(A) Schematic representation of the reprogramming protocol. The standard protocol (REP) consisted in plating the exocrine clusters in tissue culture dishes and allowing them to attach for 48h, followed by 3 day culture in serum free medium (SFM) with 10 μM SB431542 (SB), 2 μM Y27632 (Y2), 1 μM 5-Aza-2’deoxycytidine (Aza) and 1 mM sodium butyrate (NaBu). On day 4 cells were transduced with adenoviruses containing Pdx1, Ngn3, Pax4 and MafA, followed by 7 days in SFM with 1 nM betacellulin, 10 nM exendin-4 and 10 nM nicotinamide (BEN). The new protocol (REP + siARX) included the transfection with an siRNA targeting ARX at day 6 of the protocol. (B) Schematic representation of the interplay between the TFs PAX4 and ARX during the late stages of α- and β-cell development. (C) RT-qPCR analysis of endogenous ARX in untreated (N/A), REP cells, REP cells transfected with a scrambled siRNA (REP + SCRB) and REP + siARX cells. Expression was normalised to glyceraldehyde 3-phosphate dehydrogenase. Data are representative of triplicate experiments and represented as mean ± SEM. A one way ANOVA was performed followed by a Bonferroni post hoc test to compare all treatment groups, where ***P < 0.001, ** P <0.01, *P < 0.05. (D) Immunocytochemistry for ARX in untreated (N/A), REP cells, REP cells transfected with a scrambled siRNA (REP + SCRB) and REP + siARX cells. Scale bar = 50 μm. (E) RT-qPCR analysis of insulin, glucagon and somatostatin in untreated (N/A), REP cells, REP cells transfected with a scrambled siRNA (REP + SCRB), REP + siARX cells, REP + siARX in the absence of exogenous Pax4 (REP+siARX–PAX4) and in human islets. Expression was normalised to GAPDH. Data are representative of triplicate experiments. A one way ANOVA was performed followed by a Dunnet post hoc test, ***P < 0.001, ** P <0.01, *P < 0.05. (F) RT-qPCR analysis of late β-cell markers in untreated (N/A), REP cells, REP cells transfected with a scrambled siRNA (REP + SCRB), REP + siARX cells, REP + siARX in the absence of exogenous Pax4 (REP+siARX–PAX4) and in human islets. Expression was normalised to GAPDH. Data are representative of triplicate experiments. A one way ANOVA was performed followed by a Dunnet post hoc test, ***P < 0.001, ** P <0.01, *P < 0.05.