Comment on "NRF2 activation by antioxidant antidiabetic agents accelerates tumor metastasis".

on N ovem er 7, 2016 ht://stm .scencem agorg/ D ow nladed from We are writing to express concern with the conclusions drawn from the recent publication, “NRF2 activation by antioxidant antidiabetic agents accelerates tumor metastasis” (1). In a summary of the article, the editors suggested that antioxidant compounds, including first-line agents used in the treatment of type 2 diabetes, represent the potential risk of accelerating tumor metastasis. We do not find sufficient evidence in the article to support this claim and believe that a high burden of evidence should be imposed when considering that an estimated 30 million persons are receiving these therapeutic agents. The claim is based on in vitro data that suggest enhanced migration of multiple tumor cell lines. The authors show that dipeptidyl peptidase 4 (DPP4) inhibitors activate nuclear factor E2–related factor 2 (NRF2), which appears to stimulate migration by modulating cellular metabolic activity. Although interesting, it bears mentioning that the biologic effect observed in Wang et al. is modest, with most of the cells lines showing a drug-induced change in in vitro migration <1 cell per imaged field (Fig. 1A in the original article). In addition, the migration of nontumor and/or hematopoietic cells in response to drug treatment was not tested. Therefore, it is unclear whether the observed increase in migration is specific to tumor cells or may, in fact, reflect a more generalized response to drug treatment that may be shared by multiple cell types. The authors also tested their hypothesis using in vivo models. Sitagliptin or saxagliptin treatments were administered to tumor-bearing mice, and the metastatic burden was assessed using bioluminescence imaging, histopathology, and protein biomarkers of metastasis. Notably, all in vivo studies were performed in BALB/c nude mice. This is a major caveat because these mice lack T and B cells, which are essential components of antitumoral immunity. This would explain why the results differ from our recent findings that DPP4 inhibition is a strategy for improving antitumor immunity (2). Specifically, our study demonstrated that DPP4 regulates CXCL10-mediated lymphocyte trafficking into tumors, and its inhibition enhanced naturally occurring tumor immunity and tumor immunotherapy. In addition to our studies, other publications (not discussed by Wang et al.) have documented the antimetastatic (3) and antitumorigenic (4) activity of DPP4 inhibitors when used in wild-type animals. Diabetic patients are on the whole immunocompetent, and as such, the use of athymic nude mice is an inappropriate surrogate to model the impact of tumor cell metastasis. Therefore, the authors’ principal finding should be qualified as it may