Lipid Nanoparticle Packaging Is an Effective and Nontoxic mRNA Delivery Platform in Embryonic Zebrafish

Abstract

Lipid nanoparticles (LNPs) are an attractive platform for the delivery of therapeutic RNA molecules because LNPs are versatile, have been validated in clinical trials, and are well tolerated. Here, we test whether LNPs can be used to deliver a reporter green fluorescent protein (gfp) mRNA to different tissues in zebrafish embryos. We show that LNP-packaged gfp mRNA can be delivered, through injection, and taken up by cells in multiple tissues in zebrafish embryos without any apparent detrimental effects on embryonic health or survival. Zebrafish embryos injected with LNP-packaged gfp mRNA show subsequent GFP expression in neural, vascular, cardiac, and skeletal muscle tissue, depending on injection site. In contrast, comparable naked (nonpackaged) gfp mRNA injections lead to little or no GFP expression. This study shows that LNPs can be used as an mRNA delivery platform in zebrafish and thus provides a basis for testing the therapeutic functions of LNP-packaged candidate mRNAs in the increasingly diverse array of zebrafish disease models.

Introduction

A large number of diseases, including many rare diseases, are caused by loss or defective expression of proteins due to inherited or acquired genetic mutations. The use of synthetic messenger RNAs (mRNAs), for example, to replace expression of a lost protein, is thus a promising therapeutic approach.^1–3 This is particularly true for diseases that have well-understood genetic defects but lack any effective treatment options. A major hurdle for advancing mRNA therapies involves developing approaches for delivering therapeutic mRNA reagents to the affected tissue.^1,2,4

Lipid nanoparticles (LNPs) are an attractive platform for the delivery of therapeutic RNA molecules because LNPs are versatile, have been validated in clinical trials, and are well tolerated.^4–7 Recent studies illustrate the versatility of the LNP platform. In one example, LNP siRNA formulations successfully prevented mutant transthyretin (TTR) protein deposition in a mouse model of TTR-mediated amyloidosis.⁸ In addition, LNP-encapsulated modified mRNA vaccines encoding Zika virus genes provided protection against Zika virus in mice and nonhuman primates.^7,9

Zebrafish, Danio rerio, are increasingly being used as models for human diseases and for testing potential therapies.¹⁰ Studies using zebrafish are actively contributing to our understanding of diseases ranging from blood, cardiac, liver, metabolic, muscle, and vascular diseases^11–16 to complex brain disorders.^17,18 Some of the unique attributes of zebrafish as a vertebrate animal model include the very rapid, ex utero development of the embryos, transparency of the early embryos, ability to easily obtain hundreds of synchronously staged embryos in a given morning, and the ease with which one can assay the effects of genetic and pharmacological manipulations.

The zebrafish animal model provides an outstanding system with which to test whether specific mRNAs can functionally modulate a disease phenotype. To test mRNAs for potential therapeutic effects in zebrafish, the most widely used approach is microinjection into 1-cell-stage zebrafish embryos to deliver naked synthetic mRNAs. In this approach, the mRNAs are typically injected into the yolk and then, through cytoplasmic flow, the mRNAs are taken up by the embryonic blastomeres as cells undergo their early cleavage from the yolk.^19,20 One example that illustrates the use of this approach has been the rescue of the zebrafish Duchenne muscular dystrophy (DMD) model by heme oxygenase 1a (hmox1a) mRNA.²¹ In this study, overexpression of hmox1a by mRNA injection at 1-cell-stage led to restoration of normal muscle in 4-day old larval DMD zebrafish.²¹ However, 1-cell-stage mRNA injections generally lead to very broad, early, and transient expression throughout the zebrafish embryo and typically do not allow for targeted tissue expression or for expression at later developmental stages. There are many reports of the use of a variety of nanoparticle delivery systems in zebrafish.^22,23 A recent study has shown that PEG-PLA nanoparticles can be used to deliver functional siRNA molecules to cardiac cells in adult zebrafish through microinjections into the adult zebrafish thoracic cavity.²⁴ This work demonstrates that nanoparticles can be used to deliver RNA molecules across cell membranes at late stages and in a tissue-targeted manner. However, because this study targeted siRNAs specifically to injured adult hearts that had undergone ventricular resection, it is not clear whether the nanoparticle delivery of the siRNAs was somehow facilitated by the injury. To our knowledge, no studies have yet investigated the ability of nanoparticles to deliver functional mRNAs to different tissues in the zebrafish embryo.

We want to establish zebrafish as a model for evaluating potential mRNA therapeutics delivered using LNPs. The primary goal of this study was to show that LNP-packaged gfp mRNA can be delivered by injection into zebrafish embryos and show subsequent GFP expression in multiple zebrafish tissues.

Materials and Methods

Naked and packaged gfp mRNAs

Codon-optimized eGFP sequence was generated using a published algorithm.²⁵ Naked gfp mRNA was synthesized in vitro, diluted in water at 1.195 mg/mL, and stored at −80°C. LNP-packaged gfp mRNA was synthesized in vitro and packaged as previously described.⁷ The packaged mRNA was diluted in water at 1.114 mg/mL and was stored at 4°C. The packaged and naked mRNAs were diluted 1:5 from their stock concentrations in phenol red dye (0.1% phenol red and 0.2 M KCl in water) to give a final injection concentration of about 230 ng/μL and a final injection quantity of approximately 230 pg of gfp mRNA per embryo.

Zebrafish husbandry and microinjections

All experiments involving live zebrafish (Danio rerio) were carried out in compliance with IACUC guidelines at Seattle Children's Research Institute. Zebrafish were raised and staged as previously described.¹⁹ Time refers to hours postfertilization (hpf) at 28.5°C. Embryos were collected from the wild-type AB strain. Embryos were raised in ICS water (300 mg Instant Ocean/L, 0.56 mM CaCl₂, 1.2 mM NaHCO₃). Sibling control embryos were not injected with any mRNAs.

All mRNA injections were performed using a Narishige IM 300 Microinjector. Borosilicate glass microcapillary injection needles with filaments (World Precision Instruments, TW100F-4, 1 mm O.D. × 0.75 mm I.D.) were prepared using a micropipette puller device (Sutter Instruments, Inc., Flaming/Brown p-97) with settings aimed at obtaining a long needle tip.²⁶ The needle tip was broken off with fine tweezers to obtain a tip opening diameter of 5–10 μM. Microinjections were set up with an injection pressure of 30–40 psi for a duration of 30 ms to give a final injection volume of 1 nL. Injections into 1-cell stage embryos (0–1 hpf) were performed using agar wells to hold embryos in their chorions.¹⁹ Injections into 24 hpf and 48 hpf embryos were performed similarly to a previously described technique.²⁶ Injections into 24 hpf and 48 hpf embryos were performed by manually dechorionating embryos, anesthetizing embryos with tricaine methanesulfonate (MESAB¹⁹), and individually transferring anesthetized embryos onto a glass depression slide or into the groove of an agarose mold under a stereomicroscope. Most of the media were removed from around the embryo to prevent it from moving around during injections, but some media were left on the slide to prevent it from sticking to the glass. After the injections were performed, embryos were transferred back into ICS water in petri dishes.

Imaging and immunostaining

Following 1-cell stage, 24 hpf, or 48 hpf injections, embryos were allowed to develop further and GFP expression was assessed in live, anesthetized embryos using an Olympus SZX16 stereomicroscope with attached Olympus DP72 camera and the cellSens Dimension imaging software.

Embryos were processed for whole-mount immunostaining as previously described.²⁷ Embryos were fixed in 4% paraformaldehyde overnight at 4°C. Embryos were dehydrated in 100% methanol and stored at −20°C. Fixed embryos were rehydrated from 100% methanol into PBS-Tw (PBS with 0.25% Tween) using the following rinses: 1 × 5 min each: 75% MeOH in PBS-Tw, 50% MeOH, 25% MeOH, PBS-Tw. Embryos were then digested with Proteinase K at room temperature for 75 min (10 μg/mL in PBS-Tw) and washed 3 × 5 min in PBS-Tw to remove residual ProtK. Embryos were then blocked in PBS-Tw with 2% BSA and 2% Goat Serum, at room temperature for 2 h, then put into primary antibody (1:300 anti-GFP, Roche), diluted in block solution, overnight at 4°C. The second day, embryos were rinsed in PBS-Tw 5 × 15 min, then transferred into secondary antibody (goat anti-mouse AlexaFluor-488, 1:300, Life Technologies/Molecular Probes), diluted in block solution, and left at 4°C overnight. The third day, embryos were rinsed 5 × 15 min in PBS-Tw, and embryos were re-fixed in 4% PFA and stored at 4°C. Embryos were mounted in 70% glycerol in PBS under a coverslip and imaged using a Leica TCS SP5 confocal microscope. Images were assembled using Adobe Photoshop. Adjustments of brightness and contrast were applied across whole images similarly within an experimental group.

Results

1-cell-stage yolk injections

As an initial test of both toxicity of packaged gfp mRNA and ability of packaged gfp mRNA to undergo translation in zebrafish embryos, we performed injections at the 1-cell stage (0–1 h postfertilization [hpf]; see Fig. 1A). We targeted these injections to the yolk, from which mRNAs (or other dyes or nucleic acids) are taken up through cytoplasmic streaming into the early blastomeres while the blastomeres are still connected to the yolk.²⁸ 1-cell-stage injections are a standard method for expressing exogenous mRNAs in early zebrafish embryos.^19,20 Wild-type fertilized 1-cell-stage embryos were injected with 1 nL of either packaged or naked gfp mRNA, through the chorion into the yolk (Fig. 1A), and then left to develop until they were examined live for GFP expression at about 24 hpf.

FIG. 1.

Zebrafish embryo microinjection sites used to deliver gfp mRNAs. (A) In 1-cell stage embryos, injection needles target the large yolk. From the 1-cell to at least the 4-cell stage, cytoplasmic streaming will carry injected mRNAs from the yolk into the blastomeres on top of the yolk. Arrow points to 1-cell-stage blastomere. (B) In 24 hpf embryos, we used injection needles to target the hindbrain ventricle, caudal vein, trunk, and pericardial cavity. (C) In 48 hpf embryos, we used injection needles to target the hindbrain ventricle, caudal vein, trunk, and pericardial cavity. In (B, C) anterior is to the left.

Control (noninjected) embryos showed little background, except for some yolk auto-fluorescence, at 24 hpf (Fig. 2A). All embryos injected at the 1-cell stage with naked gfp mRNA showed strong GFP expression throughout most tissues at 24 hpf (Fig. 2B and Table 1). All embryos injected at the 1-cell stage with packaged gfp mRNA also showed GFP expression but, in general, exhibited less robust GFP expression than in embryos injected with a similar amount of naked gfp mRNA (Fig. 2C and Table 1). We observed excellent survival of embryos injected with either naked or packaged gfp mRNA (Table 1). We observed similar results in 2 independent experiments (Table 1). These experiments show that packaged gfp mRNA can be translated in cells of the zebrafish embryo, although possibly not as efficiently as naked gfp mRNA. These experiments also show that the mRNA packaging formulation does not appear to be toxic to early zebrafish embryos. However, because these 1-cell-stage injections are performed before the formation of complete cell membranes, these experiments do not show whether the packaged mRNA formulation can be used to deliver mRNAs across cell membranes within an embryo.

FIG. 2.

Live GFP expression in 24 hpf embryos following 1-cell-stage injections. (A) Control, noninjected embryos, show no GFP expression, with some auto-fluorescence from the yolk. (B) Embryos injected with naked gfp mRNA show very robust GFP expression throughout the embryo. (C) Embryos injected with packaged gfp mRNA show broad GFP expression.

Table 1.

Results from Zebrafish Embryo Injections of Naked or Lipid Nanoparticle-Packaged gfp mRNA

Injection type	# Experiments (Total # embryos)	Average % survival 24 h postinjection	Average % with GFP expression 24 h postinjection
1-cell stage
Noninjected controls	2 (20)	100	0
Naked gfp	2 (52)	93	100
Packaged gfp	2 (92)	94	100
24 hpf
Noninjected controls	6 (80)	100	0
Naked gfp/hindbrain ventricle	5 (85)	85	6¹
Naked gfp/caudal vein	3 (62)	71	7²
Naked gfp/trunk	4 (76)	83	33³
Naked gfp/pericardial cavity	2 (37)	100	4⁴
Packaged gfp/hindbrain	5 (101)	91	73
Packaged gfp/caudal vein	3 (71)	74	65
Packaged gfp/trunk	4 (84)	78	100
Packaged gfp/pericardial cavity	2 (53)	100	90
48 hpf
Noninjected controls	6 (82)	100	0
Naked gfp/hindbrain	5 (80)	96	1¹
Naked gfp/caudal vein	3 (52)	92	39²
Naked gfp/trunk	4 (63)	96	17³
Naked gfp/pericardial cavity	3 (60)	100	25⁴
Packaged gfp/hindbrain	5 (103)	99	86
Packaged gfp/caudal vein	3 (64)	100	63
Packaged gfp/trunk	4 (86)	93	99
Packaged gfp/pericardial cavity	3 (88)	100	97

Expression domains for packaged gfp injections are provided in Table 2.

Footnotes indicate expression domains for naked gfp injections (in addition to any background fluorescence):

About 1 neuroepithelial cell in hindbrain with GFP expression per embryo.

About 1–5 trunk skeletal muscle fibers and/or other cell types with GFP expression, and/or yolk sac GFP expression, per embryo.

About 1–5 trunk skeletal muscle fibers and/or other cell types with GFP expression per embryo.

About 1–6 cells in pericardial sac with GFP expression, and/or yolk sac GFP expression, per embryo.

GFP, green fluorescent protein.

24-hpf-stage injections

Previous studies have shown that it is possible to inject materials into zebrafish embryos through multiple routes, such as trunk skeletal muscle, caudal vein of the circulatory system, and hindbrain ventricle.^26,29,30 We wanted to test whether we could use such injections to deliver gfp mRNA into zebrafish embryos. In particular, we wanted to test whether different microinjection routes enabled delivery of LNP-packaged gfp mRNA to muscle, brain, or other tissues.

To test whether packaged gfp mRNA can be taken up by cells within the zebrafish embryo, we injected embryos at 24 hpf in four sites: the hindbrain ventricle, caudal vein, trunk dorsal to the horizontal myoseptum, or pericardial cavity (Fig. 1B). We chose the hindbrain ventricle and caudal vein as injection sites based on previous literature describing methods for injecting bacteria and nanoparticles into these sites.^26,29,30 We chose the trunk and pericardial cavity injections to test whether the packaged mRNA could be targeted to muscle and cardiac tissue, respectively. Each embryo was injected with 1 nL of either packaged or naked gfp mRNA, for a total of eight injection conditions plus noninjected control groups. Following the injections at about 24 hpf, the embryos were left to develop for another 24 h, anesthetized, and scored for survival and GFP expression at about 48 hpf (Table 1). For 1–2 sets of these 24 hpf injection experiments from Table 1, we scored GFP expression domains in detail at 48 hpf (Table 2).

Table 2.

GFP Expression Domains Observed After Lipid Nanoparticle-Packaged gfp mRNA Injections

Injection type	# Experiments scored (total # embryos with GFP expression)	Expression domains	% Embryos (# embryos) with expression domain
24 hpf
Packaged gfp/hindbrain ventricle	1 (18)	Brain	100 (18)
		Pharyngeal region	22 (4)
Packaged gfp/caudal vein	1 (16)	Vasculature^a	44 (7)
		Cells on yolk and tail	31 (5)
		Somites/skeletal muscle^b	19 (3)
		Yolk sac only	6 (1)
Packaged gfp/trunk	1 (17)	Spinal cord^c,d,e	65 (11)
		Somites/skeletal muscle^b,f	59 (10)
		Vasculature	24 (4)
		Caudal fin	24 (4)
		Skin	12 (2)
Packaged gfp/pericardial cavity	2 (44)	Heart + pericardial sac	98 (43)
		Vasculature^a	52 (23)
		Yolk sac only	2 (1)
48 hpf
Packaged gfp/hindbrain ventricle	1 (14)	Brain	79 (11)
		Spinal cord	50 (7)
Packaged gfp/caudal vein	1 (24)	Somites/skeletal muscle^b	63 (15)
		Vasculature	54 (13)
		Yolk sac	25 (6)
Packaged gfp/trunk	1 (26)	Somites/skeletal muscle^b,g	100 (26)
		Spinal cord^c,d,e	69 (18)
		Caudal fin	38 (10)
		Vasculature	31 (8)
Packaged gfp/pericardial cavity	2 (61)	Heart + pericardial sac	100 (61)
		Vasculature	5 (3)

The Table 2 patterns and frequencies of GFP expression domains, scored in detail for 1 or 2 injection experiments from Table 1, were consistent with those observed in additional experiments.

Circulating GFP+ cell seen in 1 embryo.

Somite-labeled cells include skeletal muscle fibers and other cell types. For caudal vein injections, there were usually only about 1–2 somites with GFP expression.

Three embryos exhibited GFP expression in the brain and the spinal cord.

For both 24 hpf and 48 hpf trunk injections, the average length of the neural tube (spinal cord + brain) exhibiting GFP expression was approximately 75%, with a range of about 25–100%, per embryo.

In some embryos, notochord, in addition to neural tube, exhibited GFP expression.

The average # of somites exhibiting GFP expression was about 3 per embryo, with a range of 1–12 somites per embryo.

The average # of somites exhibiting GFP expression was about 2 per embryo, with a range of 1–7 somites per embryo.

Control (noninjected) embryos showed only some background auto-fluorescence at 48 hpf (Fig. 3A), in particular around the edge of the head (arrow in Fig. 3A’). Injections of naked gfp mRNA into hindbrain ventricle, caudal vein, trunk, or pericardial cavity at 24 hpf led to similar levels of background auto-fluorescence at 48 hpf (Fig. 3B, D, F). With 24 hpf naked gfp mRNA injections, some embryos expressed GFP in a single cell or a small number of cells near the injection site, or in the yolk cell (Table 1 and Fig. 3F). This is likely due to direct injection of a cell or cells at the injection site or to nicking of cells by the injection needle. Because the yolk is a large syncytium, any nicking of the yolk during injection, and subsequent translation of the mRNA, would lead to GFP expression throughout the yolk.

FIG. 3.

Live GFP expression in 48 hpf embryos following 24 hpf-stage injections. The images in the right column are higher magnification views of embryos, shown anterior to the left, from corresponding left column panels. (A-A’) Control, noninjected embryos, show no GFP expression, with some auto-fluorescence from the yolk and along the edge of the head (arrow in A’). (B-B’, D-D’, F-F’) Embryos injected with naked gfp mRNA show little or no GFP expression. An embryo exhibiting yolk sac GFP expression following naked gfp mRNA expression is shown (arrow in F). (C-C’) Embryos injected with packaged gfp mRNA into the hindbrain ventricle show GFP expression in the forebrain, midbrain and hindbrain (demarcated by arrowheads in C’), and also in the pharyngeal region (arrow in C’). (E-E’) Embryos injected with packaged gfp mRNA into the trunk show GFP expression in the spinal cord along the body axis (arrows in E) and in somites around the injection site (arrow in E’). (G-G’) Embryos injected with packaged gfp mRNA into the pericardial cavity show GFP expression in the pericardial sac around the heart (arrow in G’) and also in the heart (the heart GFP expression is very difficult to image with the stereomicroscope due to the strength of the GFP signal in the pericardial sac).

In contrast to the naked gfp mRNA hindbrain ventricle injections, most 24 hpf embryos injected with packaged gfp mRNA in the hindbrain ventricle showed GFP fluorescence within the brain, along the forebrain, midbrain, and hindbrain (demarcated by arrowheads in Fig. 3C’ and Table 1). Some embryos injected with packaged gfp mRNA in the hindbrain ventricle also showed GFP fluorescence within the pharyngeal region (arrow in Fig. 3C’). Embryos injected at 24 hpf with packaged gfp mRNA in the caudal vein showed variable GFP expression in tissues near the injection site, including vasculature and somitic skeletal muscle cells (Table 2; not shown). Embryos injected at 24 hpf with packaged gfp mRNA into the trunk showed GFP expression centered around the injection site along the embryo axis, both in the spinal cord and in the somites, and less frequently in some other tissues as well (Fig. 3E-E’ and Table 2). Embryos injected at 24 hpf with packaged gfp mRNA into the pericardial cavity showed GFP expression in the heart, in the pericardial sac around the heart, and, in some embryos, in the vasculature (Fig. 3G-G’ and Table 2).

These results show that different microinjection routes at 24 hpf enable targeted delivery of packaged gfp mRNA to muscle, brain, and other tissues in the zebrafish embryo. We attribute most of the variability that we observed in the GFP expression domains to technical variation in the precise positioning of the injection needle at the injection sites at 24 hpf. The GFP expression patterns we observed with the 24 hpf injections (Fig. 3 and Table 2) were consistent with where we observed the distribution of the phenol red dye included in the injection solution. For example, immediately upon the 24 hpf trunk injections, we briefly observed the phenol red dye spreading into the lumen of the spinal cord in some embryos and around the somites in some embryos. Similarly, upon the 24 hpf hindbrain ventricle injections, we usually observed the phenol red dye within the hindbrain and midbrain ventricles, but, in some cases, we observed dye spreading outside the brain into the pharyngeal region. Even with this variation, our results show that LNP packaging can robustly deliver gfp mRNA to different tissues in the zebrafish embryo, unlike naked gfp mRNA, which exhibits little if any expression following 24 hpf injections.

We observed good survival of embryos injected with either naked or packaged gfp mRNA (Table 1). Because survival rates are similar after either naked or packaged gfp injections (Table 1), we attribute any reduced survival, compared to the noninjected control embryos, as likely due to damage from the injection needle and not due to toxicity of the LNP formulation. In at least two independent experiments for each of the four 24-hpf-injection sites, we observed similar results for the postinjection survival rates and for the frequencies of GFP expression (Table 1).

For at least one independent experiment of the 24 hpf naked and packaged gfp mRNA injections at each of the four injection sites (n ≥ 13 for each condition), we allowed the animals to continue developing until 7 days postfertilization, and we observed that 100% of animals remained healthy with no obvious developmental malformations or lethality (not shown). In these embryos, we observed that the GFP expression domains for the packaged mRNA-injected embryos were maintained at 7 days postfertilization, as assessed in live embryos using the stereomicroscope (not shown; Fig. 6 below), although the brightness of the GFP expression was noticeably diminished. These experiments show that packaged gfp mRNA can be taken up and translated in cells of the zebrafish embryo with little or no effect on survival.

48-hpf-stage injections

To further test whether packaged gfp mRNA can be taken up by cells within the zebrafish embryo at a more advanced stage, embryos were injected at 48 hpf in the same four sites as at 24 hpf: the hindbrain ventricle, caudal vein, trunk dorsal to the horizontal myoseptum, or pericardial cavity (Fig. 1C). As with the 24 hpf injections, each embryo was injected with 1 nL of either packaged or naked gfp mRNA, for a total of eight injection conditions plus noninjected control groups. Following injections at about 48 hpf, the embryos were left to develop for another 24 h, anesthetized, and scored for survival and GFP expression at about 72 hpf (Table 1). For 1–2 sets of these 48 hpf injection experiments from Table 1, we scored GFP expression domains in detail at 48 hpf (Table 2).

Control (noninjected) embryos showed little background auto-fluorescence at 72 hpf (Fig. 4A-A’). Injections of naked gfp mRNA at 48 hpf led to little if any GFP expression at 72 hpf (Table 1 and Fig. 4D-D’, F-F’). As with the 24 hpf naked gfp mRNA injections, with the 48 hpf naked gfp mRNA injections, some embryos expressed GFP in a single cell or a small number of cells near the injection site, or in the yolk cell (Table 1 and Fig. 4C’, F’). This is again likely due to direct injection of a cell or cells at the injection site or to nicking of cells by the injection needle.

FIG. 4.

Live GFP expression in 72 hpf embryos following 48 hpf-stage injections. The images in the right column are higher magnification views of embryos, shown anterior to the left, from corresponding left column panels. (A-A’) Control, noninjected embryos, show no GFP expression, with some auto-fluorescence from the yolk. (B-B’) Embryos injected with packaged gfp mRNA into the hindbrain ventricle show GFP expression in the forebrain-midbrain-hindbrain (arrow in B’) and in the spinal cord (arrowhead in B’). (C-D’) Embryos injected with naked or packaged gfp mRNA into the caudal vein can exhibit strong GFP expression in the yolk cell (arrowheads in C’ and D’). Caudal vein packaged gfp mRNA injections show variable GFP expression near the injection site, such as in the somites (arrow, D’). (E-E’) Embryos injected with packaged gfp mRNA into the trunk show GFP expression in somites around the injection site (arrows in E’) and also show GFP in the spinal cord (arrowheads in E’). (F-G’) Embryos injected with naked or packaged gfp mRNA into the pericardial cavity show GFP expression in the pericardial sac around the heart (arrows in F’ and G’). Embryos injected with packaged gfp mRNA into the pericardial cavity also show GFP expression in the heart (the heart GFP expression is very difficult to image with the stereomicroscope due to the strength of the GFP signal in the pericardial sac).

In contrast to the naked gfp mRNA injections, 48 hpf embryos injected with packaged gfp mRNA in the hindbrain showed strong GFP fluorescence in the forebrain, midbrain, hindbrain, and, in some cases, spinal cord (Fig. 4B-B’ and Table 2). Embryos injected at 48 hpf with packaged gfp mRNA in the caudal vein showed variable GFP expression in tissues near the injection site, including the somites and vasculature (Fig. 4D-D’ and Table 2). Some embryos injected with packaged gfp mRNA in the caudal vein also showed strong expression of GFP in the yolk sac (Fig. 4D-D’ and Table 2). Embryos injected at 48 hpf with packaged gfp mRNA into the trunk showed GFP expression centered around the injection site along the embryo axis, both in the somites and in the spinal cord, and less frequently in some other tissues as well (Fig. 4E-E’ and Table 2). Embryos injected at 48 hpf with packaged gfp mRNA into the pericardial cavity showed robust GFP expression in the heart, in the pericardial sac around the heart, and, in some embryos, in the vasculature (Fig. 4G-G’ and Table 2).

To more closely examine the cell types expressing GFP following these 48 hpf injections, we fixed embryos at about 72 hpf after the live GFP analysis, stained them with anti-GFP antibody, and then performed confocal imaging (Fig. 5). In control and in naked gfp mRNA-injected embryos, we observed background auto-fluorescence in the yolk and in what appear to be vasculature-related cells, as previously observed³¹ (Fig. 5A, B, D, E, G, H). Embryos injected with packaged gfp mRNA into the hindbrain ventricle showed GFP expression in large clusters of cells in the forebrain, midbrain, and hindbrain (Fig. 5C). Embryos injected with packaged gfp mRNA into the trunk showed GFP expression in somites around the injection site, and on average about 2 somites per embryo showed expression in muscle fibers (Fig. 5F). These embryos also showed strong GFP expression in the spinal cord and in what may be neural crest-derived cells that populate somite boundaries (Fig. 5F). Embryos injected with packaged gfp mRNA into the pericardial cavity showed robust GFP expression in the heart (Fig. 5I). Thus, while the injections of packaged gfp mRNA into the hindbrain ventricle led to broadly labeled cells within the CNS, the injections targeting the trunk and the pericardial cavity generally led to expression in trunk skeletal muscle and cardiac cells, respectively.

FIG. 5.

Confocal images of anti-GFP expression in 72 hpf embryos following 48 hpf-stage injections. (A–C) Dorsal views of optical sections through the head, anterior to the left. (A, B) Control noninjected embryos and naked gfp mRNA-injected embryos show some auto-fluorescence from blood cells and other vasculature-related cells (arrowheads in A and B). (C) Embryo injected with packaged gfp mRNA into the hindbrain ventricle shows GFP expression in clusters of cells in the forebrain, midbrain, and hindbrain (arrows). (D–F) Lateral views of optical sections through trunk, anterior to the left. (D, E) Control noninjected embryos and naked gfp mRNA-injected embryos show some auto-fluorescence from cells in the vasculature and yolk (arrowheads, D, E). (F) Embryos injected with packaged gfp mRNA into the trunk show GFP expression in skeletal muscle fibers (white arrows) in somites around the injection site, in the spinal cord (green arrow), and in cells that populate somite boundaries (yellow arrow). (G–I) Ventral views of optical sections through the pericardial cavity, anterior to the right. (G, H) Control noninjected embryos and naked gfp mRNA-injected embryos show some auto-fluorescence from the yolk, blood cells, and other vasculature-related cells (arrowheads in G and H). (I) Embryo injected with packaged gfp mRNA into the pericardial cavity shows GFP expression in the heart (arrow). The pericardial sac also exhibits GFP expression (arrowhead).

These results show that different microinjection routes at 48 hpf enable targeted delivery of packaged gfp mRNA to muscle, brain, and other tissues in the zebrafish embryo. As with the 24 hpf injections, we attribute most of the variability that we observe in the GFP expression domains to technical variation in the precise positioning of the injection needle at the injection sites at 48 hpf. The GFP expression patterns we observed with the 48 hpf injections (Fig. 4 and Table 2) were again consistent with where we observed the distribution of the phenol red dye included in the injection solution. Even with this variation, our results show that LNP packaging can robustly deliver gfp mRNA to different tissues in the zebrafish embryo, unlike naked gfp mRNA, which exhibits little if any expression following 48 hpf injections.

We again observed good survival of embryos injected with either naked or packaged gfp mRNA at 48 hpf (Table 1). In at least three independent experiments for each of the four 48-hpf-injection sites, we observed similar results for the postinjection survival rates and for the frequencies of GFP expression (Table 1).

For at least one independent experiment of the 48 hpf naked and packaged gfp mRNA injections at each of the four injection sites (n ≥ 15 for each condition), we allowed the animals to continue developing until 7 days postfertilization, and observed that 100% of animals remained healthy with no obvious developmental malformations or lethality (Fig. 6 and data not shown). In these embryos, as for the 24 hpf injections, we observed that the GFP expression domains for the packaged mRNA-injected embryos were maintained at 7 days postfertilization, as assessed in live animals using the stereomicroscope (Fig. 6 shows images of 6-day-old larvae), although the brightness of the GFP expression becomes diminished. These experiments show that packaged gfp mRNA can be taken up and translated in different tissues of the zebrafish embryo with little or no effect on survival and that the GFP expression is maintained for several days.

FIG. 6.

Live GFP expression in 6-day-old larvae following 48 hpf-stage injections. (A–C) Lateral views of 6 dpf (days postfertilization) larvae following 48-hpf hindbrain ventricle injections. (A, B) Control noninjected larvae and naked gfp mRNA-injected larvae show some auto-fluorescence from the yolk (arrowheads in A and B). (C) Larvae injected with packaged gfp mRNA into the hindbrain ventricle show GFP expression maintained in the brain (arrows). (D–F) Ventral views of 6 dpf larvae following 48-hpf pericardial cavity injections. (D, E) Control noninjected embryos and naked gfp mRNA-injected larvae show some auto-fluorescence from the yolk (arrowheads in D and E). (F) Larvae injected with packaged gfp mRNA into the pericardial cavity show GFP expression maintained in the heart and pericardial sac (arrows in F).

Discussion

These experiments show that LNP-packaged gfp mRNA, injected into four different sites, can be taken up and translated in cells of the zebrafish embryo, unlike naked gfp mRNA, which exhibits little if any expression following 24 hpf or 48 hpf injections. We also observe excellent survival following injection of LNP-packaged gfp mRNA into zebrafish embryos, with little or no evidence of toxicity. Our results show that different microinjection routes at both 24 hpf and 48 hpf enable targeted delivery of packaged gfp mRNA to muscle, brain, and other tissues in the zebrafish embryo. We attribute most of the variability that we observe in the GFP expression domains associated with each injection site to technical variation in the positioning of the injection needle within the embryo. Even with this variation, our results show that LNP packaging can robustly deliver gfp mRNA to different tissues in the zebrafish embryo.

We can draw the following conclusions from these experiments. First, these experiments show that different injection sites can be used to target different subsets of tissues in the zebrafish embryo. For example, hindbrain ventricle injections of packaged gfp mRNA readily led to GFP expression within the brain, whereas injections into the pericardial cavity led to GFP expression in the heart and pericardial sac. Second, these experiments suggest that some tissues may be more amenable to taking up the LNPs and subsequently translating the packaged mRNA than others. For example, after both hindbrain ventricle and trunk injections of packaged gfp mRNA, we see broad GFP expression along the neural tube, and not just near the injection site. However, after injections of packaged gfp mRNA at any of the four injection sites we used, we only very rarely see GFP expression in cells moving through the circulatory system (Table 2). Third, we find that, when compared with the 24-hpf packaged gfp mRNA injections, the later 48-hpf injections show similar patterns of, but generally more robust and frequent, GFP expression. This could mean that later, more differentiated cells are more amenable to taking up and translating packaged mRNA, or that the more advanced circulatory system is taking up and distributing packaged mRNA better at 48 hpf.

mRNA therapies are increasingly becoming promising treatment options for many diseases.^1–3 There are several advantages of exogenously delivered mRNA-based therapies over other nucleic acid-based approaches.^1,32,33 First, mRNA has no risk of genomic integration and thus no danger for insertional mutagenesis. Second, mRNA only needs to be delivered to the cytoplasm to be translated, in dividing or in nondividing cells, and does not need to enter the nucleus. Third, the transient nature of mRNA expression may represent a safety advantage.

There are also several challenges for mRNA therapies. Because of mRNA instability, repeated administration may be required. Different cell types may have different translation efficiencies. Also, exogenous mRNAs can trigger an innate immune response, which can negatively affect the therapeutic effects of the mRNA. To address these particular challenges, several different modifications and optimizations can be introduced into the mRNA molecules themselves to increase mRNA translatability and stability and reduce activation of the immune response.^1,2,32,33

A major hurdle for advancing mRNA therapies involves developing approaches for in vivo delivery of therapeutic mRNA reagents, into cells of the affected tissue, with efficiency, stability, and specificity.^1,4,33 Nanoparticle encapsulation of mRNA can protect mRNA from degradation and also can facilitate cellular uptake, and LNPs are a major class of nanoparticle-based delivery vehicles for mRNA. Studies in mice using injections of LNP-packaged mRNAs through different routes showed that administration of mRNA-LNP complexes results in large amounts of protein production in vivo for varying lengths of time.³³ In recent studies, LNP-encapsulated modified mRNA vaccines encoding Zika virus genes provided potent, safe, and durable protection against Zika virus in mice and nonhuman primates.^7,9 Thus, LNPs are useful and effective tools for mRNA delivery and represent a promising approach for advancing mRNA therapies.

Our studies add to the increasing number of reports of the use of nanoparticle delivery systems in zebrafish.^22,23 Lipid-coated, iron-encapsulated carbon nanoparticles, complexed with plasmid GFP DNA, appear to preferentially target the vasculature of zebrafish embryos and also exhibit some toxicity by inducing a low level of various malformations.³⁴ Nitrogen-doped fluorescent carbon quantum dots can enter the zebrafish embryo through soaking and exhibit no detectable toxicity; however, these nanoparticles appear to solely target the yolk, intestine, eye, and melanin-containing pigment cells.²² When injected into zebrafish embryos through the caudal vein, lactic acid and glycolic acid copolymer (PLGA) nanoparticles are readily taken up by macrophages and can successfully deliver antibiotics for treating zebrafish models of bacterial infections, including tuberculosis.^30,35,36 The range of different tissues and cell types in the zebrafish embryo that could be targeted by PLGA nanoparticles was not clear from these studies. In another example of the use of polymer-based nanoparticles in zebrafish, polyethylene glycol-polylactide (PEG-PLA) nanoparticles were used to deliver functional siRNA molecules to adult zebrafish hearts.²⁴ Similar to our findings with LNPs in zebrafish embryos, this study demonstrated that microinjections of PEG-PLA nanoparticles into the zebrafish thoracic cavity can efficiently deliver RNA molecules to cardiac cells.²⁴

Recently, injections of fluorescently labeled calcium phosphate LNPs into zebrafish embryos exhibited extensive labeling across the circulatory system following caudal vein injections, across the brain following brain injections, and across the spinal cord following spinal cord injections.³⁷ Another recent study found that caudal vein injections of different formulations of fluorescently labeled LNPs also led to extensive labeling across the circulatory system, and in some cases outside the vasculature, in zebrafish embryos.³⁸ In addition to these studies, a peptide-based transfection reagent has been used to deliver antisense morpholino oligos to different tissues of the zebrafish larva through intraventricular injections.³⁹ However, none of these studies have investigated the ability of nanoparticles to deliver functional mRNAs to tissues in the zebrafish embryo.

LNP packaging should be an accessible tool for use in a wide array of basic research studies in zebrafish. There are commercially available LNP-mRNA packaging systems, more typically used for cell culture transfection, that are similar to what we used here. LNP packaging could be used to target loss-of-function and gain-of-function RNA reagents to particular zebrafish tissues at different stages of development, thus providing temporal and spatial targeting advantages over the more widely used 1-cell-stage injections. In addition, we expect that LNPs should be useful for delivering CRISPR-Cas9 genome editing components to targeted zebrafish tissues. Indeed, a recent study has used LNP technology to deliver CRISPR components to mouse brains.⁴⁰

In conclusion, we show that injected LNP-packaged gfp mRNA can be translated in zebrafish embryos. These pilot experiments, therefore, demonstrate our ability to deliver expression of exogenous mRNA in zebrafish embryos. Now that we have demonstrated the ability to deliver gfp mRNA to different tissues, we expect that we will be able to deliver other potentially therapeutic mRNAs. Taking advantage of LNP mRNA packaging, and the zebrafish model, to test the efficacies of therapeutic mRNAs should help advance the development of mRNA therapies for human diseases. Our study provides a basis for using LNP delivery to test the therapeutic functions of mRNAs in various zebrafish disease models.

Footnotes

Acknowledgments

We thank Jerry Ament and the SCRI Aquatics Facility staff for expert zebrafish maintenance. This work was supported by the Seattle Children's Myocardial Regeneration Initiative and by the Seattle Children's Office of Science and Industry Partnerships.

Disclosure Statement

D.A. and P.G.V.M. are employees of Moderna Therapeutics, whose focus is the development of therapeutic mRNA approaches for human diseases. The other authors report no competing financial interests.

References

Steinle

, Behring

, Schlensak

, Wendel

, Avci-Adali

. Concise review: application of in vitro transcribed messenger RNA for cellular engineering and reprogramming: progress and challenges. Stem Cells, 2017; 35:68–79.

Kaczmarek

, Kowalski

, Anderson

. Advances in the delivery of RNA therapeutics: from concept to clinical reality. Genome Med, 2017; 9:60.

Weissman

. mRNA transcript therapy. Expert Rev Vaccines, 2015; 14:265–281.

Zatsepin

, Kotelevtsev

, Koteliansky

. Lipid nanoparticles for targeted siRNA delivery - going from bench to bedside. Int J Nanomed, 2016; 11:3077–3086.

Coelho

, Adams

, Silva

, Lozeron

, Hawkins

, Mant

, et al. Safety and efficacy of RNAi therapy for transthyretin amyloidosis. N Engl J Med, 2013; 369:819–829.

Suhr

, Coelho

, Buades

, Pouget

, Conceicao

, Berk

, et al. Efficacy and safety of patisiran for familial amyloidotic polyneuropathy: a phase II multi-dose study. Orphanet J Rare Dis, 2015; 10:109.

Richner

, Himansu

, Dowd

, Butler

, Salazar

, Fox

, et al. Modified mRNA vaccines protect against Zika virus infection. Cell, 2017; 168:1114–1125.e10.

Butler

, Chan

, Costelha

, Fishman

, Willoughby

, Borland

, et al. Preclinical evaluation of RNAi as a treatment for transthyretin-mediated amyloidosis. Amyloid, 2016; 23:109–118.

Pardi

, Hogan

, Pelc

, Muramatsu

, Andersen

, DeMaso

, et al. Zika virus protection by a single low-dose nucleoside-modified mRNA vaccination. Nature, 2017; 543:248–251.

10.

Phillips

, Westerfield

. Zebrafish models in translational research: tipping the scales toward advancements in human health. Dis Model Mech, 2014; 7:739–743.

11.

Avagyan

, Zon

. Fish to learn: insights into blood development and blood disorders from zebrafish hematopoiesis. Hum Gene Ther, 2016; 27:287–294.

12.

Bakkers

. Zebrafish as a model to study cardiac development and human cardiac disease. Cardiovasc Res, 2011; 91:279–288.

13.

Goessling

, Sadler

. Zebrafish: an important tool for liver disease research. Gastroenterology, 2015; 149:1361–1377.

14.

Maves

. Recent advances using zebrafish animal models for muscle disease drug discovery. Expert Opin Drug Discov, 2014; 9:1033–1045.

15.

Seth

, Stemple

, Barroso

. The emerging use of zebrafish to model metabolic disease. Dis Model Mech, 2013; 6:1080–1088.

16.

Wilkinson

, van Eeden

. The zebrafish as a model of vascular development and disease. Prog Mol Biol Transl Sci, 2014; 124:93–122.

17.

Haesemeyer

, Schier

. The study of psychiatric disease genes and drugs in zebrafish. Curr Opin Neurobiol, 2015; 30:122–130.

18.

Kalueff

, Stewart

, Gerlai

. Zebrafish as an emerging model for studying complex brain disorders. Trends Pharmacol Sci, 2014; 35:63–75.

19.

Westerfield

: The zebrafish book. A guide for the laboratory use of zebrafish (Danio rerio). 4th ed., Univ. of Oregon Press, Eugene, 2000.

20.

Yuan

, Sun

. Microinjection of mRNA and morpholino antisense oligonucleotides in zebrafish embryos. J Vis Exp, 2009; 27:pii: 1113.

21.

Kawahara

, Gasperini

, Myers

, Widrick

, Eran

, Serafini

, et al. Dystrophic muscle improvement in zebrafish via increased heme oxygenase signaling. Hum Mol Genet, 2014; 23:1869–1878.

22.

Khajuria

, Kumar

, Karasik

, Gedanken

. Fluorescent Nanoparticles with Tissue-Dependent Affinity for Live Zebrafish Imaging. ACS Appl Mater Interfaces, 2017; 9:18557–18565.

23.

Shaalan

, Saleh

, El-Mahdy

, El-Matbouli

. Recent progress in applications of nanoparticles in fish medicine: a review. Nanomedicine, 2016; 12:701–710.

24.

Diao

, Wang

, Chang

, Zhou

, Zhu

, Wang

, et al. PEG-PLA nanoparticles facilitate siRNA knockdown in adult zebrafish heart. Dev Biol, 2015; 406:196–202.

25.

Welch

, Govindarajan

, Ness

, Villalobos

, Gurney

, Minshull

, et al. Design parameters to control synthetic gene expression in Escherichia coli. PLoS One, 2009; 4:e7002.

26.

Benard

, van der Sar

, Ellett

, Lieschke

, Spaink

, Meijer

. Infection of zebrafish embryos with intracellular bacterial pathogens. J Vis Exp, 2012; 61:e3781.

27.

Bird

, Windner

, Devoto

. Immunocytochemistry to study myogenesis in zebrafish. Methods Mol Biol, 2012; 798:153–169.

28.

Kimmel

, Ballard

, Kimmel

, Ullmann

, Schilling

. Stages of embryonic development of the zebrafish. Dev Dyn, 1995; 203:253–310.

29.

Takaki

, Winglee

, Ramakrishnan

. Evaluation of the pathogenesis and treatment of Mycobacterium marinum infection in zebrafish. Nat Protoc, 2013; 8:1114–1124.

30.

Fenaroli

, Westmoreland

, Benjaminsen

, Kolstad

, Skjeldal

, Meijer

, et al. Nanoparticles as drug delivery system against tuberculosis in zebrafish embryos: direct visualization and treatment. ACS Nano, 2014; 8:7014–7026.

31.

Zeng

, Xu

, Li

, Wen

, Qu

. Label-free in vivo flow cytometry in zebrafish using two-photon autofluorescence imaging. Opt Lett, 2012; 37:2490–2492.

32.

Andries

, Mc Cafferty

, De Smedt

, Weiss

, Sanders

, Kitada

. N(1)-methylpseudouridine-incorporated mRNA outperforms pseudouridine-incorporated mRNA by providing enhanced protein expression and reduced immunogenicity in mammalian cell lines and mice. J Control Release, 2015; 217:337–344.

33.

Pardi

, Tuyishime

, Muramatsu

, Kariko

, Mui

, Tam

, et al. Expression kinetics of nucleoside-modified mRNA delivered in lipid nanoparticles to mice by various routes. J Control Release, 2015; 217:345–351.

34.

Vankayala

, Chiang

, Chao

, Yuan

, Lin

, Hwang

. A general strategy to achieve ultra-high gene transfection efficiency using lipid-nanoparticle composites. Biomaterials, 2014; 35:8261–8272.

35.

Ulanova

, Pinheiro

, Vibe

, Nunes

, Misaghian

, Wilson

, et al. Treatment of Francisella infections via PLGA- and lipid-based nanoparticle delivery of antibiotics in a zebrafish model. Dis Aquat Organ, 2017; 125:19–29.

36.

Vibe

, Fenaroli

, Pires

, Wilson

, Bogoeva

, Kalluru

, et al. Thioridazine in PLGA nanoparticles reduces toxicity and improves rifampicin therapy against mycobacterial infection in zebrafish. Nanotoxicology, 2016; 10:680–688.

37.

Chen

, Watson

, Morsch

, Cole

, Chung

, Saunders

, et al. Improving the Delivery of SOD1 Antisense Oligonucleotides to Motor Neurons Using Calcium Phosphate-Lipid Nanoparticles. Front Neurosci, 2017; 11:476.

38.

Sieber

, Grossen

, Detampel

, Siegfried

, Witzigmann

, Huwyler

. Zebrafish as an early stage screening tool to study the systemic circulation of nanoparticulate drug delivery systems in vivo. J Control Release, 2017; 264:180–191.

39.

Konantz

, Antos

. Reverse genetic morpholino approach using cardiac ventricular injection to transfect multiple difficult-to-target tissues in the zebrafish larva. J Vis Exp, 2014; 88:e51595.

40.

Wang

, Zuris

, Meng

, Rees

, Sun

, Deng

, et al. Efficient delivery of genome-editing proteins using bioreducible lipid nanoparticles. Proc Natl Acad Sci USA, 2016; 113:2868–2873.