Back

Isoflavones impair response to anti-PD1 therapy in murine breast cancer models, irrespective of dietary fiber and fecal short chain fatty acid levels

de Oliveira Andrade, F.; Bouker, K.; Ozgul-Onal, M.; Jin, L.; Cruz, I.; Helferich, W.; Gao, A.; Andrade, K.; Verma, V.; Staley, C.; Foley, P.; Hilakivi-Clarke, L.

2025-12-09 cancer biology
10.64898/2025.12.05.692375 bioRxiv
Show abstract

BackgroundFermentable dietary fibers, also called microbiota-accessible carbohydrates (MAC), and the consequent increase in fecal short-chain fatty acids (SCFAs) are linked to improved responsiveness to immune checkpoint blockade (ICB) therapy in human and mouse studies. However, experimental diets high in MAC also often contain estrogenic isoflavones, which may counter fibers beneficial effects by causing immunosuppression. MethodsWe studied the effects of feeding female C57BL/6Tac mice low-MAC (AIN93G), low-MAC supplemented with isoflavone genistein, high-MAC (5V5M) or high-MAC isoflavone (high-MACi; 5058D) diet on their gut microbiome and response to anti-PD1 therapy against E0771 allografted triple negative breast cancer (TNBC) and 7,12-dimethylbenz[a]anthracene (DMBA)-initiated estrogen receptor positive (ER+) mammary tumors. We also determined whether blocking ER with tamoxifen (TAM) impacted responsiveness to anti-PD1 therapy in mice fed different diets. The effect of diet and treatments on immune cell signaling pathways was investigated using NanoString PanCancer Immune Profiling Panel. ResultsHigh-MAC and high-MACi diets increased fecal microbial alpha-diversity and the abundances of SCFA producing families Lachnospiraceae, and Oscillospiraceae as well as fecal SCFA levels, compared with low-MAC diet. E0771 tumors responded to anti-PD1 in mice fed high-MAC, while mice fed high-MACi did not respond. Low-MAC fed mice with single E0771 allograft also responded to anti-PD1, but genistein supplementation eliminated responsiveness. E0771 tumors in high-MAC fed mice contained elevated levels of exhausted CD8+ T cells, which were decreased after anti-PD1 therapy. Opposite effects were seen in mice fed high-MACi diet. Mice with DMBA-initiated ER+ mammary tumors did not respond to anti-PD1. TAM converted TNBC and ER+ tumors to become sensitive to anti-PD1 therapy in mice fed high-MACi or low-MAC diets, respectively. Genes in TH17 differentiation pathways were linked to TAM-induced improved anti-PD1 response both in TNBC and ER+ mammary tumors. ConclusionsOur results highlight the role of diet in impacting the effectiveness of ICB therapies. We found that increased SCFA levels alone are not predictive of response to anti-PD1, but if tumor expresses ER or if diet contains ER activating compounds, such as isoflavones, blocking ER+ might convert unresponsive tumors responsive to anti-PD1. Word count: 339 What is already known on this topicDietary fiber is proposed to improve response to checkpoint inhibitor therapy against melanoma, but this has been challenged by a recent preclinical study in which different mouse tumor models were used. None of the studies have been done in breast cancer or preclinical breast cancer models. What this study addsOur study showed, using triple negative breast cancer (TNBC) and estrogen receptor positive (ER+) breast cancer models, that indeed high levels of microbiota accessible carbohydrates (MACs) in diet did not alone determine responsiveness to anti-PD1 therapy, but diet high in plant isoflavones/ hormones impaired anti-PD1 effectiveness, regardless of whether diet contained high fiber levels or not. We also found that the adverse effects of isoflavones were counteracted by tamoxifen, partial estrogen receptor antagonist. How this study might affect research, practice or policyOur findings could indicate that breast cancer patients, both those with TNBC and ER+ disease, should not consume diets high in isoflavones when treated with anti-PD1.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.