Supplementary Figure 1. Blood glucose and insulin levels in mice during 4-day infusion.

Transcription

1 Supplementary Figure 1. Blood glucose and insulin levels in mice during 4-day infusion. (A-B) WT and HT mice infused with saline or glucose had overlapping achieved blood glucose and insulin levels, necessitating separation into tertiles of achieved blood glucose to ascertain the effect of glucose on proliferation (C-D). Seven of thirteen KO mice (54%) were frankly hyperglycemic (>200 mg/dl), but a subset of KO mice had BG and insulin levels similar to achieved hyperglycemia in WT and HT mice, allowing comparison.

2 Supplementary Figure 2. Blood glucose, but not plasma insulin, correlates with beta cell proliferation in vivo. (A-B) For IRS2 WT, HT and KO mice, average post time-zero blood glucose (includes 24, 48, 72 and 96 hours of infusion) is plotted against the % BrdU(+) beta cells measured post mortem. In (A) only KO mice with average blood glucose <200 are included to allow better visualization of data across the mild hyperglycemia range; in (B) all mice are included. WT and HT mice were indistinguishable, and both showed a positive correlation between BG and proliferation; this relationship was lost in ice lacking IRS2. (C) Plasma insulin levels (average across hours) did not correlate with proliferation in any genotype.

3 Supplementary Figure 3. Impact of glucose, insulin and rapamycin on IRS2 expression. (A) qpcr was performed on mouse islets cultured for 72 hours in 5mM or 15mM glucose with vehicle control, insulin (100nM) or rapamycin (10nM) as indicated. Glucose increased IRS2 mrna, which was not impacted by rapamycin. Insulin did not increase IRS2 mrna. (B) 72 hour exposure of dispersed mouse islet cells to glucose (15mM) increased IRS2 protein abundance in 4 of 6 islet preps, and decreased or did not change IRS2 abundance in 2 of 6 preps. Date are from the first two columns of the immunoblot shown in Figure 7E, represented as individual mouse islet preps to allow visualization of response of each prep to glucose.

4 Supplementary Figure 4. Insulin signaling inhibitors did not induce islet cell death. (A-B) Whole mouse islets were cultured in 5 or 15 mm glucose for 72 hours in the presence or absence of rapamycin (RAP), PI-103, wortmannin (WM), PD98059 (PD), or, as a positive control, 1000 ng/ml thapsigargin (Tg). Lysates were immunoblotted for activated caspase 3 and actin. Although thapsigargin-treated islets readily induced activated caspase 3, none of the other drugs did. A representative immunoblot from n=3 experiments is shown in A; all experiments are quantified in (B).

5 Supplementary Figure 5. Kinetin Riboside, D-cyclin inhibitor, reduced cyclin D2 levels and proliferation. (A) Whole mouse islets were cultured in 5 or 15 mm glucose for 72 hours in the presence or absence of 15mM kinetin riboside and then lysed for immunoblot. Cyclin D2 induction by glucose was reduced in the presence of kinetin riboside. (B) Dispersed mouse islet cells were cultured for 72 hours as in A, with BrdU added for the final 24 hours of culture. Kinetin riboside completely eliminated beta cell BrdU incorporation, but also altered cell morphology suggestive of toxicity, so these experiments were halted.

6 Supplementary Figure 6. Images for BrdU experiments shown in Figures 7A and 7C

7 Supplementary Figure 7. Cyclin D2 levels in Adeno-cyclin D2 overexpression rescue experiments. (A) Immunoblot for cyclin D2 on islets transduced with adenovirus overexpressing cyclin D2 and/or shrna targeting IRS2, as indicated, again reconfirms increased cyclin D2 levels in 15mM glucose, and confirms high cyclin D2 levels in this rescue model. (B) Immunoblot for cyclin D2 on islets treated with rapamycin and/or adenovirus overexpressing cyclin D2, as indicated, confirms high cyclin D2 levels intended in this rescue experiment as well. High cyclin D2 levels despite IRS2 knockdown or rapamycin treatment are likely related to different regulation of the CMV-cyclin D2 viral construct than endogenous cyclin D2, as demonstrated in Figures 5D and 5M.

8 Supplementary Table 1. Primer sequences for qpcr on mouse islets Supplementary Table 2. Donor information for human islet preparations. Human islets were received from the Integrated Islet Distribution Program (IIDP). Donor and islet isolation parameters are as listed.

9 Supplementary Table 3. Experimental use of each human islet prep