Gasdermins (GSDM) genes play complex roles in inflammatory diseases and cancer. Gasdermin-B (GSDMB) is frequently upregulated in human cancers, especially in HER2-amplified breast carcinomas, and can promote diverse pro-tumor functions (invasion, metastasis, therapy-resistance). In particular, the GSDMB shortest translated variant (isoform 2; GSDMB2) increases aggressive behavior in breast cancer cells. Paradoxically, GSDMB can also have tumor suppressor (cell death induction) effects in specific biological contexts. However, whether GSDMB has inherent oncogenic, or tumor suppressor function in vivo has not been demonstrated yet in preclinical mouse models, since mice lack GSDMB orthologue. Therefore, to decipher GSDMB cancer functions in vivo we first generated a novel knock-in mouse model (R26-GB2) ubiquitously expressing human GSDMB2. The comprehensive histopathological analysis of multiple tissues from 75 animals showed that nucleus-cytoplasmic GSDMB2 expression did not clearly affect the overall frequency nor the histology of spontaneous neoplasias (mostly lung carcinomas), but associated with reduced incidence of gastric tumors, compared to wildtype animals. Next, to assess specifically the GSDMB2 roles in breast cancer, we generated two additional double transgenic mouse models, that co-express GSDMB2 with either the HER2/NEU oncogene (R26-GB2/MMTV-NEU mice) or the Polyoma middle-T antigen (R26-GB2/MMTV-PyMT) in breast tumors. Consistent with the pro-tumor effect of GSDMB in HER2+ human breast carcinomas, R26-GB2/MMTV-NEU GSDMB2-positive mice have double breast cancer incidence than wildtype animals. By contrast, in the R26-GB2/MMTV-PyMT model of fast growing and highly metastatic mammary tumors, GSDMB2 expression did not significantly influence cancer development nor metastatic potential. In conclusion, our data prove that GSDMB2 in vivo pro-tumor effect is evidenced only in specific biological contexts (in concert with the HER2 oncogene), while GSDMB2 alone does not have overall intrinsic oncogenic potential in genetically modified mice. Our novel models are useful to identify the precise stimuli and molecular mechanisms governing GSDMB functions in neoplasias and can be the basis for the future development of additional tissue-specific and context-dependent cancer models.

Background: Gasdermin-B gene (GSDMB) is frequently over-expressed in tumors, and its shortest translated variant (isoform 2; GSDMB2) increases aggressive behavior in breast cancer cells. Paradoxically, GSDMB could have either pro-tumor or tumor suppressor properties depending on the biological context. Since GSDMB gene is not present in the mouse genome, deciphering fully the functional roles of GSDMB in cancer requires novel in vivo models. Methods: We first generated by gene targeting a conditional knock-in mouse model (R26-STOP-GB2) harboring human GSDMB2 transcript within the ROSA26 locus. We next derived the R26-GB2 model ubiquitously expressing GSDMB2 in multiple tissues (confirmed by western blot and immunohistochemistry) and performed a comprehensive histopathological analysis in multiple tissues from 75 male and female mice up to 18 months of age. Additionally, we produced the double transgenic model R26-GB2/MMTV-PyMT, co-expressing GSDMB2 and the Polyoma-Middle-T oncogene, and assessed breast cancer generation and progression in GSDMB2-homozygous (n=10) and control (n=17) female mice up to 15 weeks of age. Results: In the R26-GB2 model, which showed different GSDMB2 cytoplasmic and/or nuclear localization among tissues, we investigated if GSDMB2 expression had intrinsic tumorigenic activity. 41% of mice developed spontaneous lung tumors, but neither the frequency nor the histology of these neoplasias was significantly different from wildtype animals. Strikingly, while 17% control mice developed gastric carcinomas, no GSDMB2-positive mice did. No other tumor types or additional histological alterations were frequently seen in these mice. In the R26-GB2/MMTV-PyMT model, the strong nucleus-cytoplasmic GSDMB2 expression in breast cancer cells did not significantly affect cancer formation (number of tumors, latency, tumor weight, histology or proliferation) or lung metastasis potential compared to controls. Conclusions: GSDMB2 expression alone does not have an overall tumorigenic potential in mice, but it might reduce gastric carcinogenesis. Contrary to human cancers, GSDMB2 upregulation does not significantly affect breast cancer generation and progression in mouse models. However, to evidence the GSDMB functions in cancer and other pathologies in vivo may require the presence of specific stimulus or cellular contexts. Our novel mouse strains will serve as the basis for the future development of more precise tissue-specific and context-dependent cancer models.