Drug-Selective Anesthetic Insensitivity of Zebrafish Lacking γ-Aminobutyric Acid Type A Receptor β3 Subunits

Abstract

Background

Transgenic mouse studies suggest that γ-aminobutyric acid type A (GABA_A) receptors containing β3 subunits mediate important effects of etomidate, propofol, and pentobarbital. Zebrafish, recently introduced for rapid discovery and characterization of sedative-hypnotics, could also accelerate pharmacogenetic studies if their transgenic phenotypes reflect those of mammals. We hypothesized that, relative to wild-type (WT), GABA_A-β3 functional knock-out (β3^−/−) zebrafish would show anesthetic sensitivity changes similar to those of β3^−/− mice.

Methods

CRISPR-Cas9 mutagenesis was used to create a β3^−/− zebrafish line. WT and β3^−/− zebrafish were compared for fertility, growth and craniofacial development. Sedative and hypnotic effects of etomidate, propofol, pentobarbital, alphaxalone, ketamine, MS-222, dexmedetomidine, butanol and ethanol, along with overall activity and thigmotaxis were quantified in 7-day post fertilization larvae using video motion analysis of up to 96 animals simultaneously.

Results

Xenopus oocyte electrophysiology showed that the WT zebrafish β3 gene encodes ion channels activated by propofol and etomidate, while the β3^−/− zebrafish transgene does not. Compared to WT, β3^−/− zebrafish showed similar morphology and growth, but more rapid swimming. Hypnotic EC₅₀s [mean (95%CI)] were significantly higher for β3^−/− vs. WT larvae with etomidate [1.3 (1.0 to 1.6) vs. 0.6 (0.5 to 0.7) μM; P < 0.0001], propofol [1.1 (1.0 to 1.4) vs. 0.7 (0.6 to 0.8) μM; P = 0.0005], and pentobarbital [220 (190 to 240) vs. 130 (94 to 179) μM; P = 0.0009], but lower with ethanol [150 (106 to 213) vs. 380 (340 to 420) mM; P < 0.0001] and equivalent with other tested drugs. Comparing β3^−/− vs. WT sedative EC₅₀s revealed a pattern similar to hypnosis.

Conclusions

Global β3^−/− zebrafish are selectively insensitive to the same few sedative-hypnotics previously reported in β3 transgenic mice, indicating phylogenetic conservation of β3-containing GABA_A receptors as anesthetic targets. Transgenic zebrafish are potentially valuable models for sedative-hypnotic mechanisms research.

Introduction

γ-Aminobutyric acid type A (GABA_A) receptors are the major inhibitory neurotransmitter receptors in mammalian brain. Mammalian genes for 19 subunit isotypes have been identified (α1–6, β1–3, γ1–3, δ, ε, θ, π and ρ1–3), with each subunit sharing canonical pentameric ligand-gated ion channel topology: a large (~200 amino acid) extracellular domain, a transmembrane domain with four transmembrane helices (M1 through M4), and a variable size intracellular domain between M3 and M4.^1,2 Most neuronal GABA_A receptors contain two α’s and two β’s along with γ, δ, or another β.¹ GABA_A receptors are modulated by a variety of general anesthetics, including propofol, etomidate, barbiturates, alphaxalone, halogenated volatile agents, and alcohols.^3,4 In typical synaptic αβγ GABA_A receptors, the binding sites for etomidate, propofol, and pentobarbital are formed by portions of the β subunit M1, M2, and M3 transmembrane helices that are 100% conserved among zebrafish, mice, and humans.⁵

Evidence from studies in transgenic mice indicate that major actions of etomidate, propofol, and pentobarbital are mediated by a subset of GABA_A receptors isotypes containing β3 subunits. Compared to wild-type (WT) animals, global β3 knockout (β3^−/−) mice are resistant to the sedative-hypnotic effects, measured as sleep time, of etomidate and propofol, whereas sensitivity to isoflurane is weakly affected and is unchanged for pentobarbital and ethanol.^6,7 Additionally, global β3^−/− mice are phenotypically characterized by over 90% neonatal mortality, cleft palate, epilepsy, hyperactivity, and hypersensitivity to stimuli, possibly reflecting anxiety.^8–11 A pan-neuronal β3^−/− mouse line also displays frequent neonatal mortality and resistance to loss-of-righting-reflexes after etomidate injection.¹² Another transgenic mouse line harbors the β3N265M point mutation, which obliterates sensitivity to etomidate in molecular studies.^13–15 Homozygous β3N265M transgenic mice are characterized by normal fertility, morphology, and behavior, with unchanged sensitivity to alphaxalone or alcohol, but resistance to both loss-of-righting reflexes and loss of nociceptive withdrawal actions of etomidate, propofol, and pentobarbital.^16,17

Zebrafish have recently been introduced as a vertebrate animal model with advantages for pharmacological studies of intravenous sedative-hypnotics.^18,19 Video analysis tools enable simultaneous behavioral assessments of many zebrafish larvae. Larval tissues rapidly equilibrate with aqueous drugs through transdermal and respiratory pathways. Thus, drug effect studies in zebrafish can achieve high-throughput at steady-state, avoiding pharmacodynamic variation due to complex drug pharmacokinetics following intraperitoneal or intravenous drug injections in mice. Moreover, transgenic zebrafish have provided useful models for neurological diseases, including epilepsies, and craniofacial developmental abnormalities.^20,21

For the current study, we established a global β3^−/− line of zebrafish and tested its utility as a robust and efficient animal model for studies of anesthetic drug mechanisms. We hypothesized that global β3^−/− zebrafish would display a pattern of anesthetic sensitivities similar to those of global β3^−/− mice, and might share additional phenotypic features. We therefore characterized β3^−/− in comparison to WT zebrafish for fertility and early survival, craniofacial morphology, motor activity, and sensitivity to both the sedative and hypnotic effects of a panel of anesthetic drugs varying in potency and molecular mechanisms.

Materials and Methods

Animals

Zebrafish (danio rerio, Tübingen strain) were used with approval from the Massachusetts General Hospital Institutional Animal Care and Use Committee (protocol 2014N000031) in accordance with established protocols.²² Behavioral experiments were performed on larvae at 7 days post fertilization (dpf). Sexual differentiation of zebrafish remains indeterminate at this stage of development.²³ Embryos and larvae were maintained in petri dishes (140 mm diameter) filled with buffered E3 medium (in mM, 5.0 NaCl, 0.17 KCl, 0.33 CaCl, 0.33 MgSO₄, 2.0 HEPES, pH 7.2) in a 28.5°C incubator under a 14/10hour light/dark cycle. The density of embryos and larvae was less than 100 per dish. After either use in experiments or at 8 dpf, larvae were euthanized in 0.2% MS-222 (tricaine) followed by addition of bleach (1:20 by volume). Adult zebrafish were briefly anesthetized using 0.02% tricaine before being weighted, imaged, or undergoing tail-fin clipping.

Female Xenopus laevis frogs were used as a source of oocytes with approval from the Massachusetts General Hospital Institutional Animal Care and Use Committee (protocol 2010N000002). Xenopus care and use has been previously described.²⁴

Sedative-Hypnotic Drugs

Etomidate was a gift from Douglas Raines, M.D. (Professor, Dept. of Anesthesia Critical Care & Pain Medicine, Massachusetts General Hospital, Boston, MA USA) and was prepared as a 2 mg/ml solution in 35% propylene glycol: water (by volume). Alphaxalone was purchased from Tocris Bioscience (Bristol, UK) and prepared as a 10 mM stock in DMSO. Ketamine was purchased from Mylan Pharmaceuticals (Canonsburgh, PA USA) as a 10 mg/ml aqueous solution with 0.1 mg/ml benzethonium chloride as a preservative. Dexmedetomidine was purchased from USP (Rockville, MD, USA). Propofol, pentobarbital, ethanol and n-butanol were purchased from Sigma-Aldrich (St. Louis, MO USA). Propofol was prepared as a 100 mM stock in DMSO.

CRISPR Gene Modification and Genotyping

The CRISPR guide RNA expression plasmid pDR274 and Cas9 template DNA pT3TS-nCas9n were purchased from Addgene (Watertown, MA, USA).

CRISPR guide RNA (gRNA) target sites in GABRB3 were identified using CHOPCHOP.²⁵ Complementary DNA sequences for targets in GABRB3 exon 7 were synthesized as oligonucleotides by the MGH DNA core and cloned into pDR274.²⁶ Guide RNAs (gRNAs) were generated by in vitro transcription from linearized pDR274 templates, and purified. The gRNA sequence used to generate the transgenic fish line is reported in Table 1.

Table 1.

CRISPR gRNA and Primer Sequences.

CRISPR gRNA/ Primers	Sequence
CRISPR gRNA	GGGAGGAGAAACCGCAGTGA
GABRB3 Fluorescence PCR Primer_ Forward	AATGGATCATTTACTCTTTGACTGA
GABRB3 Fluorescence PCR Primer_ Reverse	ATGCATCCATAAATTCAACGTG
GABRB3 q-RT PCR Primer_ Forward	CAAGCTAAAAAGAAACATCGGC
GABRB3 q-RT PCR Primer_ Reverse	AGCCAAGAAAACAAAGACGAAG
GABRB1 q-RT PCR Primer_ Forward	CAAGCAACATGTCATACGTCAA
GABRB1 q-RT PCR Primer_ Reverse	CGTAAAGGTCCCTACCAGTGAG
GABRB2 q-RT PCR Primer_ Forward	TTCCTCAACGACAAGAAGTCCT
GABRB2 q-RT PCR Primer_ Reverse	TACTGCAGGGTGGAGCTATCAT
β-actin q-RT PCR Primer_ Forward	GATGCCCCTCGTGCTGTTTTC
β-actin q-RT PCR Primer_ Reverse	TCTCTGTTGGCTTTGGGATTCA

The pT3TS-nCas9n plasmid encoding Cas9 was linearized with XbaI and purified (Wizard SV Gel and PCR Clean-up system, Promega Corp, Madison, WI, USA). Capped cas9 messenger RNA was synthesized in vitro (mMESSAGE mMACHINE, Thermo Fisher Scientific, Waltham, MA) on the linearized template, and purified (NucAway Spin Columns, Invitrogen/Thermo Fisher). One-cell staged zebrafish embryos were micro-injected in the cytoplasm with 2 nl of a solution containing both gRNA (50 ng/μL) and cas9 mRNA (200 ng/μL).

DNA for genotyping was isolated from either whole zebrafish embryos or tail fin clips from adults (> 2 month old), using the HotSHOT method.²⁷ PCR primers flanking the CRISPR target site are reported in Table 1. Fluorescent PCR products were synthesized using a forward primer modified with 6-carboxyflurorescein (6-FAM), and sized to determine if insertions and/or deletions were present after CRISPR mutagenesis. To determine the GABRB3 genotype in single adult fish, tail-snip derived target site PCR products were cloned into the pCR2.1-TOPO vector (Invitrogen/Thermo Fisher), which was then amplified and subjected to DNA sequencing.

Quantitative PCR of GABA_A Receptor β Messenger RNAs

Fifteen 7dpf larvae from each zebrafish line (WT, β3^+/−, and β3^−/−) were euthanized, placed on ice, and homogenized in 1 ml TRIzol (Invitrogen/ Thermo Fisher). Total RNA was isolated using phenol-chloroform exaction, treated with DNase I ²⁸ and quantified using a NanoDrop spectrophotometer (Thermo Fisher). In triplicate samples, WT, β3^+/−, and β3^−/− messenger RNA (1 μg) underwent reverse transcription (SMARTScribe Reverse Transcriptase kit, Takara Bio USA, Mountainview, CA) to produce complementary DNA (cDNA). Quantitative real-time PCR to quantify cDNA encoding β1, β2, and β3 was performed on each triplicate sample, using full-length flanking primers (sequences given in Table 1) and Platinum SYBR Green qPCR SuperMix-UDG with ROX (Invitrogen/ Thermo Fisher) on the StepOne Plus RT-PCR platform (Applied Biosystems/ Thermo Fisher). Beta-actin transcripts were used for normalization and GABRB/β-actin signal ratios were re-normalized to the average WT value. For each subunit, melt-curve analysis and gel electrophoresis were consistent with the presence of a single RT-PCR product. Negative controls were performed on samples lacking either reverse transcription or template.

Next-Generation Sequencing

To test for the presence of chimerism in β3^−/− zebrafish, we isolated genomic DNA from tailfin tissue of 3 adult β3^−/− males and 3 adult β3^−/− females, as described above. Sequences near the gRNA target site were amplified using non-fluorescent primers (Table 1). The amplicons were submitted to the Massachusetts General Hospital Center for Computational and Integrative Biology (CCIB) DNA Core for massively parallel sequencing of single DNA strands using the Illumina MiSeq platform with V2 chemistry (Illumina, Inc., San Diego, CA). Illumina-compatible adapters with unique DNA barcodes were ligated onto each sample during library construction. Multiplexed sequencing of over 100,000 single DNA copies were produced for each amplicon. The sequencing error rate is less than 0.003 per base. Next-generation sequencing results were analyzed by CCIB and reported as the frequency of unique complementary sequence pairs, including clusters of sequences with overlapping ends, from each demultiplexed sample. Sequence pairs occurring less than 100 times were not reported and assumed to be erroneous reads.

Complementary DNA Cloning and Xenopus Oocyte Electrophysiology

We cloned GABA_A β3 cDNA from WT and β3^−/− zebrafish, heterologously expressed the gene products in Xenopus oocytes, and used electrophysiology to compare the molecular function phenotypes of the gene products.

To clone cDNA, we isolated messenger RNA from zebrafish brain tissue using Trizol. RNA was reversed transcribed as described above (see quantitative RT-PCR) and β3 cDNAs were amplified using Phusion High-Fidelity DNA polymerase (Thermo Fisher) and full-length zebrafish-specific β3 cloning primers (forward: AGTTGGTACCGAGCTCGTGCCCCATTTCAAATATTCCGCCTTGG; reverse: ATGCCTCGAGTGCGAGCACGTCCGTAAAGTACATCAGAG). Full-length cDNAs were cloned into pCDNA3.1 plasmids between KpnI and XhoI endonuclease sites. Single clones carrying WT or β3^−/− were amplified and complete cDNA sequences from the purified plasmids were confirmed by Sanger sequencing at the MGH DNA core. Messenger RNA was prepared using T7 mMESSAGE mMACHINE and polyadenylation kits (both from Ambion-Thermo Fisher), purified, and stored in RNAase-free water at −80 °C.

Methods for preparation of Xenopus oocytes, injection of mRNA, and two-micro-electrode voltage-clamp electrophysiology have been previously described.²⁴ We injected oocytes with mRNA (1 ng per oocyte) encoding either WT zebrafish β3 or the β3^−/− gene product. We also tested oocytes injected with mRNA encoding human GABA_A β3 subunits as positive controls and uninjected oocytes as negative controls. To test for expression of functional β3 homomeric channels, oocytes (n = 5 per group) were voltage-clamped at −50 mV and exposed to 10 mM GABA, 100 μM propofol, and 100 μM etomidate.^29,30 In cells producing currents in response to anesthetic applications, we measured the ratio of currents elicited by 100 μM propofol and 100 μM etomidate. We also tested a separate group of oocytes expressing zebrafish β3 for inhibition of currents by 10 μM picrotoxin.

Cartilage staining and imaging processing

To analyze the craniofacial skeleton, Alcian blue staining was performed on five WT and five β3^−/− embryos euthanized at 4.5 dpf, as described previously.³¹ Embryos were mounted in 95% glycerol in 1x phosphate buffered saline with tween 20 (PBST) and images were obtained at 10x and 40x on a Nikon 80i compound microscope (Nikon Instruments Inc., Melville, NY, USA). Images were processed with an NIS-Elements advanced research image acquisition and analysis system (Nikon Instruments), using the maximum intensity projection feature applied to z-stacks. Images for each animal were examined for abnormalities and measurements recorded for palate length and width as well as lower jaw length and width.

Locomotor Activity and Thigmotaxis

Spontaneous locomotor activity of 7dpf larvae was tracked and quantified by Zebralab v3.2 software (Viewpoint Behavioral Systems, Montreal, Canada)in tracking mode, using previously described methods.³² WT and β3^−/− larvae (48 larvae/ group) were loaded individually into wells on 24-well plates (12 of each genotype per plate). After a 15min adaptation to the Zebrabox environment with white light level 110 lux, spontaneous movements were tracked for 1h with detection threshold set at 22 (scale 0 to 200). The distance and duration travelled was recorded and analyzed in three speed categories: fast movements: v ≥ 20 mm/s; slow movement: 20 mm/s ≥ v ≥ 5 mm/s; and non-moving: v < 5 mm/s. Tests were performed between 11:00 AM and 6:00 PM, in order to minimize diurnal variation.

Thigmotaxis, a measure of anxiety, is the tendency of animals to remain near the walls of their environment relative to its central area. To assess this, 1 hr locomotion tracking videos were analyzed to quantify, for each animal, both time spent and distance traveled within vs. outside a central circle with area half that of the total circular well.³³

Zebrafish Larvae Sedation and Photomotor Response Assays

Larval zebrafish from both WT and β3^−/− colonies were tested for sensitivities to both sedative and hypnotic effects of a set of 9 compounds over relevant concentration ranges: etomidate (0.03 to 5 μM), propofol (0.1 to 5 μM), pentobarbital (20 to 400 μM), alphaxalone (0.02 to 4 μM), ketamine (1 to 400 μM), dexmedetomidine (0.3 nM to 20 μM), ethyl-3-aminobenzoate methane thiosulfonate (MS-222 or tricaine; 5 to 500 μM), butanol (0.2 to 40 mM), and ethanol (30 to 400 mM). Both WT and β3^−/− larvae were tested on the same day on a given drug prepared from the same stock solution. For each genotype, six to eighteen larvae were studied at each drug concentration plus a negative (no drug) control group. Drug inhibition of both spontaneous motor activity (sedation) and photomotor responses (PMR; hypnosis) were assessed in separate groups of animals using previously described methods.¹⁸ Briefly, each larva was loaded into a well of a standard 96-well plate containing 200 μL of E3 buffer with or without drug. Plates were placed in the dark chamber of a Zebrabox maintained at 28.5 °C and incubated for 15 min, for drug equilibration and dark adaptation. Following the adaptation period, spontaneous movements of each larvae were quantified during 6 sequential 30 s epochs. In PMR assays, four test runs at 3 minute intervals were performed and recorded as digitized video. Each test run included a 10 s basal motor activity period followed by a 0.2 s white light stimulus (500 lux) and a 5 s post-stimulus period. Motor activity data (Zebralab software v3.2) during all baseline periods for each larva were normalized to 0.2 s epochs and combined to calculate mean, sd, and 95% confidence intervals. The binary PMR in each run for a given larva was considered positive if motor activity during the 0.2 s light flash or the two subsequent 0.2 s epochs exceeded the upper 95% confidence interval for that larva’s basal activity. Cumulative PMR probabilities for each animal were calculated from the four runs.

Statistical Analyses

Spontaneous movement and thigmotaxis metrics, fertility and embryonic viability, body weight and body length for groups of WT and β3^−/− animals were pooled to calculate mean ± sd. Normality of the data distributions was assessed and confirmed using D’Agostino-Pearson tests and statistical comparisons of these characteristics were based on two tailed unpaired Students t- tests. Statistical comparisons of mRNA levels from WT, β3^+/−, and β3^−/− fish were performed using ANOVA with Dunnett’s multiple comparisons tests between WT and the others. In oocytes expressing β3 homomeric channels, the within-cell ratio of electrophysiological responses to 100 μM propofol vs. 100 μM etomidate for human vs. zebrafish (n = 5 each) were compared using an unpaired two-tailed Student’s t-test. For drug-dependent inhibition of PMR probabilities and normalized spontaneous movement, results for all animals in each exposure group were combined and plotted as mean ± sd against log [drug, M]. Logistic functions (Eq. 1) were fitted to all independent data points using non-linear least squares. Sedative and hypnotic EC₅₀s are reported as mean with 95% confidence intervals. EC₅₀ comparisons for WT vs. β3^−/− larvae were based on F-tests with α = 0.05. Linear least-squares fits and statistical comparisons were performed using Graphpad Prism v7.0 (Graphpad Software, La Jolla, CA).

$$\text{Y}=\frac{\text{Max}}{1+{10}^{(\text{logEC}50-\text{log}[\text{ANES}])*\text{nH}}}$$ Eq. 1

The maximum represents either control PMR probability or normalized control spontaneous movement, [ANES] is the anesthetic concentration, EC₅₀ is the half-effect concentration, and nH is the Hill slope.

Results

CRISPR-Cas9 Targeting Zebrafish GABRB3 Exon 7 Creates a Frameshifting Insertion that Destabilizes mRNA and Eliminates β3 Subunit Function

To establish a zebrafish line lacking functional GABA_A receptor β3 subunits, GABRB3 exon 7, which encodes part of the extracellular domain of the β3 subunit, was targeted with CRISPR-Cas9 random mutagenesis. The sequences of CRISPR gRNA and flanking PCR primers are reported in Table 1. Approximately 200 one-cell stage embryos were injected with gRNA and Cas9 messenger RNA. At 24 hours post-injection, fluorescent PCR size analysis of pooled DNA from 10 embryos showed the presence of both insertions and deletions near the target sequence. Viable F0 embryos were raised to adulthood (3 months), when individual tail-snip samples were analyzed with fluorescent PCR sizing. Individuals with insertions or deletions were out-crossed with wild type zebrafish. At 3 months, heterozygous F1 offspring were genotyped at the gRNA target site (Fig 1A). Fluorescence PCR sizing identified several F1 fish of both sexes containing a 10 bp insertion in the GABRB3 exon 7 coding sequence. Subsequent DNA sequencing identified fish of both sexes with identical 10 bp insertions producing a frameshift and a premature stop codon near the gRNA target site, and predicted by in silico translation to truncate β3 peptide within the extracellular domain (Fig.1B). One pair of heterozygous F1 siblings carrying this mutation was in-crossed. The resulting F2 offspring were genotyped based on tail-snips at age 2–3 months, revealing wild type, heterozygous and homozygous mutants at approximately 1:2:1 ratios. Homozygous F2 fish were successfully in-crossed to maintain the transgenic line (Fig. 1A).

Figure 1). Generation and genotype of β3^−/− mutant zebrafish.

A) A flowchart depicting the generation of a zebrafish germline mutation. CRISPR gRNA and Cas9 mRNA was injected into one-cell stage embryos. The injected embryos were raised and outcrossed with wild-type (WT) to generate heterozygous F1 fish. Mutant fish were identified by fluorescence PCR and Sanger sequencing. F1 siblings carrying the same mutation were then crossed to generate F2 progeny, and phenotype- genotype correlations were done using F2 embryos. B) DNA sequencing at the target sequence revealed a 10bp insertion that is predicted by in silico translation to cause a frameshift and a premature stop codon (underlined sequence) downstream of CRISPR gRNA target site, truncating the protein within the extracellular domain (*). C) RT-qPCR revealed that GABRB3 mRNA in homozygotes was significantly degraded. Bar-graphs summarize normalized results (mean ± sd) of triplicate measures with 15 pooled larvae per group. D) Examples of current traces recorded from voltage-clamped oocytes exposed to 100 μM etomidate (ETO; application indicated by bars over traces). Oocytes injected with WT zebrafish (z-fish) β3 mRNA produce large currents, while oocytes injected with mRNA derived from the mutated β3^−/− gene produce no current (note 10-fold amplified current scale). Traces from an oocyte injected with human β3 mRNA and from an uninjected oocyte are shown for comparison. E) Columns represent the ratios of voltage-clamp currents elicited with 100 μM propofol (I_PRO) and 100 μM etomidate (I_ETO) recorded in the same oocyte expressing either zebrafish or human β3 subunits (n = 5 each). The P-value is based on unpaired two-tailed Student’s t-test.

To test for possible chimerism in transgenic zebrafish, 6 individual DNA samples from adults (3 males and 3 females) were amplified using primers flanking the target sequence and analyzed using massive parallel next-generation sequencing. In all 6 samples, a single sequence comprising 99.7 to 100 percent of paired sequence reads (over 50,000 paired reads per sample) was identified. This sequence contained the extra 10 base sequence identified in F2 genotyping (Fig 1B). Thus, chimerism was definitively ruled out in the β3^−/− transgenic line.

Molecular Comparisons of Wild-Type and β3^−/− Zebrafish

Wild-type (β3^+/+), β3^+/−, and β3^−/− zebrafish were characterized at 7dpf for expression of GABRB3 messenger RNA (mRNA), using reverse transcription and quantitative PCR. Negative controls lacking either reverse transcriptase or template produced no detectable signal. Compared to WT larvae, GABRB3 mRNA was significantly reduced in β3^−/−, but not in β3^+/− (Fig. 1C). The reduced mRNA level in β3^−/− larvae is probably due to nonsense-mediated RNA decay triggered by the mutation-induced premature stop codon.³⁴ Quantification of mRNAs for GABRB1 (β1) and GABRB2 (β2) was also performed in the same triplicate samples. Normalized GABRB1 mRNA levels (mean ± sd) were: WT = 1.00 ± 0.072; β3^+/− = 1.21 ± 0.088; and β3^−/− = 1.21 ± 0.16. Normalized GABRB2 mRNA levels were WT = 1.00 ± 0.043; β3^+/− = 0.92 ± 0.16; and β3^−/− = 1.18 ± 0.15. Analysis using ANOVA with Dunnett’s comparison tests indicated no significant differences in GABRB1 or GABRB2 levels between wild-type and β3^+/− or β3^−/− animals.

We used expression in Xenopus oocytes and electrophysiology to compare function of the WT and β3^−/− GABRB3 gene products. All tested oocytes injected with mRNA encoding the WT zebrafish β3 subunit (n = 5) expressed ion channels that produced small (< 50 nA) currents when exposed to 10 mM GABA, and much larger currents when exposed to 100 μM propofol (range 0.4 to 1.6 μA) or 100 μM etomidate (range 1.9 to 5.6 μA; Fig 1D). Picrotoxin (10 μM) inhibited small basal leak currents and etomidate-elicited currents in 5 separate oocytes expressing zebrafish β3 channels. Similar electrophysiological characteristics were found in oocytes injected with mRNA encoding human GABA_A β3 subunits. The within-oocyte current ratios elicited with propofol and etomidate for human vs. zebrafish β3 channels (n = 5 each) were similar (Fig 1E). In contrast, none of the tested oocytes injected with β3^−/− mRNA (n = 5) produced currents that could be distinguished from background noise. Voltage clamp recordings from these oocytes were similar to those from uninjected negative control oocytes (Fig 1D). These results confirm that the mutated GABRB3 gene in β3^−/− zebrafish encodes a non-functional peptide, presumably because the premature stop codon results in truncated subunits lacking the transmembrane domains that form ion channels.

Comparison of Growth and Gross Morphology in Wild-Type and β3^−/− Zebrafish

Neonatal mortality is high and cleft palate is common in global β3^−/− mice.⁸ Mating adult (3 month to 1 year old) β3^−/− zebrafish pairs produced viable embryos in 7 of 19 (24%) trials, compared with 15 of 19 (79%) successful WT matings (p < 0.0001 by Student’s t-test). On average, 30% fewer β3^−/− than wild-type embryos were produced per successful mating and the fraction of viable embryos surviving to 7 dpf was also 30% lower. However, at 3 months of age, no significant differences were evident between WT and β3^−/− in gross morphology, body weight, or body length of either sex (Fig. 2). Comparisons of 4.5 dpf zebrafish (WT vs. β3^−/−; mean ± sd; n = 5 per group; P-values from un-paired Student’s t-tests) for palatal length (410 ± 16 vs. 410 ± 15 μm; P = 0.97), palatal width (400 ± 9.2 vs. 380 ± 16 μm; P = 0.17), lower jaw length (400 ± 11 vs 400 ± 13 μm; P =0.84) and lower jaw width (379 ± 5.9 vs. 377 ± 9.2 μm; P = 0.78) identified no differences in craniofacial morphology (Fig. 3).

Figure 2). Comparative Morphology of 3 Month Post-Fertilization Wild-type and β3^−/− Zebrafish.

A) Representative images of wild-type (WT) and GABA_A receptor β3^−/− males (top) and females (bottom), showing that the mutants have normal overall morphology. Scales are in cm. B) Comparisons of body weight and body length by sex. Black circles are WT, red squares are mutants, each data point represents an individual fish. Mean ± sd (n = 14 per group) are indicated by horizontal lines through data. P values were calculated based on two-tailed Student’s t-tests.

Figure 3). Craniofacial Cartilage in Wild-Type and β3^−/− Zebrafish at 4.5 Days Post-Fertilization.

Cartilage was stained using Alcian blue. Craniofacial structures (A, B lateral; C, D, ventral), flat mount of the mandible (E, F) and the dissected palate (G, H) are shown. Wild-type (A, C) and β3^−/− mutants (B, D) show comparable craniofacial cartilage structures with no difference in the mandible (E, F) or the palate (G, H). Labeled structures: Meckel’s cartilage (m), palatoquadrate (pq), ceratohyal (ch) and ceratobranchial (cb). The scale bar is 200 μm for panels A-F: and 65 μm for panels G-H. WT = Wild-type.

β3^−/− Zebrafish are More Active than Wild-Type

Motor hyperactivity was observed in both global and neuron-selective β3^−/− mice.^8,12 The locomotor activity of WT vs. β3^−/− zebrafish at 7dpf was assessed for both swimming speed and distance during 1 hour. Tracking analysis (Fig. 4A) showed that β3^−/− larvae swam about 15% farther than WT larvae (Fig 4B; Total Distance). This difference was largely accounted for by increased fast swimming (Fig 4B; Fast Distance).

Figure 4). Comparison of spontaneous movement in β3^−/− and Wild-Type larvae.

A) Swimming trajectories of 7dpf larvae during 1 hour. Red: trajectories in v >= 20 mm/s, green: trajectories in 20 mm/s > v >= 5 mm/s. The insert panel shows how the central area is defined, the area of the inner circle corresponded to half of the total area of each well. B) Bars represent mean ± sd for 1 hour swimming distances (n= 48 larvae/ group). C) Bars represent mean ± sd for distance or time spent in the central half of the circular well during 1 hour, expressed as a percentage of the total (n= 48 larvae/ group). Statistical comparisons were based on two-tailed unpaired Student’s t-tests. WT = Wild-type.

Anxiety is another phenotypic feature of global β3^−/− mice.³⁵ Thigmotaxis, defined as maintenance of contact with the environmental periphery and the avoidance of open areas, is an index of anxiety, which is evolutionarily conserved in a wide range of species, including fish, rodents, and humans.^33,36,37 Using measures of both percentage distance and time spent in the central half of their individual wells, both WT and β3^−/− zebrafish larvae displayed similar preferences for swimming near walls (Fig 4C).

β3^−/− Zebrafish Show Reduced Sensitivity to Specific Sedative-Hypnotic Drugs

Sensitivities to nine different sedative-hypnotic drugs, ranging widely in potency and molecular effects, were compared in 7 dpf WT and β3^−/− zebrafish larvae. Drug-induced hypnosis was measured as reduced photomotor response (PMR) probability (Fig. 5) and sedation was measured as reduced spontaneous activity (Fig. 6). Pooled control PMR probabilities for all control WT (mean ± sd = 0.79 ± 0.23; n = 111) and β3^−/− (0.72 ± 0.23; n = 112) larvae did not differ significantly (p = 0.06 by 2-tailed unpaired Student’s t-test). Concentration-response relationships for PMR inhibition by etomidate (Fig. 5A), propofol (Fig. 5B) and pentobarbital (Fig. 5C) differed significantly between WT and β3^−/− larvae, with β3^−/− larvae exhibiting EC₅₀s about 2-fold higher than those for WT (Table 2). In contrast, concentration-dependent PMR inhibition in WT and β3^−/− larvae displayed similar EC₅₀s (Table 2) for alphaxalone (Fig. 3D), ketamine (Fig. 5E), butanol (Fig. 5F), MS-222 (Fig. 5G) and dexmedetomidine (Fig.5H). Interestingly, the EC₅₀ for PMR inhibition by ethanol in β3^−/− larvae was lower than that for WT (Fig. 5I).

Figure 5). Hypnotic Potencies of Anesthetic Drugs in Wild-type vs. β3^−/− larvae.

Photomotor response probabilities were assessed using video analysis of between 88 and 128 larvae per experiment. Plotted symbols represent mean ± 95% confidence interval (n = 8 to 16 larvae per point; error bars are symmetrical; unidirectional error bars were drawn for clarity). Wild-type is black circles; β3^−/− is red triangles. Lines through data represent logistic fits, colored to match symbols. Results of logistic fits are reported in Table 2. Significant differences in fitted EC₅₀s for wild-type vs. β3^−/− larvae were found with etomidate (panel A), propofol (panel B), pentobarbital (panel C) and ethanol (panel I). PMR = photomotor response; WT = wild-type.

Figure 6). Sedative Potencies of Anesthetic Drugs on Wild-type vs. β3^−/− larvae.

Spontaneous movement was assessed using video analysis (72 to 144 larvae per experiment) and normalized to the no-drug control group result for each drug. Plotted symbols represent mean ± 95% confidence interval (n = 9 to 18 larvae per point; error bars are symmetrical; unidirectional error bars were drawn for clarity). Wild-type is black circles; β3^−/− is red triangles. Lines through data represent logistic fits, colored to match symbols. Results of logistic fits are reported in Table 3. Significant differences in fitted EC₅₀s for wild-type vs. β3^−/− larvae were found with etomidate (panel A), propofol (panel B), pentobarbital (panel C) and ethanol (panel I). Spont. = spontaneous; WT = wild-type.

Table 2:

Hypnotic Potencies in Wild-Type vs. β3^−/− Zebrafish Larvae

Anesthetics	Wild-Type -			β3–/–
	EC50 Mean (95% CI)	Hill Slope	N*	EC50 Mean (95% CI)	Hill Slope	N*	P value (EC50s)
Etomidate	0.6 μM (0.5 – 0.7)	−3.5	88	1.3 μM (1.0 – 1.6)	−2.6	96	< 0.0001
Propofol	0.7 μM (0.6– 0.8)	−4.0	96	1.1 μM (1.0 – 1.4)	−4.2	96	0.0005
Pentobarbital	130 μM (94 – 179)	−2.1	101	220 μM (190 – 240)	−4.7	127	0.0009
Alphaxalone	1.1 μM (0.9 – 1.5)	−2.0	128	1 μM (0.8 – 1.3)	−2.6	128	0.47
Ketamine	50 μM (35 – 85)	−1.6	128	60 μM (40 – 89)	−1.3	112	0.77
Butanol	7 mM (4.1 – 11.8)	−1.6	103	8 mM (5.7– 11.6)	−1.2	96	0.66
MS-222	80 μM (56 – 111.0)	−2.1	96	80 μM (64 – 105)	−2.6	64	0.86
Dexmedetomidine	0.4 μM (0.2 – 0.6)	−1.2	96	0.3 μM (0.1 – 0.6)	−0.8	96	0.56
Ethanol	380 mM (340 – 420)	−6.0	96	150 mM (106 – 213)	−2.0	79	< 0.0001

Hypnosis was assessed as probability of photomotor responses.

^*N is the total number of larvae used in a concentration-response experiment at eight (8) different drug concentrations, including a no-drug control.

P values are based on F-tests comparing EC₅₀s in wild-type vs. β3^−/−.

Inhibition of spontaneous motor activity (sedation) was evident at lower concentrations than PMR inhibition (hypnosis) for all tested drugs except ethanol in both WT and β3^−/− larvae (Fig 6; Table 3). Comparing sedative concentration-responses in the two zebrafish lines showed a similar pattern to that for PMR inhibition. Sedative EC₅₀s for etomidate (Fig. 6A), propofol (Fig. 6B) and pentobarbital (Fig. 6C) were higher for β3^−/− than for WT, did not change for alphaxalone (Fig. 6D), ketamine (Fig. 6E), butanol (Fig. 6F), MS-222 (Fig. 6G) and dexmedetomidine (Fig. 6H), and was reduced for ethanol (Fig. 6I).

Table 3:

Sedative Potencies in Wild-Type vs. β3^−/− Zebrafish Larvae

Anesthetics	Wild-Type			β3–/–
	EC50 Mean (95% CI)	Hill Slope	N*	EC50Mean (95% CI)	Hill Slope	N*	P value(EC50s)
Etomidate	30 nM (12 – 73)	−0.8	72	160 nM (121 – 223)	−2.1	96	0.0016
Propofol	120 nM (104 – 132)	−2.2	120	220 nM (196 – 244)	−1.8	120	< 0.0001
Pentobarbital	50 μM (36 – 78)	−1.6	72	140 μM (119 – 176)	−3.0	72	< 0.0001
Alphaxalone	230 nM (186 – 287)	−5.1	144	230 nM (178 – 302)	−1.7	144	0.98
Ketamine	8 μM (5.7– 11.6)	−1.5	96	8 μM (6.1 – 10.7)	−1.0	96	0.98
Butanol	2 mM (1.9 – 2.9)	−2.9	120	2 mM (1.8 – 2.4)	−3.3	120	0.52
MS-222	20 μM (17 – 25)	−2.0	96	23 μM (18.1 – 29.5)	−1.6	72	0.62
Dexmedetomidine	10 nM (4 – 26)	−0.9	144	8 nM (5.7 – 11.5)	−1.2	144	0.69
Ethanol	670 mM (270 – 1630)	−22.9	96	370 mM (324 – 424)	−3.4	96	< 0.0001

Sedation was assessed as spontaneous motor activity normalized to no drug control groups.

^*N is the total number of larvae used in a concentration-response experiment at eight (8) different drug concentrations, including a no-drug control.

P values are based on F-tests comparing EC₅₀s in wild-type vs. β3^−/−.

Discussion

In this study, we aimed to establish whether β3^−/− zebrafish share pharmacogenetic and other phenotypes with β3 transgenic mice. Several transgenic β3 mouse lines have been used to investigate the neurobiological roles of β3, including in general anesthetic actions. These are a global β3^−/− line⁸, two neuron-specific β3^−/− lines¹², and β3N265M knock-in mice¹⁶ harboring a mutation that impairs receptor modulation by drugs that act through some of the multiple modulator sites on GABA_A receptors.⁵ However, mice are poorly suited for pharmacodynamic studies of intravenous anesthetics. Inter-individual variations in drug absorption, distribution, and metabolism after intravenous or intraperitoneal anesthetic injections increase the variation in drug-induced mouse behavioral effects. Consequently, few intravenous anesthetics, each at only a few doses, have been tested in β3 transgenic mice. And among the drugs tested in multiple transgenic β3 mouse lines, inconsistent results were reported for pentobarbital (Table 4).

Table 4:

Summary of Phenotypic Features of β3 Transgenic Mice and Zebrafish

	Transgenic Mouse Type			Zebrafish
Phenotype (relative to WT)	Global β3−/−	Neuronal β3−/−	β3N265M	Global β3−/−
Neonatal Mortality	↑↑	↑	–	↑
Craniofacial Abnormalities	↑	–	–	–
Hyperactivity	↑↑	–/↑ †	–	↑
Epilepsy	↑	–/↑ ††	–	ND
Sensitivity to Sedation, LoRR, or Nociception
Etomidate	↓	↓	↓	↓
Propofol	ND	ND	↓	↓
Pentobarbital	–	ND	↓	↓
Ethanol	–	–	–	↑
Alphaxalone	ND	ND	–	–
MS-222	ND	ND	ND	–
Ketamine	ND	ND	ND	–
Dexmedetomidine	ND	ND	ND	–
Butanol	ND	ND	ND	–
References	6,8,35	12	16,17,42

Cell contents indicate changes in β3 transgenic animal phenotypes relative to wild-type. ND indicates no data; – indicates no change.

^†Hyperactivity was not reported in pan-neuronal β3^−/− mice, but was observed in forebrain-selective β3^−/− mice.

^††Seizures were not observed in pan-neuronal β3^−/− mice, but were observed in forebrain-selective β3^−/− mice. LoRR = loss of righting reflexes.

Zebrafish larvae represent another vertebrate species for assessing the role of β3-containing GABA_A receptors in anesthetic mechanisms, while providing important advantages over mice. Zebrafish are amenable to high-throughput studies quantifying both sedative and hypnotic effects under conditions of steady-state drug exposure.¹⁸ In the current study, we assessed the effects of nine drugs, each at seven or more concentrations, in up to 18 animals per condition, in both wild-type and β3^−/− zebrafish, providing statistically robust EC₅₀ comparisons.

Similarities and Differences Between β3^−/− Zebrafish and Transgenic Mice

Fertility, Survival and Growth

The β3^−/− zebrafish showed reduced fertility and 30% lower survival at 7dpf compared to wild-type. For comparison, neonatal mortality was frequent (~ 90%) in global β3^−/− mice and 61% in neuron-specific β3^−/− mice, but infrequent in β3N265M knock-ins. Neonatal mortality in β3^−/− mice was not strongly linked to cleft palate, but possibly to poor maternal care for pups.¹² Moreover, β3^−/− mice that survived to weaning subsequently grew normally. While zebrafish survive for 7 days on yolk, β3^−/− zebrafish that survived to 3 months of age were morphologically indistinguishable from wild-type, suggesting that any post-larval developmental effects of the genotype were negligible.

Behavior

Our studies demonstrated modestly increased spontaneous motor activity in β3^−/− zebrafish larvae compared with WT (Fig 4). Overt hyperactivity was also reported in both global and forebrain β3^−/− mice (Table 4). Hypersensitivity to handling and stimuli was also reported in these transgenic mice, but in zebrafish, baseline PMR probabilities were similar in WT and β3^−/− larvae, implying normal reactivity to stimuli.

In both global and forebrain β3^−/− mice, overt seizures were observed and electroencephalography confirmed abnormal brain activity in global knock-outs (Table 4). Could the excess motor activity of β3^−/− zebrafish be due to seizures? The hyperactivity we observed in β3^−/− zebrafish was attributable to increased rapid swimming (Fig 4B), which is also increased by exposure to convulsant drugs.³⁸ Thus, β3^−/− zebrafish could experience epilepsy, but we cannot rule out a simple increase in normal fast swimming without convulsions. High-speed videography and specialized behavioral analyses or electroencephalography will be needed to detect whether intermittent seizures are occurring.³⁹

Craniofacial morphology

Craniofacial abnormalities observed in global β3^−/− mice were not evident in β3^−/− zebrafish at 4.5 dpf, when bony structures are established (Fig 3). Interestingly, cleft palate is not evident in neuron-specific β3^−/− mice, and it appears that non-neuronal β3 GABA_A receptors influence murine palate development.⁴⁰ Significant linkage disequilibrium has been reported between the human GABRB3 gene and cleft lip/palate.⁴¹ The lack of such linkage in zebrafish illustrates the limitations inherent in animal models.

Sensitivities to Sedative-Hypnotics

Both global β3^−/− mice and zebrafish larvae share one pharmacogenetic characteristic: insensitivity to etomidate (Table 4). However, the β3^−/− genotype in mice vs. zebrafish is associated with divergent sensitivities for pentobarbital and ethanol. After either pentobarbital or ethanol exposure, global β3^−/− and wild-type mice sleep for similar durations, while β3^−/− zebrafish displayed lower sensitivity to pentobarbital and higher sensitivity to ethanol in comparison to wild-type.

The drug sensitivity profile of β3^−/− zebrafish larvae more closely resembles that of β3N265M mice (Table 4). In comparison to the respective wild-type controls, both β3^−/− zebrafish and β3N265M mice displayed reduced sensitivities to etomidate, propofol, and pentobarbital, but similar sensitivity to alphaxalone. The only notable difference among drug sensitivities is for ethanol. While β3^−/− zebrafish larvae exhibited hypersensitivity to ethanol in assays for both sedation and hypnosis, WT and β3N265M mice displayed similar sensitivities to ethanol-induced loss-of-righting-reflexes (hypnosis).⁴²

At the molecular level, β3N265M mutations impair GABA_A receptor modulation by etomidate and propofol, drugs that bind in the two β+/α–outer transmembrane inter-subunit sites in typical synaptic αβγ receptors.⁵ Importantly, αβ3N265Mγ receptors retain sensitivity to alphaxalone, which binds at distinct β+/α–interfacial sites that are adjacent and intracellular to etomidate/propofol sites.⁴³ The case of pentobarbital is more complex. Pentobarbital inhibits photolabeling by both etomidate analogs (in β+/α–sites) and a potent barbiturate, R-mTFD-MPAB that selectively binds in α+/β–and γ+/β–transmembrane sites.⁴⁴ Pentobarbital effects in β3 homomers, but not in α1β3 heteromeric receptors are reduced by β3N265S mutations³⁰, while the effect of β3N265M on pentobarbital modulation of αβγ receptors remains unknown. However, β3N265M modestly reduces R-mTFD-MPAB modulation of α1β3γ2L receptors⁴⁵ and β3N265M mice exhibit normal sensitivity to hypnosis induced by R-mTFD-MPAB injections.⁴⁶

The pharmacogenetic phenotype of β3^−/− zebrafish larvae both confirms and extends inferences drawn from previous molecular and transgenic animal studies. Our results confirm that β3-containing GABA_A receptors mediate important behavioral effects of etomidate, propofol, and pentobarbital. We observed no reduction in sensitivity to MS-222, ketamine or dexmedetomidine in β3^−/− larvae, consistent with evidence that these anesthetics minimally modulate GABA_A receptors while affecting other molecular targets.⁴ While ethanol, butanol, and other alcohols modulate GABA_A receptors and multiple other anesthetic targets, our current results indicate that β3-containing GABA_A receptors are not important mediators of their sedative-hypnotic effects. At present, we have no basis for speculating on why the β3^−/− genotype differentially affects ethanol sensitivity in fish vs. mice. The normal sensitivity of β3N265M mice to alphaxalone was previously interpreted as reflecting the unaltered molecular sensitivity of β3N265M receptors to neuroactive steroids.¹⁶ However, our β3^−/− zebrafish studies suggest that alphaxalone effects are mediated through mechanisms entirely independent of the β3 subunit. Other transgenic mouse studies implicate GABA_A receptors containing δ subunits, which are usually extra-synaptic, in anesthetic actions of neurosteroids.⁴⁷

Limitations of This Study

One limitation in comparing our results with those in transgenic mice is the lack of tests for nociceptive responses. Manual methods for this outcome in zebrafish have been reported.¹⁹ High throughput methods using electric shock or photo-switchable irritant chemicals may be employed for this purpose.⁴⁸ We did not test zebrafish sensitivities to volatile anesthetics, which are readily studied at steady-state in mice. Modification of zebrafish test equipment may extend our experimental repertoire to accommodate metered anesthetic gas exposure. Additionally, we have not tested drug sensitivities in more mature WT and β3^−/− zebrafish, which could differ from those in larvae. Our high-throughput testing approach is not amenable to studies of mature adult zebrafish.

Animals may compensate for gene knock-outs by over-expressing similar gene products. We tested β3^−/− zebrafish for compensation by quantifying β1 and β2 mRNAs. In WT zebrafish, levels of mRNA for β2 are around 3-fold higher than for β3, so compensatory overexpression of β2 may be difficult to detect.⁴⁹ Zebrafish also possess another gene for a GABA_A receptor β4 subunit. We aim to determine whether silencing other β subunits affect zebrafish sensitivity to intravenous anesthetics. Another strategy that minimizes genetic compensation is to incorporate point mutations like β3N265M that selectively alter anesthetic sensitivity of GABA_A receptors. Creating knock-ins in zebrafish with CRISPR-Cas9 has proven extremely difficult, and reliable strategies are being sought.⁵⁰

Summary and Conclusions

Global β3^−/− zebrafish and mice display more differences than similarities among the morphological, behavioral, and pharmacodynamic phenotypes that we assessed (Table 4). Unexpectedly, the phenotype for β3^−/− zebrafish most closely resembles that of homozygous β3N265M knock-in mice, which are characterized by normal growth and neuromotor behavior, while displaying insensitivity to specific general anesthetics including etomidate, propofol, and pentobarbital.

Several other transgenic mouse lines have provided information on molecular targets mediating important anesthetic actions in vertebrates.^3,47,51 However, because of limitations discussed above, few intravenous anesthetics have been tested in these mouse lines. Development of additional transgenic zebrafish could help determine the molecular targets that mediate anesthetic drug actions.

Our current results, along with other recent studies,^18,19 show how transgenic zebrafish may accelerate both drug discovery and mechanisms research related to general anesthesia. A direct extension of this current work is identification of novel sedative-hypnotics that depend on β3-containing GABA_A receptors. Zebrafish also represent a model for studying neural circuit activity, and experiments in transgenic fish may help reveal where specific targets in specific neural circuits mediate key anesthetic actions.

Summary Statement.

Zebrafish larvae lacking GABA_A receptor β3 subunits displayed normal development and selective insensitivity to sedation and hypnosis induced by etomidate, propofol, and pentobarbital. Thus, β3-containing GABA_A receptors mediate important actions of these intravenous anesthetics.