The scale of the screening was ambitious. The researchers measured 975 sweetener–bacterium pairs, growing each bacterial strain in modified Gifu anaerobic medium under strictly anaerobic conditions while exposing it to 50 micromolar concentrations of each sweetener—a level the team estimated to be within the range relevant to the human colon. Growth was tracked hourly for 24 hours, quantified as the area under the growth curve, and any apparent interaction had to pass stringent criteria: statistical significance after correction for multiple testing, a change in growth of at least 20 percent, and confirmation in an independent experiment across five different concentrations. After this rigorous validation, 30 genuine interactions emerged between 26 sweeteners and 5 bacterial strains, none of which had been reported before. Three-quarters of the tested sweeteners individually affected the growth of at least one strain.

Some bacteria proved especially sensitive. Clostridium symbiosum and Lacrimispora saccharolytica were hit by the largest number of compounds, while the sweetener component isosteviol—a stevia-derived molecule widely used in commercial formulations—emerged as the most potent single agent, inhibiting three bacterial species while paradoxically promoting the growth of Lactobacillus gasseri. But the most striking findings came when the researchers moved beyond individual compounds. Because sweeteners are rarely consumed in isolation—they appear alongside caffeine in soft drinks, vanillin in processed foods and even baby formula, and are co-formulated into pharmaceutical tablets to mask bitterness—the team screened 156 sweetener–xenobiotic combinations against all 25 bacteria, yielding 3,900 potential interactions, plus another 300 combinations involving common tablet drugs such as ibuprofen, acetaminophen, and cetirizine.

Using the Bliss model of independence, a classic pharmacological framework in which two compounds are considered synergistic if their combined effect exceeds the product of their individual effects and antagonistic if it falls short, the team identified 102 significant interactions involving nine bacterial species—68 antagonistic and 34 synergistic. Isosteviol was again central, showing combinatory effects with three of the four co-consumed compounds across five bacterial species. The strongest synergy by far was between isosteviol and duloxetine, which together inhibited Roseburia intestinalis, a butyrate-producing bacterium repeatedly linked to glucose homeostasis and protection against intestinal inflammation, and Parabacteroides merdae, a prevalent commensal associated with a healthy microbiota. On the antagonistic side, vanillin buffered the effect of saccharin against Bifidobacterium adolescentis—two compounds that frequently co-occur in processed foods.

To translate single-species effects into community-level consequences, the researchers assembled a synthetic gut community from all 25 strains and passaged it serially in the presence of the solvent control, isosteviol, duloxetine, or the combination. All compound treatments reduced species diversity, but the isosteviol–duloxetine combination was the most disruptive, significantly depleting R. intestinalis, P. merdae, Segatella copri, Bacteroides uniformis, and Phocaeicola vulgatus. Intriguingly, the picture did not simply mirror the monoculture results. R. intestinalis showed synergy in isolation but not in community, while P. merdae showed the reverse—a pattern the authors describe as emergent, meaning that ecological interactions among community members generated effects that could not be predicted from single-species data. Of the 25 strains, 24 grew worse under the combination in monoculture, yet 7 actually benefited in the community setting, underscoring how metabolic cross-feeding can reshape drug and sweetener responses.

What is the molecular mechanism behind this synergy? Proteomic analysis of R. intestinalis and P. merdae after four hours of exposure revealed that the combination produced protein abundance changes that deviated from what would be expected from purely additive single-compound effects. Eleven proteins in R. intestinalis and 29 in P. merdae responded specifically to co-exposure, with no overlap between the two species. In R. intestinalis, many changes clustered in vitamin B5 biosynthesis and fatty acid metabolism, but supplementing cultures with pantothenate, coenzyme A, and 18 other metabolic additives failed to rescue growth, ruling out this pathway as the primary target. In P. merdae, the response was dominated by cell envelope proteins: twelve upregulated proteins were transporters, efflux pump subunits, or envelope components, while the saccharide uptake system subunits SusC and SusD were downregulated—a clear signature of altered small-molecule transport across the membrane.

A genome-wide transposon mutant screen in P. merdae, using a barcoded library covering roughly 3,000 non-essential genes, reinforced this conclusion. Mutants losing the strongest fitness under co-exposure mapped to genes encoding membrane proteins, including a CorA-family putative magnesium transporter, a FadL-family hydrophobic compound transporter, a domain of the outer membrane assembly factor BamD, and a c-di-AMP phosphodiesterase ortholog implicated in membrane homeostasis in the oral pathogen Porphyromonas gingivalis. The CorA transporter stood out because it showed both strong negative fitness and increased protein abundance under duloxetine exposure—yet adding magnesium to the medium did not rescue growth, suggesting the protein performs an unknown homeostatic role rather than simply importing the ion. Bioaccumulation assays added another layer: R. intestinalis accumulated isosteviol only when duloxetine was present, while P. merdae accumulated it even without the drug, consistent with transport changes driving the synergy.

The metabolic consequences were substantial. Untargeted liquid chromatography–mass spectrometry showed that co-exposure reshaped the secreted metabolome of all three tested species, with 16 of 19 significantly altered peaks annotated as di- and tripeptides—microbial products previously linked to blood pressure regulation, Listeria infection, and intestinal inflammation. Targeted quantification of 46 metabolites revealed that in R. intestinalis, co-exposure raised glutamine secretion by roughly 50 percent (about 65 micromolar) and cut butyric acid by more than 25 percent (about 0.6 millimolar), while isovaleric acid—a metabolite correlated with depression in fecal studies—also increased. In P. merdae, asparagine concentrations shifted by more than 200 micromolar. Because butyrate fuels colonocytes, supports insulin sensitivity, and helps maintain anti-inflammatory immune tone, while glutamine can be either protective or harmful depending on context, these shifts carry real implications for host physiology.

The final experiments connected microbial chemistry back to human cells. When HeLa cells were incubated with supernatants from the 25-member community, supernatants from combination-treated communities were significantly more cytotoxic than controls—even controls where the compounds were added to supernatant only after harvesting, proving the toxicity stemmed from altered bacterial metabolism rather than the compounds themselves. In Caco-2 intestinal epithelial cells, the combination-treated supernatant suppressed secretion of the inflammatory cytokines IL-6 and IL-8 by more than 75 percent while leaving IP-10 unaffected. Because IL-6 and IL-8 coordinate neutrophil recruitment and gut defense, the authors caution that such dampening could reflect either anti-inflammatory effects or an impaired capacity to fight pathogens. The team links the cytokine changes to reduced short-chain fatty acids in the community supernatant, and shows that butyrate plus propionate can itself stimulate IL-8 secretion.

The authors are candid about limitations: the screens used a single concentration of 50 micromolar in a single growth medium, the synthetic community contains only 25 taxa, and real-world exposures are far more variable across individuals. Yet prior work suggests roughly 60 percent of monoculture drug effects are preserved in communities, making these findings a credible starting point. What the study makes unmistakably clear is that sweetener safety cannot be evaluated in isolation. Given that duloxetine alone accounted for more than 18 million prescriptions in the United States in a single year, and that sweeteners, caffeine, and vanillin saturate the modern food supply, the sweetener–xenobiotic–microbiome axis may represent an overlooked dimension of public health—one in which the bacteria in our intestines act as silent arbiters of what we eat, what we medicate with, and how our bodies respond.

Subject of Research: Interactions between low-calorie sweeteners, co-consumed xenobiotics, and human gut bacteria in vitro

Article Title: Common xenobiotics modulate gut microbial responses to low‑calorie sweeteners in vitro

Article References: Common xenobiotics modulate gut microbial responses to low‑calorie sweeteners in vitro. (n.d.). https://doi.org/10.1038/s44320-026-00225-6

Image Credits: AI Generated

DOI: 10.1038/s44320-026-00225-6

Keywords: low-calorie sweeteners, gut microbiota, xenobiotics, duloxetine, isosteviol, Roseburia intestinalis, Parabacteroides merdae, bliss synergy, metabolomics, proteomics, cytokines, microbiome

The scale of the problem is enormous. The review’s authors report that roughly 34 percent of patients worldwide experience CIPN six months or longer after finishing chemotherapy, with prevalence peaking at around 68 percent during the first month after treatment. Symptoms arise because neurotoxic drugs damage the myelin sheaths, axons, and neuronal cell bodies of the dorsal root ganglia, triggering neuroinflammation, elevated pro-inflammatory cytokines, apoptosis, and altered neuronal excitability. Clinicians distinguish positive symptoms, such as allodynia, hyperalgesia, dysesthesia, and paresthesia, from negative symptoms like numbness and motor dysfunction. In severe cases, the condition forces oncologists to reduce chemotherapy doses or abandon potentially curative regimens altogether, making effective symptomatic treatment a genuine clinical priority.

International guidelines from the American Society of Clinical Oncology and the European Society for Medical Oncology currently identify duloxetine, a serotonin-norepinephrine reuptake inhibitor, as the only pharmacological option supported by strong evidence for painful CIPN, though its use remains off-label. The foundational evidence came from a randomized placebo-controlled trial by Smith and colleagues, later confirmed by Hirayama and coworkers in a large open-label Japanese study. Yet clinical practice has moved on, with growing interest in combination and multimodal approaches, and the comparative evidence base had never been systematically consolidated. The new review set out to fill exactly that gap, synthesizing studies published between 2019 and 2025 under contemporary diagnostic and reporting standards.

Methodologically, the review was rigorous and transparent. Two investigators independently searched PubMed, Web of Science, and Scopus following PRISMA guidelines and the PICO framework, with the protocol registered on PROSPERO. From 892 initial records, 374 duplicates were removed, and successive rounds of screening excluded pre-2019 publications, preclinical studies, reviews, editorials, and inaccessible or irrelevant articles. Eleven clinical studies survived the selection process, spanning randomized double-blind placebo-controlled trials, a non-inferiority trial, prospective pilot work, retrospective observational cohorts, and individual case reports. Risk of bias was assessed with the Cochrane RoB 2 tool for randomized trials and the Newcastle-Ottawa Scale for observational studies, and most randomized trials showed low risk across the key domains of randomization, blinding, and outcome measurement.

The patient populations were heterogeneous, reflecting the diversity of oncology practice. Enrolled individuals had breast, gastrointestinal, lung, hematological, head and neck, genitourinary, and other malignancies, with sample sizes ranging from single-patient case reports to more than 230 participants. The antidepressants evaluated included duloxetine at daily doses of 20 to 60 milligrams, oral amitriptyline at 10 to 20 milligrams, topical 10 percent amitriptyline gel, and venlafaxine, alongside combination regimens pairing duloxetine with mirogabalin at 5 to 25 milligrams daily or tapentadol at 50 to 500 milligrams daily. Pain outcomes were tracked with the Numerical Rating Scale, the Visual Analogue Scale, and the Brief Pain Inventory, while neuropathy severity was graded using CTCAE criteria, the Total Neuropathy Score, and the DN4 questionnaire.

The results for duloxetine monotherapy were strikingly consistent. In one double-blind placebo-controlled trial in non-metastatic breast cancer patients receiving paclitaxel, duloxetine cut average pain scores on the Numerical Rating Scale from 4.19 to 2.63 over eight weeks, while CTCAE neuropathy grades fell from 1.48 to 0.79, a statistically significant difference against placebo. A second placebo-controlled trial in gastrointestinal cancer patients treated with oxaliplatin similarly found that duloxetine prevented worsening of peripheral sensory neuropathy, with grade scores improving from 0.86 to 0.62. In a broader randomized trial across multiple tumor types, 44.1 percent of duloxetine-treated patients achieved a clinically significant pain reduction of at least 30 percent, compared with only 18.2 percent on placebo, and 32.4 percent achieved reductions of 50 percent or more versus 3.0 percent of placebo patients. A retrospective observational study was the outlier, reporting only a modest, statistically insignificant Visual Analogue Scale change, likely reflecting uncontrolled dosing and confounding.

Combination therapies generated the most eye-catching numbers, though the authors are careful about interpretation. In a randomized study of 72 breast cancer patients, duloxetine paired with electrostimulation reduced average pain scores from about 6.1 to 3.6 in two weeks. A single case report described mirogabalin plus duloxetine lowering pain scores from 8 to 1 and CTCAE grade from 3 to 1. In a randomized non-inferiority trial of 114 patients, tapentadol combined with duloxetine reduced the Numerical Rating Scale from 7.51 to 2.87, statistically equivalent to tapentadol alone. A prospective randomized comparison of 89 patients found duloxetine outperforming pregabalin on both pain and DN4 neuropathy scores, with duloxetine dropping scores from roughly 7 to 4 over four weeks. Meanwhile, a retrospective cohort in lung cancer showed duloxetine plus low-dose mirogabalin reducing median pain scores from 5.5 to 4.0. Importantly, adverse events remained mild across all regimens, limited mainly to nausea, constipation, dizziness, and transient somnolence, and combination treatments did not increase side-effect burdens relative to monotherapy.

One of the more intriguing findings concerned topical amitriptyline. In a prospective pilot study of 44 patients treated with bortezomib or oxaliplatin, a 10 percent topical amitriptyline emulsion applied twice daily for six months reduced Visual Analogue Scale scores from 7 to 2, with the difference reaching high statistical significance. Topical delivery may offer localized analgesia with far better tolerability than systemic tricyclic therapy, which is otherwise limited by dry mouth, weight gain, and drowsiness. The evidence base remains small, but the result hints at a practical option for patients who cannot tolerate oral agents, and it underscores how routes of administration may reshape the pharmacology of otherwise old drugs.

Why should antidepressants relieve nerve pain at all? The review explains the mechanistic logic. SNRIs like duloxetine and tricyclics like amitriptyline enhance both serotonergic and noradrenergic transmission, strengthening descending inhibitory pathways, particularly noradrenergic projections from the locus coeruleus to the spinal dorsal horn that suppress pain signals before they reach consciousness. Selective serotonin reuptake inhibitors, by contrast, fail to engage these noradrenergic circuits, which may explain their inconsistent efficacy in neuropathic pain. Preclinical work supports these mechanisms, with duloxetine, amitriptyline, and pregabalin all reducing pain behaviors in animal models, although translational gaps persist, especially for pregabalin and venlafaxine, which perform better in the laboratory than in the clinic, likely because animal models incompletely reproduce human CIPN and because dosing and timing differ.

The authors are candid about limitations. Study heterogeneity in design, populations, and outcome measures precluded quantitative meta-analysis, several trials were small, and the restriction to English-language open-access studies published after 2019 may have introduced bias and excluded foundational duloxetine trials, which the authors discussed as context rather than primary evidence. Combination benefits remain hypothesis-generating, since most multimodal studies lacked appropriate comparator arms to disentangle whether effects are additive, synergistic, or driven by a single component. Prevention, too, remains unproven; current evidence supports antidepressants only for treating established CIPN. Still, the message for patients and clinicians is clear and cautiously optimistic: duloxetine stands as the first-line pharmacological option, topical amitriptyline deserves larger trials, and well-designed randomized studies of combination strategies could finally deliver the relief that millions of cancer survivors still lack.

Subject of Research: The efficacy of antidepressants, particularly duloxetine, in managing chemotherapy-induced peripheral neuropathic pain

Article Title: The Role of the Antidepressants in Managing Chemotherapy‐Induced Neuropathic Pain: A Systematic Review

Article References: Caminiti, R., Mazza, V., Nucera, S., Oppedisano, F., Passacatini, L. C., Maiuolo, J., Malafoglia, V., Giorgio, A., Soluri, A., Pileggi, C., Tomino, C., Mollace, V., Ilari, S., & Muscoli, C. (2026). The Role of the Antidepressants in Managing Chemotherapy‐Induced Neuropathic Pain: A Systematic Review. Pharmacology Research & Perspectives, 14(5), Article e70327. https://doi.org/10.1002/prp2.70327

Image Credits: AI Generated

DOI: 10.1002/prp2.70327

Keywords: chemotherapy-induced peripheral neuropathy, CIPN, duloxetine, antidepressants, neuropathic pain, systematic review, SNRI, amitriptyline, cancer survivors, pain management, mirogabalin, tapentadol