The murine CT26 colon carcinoma cell line was a kind gift from Dr. Siegfried Weiss (Helmholtz Centre for Infection Research, Braunschweig, Germany). The colon adenocarcinoma MC38 cell line was provided by Prof. Jo Van Ginderachter (Vrije University Brussel, Belgium). Cells were maintained as previously described (7, 22). Manuka honey (UMF^® 20+ from ApiHealth, Auckland, New Zealand) was used in the current study and diluted in distilled water under aseptic conditions. MH is composed mainly (~76%) of a mixture of sugars (fructose, glucose, maltose, sucrose and galactose) together with a significant component of bioactive compounds, including phenolics and flavonoids (6). As a control for MH, a sugar solution, designated sugar control (SC), containing equivalent concentrations of the three major sugars in honey (38.2% fructose, 31.3% glucose, and 1.3% sucrose) was used (Sigma, St. Louis, MO, USA) (8).

BALB/c and C57BL/6 mice were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). IFNγ-deficient (IFNγ^-/-) mice were purchased from the Jackson laboratories (USA) and have been described (23). All animals were bred in the animal facility of the College of Medicine and Health Sciences, United Arab Emirates University. For the current study, male mice at the age of 8-10 weeks were used. Female mice were not used to avoid potential physiological variability associated with the estrous cycle. Mice received rodent chow and water ad libitum and were maintained 5-6 mice per cage in a standard 12 h light/12 h dark cycle at a temperature of 21°C with 40–57% humidity. All studies involving animals were carried out in accordance with, and after approval of the Animal Research Ethics Committee of the United Arab Emirates University (Protocols #A12-13 and ERA-2019-5853).

BALB/c mice of comparable age and weight were randomly divided into two groups. Mice were gavaged daily with 0.2 mL of a water solution containing 70% SC or 70% MH (w/v). After 4 weeks of treatment, mice were euthanized, and their mesenteric lymph nodes (MLNs), inguinal lymph nodes (ILNs), and spleens were excised for further analysis.

For the tumor model studies, BALB/c or C57BL/6 mice were treated with SC or MH for 4 weeks, then subcutaneously inoculated with CT26 (2×10⁵/mouse) or MC38 (1×10⁵/mouse) cells, respectively, in the right flank. Tumor dimensions (width and length) were measured using a digital caliper twice a week, and tumor volume was calculated using the formula: tumor volume = (L×W²)/2, as detailed elsewhere (24, 25). Mice were euthanized 21 days post-implantation, and their tumors were excised for further analysis. In other experiments, oral MH was administered as a therapeutic regimen by first implanting tumor cells and then starting daily MH administration from day 3 post implantation for a total period of up to 3 or 4 weeks, as indicated. For these studies, we utilized wild-type C57BL/6 and IFNγ^-/- mice.

Single cell suspensions were prepared from the spleens, MLNs, and ILNs by mechanical dissociation as previously described (26). Tumors were processed using a previously described method, with modification (27). Briefly, dissected tumor tissues were subjected to mechanical and enzymatic digestion in gentleMACS C-tubes (Miltenyi Biotec, Germany) using a tumor dissociation kit (Miltenyi Biotec) and the GentleMACS dissociator (Miltenyi Biotec), according to manufacturer’s instructions. Tumor-infiltrating leukocytes (TILs) were subsequently purified from tumor cell suspensions using magnetic CD45⁺ microbeads and the autoMACS cell separator, according to the manufacturer’s protocol (Miltenyi Biotec).

Analysis of MLN, ILN, spleen, and tumor cells was carried out using multi-color flow cytometry, following our standard protocol (22, 27). The following antibodies (all purchased from Biolegend, San Diego, CA, USA) were used in the current study: anti-CD45-APC (Cat# 103112), anti-CD19-PE (Cat# 115508), anti-CD19-PE-Texas Red (Cat# 115554), anti-CD3-BV785 (Cat# 100232), anti-CD4-FITC (Cat# 100509), anti-CD8-APC-Cy7 (Cat# 100714), anti-CD8-APC (Cat# 100712), anti-CD11b-Alexa Flour-488 (Cat# 101217), anti- CD11c-PE (Cat# 117308), anti- Ly6G-BV605 (Cat# 127639), Ly-6A/E (Sca-1)-PE-Texas Red (Cat# 108138), anti-MHC II (I-A/I-E)-BV785 (Cat# 107645), anti-MHC I H-2K^d -BV421 (Cat# 116623). Non-viable cells from tumors were excluded using 7-AAD viability dye (Biolegend) and non-viable cells from spleens, MLNs, and ILNs were excluded using Zombie Aqua dye (Biolegend). Data were collected on 10,000-50,000 cells (depending on the organ) using a FACSCelesta flow cytometer (BD Biosciences, Mountain View, CA, USA) and analyzed using FlowJo software (BD Biosciences).

Immunohistochemical staining was performed on tumor tissue sections as per established protocols in our laboratory (22, 28). Sections were incubated overnight with specific monoclonal antibodies to CD8 (ab209775; Abcam, UK), CD4 (ab183685; Abcam), or granzyme-B (44153S; Cell Signaling Technology, Danvers, MA, USA), after which they were incubated with HRP-conjugated goat polyclonal secondary antibody for 45 min at room temperature. Sections were then developed using DAB chromogen substrate (Dako, Carpinteria, CA, USA), counterstained with hematoxylin, and examined using an Olympus BX51 microscope (Olympus Corporation, Japan) at 40× magnification. The positive cells were counted in 10-20 randomly selected high-power fields (HPF), and the average count was calculated.

qRT-PCR was carried out essentially as previously detailed (22, 27). We used premade TaqMan primers and probes (Applied Biosystems, Foster City, CA, USA) for the following genes, CXCL1 (Mm04207460_m1), CXCL2 (Mm00436450_m1), CXCL10 (Mm99999072_m1), IFN-γ (Mm01168134_m1), and granzyme B (Mm00442834_m1). mRNA levels of target genes were normalized according to the comparative ΔCq method to respective mRNA levels of the housekeeping gene HPRT (Mm01545399_m1). The expression of the target gene is reported as the level of expression relative to HPRT and presented as fold change relative to control mice.

DNA was extracted from stool samples using the QIAamp Fast DNA Stool Mini Kit (Qiagen, Valencia, CA, USA), following standard protocol. DNA concentration was measured using the NanoDrop 1000 Spectrophotometer (ThermoFisher Scientific, Waltham, MA, USA).

16S Metagenomic Sequencing kit (Illumina, San Diego, CA, USA) was used for library preparation. V3–V4 hypervariable regions of the bacterial 16S rRNA gene were amplified using the primers (5′-CCTACGGGNGGCWGCAG-3′ and 5′ GACTACHVGGGTATCTAATCC-3′) provided by the manufacturer and following the recommended protocol as described before (29). Library concentration was assessed by Qubit Fluorometric Quantitation (Invitrogen, Waltham, MA, USA). Short-read paired-end amplicon sequencing was performed using Illumina^® MiSeq Instrument for 600 cycles.

Processing of sequencing reads (adaptor trimming and filtering of low-quality reads) followed by taxonomic classification were done using Quantitative Insights Into Microbial Ecology version 2 (QIIME2) software suite (30). After the identification of Operational Taxonomic Units (OTUs), downstream analyses were carried out in RStudio (v 4.1.2). Diversity was measured using BiodiversityR (v 2.15-2) and plotted by ggplot2 (v. 4.1.3). Alpha diversity measures (Observed OUT, CHAO1, Shannon’s Diversity, and Simpson’s Diversity indices) were compared between the groups using Mann-Whitney U test. For beta diversity, principal coordinate analysis based on Jaccard and Bray Curtis dissimilarity metrics was used to assess differences between the groups using non-parametric multivariate analysis of variance (PERMANOVA). Linear discriminant analysis (LDA) effect size (LEfSe) was used to detect biomarkers from microbial profiles (31) using the Microbiome Analyst 2.0 platform (McGill, Canada), which was also used to generate the graph of relative abundance and the heatmap for groups comparison (32). Venn diagrams were generated to compare the taxa exhibiting significant differences based on the LDA analysis for the identification of shared and unique OTUs (33).

All statistical analyses were performed using GraphPad Prism 9.0 (San Diego, CA, USA). Statistical significance between control and treated groups was determined using 2-way ANOVA or the unpaired, two-tailed Student’s t-test, as indicated. In all analyses, p < 0.05 was considered statistically significant * (p < 0.05), ** (p < 0.01), *** (p < 0.001) and **** (p < 0.0001).

We have previously demonstrated the ability of i.p. administration of MH to effect changes in the immune system via inducing the recruitment of neutrophils into the peritoneal cavity and the maturation of peritoneal macrophages (15). In our efforts to apply a more physiological route of administration that would be safe and more applicable to humans, we investigated the effect of repeated oral administrations of MH on the immune system of BALB/c mice. Based on our previous experience, a solution of 50-70% MH (w/v) is suitable for in vivo use in mice (7). Naïve BALB/c mice were orally gavaged with water (control group), sugar control (SC) solution, or MH for 4 weeks. To address if repeated oral doses of MH are associated with any adverse events, body weights were determined in treated animals over the 4-week period. Baseline body weights were recorded before starting the treatment (Means ± SEM = 22.97 ± 0.397, 23.63 ± 0.783, 23.54 ± 0.818 g for the water, SC and MH groups, respectively) and at weekly intervals after treatment. The percentage change in body weight from baseline was then calculated. The results indicated that, in comparison with water-treated and SC-treated groups, treatment with MH over 4 weeks did not affect normal weight gain, with all 3 experimental groups showing comparable levels of body weight gain over the treatment period ( Supplementary Figure 1A ). Accordingly, we selected SC as the control for all subsequent experiments. The potential effect of oral MH on blood glucose levels was also investigated by determining non-fasting glucose levels in blood samples of randomly-selected mice collected on a weekly basis in SC or MH-treated groups. The average random glucose blood level in untreated, age-matched, control mice is 172.0 ± 14.9 mg/dL (mean ± SD). At the end of the treatment period, we observed that MH administration did not alter the blood glucose levels, with both SC and MH groups showing comparable glucose levels that lie within the normal range (<200 mg/dL) ( Supplementary Figure 1B ). Thus, no apparent negative effects were evident following oral administration of MH.

The capacity of MH to effect changes in the immune system was next investigated. Different peripheral lymphoid tissues including gut-draining MLNs, ILNs, and spleens were collected, and their weights and absolute cell counts were recorded. Our results indicated that MH administration did not alter the weights or the total cell counts of the collected tissues in comparison to SC-treated mice ( Supplementary Figures 1C–H ).

Next, multi-color flow cytometry was utilized to analyze the cellular changes in the collected tissues following MH administration. The gating strategy employed to identify the major immune subpopulations is shown in ( Supplementary Figure 2 ). FACS data indicated that MH administration did not lead to alterations in the cellular landscape of MLNs, ILNs, or spleens between SC or MH-treated mice ( Supplementary Figure 3 ).

In the context of our previous findings demonstrating a functional maturation of macrophages that was observed following i.p. administration of MH, we next sought to investigate if similar alterations were induced following oral MH administration. Upregulation of MHC class II proteins on myeloid cells is a key event that is induced in response to activation through type I/II interferon signaling pathways (34, 35). Therefore, we investigated whether oral administration of MH can induce any alterations within the cellular landscape of the peripheral lymphoid tissues, including MLN, ILN, and spleen. Given that the majority of cells in these tissues comprise T and B lymphocytes, we focused on analyzing changes in the expression of proteins known to be induced by type I/II IFNs. One of the well-known IFN-inducible genes is Ly6a, which encodes for Sca-1 protein on T lymphocytes (36–38). The results of the flow cytometric analysis showed that oral administration of MH led to a significant increase in Sca-1 expression on both CD8⁺ and CD4⁺ T cells ( Figure 1 ). The percentage of Sca-1-positive CD8⁺ T cells in MLNs, ILNs, and spleens increased by 21.6%, 24.4%, and 24.7%, respectively in MH-treated mice in comparison to SC-treated mice ( Figures 1A, B, D, E, G, H ). Similarly, the percentage of Sca-1-positive CD4⁺ T cells increased by 36.5%, 39.8%, and 62.9% in the same three organs, respectively ( Figures 1A, C, D, F, G, I ). These results show that oral MH administration induced IFN-dependent responses in T cells, both at the level of the gastrointestinal tract as well as in systemic lymphoid organs.

The demonstration of the capacity of oral MH administration to activate T lymphocytes supports its potential role as an immunomodulatory agent. We reasoned that oral MH could potentially be used to boost immune responses preventatively in different disease settings. To test this hypothesis, we investigated the capacity of MH to modulate anti-tumor immune responses using two different syngeneic murine CRC models, namely CT26 tumors in BALB/c mice and MC38 tumors in C57BL/6 mice. The treatment protocol followed in this study is illustrated in Figure 2A . Our findings revealed that pre-treating mice with MH resulted in a significant growth retardation of both CT26 ( Figure 2B ) and MC38 ( Figure 2C ) tumors. In the CT26 model, tumors grew continuously and rapidly in mice given vehicle (H₂O) or SC solution, reaching a mean volume ± SEM of 897 ± 169 mm³ and 916 ± 114 mm³, respectively on day 21 post-implantation ( Figure 2B ). On the other hand, mice treated with MH exhibited a significant reduction in tumor volume, with a mean of 511 ± 90 mm³ on day 21 post-implantation. The suppression in tumor growth was observed as early as 9 days post-implantation, and by day 21, it reached 44% compared to SC-treated mice (p = 0.006) ( Figure 2B ). Tumor growth in individual mice of each of the three experimental groups is shown in Supplementary Figures 4A, C, E . Very similar findings were observed using the MC38 tumor model ( Figure 2C ). Pre-treatment with MH resulted in 55% suppression (p = 0.002) in the growth of MC38 tumors by day 20 post implantation ( Figure 2C ). These results highlight a potential immune-boosting anti-tumor role for MH when given preventively.

To investigate the underlying mechanism for the observed MH-mediated retardation of tumor growth, we analyzed the tumor microenvironment (TME) for alterations in the cellular landscape by flow cytometry. Tumors were excised at the end of the observation period, subjected to mechanical and enzymatic digestion, and processed to a single cell suspension. The cells were then stained with different panels of mAbs to quantify the cellular constituents within the tumor tissue. The gating strategies employed to identify the cellular subpopulations are illustrated in ( Supplementary Figure 5 ).

Tumor-infiltrating immune cells were identified by being positive for the pan-hematopoietic CD45 cell surface marker. There was a significant 36% increase in CD45⁺ immune cells in the tumors of MH-treated mice compared to the control group (28% vs. 20%, respectively) ( Figures 2D, E ). FACS analysis revealed alterations in both the phenotypic and functional characteristics of the lymphoid and myeloid subpopulations in the TME. Regarding the CD3⁺ T cell population, we observed a ~2-fold increase in the percentages of both CD8⁺ and CD4⁺ T cells ( Figures 2F–H ) following MH treatment. The increase in the infiltration of T cells was also demonstrated morphologically by immunohistochemistry, where the number of both CD8⁺ and CD4⁺ T cells was substantially increased in tumor tissue sections of MH-treated mice ( Figures 2I–L ).

Further analysis using myeloid cell-specific antibodies showed that the majority (70-80%) of the gated CD45⁺ population in the TME were CD11b⁺ myeloid cells ( Figure 3A ). Interestingly, there was a significant decrease (~18%) in the percentage of intratumoral myeloid cells in the MH-treated group compared to the SC-treated group ( Figures 3A, D ). This was largely accounted for by a 50% reduction in the percentage of Ly6G⁺ granulocytes ( Figures 3B, E ), most likely representing myeloid-derived suppressor cells (39). In contrast, the percentage of Ly6C^hi cells increased significantly (~1.7-fold) in MH-treated mice ( Figures 3B, F ). These cells have been described to be pro-inflammatory in function and are recruited to tumor tissue in response to CCL2/CCR2 signaling (40). They further differentiate into MHC class II (MHC-II) positive or negative tumor-associated macrophages (TAM) dependent on macrophage colony-stimulating factor-mediated signals (41). In terms of the other myeloid subpopulations, there was no major change in the percentages of Ly6C^lo/Neg cells ( Figures 3B, G ) and dendritic cells (CD11c⁺ cells) ( Figures 3C, H ) following MH treatment.

To gain insight into the functionality of intratumoral myeloid cells, we analyzed the level of expression of MHC-II proteins on CD11b⁺ Ly6G^- subpopulation. There was a significant increase in the percentage of myeloid cells expressing MHC-II proteins (2.4-fold) in MH-pretreated mice ( Figures 3I, J ). Furthermore, the overall level of expression of MHC-II on myeloid cells in MH-treated group tended to be slightly elevated compared to SC-treated mice, but this difference did not reach statistical significance (p = 0.083) ( Figures 3I, K, L ). These data are suggestive of an enhancement in the antigen-presentation capacity of myeloid cells within the TME of MH-treated mice.

Given the evidence of the involvement of type I/II IFN pathways in the observed functional changes in cellular function, we next analyzed whether similar alterations could be observed on the tumor cells. It is well known that the expression of MHC class I (MHC-I) proteins is regulated by type I/II IFN signaling pathways (42). Therefore, we analyzed the expression of MHC-I proteins on CD45^- tumor cells grown in mice after pretreatment with MH in comparison with tumor cells grown in control mice given SC solution. The results of this analysis showed that tumor cells grown in control mice showed bimodal levels of MHC-I expression, with 2 clearly discernible subpopulations being observed. The majority of these tumor cells (~62%) expressed low levels of MHC-I proteins, while the remaining population (~38%) showed high levels of MHC-I ( Figures 4A–C ). In sharp contrast, approximately 70% of tumor cells grown in mice pre-treated with MH exhibited high levels of MHC-I proteins ( Figures 4A, C ). Furthermore, a 7-fold increase in the MFI level of MHC-I proteins on tumor cells was observed following MH administration ( Figures 4D, E ). These results suggest that MH treatment indirectly enhanced the immunogenicity of tumor cells, rendering them more susceptible to killing by anti-tumor CD8⁺ T effector cells. Taken together, our findings indicate that the ability of MH to effect changes in tumor growth is linked to a series of immunomodulatory alterations within the TME.

To elucidate the mechanism by which MH modulates the cellular components of the TME and exerts the observed anti-tumor response, RNA was extracted from purified, intratumoral, CD45⁺ cells, or whole tumor tissue, of SC-treated or MH-treated mice. The RNA was then used to determine the gene expression levels of key inflammatory chemokines and cytotoxic effector molecules by qRT-PCR. At the level of tumor-infiltrating leukocytes, MH treatment led to a small (1.6-fold) but insignificant increase in the expression level of the chemokine CXCL10 ( Figure 5A ). At the whole tumor level, there was also a 2.3-fold increase in the expression of CXCL10 in MH-treated mice ( Figure 5F ; p = 0.0131). CXCL10 is secreted in response to IFN-γ and preferentially regulates the recruitment of inflammatory T lymphocytes (43). The expression levels of CXCL2 and CXCL1 chemokines, which are potent chemoattractants that control the recruitment of polymorphonuclear leukocytes in inflammation and tissue injury (44), were also examined. qRT-PCR results indicated a statistically significant 1.6-fold decline in the expression level of CXCL2 in MH-treated mice ( Figure 5B ). A trend toward a decrease in the expression levels of CXCL1 was also observed, but this did not reach statistical significance ( Figure 5C ). These findings may underlie the observed reduction in the proportion of intratumoral Ly6G⁺ granulocytes following MH treatment.

MH treatment also resulted in a significant increase in the expression levels of IFN-γ (~1.7-fold) and granzyme B (~2.8-fold), as detectable at the level of TILs ( Figures 5D, E ). Both of these mediators are secreted by effector immune cells to induce tumor cell apoptosis (45, 46). A significant increase (2.6-fold) in the expression of IFN-γ (p = 0.0148) and granzyme B (p = 0.0261) was also observed at the level of the whole tumor tissue ( Figures 5G, H ). Consistent with these findings, immunohistochemical staining of tumor tissues indicated an increase in the number of granzyme B-positive cells following MH treatment ( Figures 5I, J ), reflecting the presence of activated, anti-tumor, cytotoxic lymphocytes, most likely T cells and/or NK cells.

Having demonstrated the capacity of oral MH as a preventive treatment against cancer growth, we assessed its potency in a therapeutic model. Mice were implanted with MC38 CRC cells and oral MH was subsequently administered on a daily basis starting on day 3 post implantation. The data affirm that daily MH administration leads to a significant inhibition of tumor growth in normal, immunocompetent, mice ( Figure 6A ). The effect of MH on tumor growth was first apparent at about 10 days after the initiation of treatment. In sharp contrast, MH-induced curtailment of tumor growth was completely abrogated in IFNγ-deficient mice ( Figure 6B ), demonstrating mechanistically that MH most likely exerts its immunomodulatory effect via the activation of the IFNγ pathway.

We hypothesized that oral administration of MH could induce changes in gut microbiota that would underlie the enhanced anti-tumor immune responses observed in these mice. To address this possibility, we determined the composition of gut microbiota in fecal samples collected from mice either before treatment or after 4 weeks of treatment with MH or SC solution. Microbiota were profiled at the genus level to detect the alterations in abundance and diversity caused by each treatment. The stacked area plot ( Figure 7 ) shows the relative abundances of genera ranked based on their prevalence in the samples (listed below in the graph) collected from mice before and after treatment with SC or MH. Variations were obvious among the samples. The microbiota fingerprint in the control group was maintained between week 0 and week 4. As for the MH group, microbiota profiles looked more homogenous after treatment and with more similarity compared to the variability seen in week 0. To identify the genera with significant differences before and after each treatment, linear discriminant analysis (LDA) was done. As shown in Figures 8A, B , treatment with either SC or MH caused significant changes in microbiota profiles, with depletion of some genera (red color in the graphs) and enrichment of others (blue color in the graphs). These findings confirm that the microbiota were changed after 4 weeks of either treatment. It is noteworthy that treatment with SC caused depletion of Lactobacillus which is generally considered a beneficial bacteria. In sharp contrast, MH treatment caused depletion of pathogenic bacteria namely, Staphylococcus, Enterococcus, and Bacteroides.

Next, we investigated if microbiota alteration induced by treatment was similar in MH versus SC groups. Significantly changed microbiota identified by LDA analysis in MH and SC groups (from Figures 8A, B ) were compared, and Venn diagram ( Figure 8C ) was used to identify unique and shared genera. Most of the genera altered in response to treatment with MH and SC were unique for each group. Only two genera were shared between the two groups, namely Methylarcula (more in MH in week 4, and more in SC in week 0, i.e. enriched after treatment with MH and depleted after treatment with SC), and Geovibrio (less in week 4 in both groups, i.e. reduced due to treatment with both SC and MH). Moreover, we compared the genera detected in MH and SC groups after 4 weeks of treatment to explore microbiota differences between these groups post-treatment. The heatmap also shows the genera with significant difference between week 0 and week 4 after each treatment (shown in Figures 8A, B ) and reveals the distinct microbiota fingerprints per group. As shown in Figure 8D , significant differences were found in the genera after 4 weeks of treatment with SC or MH (marked with asterisk in the heatmap). For example, MH induced the depletion of several genera, including Bacteroides, Staphylococcus and Enterococcus, which have been described to have pro-tumorigenic activities (47–51). In contrast, two of the microbiota genera enriched after MH treatment were Ruminococcus and Clostridium cluster IV, both of which encompassing members that have been recognized for their anti-tumorigenic potential (52, 53). The LDA analysis of the significantly different genera is shown in ( Supplementary Figure 6 ). The relative abundance of each significantly different genus among the groups is shown in ( Supplementary Figure 7 ).

Additionally, we have explored the effect of MH and SC treatment on microbiota diversity. The difference was not statistically significant in alpha and beta diversity ( Figures 9A–E ), but the only exception was seen in Shannon’s index, which is a widely used alpha diversity metric. The latter index is the negative sum of proportional microbiota abundance multiplied by the log of its proportional abundance (54). It is generally useful in predicting diversity by reflecting both richness (the number of microbiota) and evenness (the uniformity of distribution of microbiota) within a community (55). Shannon’s index was significantly higher after treatment with MH for 4 weeks compared to the baseline in week 0 ( Figure 9A ), suggesting an increase in the richness and evenness of microbiota after treatment with MH. This effect was not evident after treatment with SC. Nevertheless, pairwise comparison between SC and MH in week 0 and week 4 did not reveal any significant difference between these groups. Altogether, our findings demonstrate that MH treatment led to distinct changes in microbiota composition that are significantly different from the effect of SC, with identification of key microbiota that were increased or decreased following treatment.