Phenotypic screening in zebrafish larvae identifies promising cyanobacterial strains and pheophorbide a as insulin mimetics

Abstract

Diabetes is a pandemic disease that causes the loss of control of glucose regulation in the organism, in consequence of dysfunction of insulin production or functionality. In this work, the antidiabetic bioactivity of 182 fractions from 19 cyanobacteria strains derived from the LEGE Culture Collection were analysed using the 2-NBDG assay in zebrafish larvae. From this initial screening, two fractions (57 (06104_D) and 107 (03283_B)) were identified as promising insulin mimetics. These were further characterized by measuring glucose levels in whole larvae, the expression of glucose transporters (GLUT 1-3) using western blot, and the mRNA expression levels of the glut2, pepck, and insa genes using real-time qPCR. Both fractions showed a decrease in free glucose levels. Furthermore, exposure to fraction 06104_D decreased GLUT1 and increased insa mRNA levels. The chemical composition of these fractions was determined using LC-HRESIMS/MS and compared to inactive fractions of the same polarity in order to identify the unique bioactive molecules. The molecular networks constructed using the GNPS platform revealed that fraction 06104_D contained mass clusters primarily composed of chlorins, lipids, and terpenoids, while fraction 03283_B contained xanthophylls, peptides, and terpenoids. To correlate the observed activity with the chemical composition of fraction 06104_D, pheophorbide a was chosen as a representative of chlorophyll derivatives. Exposure to zebrafish larvae at 10 and 20 µM confirmed the increased glucose uptake on the 2-NBDG assay. These findings highlight the bioactivity of chlorophyll derivatives as insulin mimetic compounds, as well as cyanobacteria as a source of potential therapeutic diabetes applications.

Keywords: Cyanobacteria; Diabetes; Glucose; Metabolomics; Pheophorbide a; Zebrafish.

Fig. 1

Screening for insulin mimetic activity in zebrafish larvae exposed to cyanobacterial fractions using the 2-NBDG assay. Percentage of glucose uptake relative to solvent control (DMSO 0.1%) was evaluated. 182 fractions were screened and fluorescence was quantified in the yolk sac (A) and in the eye (B). Fractions with an increase > 20% are shown in blue, and > 50% in green. Promising fractions from the initial screening were tested in an additional assay. Fractions marked in green increased significantly 2-NBDG uptake in the yolk sac (C) or eye (D). The data is represented as box-whisker plots from the fifth to 95th percentiles. Asterisks highlight significantly altered fluorescence intensities that indicate changes in 2-NBDG uptake (* p < 0.05), Kruskal-Wallis and Dunn’s test. E) Representative images from fluorescence microscopy: solvent control (DMSO, 0.1%); positive control (emodin, 10 µM) and fraction 57, 10 µg/mL.

Fig. 2

Glucose measurement on 4 dpf zebrafish embryos, exposed to Fraction 06104_D and 03283_B (at 10 µg/mL and 20 µg/mL), as well to the solvent control, DMSO (0.1%), and positive control, T3 (10 µM). Three replicates were used in two independent assays, n = 6. The data is represented as box-whisker plots from the fifth to 95th percentiles. Asterisks show significant differences vs. the solvent control (*** p < 0.001; ** p < 0.01; * p < 0.05), Kruskal-Wallis and Dunn’s test.

Fig. 3

The evolutionary history of glucose transporters (GLUT 1–3) with the sequences of Mus musculus, Homo sapiens and Xenopus tropicalis was inferred by using the Maximum Likelihood method based on the JTT matrix-based model. The tree with the highest log likelihood (−4071.74) is shown. The percentage of trees in which the associated taxa clustered together is shown next to the branches. The analysis involved 23 amino acid sequences. All positions containing gaps and missing data were eliminated. There was a total of 260 positions in the final dataset. Evolutionary analyses were conducted in MEGA7 ¹⁷.

Fig. 4

Western blot analysis of glucose transporters (GLUTs). A and B: GLUT expression from whole zebrafish embryos (3 dpf) and from head, and body, including the yolk sac. The expression of each GLUT and respective β-actin on the membrane is shown on A (top GLUT, bottom β-actin). B: quantification of the relative expression of each GLUT normalized by β-actin. C: GLUT protein level from zebrafish embryos (3 dpf) normalized to β-actin, after exposure to fractions 06104_D and 03283_B at 10 µg/ml and Emodin at 10 µM, using DMSO (0.1%) as solvent control. Three replicates consisting of a pool of 16 larvae each were used per condition in two independent assays (n = 6). The data is represented as box-whisker plots from the fifth to 95th percentiles, and asterisks highlight significant differences vs. the solvent control (*** p < 0.001; ** p < 0.01; * p < 0.05), One-Way ANOVA and Dunnett’s posthoc test.

Fig. 5

Relative mRNA expression of insa, pepck and glut2 from zebrafish embryos (Danio rerio), after 24 h of exposure to solvent control (DMSO 0.1%) and fractions 06104_D and 03283_B at 10 µg/mL. Data are expressed as mean ± SD (n = 6 for each group), using box-whisker plots from the fifth to 95th percentiles. Asterisks highlight significantly altered mRNA expression (* p < 0.05), One-Way ANOVA and Dunnett’s posthoc test.

Fig. 6

Cytoscape visualization of the main clusters of the final molecular network of fraction 06104_D (TOP) and 03283_B (BOTTOM), for each GNPS analysis approach. Compound families’ identifications were done using ClassyFire. In the classic GNPS approach, larger nodes are masses that are unique to the bioactive fraction, whereas in the bioactivity-based molecular network approach, the size of the nodes is proportional to their significance to the observed bioactivity. Each color represents a fraction with the same solvent polarity from cyanobacteria strains belonging to the genus Nodosilinea. The dark blue color corresponds to the positive hits, fractions 06104_D and 03283_B, respectively, while the others are from non-active fractions. Networks derived from Cytoscape 3.10.0 23.

Fig. 7

2-NBDG uptake on zebrafish larvae exposed to pheophorbide a, a member of the chlorophyll derivatives. Zebrafish larvae were exposed to pheophorbide a at 10 µM and 20 µM for 1 h, and 2 independent assays were done for each sample (n = 16). The data is represented as box-whisker plots from the fifth to 95th percentiles. Asterisks highlight significant differences vs. the solvent control, DMSO 0.1% (** p < 0.01; * p < 0.05), One-Way ANOVA and Dunnett’s posthoc test.

Fig. 8

Solvent gradient used for VLC fractionation of crude extracts of the 19 cyanobacteria strains. Solvent proportions are identified by the letters A-I as shown in the figure.