My research focus for about 10 years has been on the safety evaluation of pesticides, publishing around 80 scientific articles cited more than 6,500 times. My main objective is to identify gaps in the evaluation of pesticide toxicity, and to propose solutions to better predict their effects in human populations.
Health effects of pesticides are not always correctly predicted before products are released for marketing. Laboratory studies are often insufficient to detect effects on the gut microbiome, which are generally not measured in the battery of tests performed to regulate pesticide adverse effects.
We combined shotgun metagenomics and metabolomics to assess effects of glyphosate on the gut microbiota of Sprague-Dawley rats (Mesnage et al., 2021). Glyphosate did not act as an antibiotic in the gut — while it does inhibit the shikimate pathway in gut microorganisms, bacteria can obtain aromatic amino acids from diet rather than synthesis, so the antibiotic effect is minimal. However, glyphosate's phosphorus atom may serve as an energy source for some bacteria, altering microbiome composition.
We further studied 65 adult twins in the UK and evaluated if urinary glyphosate excretion associates with faecal microbial community composition (Mesnage et al., 2022). Glyphosate was found in 53% of individuals, and its excretion was positively associated with bacterial species richness.
We also described for the first time the effects of glyphosate on infant gut microbiota using SHIME technology (Mesnage et al., 2022), and studied fungal population changes as part of the Global Glyphosate Study (Mesnage et al., 2022).
Despite widely recognised benefits of organic farming, endorsement by government food standards agencies is generally lacking due to the scarcity of studies showing direct health benefits. Our TwinsUK cohort study (65 monozygotic twin pairs discordant for organic food consumption) showed inconsistencies in dietary questionnaire responses, highlighting the need for controlled dietary intervention studies (Mesnage et al., 2020).
We subsequently launched the ORGAMIC study — the first controlled dietary intervention comparing organic vs non-organic plant-rich diets on urinary polyphenols, pesticide levels, gut microbiota, and cardiovascular biomarkers in healthy young individuals.
We performed the first in-depth comparative toxicogenomic evaluation of glyphosate and Roundup formulations (Mesnage et al., 2021). Roundup altered the expression of 96 genes in rat liver linked to DNA damage and oxidative stress. Direct DNA damage measurement by apurinic/apyrimidinic lesion formation was increased with glyphosate exposure, and the Roundup formulation was more toxic than glyphosate alone.
In vitro tests found that two Roundup formulations (but not glyphosate) activated oxidative stress and misfolded protein responses — clear markers of carcinogenicity — suggesting that surfactants in formulations have their own toxic effects that are not predictable from glyphosate testing alone.
We tested a mixture of six pesticides frequently detected in foodstuffs (azoxystrobin, boscalid, chlorpyrifos, glyphosate, imidacloprid, thiabendazole) in Sprague-Dawley rats (Mesnage et al., 2021). Standard histopathology and serum biochemistry showed little effect, but serum and caecum metabolomics revealed oxidative stress response activation. Transcriptomics and DNA methylation analysis showed pathway disturbances undetectable by standard measures, demonstrating the value of omics approaches for low-dose mixture toxicity assessment.
Following findings that glyphosate formulated products caused NAFLD-like phenotype in rats (Mesnage et al., 2015), we developed a sensitive transcriptomics/metabolomics assay using the human HepaRG liver cell line to reveal molecular signatures of NAFLD (Mesnage et al., 2018a). This system was then used to investigate herbicides including glyphosate, quizalofop-p-ethyl, isoxaflutole, and mesotrione (Mesnage et al., 2018b).
We evaluated the obesogenic potential of common herbicides in adipogenesis assays, finding that quizalofop-p-ethyl has the most pronounced adipogenic effect (Biserni et al., 2019). We also performed the first evaluation of seven neonicotinoid insecticides for oestrogenic, thyroid, and adipogenic effects, finding imidacloprid is a potential obesogen (Mesnage et al., 2018c).
The estrogenic potential of glyphosate in breast cancer cells was assessed in collaboration with the US EPA using RNA-sequencing technologies. Glyphosate only activated oestrogen receptor alpha at very high concentrations (Mesnage et al., 2017a).
Starting in 2008, I published the first study linking ethoxylated surfactant concentration to the toxicity of glyphosate-based herbicides (Mesnage et al., 2013). Seven formulations were tested — some were approximately 300 times more toxic than glyphosate alone, linked to ethoxylated tallowamine surfactants. This observation contributed to their banning from glyphosate-based herbicides in Europe.
A follow-up study of 9 active ingredients found that 8 out of 9 formulations were up to 1,000 times more toxic than their active principles (Mesnage et al., 2014). In 2021, I led a consortium calling for more transparent pesticide use data, resulting in a Nature Ecology & Evolution publication (Mesnage, Straw et al., 2021) that contributed to EU regulatory reform. I also addressed co-formulant toxicity in my Herbicides book and have spoken at the European Parliament PEST hearing.