Whole organism high content screening identifies stimulators of pancreatic beta-cell proliferation

Abstract

Inducing beta-cell mass expansion in diabetic patients with the aim to restore glucose homeostasis is a promising therapeutic strategy. Although several in vitro studies have been carried out to identify modulators of beta-cell mass expansion, restoring endogenous beta-cell mass in vivo has yet to be achieved. To identify potential stimulators of beta-cell replication in vivo, we established transgenic zebrafish lines that monitor and allow the quantification of cell proliferation by using the fluorescent ubiquitylation-based cell cycle indicator (FUCCI) technology. Using these new reagents, we performed an unbiased chemical screen, and identified 20 small molecules that markedly increased beta-cell proliferation in vivo. Importantly, these structurally distinct molecules, which include clinically-approved drugs, modulate three specific signaling pathways: serotonin, retinoic acid and glucocorticoids, showing the high sensitivity and robustness of our screen. Notably, two drug classes, retinoic acid and glucocorticoids, also promoted beta-cell regeneration after beta-cell ablation. Thus, this study establishes a proof of principle for a high-throughput small molecule-screen for beta-cell proliferation in vivo, and identified compounds that stimulate beta-cell proliferation and regeneration.

Figure 1. Development and characterization of fluorescent ubiquitylation-based cell cycle indicator (FUCCI) for pancreatic beta-cells in zebrafish.

(A) Schematic diagrams of FUCCI constructs for pancreatic beta-cells. The S/G2/M reporter mAG-zGeminin(1/100) and the G1 indicator mKO2-zCdt1(1/190) are expressed under the zebrafish insulin promoter. For efficient selection of transgenic animals, an eye-marker cassette, cryaa:RFP or cryaa:CFP, was introduced into Tg(ins:mAG-zGeminin(1/100)) and Tg(ins:mCherry-zCdt1(1/190)), respectively . (B) Tg(ins:mCherry-zCdt1(1/190),cryaa:CFP)^s948;Tg(ins:mAG-zGeminin(1/100),cryaa:RFP)^s947 larvae were examined at 4.5 dpf using fluorescence microscopy. A close up of the islet is shown in the inset. A majority of the beta-cells are Tg(ins:mCherry-zCdt1(1/190))⁺ indicating that they are in the G1 phase of the cell cycle. Only four beta-cells are Tg(ins:mAG-zGeminin(1/100))^{s947 +} indicating that they are in the S/G2/M phase of the cell cycle. Note that the animals are expressing the eye-marker, e.g., cryaa:CFP fluorescence can be observed through the GFP filter. (C) Time-lapse imaging of Tg(ins:mCherry-zCdt1(1/190))^s948;Tg(ins:mAG-zGeminin(1/100))^s947 larvae at 4 dpf. Arrowheads point to dividing Tg(ins:mAG-zGeminin(1/100))^{s947 +} beta-cells. (D) Tg(ins:mCherry-zCdt1(1/190))^s948;Tg(ins:mAG-zGeminin(1/100))^s947 larvae were incubated with EdU from 3 to 4 dpf. The white arrow points to a Tg(ins:mAG-zGeminin(1/100))^{s947 +} single-positive beta-cell. This cell exhibits high levels of EdU incorporation. The yellow arrow points to a Tg(ins:mCherry-zCdt1(1/190))^s948;Tg(ins:mAG-zGeminin(1/100))^s947 double positive beta-cell which exhibits low levels of EdU incorporation indicating that this cell entered S phase at the end of the EdU labeling period. (E) Confocal stacks of Tg(ins:mAG-zGeminin(1/100))^{s947 +} (green) beta-cells stained for Insulin (blue). The animals were fixed at 12 h intervals until 5 dpf. (Scale bar = 20 µm.). (F) The graph shows a quantification of the number of Tg(ins:mAG-zGeminin(1/100))^{s947 +} beta-cells. Error bars represent SEM; n = 13–15 larvae for each time point. B is a lateral view, anterior to the left and dorsal to the top. C–E show lateral views, anterior to the top and dorsal to the left.

Figure 2. Retinoic acid and Trazodone promote beta-cell proliferation without inducing hyperglycemia.

(A) Confocal images of Tg(ins:mAG-zGeminin(1/100))^{s947 +} beta-cells (green) stained for Insulin (blue). The animals were treated with 1 µM retinoic acid, 10 µM trazodone or 10 µM prednisolone in 1% DMSO (scale bar = 20 µm). (B) Dose–response curves for Tg(ins:mAG-zGeminin(1/100))^{s947 +} beta-cells showing the relationship between the concentration of the compounds and the number of proliferating beta-cells in larvae treated with retinoic acid, trazodone, or prednisolone in 1% DMSO from 3 to 4 dpf. Error bars represent SEM. *P<0.05, and ***P<0.005 compared to vehicle-treated controls; n = 16–20 larvae for each group. (C) Absolute glucose values in zebrafish larvae treated with retinoic acid, trazodone, or prednisolone in 1% DMSO. Error bars represent SEM. *P<0.05, and ***P<0.005 compared to vehicle-treated controls; n = 16–20 larvae for each group. (D) Blood glucose concentration in adult zebrafish treated with 30 µM prednisolone for 24 h. Error bars represent SEM. ***P<0.005 compared to vehicle-treated controls; n = 10 fish for each group.

Figure 3. RA and Prednisolone effectively increase beta-cell proliferation under feeding metabolism.

(A) Schematic diagram for assessment of beta-cell proliferation. At 21 dpf, after feeding from 5–21 dpf with Kyowa N-250 (Kyowa), Tg(ins:mAG-zGeminin(1/100))^s947 larvae were treated with 1 µM retinoic acid or 10 µM prednisolone in 1% DMSO. The number of Tg(ins:H2BGFP)⁺ cells was counted at 22 dpf. (B) Tg(ins:mAG-zGeminin(1/100))^{s947 +} beta-cells were examined at 22 dpf using an epifluorescence microscope. Scale bar = 200 µm. A close up of the islet examined using a confocal microscope is shown in the inset. Scale bar = 10 µm. (C) Quantification of proliferating beta-cells per larva at 22 dpf. Error bars represent SEM. *P<0.05, and ***P<0.005 compared to DMSO-treated controls; n = 13–16 animals for each group.

Figure 4. RA and Prednisolone but not Trazodone, enhance beta-cell regeneration.

(A) Schematic diagram for assessment of beta-cell regeneration. To examine beta-cell regeneration, we made use of the NTR/MTZ beta-cell ablation model. At 80 hpf, after ablating the beta-cells with MTZ from 50–80 hpf, Tg(ins:H2BGFP);Tg(ins:CFP-NTR) larvae were treated with the compounds for 48 h. The numbers of Tg(ins:H2BGFP)⁺ cells were counted at 128 hpf. (B) Confocal images of Tg(ins:H2BGFP)⁺ beta-cells in larvae treated with 1 µM retinoic acid, 10 µM trazodone, or 10 µM prednisolone in 1% DMSO at 128 hpf. Each image is a lateral view, anterior to the bottom and dorsal to the right. (C) Quantification of beta-cell regeneration per larva at 128 hpf, following treatment with hit compounds from 80–128 hpf. Error bars represent SEM. *P<0.05 compared to DMSO treated controls; n = 13–16 larvae for each group.