Fig. 7. Experimental metastasis is dependent on PGD and suppression of TXNIP in vivo.
a (Left) PGDhigh 13Pr cells implanted into the spleens of athymic mice preferentially develop liver metastasis (arrows, n = 4/4) without involvement of the pancreatic parenchyma (n = 0/4), as shown by representative gross photographs and H&E-stained sections. (Right) In contrast, PGDlow 38Per cells formed grossly evident pancreatic tumors (n = 5/5), with small tumor implants on the outside surface of the liver as seen on H&E-stained sections (n = 3/5). Gross photograph scale bars: 1 cm, H&E scale bars: 400 µm. L liver, P pancreas, T tumor. b, c Engineering 13Pr cells with PGD Crispr/Cas sgRNAs (b) or exogenous TXNIP (c) reversed TXNIP suppression (top), surface GLUT1 retention (middle), and H3K27 hyperacetylation (bottom) by confocal IF (n = 3 biological replicates, error bars: s.e.m., indicated p values calculated by two-tailed t tests, scale bars: 20 µm). d, e 13Pr cells engineered with Crispr/Cas PGD inactivating sgRNAs (d) or exogenous TXNIP (e) impaired 3D in vitro tumoroid growth relative to 13Pr control cells (n = 3 technical replicates, error bars: s.d.m., indicated p values calculated by two-tailed t tests, scale bars: 400 µm). f Representative gross and H&E-stained sections demonstrate that 13Pr cells engineered with exogenous TXNIP (n = 0/6) or Crispr/Cas inactivation of PGD (n = 0/6) failed to develop liver metastases compared to 13Pr control cells (n = 2/3). Gross photograph scale bars: 1 cm, H&E scale bars: 400 µm. L liver parenchyma, T tumor.