Pancreatic ductal adenocarcinoma (PDAC) exhibits one of the worst survival rates of all cancers. While death rates show declining trends in the majority of cancers, PDAC registers rising rates. Based on the recently described crosstalk between TGF-β1 and Nrf2 in the PDAC development, the involvement of ATF3 and its splice variant ΔZip2 in TGF-β1- and Nrf2-driven pancreatic tumorigenesis was investigated. As demonstrated here, PDAC (Panc1, T3M4) cells or premalignant H6c7 pancreatic ductal epithelial cells differentially express ΔZip2- and ATF3, relating to stronger Nrf2 activity seen in Panc1 cells and TGF-ß1 activity in T3M4 or H6c7 cells, respectively. Treatment with the electrophile/oxidative stress inducer tBHQ or the cytostatic drug gemcitabine strongly elevated ΔZip2 expression in a Nrf2-dependent fashion. The differential expression of ATF3 and ΔZip2 in response to Nrf2 and TGF-ß1 relates to differential ATF3-gene promoter usage, giving rise of distinct splice variants. Nrf2-dependent ΔZip2 expression confers resistance against gemcitabine-induced apoptosis, only partially relating to interference with ATF3 and its proapoptotic activity, e.g., through CHOP-expression. In fact, ΔZip2 autonomously activates expression of cIAP anti-apoptotic proteins. Moreover, ΔZip2 favors and ATF3 suppresses growth and clonal expansion of PDAC cells, again partially independent of each other. Using a Panc1 tumor xenograft model in SCID-beige mice, the opposite activities of ATF3 and ΔZip2 on tumor-growth and chemoresistance were verified in vivo. Immunohistochemical analyses confirmed ΔZip2 and Nrf2 coexpression in cancerous and PanIN structures of human PDAC and chronic pancreatitis tissues, respectively, which to some extent was reciprocal to ATF3 expression. It is concluded that depending on selective ATF3-gene promoter usage by Nrf2, the ΔZip2 expression is induced in response to electrophile/oxidative (here through tBHQ) and xenobiotic (here through gemcitabine) stress, providing apoptosis protection and growth advantages to pancreatic ductal epithelial cells. This condition may substantially add to pancreatic carcinogenesis driven by chronic inflammation.