in the induction of trained
   immunity. Moreover, we demonstrated that arachidonic acid, the
   hydrolysis product of phosphatidylcholine, contributes to the induction
   of trained immunity. Recent studies have pinpointed various endogenous
   sterile damage factors, such as oxidized low‐density lipoprotein
   (oxLDL), lipoprotein(a), uric acid, hyperglycemia, and Western dietary
   patterns, which induce trained immunity through epigenetic
   reprogramming in human monocytes.^[ [195]^48 ^] Our results further
   support the idea that enhancing reuterin‐induced trained immunity
   through the activation of arachidonic acid synthesis or direct
   supplementation with arachidonic acid may represent an effective
   strategy. This suggests that the signaling pathway could serve as a
   potential target for trained immunity‐based immunotherapies, with
   arachidonic acid or its metabolic agonists potentially serving as
   vaccine adjuvants.

   Trained immunity, a functional state of myeloid cells, has emerged as a
   promising target in immune oncology. Research has demonstrated that
   inducing trained immunity with a bone marrow‐avid nanoimmunotherapeutic
   effectively inhibits tumor growth and enhances the immune system's
   responsiveness to checkpoint blockade therapy.^[ [196]^49 ^]
   Furthermore, studies indicate that nanovaccine formulation induces
   trained innate immunity and elicits strong antitumor adaptive immunity,
   significantly reversing tumor growth and even completely eliminating
   established tumors.^[ [197]^50 ^] Probiotics are increasingly
   recognized for their potential in cancer immunotherapy, with
   demonstrated immunomodulatory effects across various cancers.
   Trimethylamine N‐oxide (TMAO) derived from Clostridiales induces
   pyroptosis in tumor cells, promoting cell death and enhancing CD8^+ T
   cell‐mediated antitumor immunity.^[ [198]^51 ^] Engineered Escherichia
   coli Nissle 1917‐derived L‐Arginine increases tumor‐infiltrating T
   cells in colorectal cancer, boosting immune activation and enhancing
   the efficacy of PD‐L1 blocking antibodies.^[ [199]^52 ^] Additionally,
   butyric acid from Ruminococcus and related genera activates cytotoxic
   CD8^+ T cells in gastric cancer, improving the response to anti‐PD1
   therapy.^[ [200]^53 ^] These findings suggest that probiotics, by
   modulating immune function and the tumor microenvironment, may offer
   novel strategies for cancer immunotherapy as adjuvant therapies.

   Porous materials, such as metal‐organic frameworks (MOFs) and Covalent
   Organic Frameworks (COFs), have demonstrated efficacy in drug delivery
   by encapsulating drugs and enabling sustained release. Specifically,
   Zeolitic Imidazolate Framework‐8 (ZIF‐8) nanoparticles, functionalized
   with polydopamine (PDA), have been incorporated into a dual‐network
   hydrogel composite, which is designed to accelerate the healing of
   diabetic bone defects.^[ [201]^54 ^] COFs are crystalline organic
   polymers connected by dynamic covalent bonds, which allow them to load
   small molecules. Their metal‐free composition is expected to enhance
   biocompatibility.^[ [202]^55 ^] Hydrogels are biocompatible materials
   with flexible properties that make them useful for both sensing and
   therapy. They can be designed to assemble in response to disease,
   degrade in a controlled way, and release drugs locally in response to
   stimuli.^[ [203]^56 ^] A biphasic system combining hydrogels and
   nanoparticles can simultaneously deliver antigens and immune training
   inducers, enhancing both adaptive and trained immunity for long‐lasting
   antitumor immune memory in cancer vaccines.^[ [204]^50 ^] Hydrogels can
   modulate the balance between M1 and M2 macrophage phenotypes, thereby
   suppressing inflammation induced by reactive oxygen species.^[ [205]^54
   , [206]^57 ^] Macrophages trained with COF‐Reuterin exhibit
   significantly increased tumoricidal activity, demonstrating enhanced
   efficacy in killing tumor cells. Inducing trained immunity in innate
   immune cells holds great promise for cancer therapy by amplifying the
   antitumor immune response. However, whether trained immunity can be
   activated in TAMs within the TME to trigger antitumor immune responses
   remains underexplored. Macrophages exhibit metabolic plasticity,
   allowing them to reprogram their metabolic pathways in response to
   environmental cues. The bioactive and smart‐responsive composite
   hydrogel can selectively induce macrophage polarization toward the M2
   phenotype by modulating the local microenvironment to promote tissue
   repair and regeneration.^[ [207]^58 ^] M1‐TAMs are effective at
   phagocytosis and cytokine secretion, whereas M2‐TAMs promote immune
   suppression and tumor evasion. Our results show that COF‐Reuterin
   increases the proportion of M1‐TAMs and decreases M2‐TAMs in melanoma,
   along with a significant reduction in myeloid‐derived suppressor cells
   (MDSCs). These findings underscore the potential of
   COF‐Reuterin‐induced trained immunity as a novel strategy to reprogram
   TAMs, offering a promising approach to enhance antitumor immunity and
   improve cancer therapy.

   Despite the relatively low abundance of resident microbes within
   tumors, they play an essential role in tumor progression.^[ [208]^59 ^]
   These microbes are primarily localized within tumor and immune cells,
   where they modulate cellular functions and activities. Each tumor has a
   unique bacterial community composition. For example, Staphylococcus
   aureus is a common pathogen in breast tumors and melanoma.^[ [209]^35b
   ^] Although the precise contributions and molecular mechanisms of
   tumor‐associated microbes in tumorigenesis remain poorly understood,
   existing evidence indicates that antibiotic treatment can effectively
   reduce bacterial presence in tumor tissues, thereby inhibiting the
   progression of lung cancer.^[ [210]^60 ^] Additionally, bacterial
   depletion in breast tumors has been shown to significantly decrease
   metastasis, highlighting the potential of targeting tumor‐associated
   microbes as a therapeutic strategy.^[ [211]^61 ^] In line with this,
   our results demonstrate that COF‐Reuterin exerts a multifaceted effect
   within the tumor microenvironment. Firstly, it enhances antitumor
   immunity by inducing trained immunity and improving the
   immunosuppressive tumor microenvironment. Secondly, COF‐Reuterin
   directly targets the tumor microbiota through the bactericidal activity
   of reuterin, resulting in a significant reduction in bacterial load
   within the tumor. Lastly, by leveraging the cytotoxic properties of
   reuterin, COF‐Reuterin holds promise for reversing established tumor
   growth and may even contribute to the complete eradication of tumors.

   In summary, we identified a gut commensal Limosilactobacillus
   reuteri‐derived small molecule trained immunity agonist, reuterin,
   which enhances glycerophospholipid metabolism through the
   AHR‐ROS‐HIF‐1α signaling axis. This leads to increased arachidonic acid
   production, thereby inducing trained immunity and promoting the host's
   ability to suppress melanoma growth in mice. In vivo experiments
   confirmed the tumor therapeutic efficacy of COF‐Reuterin, demonstrating
   its capacity to induce trained immunity, reduce intratumor bacterial
   load, and enhance tumor cell killing. Consequently, our study validates
   the successful strategy of inducing trained immunity in TAMs for
   melanoma treatment in mice. These findings provide valuable strategies
   for the development of tumor vaccines that do not rely on specific
   antigens, offering the potential for broad‐spectrum vaccine efficacy
   and guiding the development of next‐generation cancer vaccines.

4. Experimental Section

Bacteria Strains and Oral Administration

   The Limosilactobacillus genus strains L. reuteri DSM 20 016, L. reuteri
   SKB1241, L. caviae MOZM2, L. pontis LTH 2587, and L. fermentum NBRC
   15 885 were maintained at −80 °C in 25% glycerol prior to use. For
   experimental preparation, the strains were subcultured twice in MRS
   broth (Hopebio, China) and incubated overnight at 37 °C. Following
   cultivation, bacterial cells were harvested and resuspended in PBS to a
   final concentration of 1 × 10^9 CFU mL^−1. Mice were then gavaged daily
   with 200 µL of the bacterial suspension of different
   Limosilactobacillus strains for 7 consecutive days as indicated in the
   experimental protocol.

Mice

   Female C57BL/6J mice (6–8 weeks old) were purchased from Liaoning
   Changsheng Biotechnology Co., Ltd. and maintained in specific
   pathogen‐free (SPF) conditions at the Laboratory Animal Center of Jilin
   University. Mice were randomly assigned to experimental groups,
   ensuring age and sex matching.

Trained Immunity Experimental Protocol

   For in vitro experiments, Peritoneal macrophages (PMs) were isolated
   from 8‐week‐old female C57BL/6J mice by lavage of the peritoneal
   cavity. The cells were exposed to 10 × 10^−6 m reuterin for 24 h, with
   heat‐killed Candida albicans (HK‐C. a.) as a positive control. After
   incubation, the cells were washed twice with PBS and cultured in RPMI
   1640 medium supplemented with 10% fetal bovine serum (FBS) for 5 days.
   On day 6, the cells were restimulated with PBS, LPS (100 ng mL^−1),
   Pam3C (1 µg mL^−1), or 50% conditioned medium from B16‐F10 tumor cells.
   For in vivo experiments, 8‐week‐old female C57BL/6J mice were orally
   administered PBS, live L. reuteri, or heat‐killed L. reuteri (2 × 10⁸
   CFU) daily for 7 days. For inhibitor experiments, cells were pretreated
   with stemregenin 1 (MCE, 10 × 10^−6 m), or Ascorbic acid (MACKLIN,
   Shanghai, China, 25 × 10^−6 m) for 24 h prior to exposure to reuterin.
   For nanoparticle experiments, 8‐week‐old C57BL/6J mice were injected
   intraperitoneally with PBS, COF, or COF‐Reuterin. Seven days
   post‐injection, PMs were isolated and restimulated as described above.

Prophylactic and Therapeutic Antitumor Effects

   In the prophylactic study, 7‐week‐old female C57BL/6 mice were
   subcutaneously vaccinated with saline, OVA (100 µg), Reuterin (100 µg),
   COF‐OVA (100 µg COF + 100 µg OVA), or COF‐OVA‐Reuterin (100 µg
   COF‐Reuterin + 100 µg OVA) per mouse. Vaccination was administered once
   a week for three weeks, followed by a challenge with 5 × 10^5 B16‐OVA
   cells. Tumor progression was monitored daily using caliper
   measurements, and tumor volume was calculated using the formula V =
   (length × width2) / 2. In the therapeutic study, 7‐week‐old female
   C57BL/6 mice were injected with B16‐OVA cells. When the tumor volume
   reached ≈50 mm^3, mice were treated with saline (control), cisplatin
   (80 µg), COF‐OVA (100 µg COF + 100 µg OVA), COF‐OVA‐Reuterin (Low)
   (50 µg COF‐Reuterin + 100 µg OVA), or COF‐OVA‐Reuterin (High) (250 µg
   COF‐Reuterin + 100 µg OVA). Treatments were administered on days 7, 9,
   and 11 post‐inoculation. Mice were euthanized on day 14 for tumor
   analysis. In the breast tumor therapy study, 8‐week‐old female BALB/c
   mice were orthotopically injected with 4T1 cells to establish the tumor
   model. When tumor volumes reached ≈30–50 mm^3, mice were randomly
   assigned to receive one of the following treatments: saline (control);
   intravenous administration of 300 µL antibiotic suspension containing
   vancomycin (10 mg mL^−1), imipenem/cilastatin (4 mg mL^−1), and
   neomycin (1.5 mg mL^−1); intravenous cisplatin (80 µg); COF combined
   with tumor antigens derived from 4T1 tumor cell membranes (COF‐TAs;
   250 µg COF + 100 µg tumor cell membranes); or COF‐TAs combined with
   Reuterin (COF‐Reuterin‐TAs; 250 µg COF‐Reuterin + 100 µg tumor cell
   membranes). Treatments were administered on days 1, 4, 7, and 14
   post‐randomization. Tumor volumes were measured every three days
   throughout the duration of the study.

16S rRNA Gene Sequencing

   Fecal genomic DNA was extracted using a commercial DNA extraction kit,
   followed by library construction and high‐throughput sequencing. The
   16S rRNA gene library was constructed by amplifying the V3‐V4 regions,
   with forward primer 5′‐ACTCCTACGGGAGGCAGCA‐3′ and reverse primer
   5′‐CGGACTACHVGGGTWTCTAAT‐3′. Data analysis was performed using QIIME2
   and R (v3.3.2), with amplicon sequence variants (ASVs) classified using
   the Silva 138 reference database. Alpha diversity, including Shannon's
   index and observed species, was assessed via the Wilcoxon test and
   visualized with ggplot2. Beta diversity was evaluated using Bray‐Curtis
   dissimilarity and visualized through principal coordinate analysis
   (PCoA). Taxonomic composition at the phylum and family levels was
   visualized using QIIME2's taxa barplot function.

Metabolome Analysis

   Macrophages were treated with PBS or reuterin (10 × 10^−6 m) for 24 h
   and rested for 5 days. Metabolites were extracted using a
   methanol‐acetonitrile‐water solution (2:2:1, v/v) and analyzed by
   UHPLC‐MS on a Vanquish system coupled to an Orbitrap Exploris 120.
   Chromatography was conducted on a Waters ACQUITY UPLC BEH Amide column
   with mobile phases A (25 mmol L^−1 ammonium acetate and ammonia
   hydroxide, pH 9.75) and B (acetonitrile). The Orbitrap operated in IDA
   mode with full MS and MS/MS resolutions of 60 000 and 15 000,
   respectively. Data processing was performed using Compound Discoverer
   3.3, and statistical analysis was conducted with R and Python.

Macrophage‐Mediated Cytotoxicity Assay

   The cytotoxic activity of macrophages against B16‐F10 melanoma cells
   was assessed using a Transwell co‐culture system (Corning, USA).
   B16‐F10 cells were seeded into the lower chamber of a Transwell plate
   at a density of 2 × 10⁴ cells per well in DMEM containing 10% fetal
   bovine serum (FBS) and incubated overnight to allow adherence.
   Activated macrophages were added to the upper chamber (0.4 µm pore
   inserts) in RPMI 1640 medium supplemented with 10% FBS. The co‐culture
   was maintained for 48 h under standard conditions. After the incubation
   period, the lower chamber was washed twice with PBS to remove debris
   and nonadherent cells. Tumor cell viability was determined using the
   Cell Counting Kit‐8 (CCK‐8, Beyotime, China) according to the
   manufacturer's protocol.

Nitric Oxide Production

   Nitric oxide (NO) levels in cell culture supernatants were quantified
   using a NO assay kit (Beyotime) in accordance with the manufacturer's
   instructions. Supernatants were collected post‐treatment, and NO
   production was evaluated by the Griess reaction, which measures
   nitrite, a stable end product of NO metabolism. Absorbance was recorded
   at 540 nm using a microplate reader, and the data were normalized to a
   standard curve derived from known sodium nitrite concentrations.

ROS Measurement

   Macrophages were seeded in 96‐well plates at a density of 2 × 10⁵ cells
   per well and treated with reuterin or inhibitor for 24 h. After
   treatment, the medium was removed, and the cells were washed with PBS.
   The cells were then incubated with 10 × 10^−6 m DCFH‐DA in RPMI 1640 at
   37 °C for 30 minutes. Following incubation, the cells were washed with
   PBS and ROS levels were assessed by fluorescence detection, using
   excitation and emission wavelengths of 488 nm and 525 nm, respectively.

RT‐qPCR Analysis

   RNA was extracted from cells using TRIzol reagent (TaKaRa) according to
   the manufacturer's instructions, and complementary DNA (cDNA) was
   synthesized using the PrimeScript Reverse Transcriptase Kit (TaKaRa).
   Quantitative real‐time PCR (RT‐qPCR) was performed with SYBR Green
   reagents on a QuantStudio 3 Detection System (Applied Biosystems,
   Thermo Fisher Scientific). Gene expression levels were normalized to
   β‐actin as the internal control and calculated using the comparative
   cycle threshold (2^−ΔΔCt) method. Primer sequences are listed in Table
   [212]S1 (Supporting Information). Each sample was analyzed in technical
   triplicates to ensure data reliability.

Reuterin Detection

   Reuterin production was quantified as described by Circle et al.^[
   [213]^62 ^] Briefly, overnight cultures of Limosilactobacillus strains
   were harvested by centrifugation to collect the cells. The cells were
   washed twice with 50 × 10^−6 m potassium phosphate buffer (PBS) and
   resuspended in 25 mL of 250 × 10^−6 m glycerol, corresponding to 500 mg
   of the cells. The suspension was incubated at 37 °C for 2 h to allow
   reuterin production. After incubation, the cells were pelleted by
   centrifugation, and the supernatant was filtered through a 0.45 µm
   syringe filter. For reuterin detection, 300 µL of each supernatant was
   mixed with 225 µL of a 10 × 10^−6 m tryptophan solution, followed by
   the addition of 900 µl of concentrated HCl. The reaction mixtures were
   incubated at 37 °C for 20 minutes, and absorbance was measured at
   560 nm using a Biotek microplate reader.

Synthesis of COF‐Reuterin

   The COF was synthesized using a Schiff base condensation reaction.
   Briefly, 1 mg of 1,3,5‐tris(4‐aminophenyl)benzene (TAPB) and 2 mg of
   benzene‐1,3,5‐tricarbaldehyde (BTCA) were dissolved in 4 mL of
   acetonitrile with ultrasonic assistance. Glacial acetic acid (0.3 mL)
   was added, and the mixture was stirred at room temperature for 24 h.
   After 12 h, 100 µL of reuterin solution was added, and the reaction
   continued for another 12 h. The product was isolated by centrifugation
   at 10 000 rpm for 10 min, washed three times with anhydrous ethanol,
   and dried at 60 °C for 24 h, yielding a pale yellow powder
   (COF‐Reuterin). Reuterin was dissolved in DMSO to prepare a stock
   solution, serially diluted for HPLC analysis, and used to generate a
   standard curve of peak area versus concentration to determine drug
   loading and encapsulation efficiency in the nanoparticles. To evaluate
   acidic degradation, 1 mg of COF‐Reuterin was dispersed in 4 mL of PBS
   (pH 5.0) and incubated at 37 °C with shaking (220 rpm). Samples
   (0.2 mL) were taken at 0, 4, 8, 24, and 48 h for morphological
   observation by scanning electron microscopy.

Trained Immunity Assay for TAMs

   Tumor‐associated macrophages (TAMs) were isolated following a
   previously established protocol^[ [214]^63 ^] Briefly, B16‐F10
   tumor‐bearing mice were administered various nanoparticle formulations,
   after which tumor tissues were harvested. The tumors were finely minced
   and enzymatically digested with collagenase type IV at 37 °C for 1 h. A
   single‐cell suspension was obtained by passing the digested tissue
   through a 70‐µm cell strainer. The cells were incubated with red blood
   cell lysis buffer at 4 °C for 2 minutes, followed by washing with
   (PBS). The resuspended cells were mixed with 40% Percoll solution and
   carefully layered with 80% Percoll solution. After centrifugation at
   750 × g for 30 min, the macrophage‐enriched population at the
   interphase was collected and cultured in RPMI‐1640 medium supplemented
   with 10% fetal bovine serum (FBS) and 1% penicillin‐streptomycin.
   Following an overnight incubation, fresh medium was replenished for
   subsequent stimulation and trained immunity analysis.

Statistical Analysis

   Data are presented as means ± standard deviation (SD). All statistical
   analyses and graphical representations were conducted using GraphPad
   Prism 9.0 software. The Shapiro–Wilk test was employed to evaluate the
   normality of the data distribution.The F‐test was used to evaluate
   whether the variances between the two groups were equal, while the
   paired t‐test with unequal variances was applied to compare the means
   of the two groups. The Brown–Forsythe test was used to assess the
   homogeneity of variances across multiple groups. Depending on the data
   distribution, independent‐samples Student's t‐test, one‐way ANOVA or
   two‐way ANOVA was used to compare two or more groups. Significance was
   determined at *p < 0.05, **p < 0.01, ***p < 0.001.

Conflict of Interest

   The authors declare no conflict of interest.

Ethics declarations

   All in vivo experiments were performed in accordance with the
   institutional guidelines and regulations for animal experimentation of
   Jilin University. The protocols were approved by the Jilin University
   Animal Center's Ethics Committee (Approval No. KT201902111).

Supporting information

   Supporting Information
   [215]ADVS-12-e04784-s001.docx^ (22.8MB, docx)

Acknowledgements