Fig. 2. TXNIP suppression is PGD-dependent.

a Illustration depicting how PGD-driven suppression of TXNIP can allow retention of surface GLUT1 with increased glucose uptake. b Confocal IF detected high surface GLUT1 signals (top, cobblestone pattern) and reduced cytosolic TXNIP signals (bottom) in PGD^high liver and lung metastases compared with a control PGD^low peritoneal deposit from the same patient. c PGD Crispr/Cas⁵ (#1, #2) reduced glucose consumption rates (top) and increased TXNIP expression (bottom qRT-PCR) relative to wild-type (WT) baseline controls (n = 3 technical replicates per sample, error bars: s.d.m., indicated p values calculated by two-tailed t tests). d Confocal IF detected reductions in surface GLUT1 (top) with increased cytosolic TXNIP (bottom) in response to PGD Crispr/Cas (#1, #2), compared to wild-type baseline controls (WT). e Similar confocal IF experiments show that siRNAs against TXNIP (#1, #2) prevents upregulation of TXNIP protein and rescues surface GLUT1 localization during PGD inactivation (PGD Crispr/Cas, right three panels), as compared to both nontargeting control siRNAs (siNEG) and wild-type control cells (left three panels) (n = 2 biological replicates, error bars: s.e.m.). f Treatment of the indicated PGD-dependent subclones with PGD inhibitor (PGDi: 50 µM 6AN) increased TXNIP expression by RT-qPCR (bottom) and slowed of glucose consumption rates (top) (n = 3 technical replicates per sample, error bars: s.d.m., indicated p values for glucose consumption rates calculated by two-tailed t tests). For all IF experiments (unless indicated otherwise): green, antibody signal; blue, Hoechst signal; n = 3 biological replicates, error bars: s.e.m., indicated p values calculated by two-tailed t tests, scale bars: 20 µm.