Experiments were performed using 18–19-day old Hy-Line Brown embryos (Gallus gallus) obtained from the National Avian Research Facility, Edinburgh, UK. Embryos were humanely culled under the authority of UK Home Office Project Licences (PE263A4FA) following the guidelines and regulations of the UK Home Office ‘Animals (Scientific Procedures) Act’ 1986. The small intestine (duodenum, jejunum and ileum) was removed, cut open longitudinally then into 3 mm sections and collected in Mg²⁺ and Ca²⁺ free phosphate buffer saline (PBS). For each independent batch, the intestines from five embryos were pooled and a total of eight independent batches were tested. The villi were released from the tissue as previously described (7). In brief, the tissues were digested with Clostridium histolyticum type IA collagenase (0.2 mg/mL, Merck, Gillingham, UK) at 37°C for 50 min with agitation at 200 rpm. Single cells were removed by filtering the digestion solution through a 70 µM cell strainer (Corning, Loughborough, UK). The villi were collected by washing the inverted strainer. Villi were collected and pelleted at 100 g for 4 min. The villi were resuspended in Floating Organoid Media (FOM media; Advanced DMEM/F12 supplemented with 1X B27 Plus, 10 mM HEPES, 2 mM L-Glutamine and 50 U/mL Penicillin/Streptomycin; ThermoFisher Scientific, Paisley UK (TFS)) and seeded at ~3000 organoids/well in 6 well plates and incubated overnight at 37°C, 5% CO₂.

Salmonella enterica serovar Typhimurium strain 4/74 (STm) was engineered previously to constitutively express GFP by transformation with a derivative of pFVP25·1 (17). Bacteria were cultured in Luria-Bertani broth supplemented with 50 µg/mL of ampicillin (Merck) and 20 µg/mL of naladixic acid (TFS) and incubated for 18 h at 37°C with shaking at 180 rpm. Bacteria were grown to the optical density of 1 at 600 nm and pelleted at 3220 g for 10 min at 4°C. Bacteria were washed twice with PBS and resuspended in 10 mL of antibiotic-free FOM. Tenfold serial dilutions were plated on naladixic acid containing LB agar in duplicate and incubated at 37°C overnight to determine colony forming units (CFU) and expression of GFP was checked under blue light. A titration of the challenge dose was performed using 1000, 500, 250 and 125 CFU of STm per organoid. Infection of organoids for 3 h with 500 CFU or more of STm provided the most reproducible CFU measurements based on CFU counts from homogenized organoids (data not shown). Therefore, 500 CFU of STm was used as the challenge dose in this study.

A proprietary blend of organic acids and essential oils (OA+EO) (Jefo Nutrition Inc. Canada) was prepared at a concentration of 1 mg/mL in antibiotic-free FOM and dissolved at 37°C with regular inversion. After 24 h in culture the organoids were collected by centrifugation at 100 g for 4 min and resuspended in antibiotic-free FOM media or antibiotic-free FOM media supplemented with 0.5 mg/mL (high dose) or 0.25 mg/mL (low dose) of OA+EO in 6 well plates at 37°C, 5% CO₂. After 24 h, the untreated and the OA+EO treated organoids were collected by centrifugation at 100 g for 4 min and reseeded at 200 organoids per well on 24 well plates in a final volume of 350 µL of antibiotic-free FOM media or antibiotic-free FOM media supplemented with 0.5 mg/mL or 0.25 mg/mL of OA+EO for 24 h at 37°C, 5% CO₂. The treated organoids were exposed to OA+EO a total of 48 h before inoculation with STm.

STm was resuspended with antibiotic-free FOM media supplemented with high (0.5 mg/mL) or low (0.25 mg/mL) OA+EO and incubated at 4°C for 24 h before inoculation of the organoids.

For bacterial invasion assays, control organoids were infected with 500 CFU/organoid of STm not pre-treated with OA+EO or remained uninfected. Organoids treated for 48 h with OA+EO were infected with 500 CFU/organoid of OA+EO treated STm. To encourage bacterial:organoid interaction, plates were centrifuged for 5 min at 10 g and subsequently incubated at 37°C, 5% CO₂ for 3 h. For each batch of organoids, 3 wells on a 24 well plate were prepared for bacterial enumeration, 3 wells for RNA isolation, and 3 wells for protein isolation. The experiments were repeated twice with 4 independent batches of organoids in each experiment (total N=8 independent batches).

To quantify the number of bacteria that invaded the untreated and OA+OE treated organoids, organoids were treated with gentamycin (Gibco, 50 μg/mL) for 30 min at 37°C. Three wells were pooled, and centrifuged at 400 g for 4 min. Organoids were washed twice in PBS and homogenized in 300 μL PBS using steel beads in a Tissue-Lyser at 25 Hz for 1 min. Ten-fold serial dilutions were plated on naladixic acid (20 µg/mL) containing LB agar in duplicate and incubated at 37°C overnight to enumerate intracellular bacteria. All input STm inocula were plated out on the day of infection to measure the direct effect of OA+EO on Salmonella invasion. The mean and standard deviation were calculated and significant differences between groups were determined by the nonparametric Mann-Whitney U test, with a P-value of <0.05 considered statistically significant.

Organoids were collected from 3 wells of each treatment group and centrifuged at 2300 g for 4 min. The pellets were lysed in 350 µL RLT Plus buffer with 2β-mercaptoethanol. Total RNA was extracted using a RNeasy Plus Mini Kit (Qiagen) consisting of a genomic DNA column eliminator according to the manufacturer’s instructions and quantified spectrophotometrically. Twenty-five ng/µl of RNA was reversely transcribed using a High-Capacity Reverse Transcription kit (Life Technologies, Paisley, UK), according to the manufacturer’s instructions, with a random hexamer primer and oligo(dT). The cDNA was stored at −20°C until future use.

Pre-amplification of cDNA was performed as previously described (18) and unincorporated primers were digested from the pre-amplification samples using 16 U/μl Exonuclease I (E. coli, New England Biolabs) at 37°C for 30 min. High-throughput qPCR was performed using 96x96 Integrated Fluid Circuits (IFC) arrays (Standard BioTools) as previously described (19). Samples were amplified in duplicate reactions using primers for 88 genes of interest and seven reference genes. Quantitative PCR was performed on the BioMark HD instrument (Fluidigm) using the thermal cycling conditions as previously described (19). The fluorescence emission was recorded after each cycling step. Raw qPCR data quality threshold was set to 0.65-baseline correction to linear (derivative) and quantitation cycle (Cq) threshold method to auto (global) using the Real-Time PCR Analysis software 3.1.3 (Standard BioTools).

The raw Cq values were processed with GenEx.v6 MultiD Analyses AB, with correction for primer efficiency. IL1R2 gene was removed from the analysis due to missing over 50% of the data. The stability of the expression of seven putative reference genes - TATA box binding protein (TBP), Tubulin alpha chain (TUBA8B), beta-actin (ACTB), beta-glucuronidase (GUSB), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), Beta-2-Microglobulin (B2M) and ribosomal 28S (r28S) - was evaluated via the NormFinder tool in GenEx. The geometric mean of the most stable genes (ACTB, B2M, TBP) was used to normalize all samples. Technical replicates were averaged, and the relative quantification values were assessed to the maximum Cq value obtained per gene, transformed to the logarithmic scale.

Statistical analysis of the gene expression of organoids treated with OA+EO and STm from the IFC array was conducted to identify significantly differentially expressed genes (DEGs) between stimulated groups and was performed using GenEx6, with group means compared with two-way t-tests adjusted for multiple comparisons with post hoc Bonferroni correction, with significant DEGs having a fold change >1.5 and <−1.5, illustrated in heat maps, and the shared or unique genes annotated in Venn diagrams. For all statistical analyses, P-values < 0.05 were considered significant. All statistical analyses were conducted using GraphPad Prism 9 or GenEx v6.

Organoids were centrifuged at 2300 g for 4 min and the pellet was snap-frozen in liquid nitrogen and stored at -80°C. Samples were shipped on dry ice to the University of Delaware, for kinome peptide array analysis.

The kinome peptide array was performed as described by (20). Forty mg of samples were lysed using bead-based homogenization in 100 μL of lysis buffer containing protease and phosphatase inhibitors. The lysed organoids were incubated pelleted at 14,000 g for 10 min at 4°C. An aliquot of supernatant was mixed with 10 μL of activation mix containing ATP as the phosphate group donor. Eighty μL of the supernatant-activation solution was applied to the peptide microarray. The custom-designed peptide arrays were obtained from JPT Peptide Technologies (Berlin, Germany), based on in-house sequence designs. A 25 × 60 mm, glass lifter slip was then applied to the microarray to sandwich and disperse the applied lysate.

The microarrays were incubated in a humidity chamber at 40°C and 5% CO₂. Arrays were placed in a 50 mL centrifuge tube containing PBS-1% Triton, to remove the lifter slip from the microarray surface, and submerged in 2M NaCl-1% Triton and agitated for a minimum of 30 s. This process was repeated with fresh 2M NaCl-1% Triton and arrays were washed in double distilled water with agitation.

Array slides were submerged in phosphospecific fluorescent ProQ Diamond Phosphoprotein Stain (Life Technologies, Carlsbad, CA) on a shaker table at 50 rpm for 1 h and destained twice for 10 min with agitation at 50 rpm in 20% acetonitrile (EMD Millipore Chemicals, Billerica, MA) and 50 mM sodium acetate (Sigma Aldrich, St. Louis, MO). The arrays were then washed with double distilled water, spun dried, and scanned using a Tecan PowerScanner microarray scanner (Tecan Systems, San Jose, CA) at 532 to 560 nm with a 580 nm filter to detect dye fluorescence.

Images were gridded using GenePix Pro software, and the spot intensity signal was collected as the mean of pixel intensity using local feature background intensity calculation with the default scanner saturation level. The resultant data was analyzed by the PIIKA2 peptide array analysis software (http://saphire.usask.ca/saphire/piika/index.html) (21). Briefly, the resulting data points were normalized to eliminate variance due to technical variation, for example, random variation in staining intensity between arrays or between array blocks within an array. Variance stabilization normalization was performed. Using the normalized data set comparisons between treatment and control groups were performed, calculating fold change and a significance P-value. The P-value was calculated by conducting a one-sided paired t-test between treatment and control values for a given peptide. The resultant fold change and significance values were then used to generate higher order analysis (heat maps, hierarchical clustering, principal component analysis, pathway analysis, etc.).

As described by Perry et al. (22), post PIIKA2 analysis was performed using the following online databases and tools; STRING database (23) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and KEGG color and search pathways (24), PhosphoSitePlus (25), Uniprot (26), and Venny 2.1 (27).

To investigate the effect of OA+EO on Salmonella enterica serovar Typhimurium (STm) invasion and its effects on the innate immune responses, we mimicked the in vivo circumstances by pre-treating the organoids for 48 h and STm with OA+EO for 24 h prior to infection similar to expose in the intestinal tract. The treatment of organoids with low (0.25 mg/mL) or high (0.5 mg/mL) OA+EO did not alter the morphology of the organoids compared to the untreated organoids ( Supplementary Figure 1 ). The effect of OA+EO on the viability of STm has been described previously (14) and in our study, the OA+EO treatment of STm for 24 h at 4°C also reduced the viability, on average by 28%, compared to storage in a floating organoid medium (FOM) for 24 h at 4°C (data not shown).

The number of live bacteria that invaded the organoids was calculated as relative to the number of inoculated bacteria within each independent experiment, i.e. the number of invaded bacteria compared to its input inoculum. The treatment of organoids with 0.5 or 0.25 mg/mL OA+EO resulted in a significantly lower number of invaded bacteria ( Figure 1 ) compared to the untreated organoids, suggesting that OA+EO has a minor effect on the bacteria while having a major effect on the intestinal organoids.

To determine the effects the OA+EO on the immune responses of organoids, a high throughput qPCR array was used to analyze the mRNA expression levels of 88 innate immune genes. The organoids were treated for 48 h with 0.5 or 0.25 mg/mL OA+EO and compared to untreated (0 mg/mL) organoids. The number of significantly differentially expressed genes (DEGs) with a fold change (FC) ≥1.5 at p<0.05, was higher after treatment with 0.5 mg/mL compared to 0.25 mg/mL, 24 and 7 respectively, with 5 genes in common ( Figure 2A ).

The higher dose of OA+EO upregulated 16 genes and downregulated 10 genes, whereas the lower dose of OA+EO upregulated one gene and downregulated six genes compared to untreated samples ( Figure 2B ). Five genes were regulated by both concentrations of OA+EO of which one gene is upregulated, the chemokine ligand CXCLi2 which is an orthologue of human CXCL8. GLUL (Glutamate-Ammonia Ligase), LYG2, and UPP1 were all downregulated. UPP1 or Uridine Phosphorylase 1 in human studies was profoundly associated with immune and inflammatory response and correlated with MHC-II and LCK, and is expressed in macrophages and distal enterocytes (28). Lysozyme G like 2 (LYG2) is an antibacterial peptide expressed by heterophils (29). The C-type lectin superfamily member CD72 was affected in a dose dependent manner, with a higher concentration of OA+EO upregulated CD72 whereas the lower dose downregulated CD72 albeit at a low level (FC -1.8). CD72 is involved in B cell activation and signaling and the cytoplasmic domain contains two immunoreceptor tyrosine-based inhibitory motifs (ITIM1 and 2; 30). In mice, CD72 was shown to be an inhibitory receptor on NK cells regulating cytokine production (31).

Next, we analysed the FC levels of the genes specifically regulated by the different doses of OA+EO. Treatment with the higher dose of 0.5 mg/mL resulted in the upregulation of three genes with an FC >3, EDN1, DTX2 and IL10RA. Although Endothelin 1 (EDN1) is associated with endothelial cells and vasoconstriction, but recent human single-cell analysis showed high RNA expression in enteroendocrine cells and enterocytes (The Human Protein Atlas, version 23.0). DTX2 encodes the enzyme Deltex E3 ubiquitin ligase 2 and regulates Notch signaling, a signaling pathway involved in cell-cell communications that regulate a broad spectrum of cell-fate determinations, controlling the homeostasis of occludin and ER stress (32). Only two genes, CCL20 (FC -1.5) and IL12β (FC -2.9), were specially downregulated by a low dose of OA+EO treated enteroids ( Figure 2D ). In summary, treatment of intestinal organoids with a high dose of OA+EO resulted in the activation of innate immune responses, whereas the low dose of OA+EO mostly downregulated genes, while only CXCLi2 was slightly upregulated by both concentrations of OA+EO.

To determine the effects of OA+EO on altering the innate-immune responses to STm infection in chicken 3D enteroids, we first determined the induction of inflammatory responses to STm without treatment. The infection of organoids with STm resulted in a strong upregulation of (pro)-inflammatory genes at 3 hpi ( Figure 3 ). Especially the cytokine (IL1B, IL6) and chemokine genes (CCLi2, CXCLi2, CCL20) were upregulated with a fold change of 14–79 compared to uninfected controls. A molecular pathway promoting cell activation is the nuclear factor-κB (NF-κB) signaling pathway and STm upregulated several signaling genes such NFKB2, NFKBIZ and TNFAIP3 a gene involved in tightly regulating NF-κB activity (33).

In contrast, when the organoids were treated with OA+EO and then infected with STm the inflammatory responses were prevented ( Figure 4 ). A total of 50 significant DEGs were found after OA+EO treatment and STm challenge, 15 DEGs were regulated by the lower dose and 13 DEGs by the higher dose while 22 DEGs were in common between the high and low doses ( Figure 4 ). However, compared to the expression in the STm challenge control group, all DEGs were downregulated after OA+EO treatment of the organoids, with the exception of PPARG in the high dose group.

Another striking difference between treatment and bacterial challenge versus bacterial challenge only was the level of expression of DEGs. After STm infection 19 DEGs had a fold change of >4 or <4 and a maximum fold change of 78. In contrast, the OA+EO treated and challenged organoids had 7 DEGs with a fold change of >4 or <4 and a maximum FC of -9 ( Figures 4B, C ). In conclusion, treatment of organoids with OA+EO downregulated the inflammatory responses induced by STm infection aiming to balance homeostatic status.

The kinome array was used to study immune signal transduction pathways occurring in 3D organoids after treatment with high or low dose OA+EO and STm challenge. All the samples were compared with the uninfected and untreated control (0 mg/mL) using the Platform for Integrated, Intelligent Kinome Analysis 2 (PIIKA2) online software (21). The heatmap of phosphorylation changes and experimental group clustering data are shown in Figure 5 . We observed that high dose OA+EO+STm shows tighter clustering with the low dose OA+EO while these groups do not cluster with high dose OA+EO compared to the untreated control. Low dose OA+EO+STm shows tighter clustering with low dose OA+EO and does not cluster with the high dose OA+EO. These results show that the addition of OA+EO has a significant effect on the signaling of the organoids in the context of the Salmonella challenge, especially at the high dose.

Similar to the gene expression data, we compared the proteins that were significantly altered by the high dose and low dose OA+EO compared to the control ( Figure 6A ). While the gene expression data showed that most of the changes that occurred were unique to the high dose ( Figure 2A ), the kinome data showed that most of the changes at the protein phosphorylation level (75%) occurred uniquely in the low dose OA+EO group ( Figure 6A ). Most of the changes detected by the peptide array showed a relative decrease in phosphorylation of the proteins unique to the low dose OA+EO while in those proteins unique to the high dose OA+EO, the changes were an increase in phosphorylation ( Supplementary Table ). These results are similar to the gene expression data, where the high dose induced increased gene expression and the low dose mostly reduced gene expression ( Figure 2 ). When we considered the Salmonella inoculated groups treated with high or low dose OA+EO the majority of the differential protein phosphorylation was shared between the two groups at 62% ( Figure 6B ). The trend in the OA+EO+STm groups was that much of the phosphorylation change relative to control was decreased phosphorylation ( Supplementary Table ).

The pathway overrepresentation analysis of the data generated a list of pathways for each experimental group relative to the control. For all groups, the PI3K-Akt pathway was highly represented in the data ( Tables 1 – 5 ). The PI3K-Akt pathway is a central signaling pathway that leads to a number of immune and metabolic responses. Though this pathway is overrepresented in all groups, there are distinct differences in the number of peptides displaying differential phosphorylation and the direction of that phosphorylation (increased or decreased relative to control) ( Supplementary Table ). The Salmonella challenge group showed a relatively small number of phosphorylation changes in PI3K-Akt thought many of those were increased phosphorylation, among those several proinflammatory and proliferative response proteins including, Raptor, IRS1, HSP90, FGFR2, ERK, Raf1 ( Supplementary Table ). Interestingly, the high dose OA+EO treated group showed a similar profile to the infection group, with a limited number of phosphorylation changes but among those pro-inflammatory and proliferative (for example NFkB, mTOR, IRS1, HSP90). This contrasts with the low dose OA+EO treated group, which showed a larger number of changes in the PI3K-Akt pathway members, but these were predominantly decreased in phosphorylation, especially amongst the proinflammatory proteins ( Supplementary Table ). Both low and high dose OA+EO+STm showed a relatively large number of changes in the PI3K-Akt pathway, again predominantly as decreased phosphorylation. Therefore, while there was a split in the number of peptides affected and directionally of change between the low and high dose OA+EO, the effects of these two doses merged when in the context of Salmonella.