Carrot Rhamnogalacturonan-I: effects on gut microbiota and immunity

The main result indicates that a small amount of this pectin modified a specific microbial target, rather than the entire gut microbiota. In the cRG-I group, fecal counts of Bifidobacterium grew more over time than in the placebo group, reaching approximately 1.5 times the initial value by the third week. By the fourth week, 57% of the treated individuals and 43% of those in the placebo group showed an individual increase. This signals an average group response, but does not prove that every participant responded in the same way.

Regarding the immune system, myeloid dendritic cells—cells that help present signals to other elements of the immune response—showed a higher proportion of cells with CD86 and HLA-DR. These are markers of cell activation: they indicate that the cells were expressing more signals involved in immune communication. The comparison with the placebo was significant for the proportion of cells positive for both markers (p < 0.005); the intensity of CD86 and HLA-DR increased with p < 0.02 and p < 0.05, respectively. However, this is not equivalent to demonstrating greater protection against infections or the improvement of a disease.

Only fecal isobutyric acid was higher after four weeks, increasing from 5.96 ± 1.04 to 6.75 ± 1.91 micromoles per gram of stool. The other short-chain fatty acids, molecules produced by bacterial fermentation, did not differ. This is important because it prevents us from inferring a generalized increase in fermentation from a single metabolite. The diarrhea score decreased from 4.50 ± 1.94 to 3.67 ± 1.20; in healthy adults, this should be interpreted as slightly firmer stools, not as a treatment for diarrhea.

The value of this research lies in demonstrating that fibers with different structures can have selective effects, even at doses much lower than those that contribute significantly to the total daily intake of dietary fiber. In this case, 500 mg per day coincided with an increase in bifidobacteria, even in individuals with an already relatively high habitual fiber intake. This finding therefore supports the idea that the composition and structure of fiber matter, in addition to the overall quantity.

Preclinical models complete the biological picture. In Caco-2 and THP-1 cells, used to simulate an intestinal epithelium exposed to inflammation, the cRG-I fermentate preserved the transepithelial electrical resistance, or TEER, an indicator of the integrity of the cell layer. In intestinal biopsies subjected to stress, cRG-I or its fermentate limited certain signals of loss of barrier function. This step suggests a plausible chain, from fiber to fermentation and the epithelial response. However, biological plausibility does not prove that the same chain produces measurable health benefits in healthy people or those with intestinal disorders.

The study is brief, involved a limited number of healthy adults, and was not designed to evaluate clinical outcomes such as the prevention of diseases, infections, or chronic intestinal disorders. No hazard ratios (HR), relative risks (RR), or odds ratios (OR) are available because this was not a study on clinical events over time. The results describe variations in biomarkers and questionnaires, not a reduction in the risk of pathology.

Furthermore, the increase in Bifidobacterium does not, by itself, prove an improvement in health. The composition of the gut microbiota varies over time, and stool samples represent only a portion of the intestinal environment. The absence of differences for the majority of short-chain fatty acids, fecal calprotectin, and self-reported well-being reduces the possibility of generalizing the signal to a broad transformation of intestinal physiology. Models using cells and biopsies employ controlled concentrations and experimental conditions: they are useful for formulating hypotheses, but they do not replace a clinical trial with patient-relevant outcomes.

Confirmation in humans comes from the 2024 trial by Jian and colleagues, which analyzed intestinal samples from 146 healthy adults exposed to 0.3 or 1.5 g per day of cRG-I or a placebo for eight weeks. Here, too, the relative abundance of Bifidobacterium increased, especially B. adolescentis and B. longum. The study supports the initial step of the chain—that is, the selection of specific gut bacteria by cRG-I—but it does not demonstrate that this variation causes the respiratory outcomes analyzed in the same research program. The confirmation is direct for the bifidogenic signal, with the limitation of a healthy population and analyses based on stool samples [1].

A second trial, conducted by Kerezoudi and colleagues in 2026 on 55 healthy adults, tested a rhamnogalacturonan-I derived from chicory. The fiber was not identical to that from carrots, but at 500 mg per day for four weeks, it reproduced the increase in bifidobacteria and the increase in HLA-DR on myeloid dendritic cells. It therefore supports, with moderate strength, that some RG-I variants may link a microbial response to systemic immune signals. However, the different botanical origin prevents attributing the effect solely to the common structure [2].

The transition from fermentation to metabolites is supported by an ex vivo experiment on fecal material from 24 adults. Van den Abbeele and colleagues observed increases in acetate and propionate and growth of taxa linked to Bifidobacterium with cRG-I; compared to inulin, the gas produced was 44% lower. This is a useful mechanistic confirmation, but the model outside the organism does not include absorption, diet, intestinal motility, or a complete immune response [3].

The available evidence suggests that one should not expect a linear dose-response. In the rhinovirus trial by Lutter and colleagues, 177 healthy adults took 0, 0.3, or 1.5 g per day of cRG-I before experimental exposure to the virus. The 0.3 g dose was associated with a 20% reduction in symptom severity and a 25% reduction in duration, while with 1.5 g, the clinical results were intermediate and more similar to the control group. The authors hypothesized an immune feedback mechanism; however, this remains an interpretation, and the respiratory context is not comparable to the intestinal research by Kerezoudi [4].

There is also a significant difference between experimental models and fecal measurements. In an in vitro system, Van den Abbeele and colleagues detected increases in acetate, propionate, lactate, and, in some phases, butyrate during the fermentation of cRG-I. In the 2025 trial, however, fecal short-chain fatty acids did not change, with the exception of isobutyrate. The two results are not necessarily incompatible, because the molecules produced in the colon can be absorbed, and feces do not measure their total production. They do show, however, that an effect observed in vitro cannot be automatically transferred to humans [5].