Glucose inhibits haemostasis and accelerates diet-induced hyperlipidaemia in zebrafish larvae

Hyperglycaemia damages the microvasculature in part through the reduced recruitment of immune cells and interference with platelet signalling, leading to poor wound healing and accelerated lipid deposition in mammals. We investigated the utility of zebrafish larvae to model the effect of exogenous glucose on neutrophil and macrophage recruitment to a tail wound, wound-induced haemostasis, and chicken egg yolk feed challenge-induced hyperlipidaemia by supplementing larvae with exogenous glucose by immersion or injection. Neither method of glucose supplementation affected the recruitment of neutrophils and macrophages following tail transection. Glucose injection reduced thrombocyte retention and fibrin plug formation while only thrombocyte retention was reduced by glucose immersion following tail transection. We observed accelerated lipid accumulation in glucose-injected larvae challenged with high fat chicken egg yolk feeding. Our study identifies conserved and divergent effects of high glucose on inflammation, haemostasis, and hyperlipidaemia in zebrafish larvae compared to mammals.

www.nature.com/scientificreports/ immune systems to mammals, and these have been used to provide insight into the shared function of these systems across vertebrates 19,20,23 . Here we have combined these models to investigate the role of exogenous glucose supplementation in disrupting thrombosis and lipid accumulation, but not immune cell recruitment to a wound, in zebrafish larvae.

Results
Exogenous glucose exposure increases glucose in zebrafish larvae. To establish the efficacy of the injection and immersion techniques to increase glucose concentrations in 5 dpf zebrafish larvae, we conducted a glucose oxidase assay on homogenised whole zebrafish larvae using the Sigma Amplex Red Glucose oxidase kit. Consistent with past literature, we observed an increase in the glucose concentration contained within the glucose-injected and -immersed larvae compared to controls (Fig. 1A, B) 24,25 . Glucose immersion caused sporadic microbial overgrowth and we observed reduced growth of larvae immersed in either glucose or mannitol, but not in glucose-injected larvae, as measured by total body area (Fig. 1C, D) or eye area (Fig. 1E, F).
We next analysed the effect of glucose supplementation on the development of key innate immune cells. We estimated the quantity of macrophages in transgenic Tg(mfap4:turquoise) xt27 larvae, where macrophages are marked by turquoise fluorescent protein, and found similar numbers of macrophages in larvae that were injected with or immersed in glucose (Fig. 1G, H). We estimated the quantity of neutrophils in Tg(lyzC:DsRed) nz50 larvae, where neutrophils are marked by DsRed fluorescent protein, and found reduced numbers of neutrophils in glucose-injected larvae but similar numbers of neutrophils in glucose-immersed larvae (Fig. 1I, J).
Glucose does not affect neutrophil and macrophage recruitment to wounds in zebrafish larvae. Altered innate immune cell recruitment to wounds is a conserved feature of hyperglycaemia in mammals 4,5,[26][27][28] . To determine if this phenomenon was conserved in zebrafish larvae, we utilised the tail transection wound model which causes reproducible leukocyte recruitment 29 ( Fig. 2A). We first performed this assay using transgenic Tg(mfap4:turquoise) xt27 larvae to quantify macrophage recruitment to the tail wound (Fig. 2B). Surprisingly, we observed no difference in macrophage recruitment between the glucose-injected and control larvae at 6 h post wounding (hpw) (Fig. 2C). We also observed no difference in macrophage recruitment in the glucose immersion model (Fig. 2D).
We then used Tg(lyzC:DsRed) nz50 larvae to quantify the recruitment of neutrophils to the tail wound (Fig. 2E). We observed great variability between experiments with two out of four experiments finding significantly reduced neutrophil recruitment between the control and glucose-injected larvae at 6 hpw and two out of four experiments finding significantly increased neutrophil recruitment (Fig. 2F). Consistently, we did not observe any difference in neutrophil recruitment in the glucose immersion model (Fig. 2G).
Together, these results indicate that zebrafish neutrophil and macrophage recruitment is not affected by exogenous glucose supplementation in zebrafish larvae.
To determine if the reduction in thrombocyte recruitment was mirrored by perturbed clotting, we conducted tail transections on Tg(fabp10a:fgb-gfp) mi4001 larvae (Fig. 3E), which express fluorescently tagged fibrinogen and allow visualisation of clots 32 . Glucose-injected Tg(fabp10a:fgb-gfp) mi4001 larvae had reduced fibrin accumulation www.nature.com/scientificreports/ following tail transection compared to PBS-injected larvae (Fig. 3F). We did not observe any effect of glucose immersion on the deposition of fluorescently tagged fibrinogen (Fig. 3G).
Together, these results demonstrate that a exogenous glucose supplementation inhibits the thrombocyte component of haemostasis in zebrafish larvae.
Injection of glucose accelerated hyperlipidaemia in zebrafish larvae challenged with a high fat diet. Hyperglycaemia and hyperlipidaemia are intimately associated in mammals 33 . To determine if this interaction is conserved in zebrafish larvae, we fed glucose-injected larvae a high fat diet consisting of emulsified chicken egg yolk from 5 to 7 dpf (Fig. 4A). Glucose-and PBS-injected larvae had similar Oil Red O vascular staining prior to the initiation of chicken egg yolk feeding at 5 dpf (Fig. 4B, C). Quantification of Oil Red O staining revealed glucose-injected larvae had increased vascular lipid content at one and two days post feeding (Fig. 4D, E).
These results recapitulate the interaction between hyperglycaemia and hyperlipidaemia seen in mammals 15 , demonstrating that zebrafish larvae are a robust model to study this conserved interaction. www.nature.com/scientificreports/

Discussion
In this study, we explored the effect of exogenous glucose supplementation on inflammation, thrombosis, and lipid accumulation in the zebrafish embryo model. We found that exogenous glucose reduces haemostasis around wound sites. In addition, we determined that injection of glucose accelerates lipid accumulation when larvae are challenged with a high fat diet. The reduced size of glucose and mannitol immersed larvae compared to control larvae is a major caveat when interpreting datasets generated with the glucose immersion model. We thus preferred the injection method for the chicken egg yolk challenge experiments and for the interpretation of the effects of glucose on larval zebrafish immunity and haemostasis. Despite daily changes of media, we experienced further difficulty rearing zebrafish larvae from 2 to 5 dpf in 5% glucose solutions due to microbial overgrowth which may have slowed larval development by consuming oxygen or damaging larval mucosal surfaces. Although the glucose injection model did not achieve as high a fold change in glucose at 5 dpf, we did not observe developmental delays and haematopoiesis was largely comparable to PBS-injected control larvae.
Hyperglycaemia in mammals causes vascular dysfunction that restricts the recruitment of a broad range of immune cells to wounds 16 . It was therefore surprising that exogenous glucose appeared to have no effect on the recruitment of leukocytes to the tail wound in this study. It is possible that exogenous glucose supplementation www.nature.com/scientificreports/ to zebrafish larvae does not affect abluminal crawling of leukocytes along blood vessels, since wound-responsive leukocytes have been demonstrated to move predominantly through interstitial tissue in zebrafish larvae 5,34 . Overall our findings suggest exogenous supply of glucose to zebrafish larvae may not be a suitable platform for studying the impact of hyperglycaemia on leukocyte biology. The relationship between hyperglycaemia and hyperlipidaemia has been previously reported in various mammalian species 15,25,35 , and a recent study by Wang et al. has demonstrated a similar the interaction of a high cholesterol diet with glucose immersion on vascular lipid accumulation in zebrafish after 10 days of feeding 25 . Our study using chicken egg yolk as a high fat diet challenge demonstrates a dramatically accelerated accumulation of lipid after just one day of feeding. This chicken egg yolk feeding-based model, therefore provides a more rapid model to investigate lipid accumulation.
In summary, we report glucose-supplemented zebrafish larvae as a tractable platform to investigate the conserved interactions between glucose and haemostasis, and glucose and diet-induced hyperlipidaemia.

Induction of high glucose concentrations in the zebrafish. Injection method: Eggs were injected
with approximately 15 nmol of glucose from a filter-sterilised 2 M solution dissolved in PBS, or an equal volume of PBS as a control within four hours of fertilisation.
Immersion method: Dechorionated 2 dpf embryos were immersed in 5% mannitol (osmolarity control) or 5% glucose dissolved in E3 media, media was changed daily to reduce microbial growth.
Glucose oxidase assay. Whole larvae were snap frozen and stored at − 20 °C for quantification of glucose content. Larvae were lysed in the buffer solution from the Amplex Red Oxidase kit (Sigma: A22188), lysates were then sedimented by centrifugation, and the Amplex Red Glucose oxidase kit was used to measure the glucose content of the supernatant with a microplate reader in accordance with the manufacturer's instructions.
Tail transection assays. Caudal fin amputations were performed on larvae at 5 dpf. Larvae were anesthetised with 2.5% (v/v) ethyl-3-aminobenzoate methanesulfonate (tricaine) (Sigma, E10521), wounded posterior to the notochord using a sterile scalpel and kept in a 28 °C incubator to recover as previously described 29 . Wounded larvae were imaged at 6 h (neutrophil and macrophages) or 2.5 h (fibrin and thrombocytes).