Fig 2: GPR120 expression preserves inner mucus layer integrity. (a) The expression levels of Muc2 gene was quantified by qRT-PCR in healthy GPR120ΔIEC(n = 3) and WT (n = 3) mice. (b) Upper panel shows a representative band of Muc2 obtained by western blot on GPR120ΔIEC and WT mucus samples. The relative density of these bands is reported in the lower panel (n = 3). (c) Mucus layer samples from healthy GPR120ΔIEC (n = 3) and WT (n = 3) mice were scraped off the colon and subjected to two-dimensional electrophoresis. Gels from transgenic and WT mucus were overlapped in a high master gel (HMG) by software to identify proteins differentially expressed in the two groups. Red spots belong to WT samples, blue spots belong to GPR120ΔIEC samples, and green spots are in common between two groups. (d) Scatter plot showing all proteins detected by mass-spectrometry and differentially expressed (in red) between GPR120ΔIEC(TG) and WT mice, according to DAVE score + 2/-2. Among dysregulated proteins, Muc2 was significantly downregulated in GPR120ΔIEC mucus samples. (e) A metagenomic analysis was performed on fecal samples of healthy WT (n = 5) and GPR120 ΔIEC (n = 6) mice. Plot showing alpha diversity (Shannon index) for microbial communities in the two groups, upon16S rRNA gene sequencing analysis. (f) Principal component analysis (PCA) plots showing Weighted (right panel) and Unweighted (left panel) beta diversity in fecal samples from the two groups. (g) DESeq2 analysis showing operational taxonomical units (OTUs) in GPR120 ΔIEC versus WT fecal samples. Each spot represents a taxonomical unit; red spots represent taxonomical units with statistically significant differences between the two groups. (h) OTUs distinguishing WT and GPR120 ΔIEC mice determined using the DESeq2 negative binomial distribution method on the 16S rRNA gene profiling data of fecal samples. The taxonomic lineage of each taxon is shown: p, phylum; c, class; o, order; f, family; g, genus; s, species. The black-yellow heatmap represents the mean normalized relative abundances of the reported OTUs. Positive fold changes (shown on a red background) designate OTU overrepresentation in GPR120 ΔIEC mice (KO); negative fold changes (shown on a blue background) designate the OTU overrepresentation in WT mice. padj, adjusted p values were represented in a heatmap. (i) Bar plots showing quantification of SCFAs, including acetate, butyrate, propionate and valerate in fecal samples of healthy GPR120ΔIEC (n = 10) and WT (n = 9) littermates. Values are expressed as median ± 95% CI. *p.value < 0.05 by Mann Whitney test.

Fig 3: GPR120 expression in intestinal epithelial cells. (a, b) Representative immunofluorescence images from the human (a) and mouse (b) frozen colon sections stained with antibodies against GPR120, cytokeratins (Pan CK) and junctional adhesion molecule A (Jam A). In the panel (a), arrows indicate the basolateral side of the epithelium, while the star (asterisk) is in the lumen and is delimited by dashed line. Magnification: 40X. Scale bar: 50 µm. (c) Reanalysis of single-cell RNA-seq data from human epithelial cells (GEO ID GSE125970). Cell dispersion within the UMAP multidimensional scaling space, colored by cluster or by FFAR4 normalized expression, and labeled by cell type. FFAR4 is clearly expressed by Goblet cells and a subset of progenitors. (d) Representative immunofluorescence images from frozen colon tissues of transgenic (GPR120?IEC) mice stained with antibodies against GPR120 and Jam A. Magnification: 20x. Scale bar: 50 µm. (e) Intestinal permeability was evaluated in healthy GPR120 ?IEC mice (n = 4) and WT (n = 4) by perfusion in the mouse intestine with Evans Blue dye for 60 min. The amount of dye eluted was quantified using a spectrophotometer at a wavelength of 620 nm and results are expressed as OD per gram of colon tissue. (f) Frozen colon sections from healthy GPR120 ?IEC mice (n = 3) and WT (n = 3) littermates were stained with FISH using a bacterial probe (red) and an anti-Muc2 antibody (green). The inner mucus layer is indicated with white dashed line. Magnification 40x. Scale bar: 50 µm. (g-h) TEER was measured in siRNA-treated Caco-2 and LoVo cells and compared with scramble siRNA controls. Values are expressed as median ± 95% CI. *p value < 0,05; **p.value < 0,001 by Mann Whitney test.

Fig 4: GPR 120 protein induces Wnt/ß-catenin signaling. (a) Caco-2 (left panel) and LoVo (right panel) cells were transfected with siRNA against FFAR4 and scramble siRNA controls, after 48 h from transfection proliferation was assessed by using the Burker chamber and expressed as absolute cell number. (b) The cell cycle distribution of siRNA-treated Caco-2 and LoVo cells was performed by FACS 48 h after siRNA transfection, and upon DAPI staining. Quantifications of Caco-2 and LoVo cells labeled with DAPI in each phase of the cell cycle are shown. Values are median ± 95% CI of biological triplicates. (c) Quantification of CCND1 gene transcripts by qRT-PCR in Caco-2 and LoVo cells treated with either siRNA against FFAR4 or scramble siRNA. (d) Intracellular calcium was measured 48 h after siRNA transfection, by using a colorimetric kit and results are expressed as mM of calcium concentration. Biological triplicates. (e) The expression of ß-catenin gene transcript (Ctnnb1) was extrapolated from our RNA-seq analysis performed on the epithelium of healthy GPR120 ?IEC (n = 3) and WT (n = 3) mice, and expressed as log2 RPKM. (f) Ctnnb1expression levels were quantified by qRT-PCR in AOM/DSS-induced GPR120 ?IEC (n = 7) and WT (n = 7) mice. (g,h) Paraffin-embedded colon sections from GPR120 ?IEC (n = 7) and WT (n = 7) mice were immunostained using an anti-ß-catenin antibody. Magnification: 20X. Scale bar: 100 µm. Representative immunostainings are shown in (g), with relative quantification of immune-positive areas in (h). Values are expressed as median percentage of positive over total area ± 95% CI. (i) Quantification of CTNNB1 gene transcripts by qRT-PCR in Caco-2 and LoVo cells treated with either siRNA against FFAR4 or scramble siRNA. Values are expressed as median ± 95% CI of biological triplicates. *p.value < 0.05; **p.value < 0.001; ****p.value < 0.00005 by Mann Whitney test.

Fig 5: The loss of epithelial GPR120 impacts on fecal lipids metabolites. (a,b) ω-3 PUFAs (a) and DHA-derived metabolites (b) were quantified through LC–MS/MS analysis and expressed as area ratio/feces concentration. Values are expressed as median ± 95% CI. *p.value < 0.05; **p.value < 0.001; ****p.value < 0.00005 by Mann Whitney test.