Experimental flow chart. The flow chart describes the workflow, exclusion criteria, and the final number of larvae included in the statistical analysis in all experiments. (a) Transgenic (Tg[−1.2insH2b:mCherry; 2.8fabp10a:GFP]) larvae (AB) were fed on either a control amount (n = 350), 3x as much (n = 366) from day 5 to day 9 post fertilization. (b) Zebrafish orthologues of 12 established human obesity genes were targeted using a multiplexed CRISPR/Cas9 approach. Two sets of CRISPR/Cas9 founders for eight (arid5b, mc4r, lepr, negr1, pcsk1, pomca, pomcb, sec16b) and seven (bdnf, irs1, irs2a, irs2b, sh2b1, sim1a, sim1) orthologues were generated using fertilized eggs from parents with fluorescently labelled liver and pancreatic beta cell nuclei. Founders were raised and in-crossed, and F₁ larvae were overfed and phenotypically characterized at 10 days post-fertilization (dpf) for 14 traits, using semi-automated fluorescence microscopy and biochemistry. (c) Fertilized eggs from parents with fluorescently labelled liver and pancreatic beta cell nuclei were microinjected at the single cell stage with a duplex gRNA targeting kita (controls) or additionally with a duplex gRNA to generate affected larvae for the candidate gene of interest. Larvae were overfed from day 5 to day 9. Larvae were imaged on day 10 or 11 and/or collected for biochemistry. (d) A duplex gRNA targeting the kita gene was injected in fertilized eggs from AB parents to generate controls; while affected larvae were generated by micro-injecting up to three duplex gRNAs targeting early exons or domains of each orthologue of the human gene of interest, alongside the duplex gRNA targeting kita. Larvae were overfed from day 5 to day 7. At day 8, larvae were fed with fluorescently labeled food and imaged to quantify food intake.

Heatmap visualizing the effects of all genetic and dietary perturbations on obesity-related traits in zebrafish larvae. Numbers in cells are effects of exposure vs. controls expressed in z-scores, so values <−1.96 and > 1.96 reflect effects with p < 0.05. The study is divided into five main parts, examining the effect of: (1) overfeeding in larvae free from CRISPR/Cas9-induced mutations; (2) genetic burden scores comprising the summed number of mutated alleles across the eight (Multiplex score 1) and seven (Multiplex score 2) genes targeted simultaneously. Mutations in each allele are weighted by their predicted effect on protein function. Larvae from both multiplexes are included in the regression analysis and the effect of Multiplex score 1 is adjusted for the weighted effect of mutations in genes targeted in Multiplex 2 and vice versa; (3) the effect of CRISPR/Cas9-induced mutations in zebrafish orthologues of human obesity genes on obesity-related traits in offspring of CRISPR/Cas9 founders of Multiplexes 1 and 2; (4) the effect of CRISPR/Cas9-induced mutations on cardiometabolic outcomes in CRISPR/Cas9 founders targeted at orthologues of one human gene at a time; and (5) the effect of CRISPR/Cas9-induced mutations on food intake in CRISPR/Cas9 founders (AB). * effect of each additional mutated allele weighted by mutations’ predicted effect on protein function in the F₁ generation. ** effect of frameshift and/or premature stop codon-introducing mutations in both alleles vs. no mutated alleles in the F₁ generation. *** effects in CRISPR/Cas9-edited larvae vs. controls in the F₀ generation.

Lipid deposits observed in several anatomical regions in 10-day-old zebrafish larvae. a) Whole-body bright field image of a zebrafish larva in lateral orientation. The white rectangular box indicates the location of magnified images in the lower panel; the white dotted line depicts the liver area and the yellow dashed line indicates the edge of the operculum. b-e) Fluorescence microscopy images at 10x magnification of 10-day-old, Tg:2.8fabp10a:GFP positive zebrafish larvae illustrating the location of dye-stained lipid droplets (highlighted by arrows). The majority of the lipids were stored in the visceral region (abdominal or pancreatic, blue arrow; cardiac, red arrow), followed by the cranial region (yellow arrow), and the appendicular region (grey arrow). f) Results from inspection of 275 10-day-old imaged larvae raised under normal feeding conditions (n = 129) or overfed three-times more (n = 146) for the presence of lipids in adipocytes. Percentages are shown in parenthesis. g) In Multiplexes 1 (left) and 2 (right), a total of 108 (28%) and 62 (16%) larvae had quantifiable lipid deposits in at least one anatomical region, respectively. Most lipids were stored in the abdominal region, followed by the opercular region.

The effect of CRISPR/Cas9-induced mutations on lipid deposition in adipocytes and ectopic regions in 10-day-old zebrafish larvae. Results are shown for genes with data from (1) a comparison of CRISPR/Cas9 edited founders (F0 generation) for all zebrafish orthologues of a human gene vs. sibling controls (open circles) and/or (2) from a comparison of offspring from an in-cross of CRISPR/Cas9 founders (F₁ generation) with frameshift and/or premature stop codon introducing mutations vs. larvae free from CRISPR/Cas9-induced mutations in the gene (filled circles). Dots and error bars show odds ratios (logistic regression, for lipid deposition in adipocytes, left), effect sizes expressed in standard deviation units (linear regression, for hepatic lipid deposition, middle), and incident rate ratios (negative binomial regression, for vascular lipid deposition, right) and their 95% confidence intervals. In the F₀ experiment, effects are adjusted for batch, tank and time of day at imaging. In the F₁ experiment, effects are adjusted for multiplex (1 or 2), batch, time of day at imaging, and the number of mutated alleles in the remaining CRISPR/Cas9-targeted genes weighted by the predicted effect of such mutations on protein function (synonymous 0.33, missense 0.66, frameshift and/or premature stop codon introducing 1.00). Effects with P < 0.05 are shown in red. Of note: F₁ results for irs2 (filled circles) are for irs2b only, since the data does not facilitate a comparison of larvae with frameshift and/or premature stop codon introducing mutations in both irs2a and irs2b vs. controls.

The effect of CRISPR/Cas9-induced mutations in 13 obesity genes on food intake in 8-day-old CRISPR/Cas9 founders. Dots and error bars show effect sizes and 95% confidence intervals for affected larvae (number on the left) vs. controls (number on the right). The effect of being in the first quartile of food intake [Q1]; left); the effect of being in the third quartile (Q3; middle); the effect of being in the top quartile (Q4; right). All are compared with the second quartile (Q2) as a reference. The results for irs2 and lep are derived from the concurrent targeting of both orthologues (i.e., irs2a and irs2b; lepa and lepb). Effects have been adjusted for batch, tank and time of day at imaging. Effects with P < 0.05 are shown in red.