Four similar flavonoids, one colitis model
The paper asks whether closely related dietary flavonoids differ in protecting the brain when the gut is inflamed. The four compounds share a flavone backbone but differ in hydroxylation: quercetin carries five hydroxyl groups and apigenin three; luteolin and quercetin have a 3',4'-catechol group on the B-ring, whereas apigenin and kaempferol carry a single 4'-hydroxyl group; and only quercetin and kaempferol have a hydroxyl at the C-ring 3-position. Those structural differences are the paper's starting point for expecting different effects.
The design was a six-group mouse experiment (n = 6 per group): a normal control, a DSS model group, and four flavonoid groups. Mice received 2.8% (w/v) dextran sodium sulfate in drinking water for one week to induce colitis, and flavonoid groups were gavaged with 0.2 mL of the corresponding compound at 200 mg kg−1 daily across the two-week protocol, while control and DSS mice received vehicle. Outcomes spanned colitis phenotype, two behavioral tests, glial and synaptic markers in the hippocampus, brain oxidative and inflammatory markers, gut microbiota by 16S rRNA sequencing, and intestinal barrier measures.
Because every comparison is against the DSS group in one mouse strain and one induction protocol, the results describe differential potency within this model, not a ranking that transfers directly to people with IBD.
Kaempferol and quercetin lead the memory rescue
Colitis was established as expected: DSS mice lost weight from day 3 and averaged a 21.07% body-weight decrease by the end, and their disease activity index reached 7.18 on day 7. All flavonoid groups significantly ameliorated weight loss (p < 0.001), lowered diarrhea and fecal occult blood scores (p < 0.05), and significantly restored colon length (p < 0.05), with the spleen index partially alleviated.
Cognitive impairment followed. In the novel object recognition test, DSS mice spent more time with the familiar object and less with the novel one than controls (p < 0.001), and flavonoid treatment increased novel-object preference (p < 0.01). In the Y-maze, spontaneous alternation fell from 66.01% in healthy mice to 43.58% in DSS mice, and oral flavonoids counteracted the decline, especially kaempferol and quercetin (p < 0.001); novel-arm entries confirmed the same direction (p < 0.01).
The two tests measure different things — object-recognition preference versus spontaneous alternation as a short-term and spatial working memory readout — so the shared direction of effect is more informative than either percentage alone. The paper's ranking of kaempferol and quercetin as most effective rests on these behavioral comparisons within the DSS model.
Glial and synaptic effects split by compound
The paper used IBA1 as a marker whose increased expression typically indicates microglial activation, and GFAP as an astrocyte marker, then added Sholl analysis to describe cell branching. In DSS mice, IBA1+ and GFAP+ cell numbers in the hippocampus rose, and microglial intersection counts, ramification index and ending radius fell (p < 0.05); all four flavonoids mitigated these microglial changes, meaning all four inhibited the activation pattern.
Astrocytes behaved differently. Up-regulated intersection counts, ramification index and ending radius indicated astrocyte overactivation in colitis mice, and only apigenin significantly reversed this phenotype (p < 0.05). This is the clearest cell-type-specific split in the paper: microglial suppression was shared, astrocytic reversal was not.
Neuronal structure also changed. NeuN-positive area in hippocampal CA1 and CA3 fell in DSS mice (p < 0.01) and was largely reversed by flavonoid treatment, while presynaptic synaptophysin and postsynaptic PSD-95 densities dropped (p < 0.05) and colocalized puncta confirmed synaptic depletion (p < 0.001). Apigenin and quercetin were particularly effective at restoring synaptic puncta (p < 0.001).
Barrier and microbiota effects diverge by compound
Intestinal barrier readouts separated the compounds more sharply than behavior did. Apigenin and luteolin upregulated tight junction proteins, and quercetin reduced intestinal permeability, reflected in notably decreased FITC-dextran flux (p < 0.05). Histology showed the contrast plainly: control mice had a mean damage score of 0.67 with intact wall layers and preserved crypt architecture, DSS mice scored 9.83 with crypt loss and extensive inflammatory infiltration, and apigenin, luteolin and quercetin significantly mitigated these changes.
Microbiota composition was reshaped rather than simply normalized. At the phylum level, Bacteroidota and Bacillota dominated all groups at over 75% of the total; DSS increased Bacteroidota, Verrucomicrobiota and Proteobacteria and decreased Bacillota. Alpha diversity improved in Shannon and InvSimpson indices without a corresponding increase in observed species, which the authors interpret as restructuring and improved evenness rather than expansion of species number. Kaempferol was most effective at restoring diversity (Shannon index, p < 0.001), and quercetin and kaempferol enriched Bifidobacterium and Lactobacillus.
The paper is explicit that the flavonoid increase in beneficial taxa such as Lactobacillus and Bifidobacterium was not statistically significant, and that links between microbiota, barrier and cognition come from correlation analyses, including redundancy analysis. Those associations are hypotheses about mechanism, not demonstrations of cause.
Quercetin carries the antioxidant and TNF-α signal
Brain oxidative and inflammatory markers also diverged. Total antioxidant capacity and total superoxide dismutase activity both fell in colitis mice; quercetin counteracted both (p < 0.05), while luteolin, kaempferol and quercetin were effective in restoring SOD activity (p < 0.05).
Among cytokines, none of the flavonoids changed the anti-inflammatory cytokine IL-10 in the brain. IL-6, TNF-α and IL-1β were significantly elevated in DSS mice versus controls (p < 0.01). All compounds except luteolin substantially inhibited the IL-6 and IL-1β increases, but only quercetin reduced TNF-α levels (p < 0.01). In colon tissue, apigenin and luteolin markedly elevated Il-10 expression.
This is where the paper's claim of complementary rather than unified action is easiest to see: quercetin is distinguished by antioxidant capacity and TNF-α suppression, apigenin by astrocyte reversal, kaempferol by microbial diversity, and luteolin by tight junction upregulation without the same cytokine coverage.
Read the ranking as model-specific specialization
The practical takeaway is a division of labor, not a single winner. If the question is which compound best restored the behavioral readouts used here, kaempferol and quercetin lead. If the question is which one touched astrocyte overactivation, only apigenin did. If the question is tight junction protein upregulation, apigenin and luteolin did it; if it is reduced permeability measured by FITC-dextran flux, quercetin did. If it is alpha diversity, kaempferol was optimal.
The authors propose a shared systemic mechanism centered on restoring intestinal barrier integrity, modulating the microbiome and attenuating neuroinflammation, with compounds differing in their primary regulatory pathway. They also note that the exact structure–activity relationships — such as how hydroxyl substitution affects blood–brain barrier crossing or confinement to the gut — remain unresolved, and that causal designs such as fecal microbiota transplantation or targeted manipulation of specific taxa and barrier function are still needed.
The most defensible reading is therefore a hypothesis about multi-targeted dietary combinations: structurally similar flavonoids appear to cover different nodes of the gut–brain axis in this DSS mouse model, and the value of combining them is a proposal the paper raises rather than a tested intervention.
Sources used in this answer
Four dietary flavonoids differentially modulate the gut microbiota–barrier interaction against colitis-driven cognitive deficits
In DSS colitis mice, all four flavonoids reduced colitis severity, but kaempferol and quercetin most effectively rescued novel object recognition and Y-maze performance, apigenin uniquely reversed astrocyte overactivation, quercetin and apigenin raised hippocampal synaptic proteins, and the compounds diverged on barrier, microbiota and cytokine measures.