Propofol Immersion as a Euthanasia Method for Adult Zebrafish (Danio rerio)

Abstract

The exponential rise of the zebrafish (Danio rerio) as a model organism in biomedical research has far outstripped our understanding of basic husbandry and welfare for this species. As a case in point, here we investigate the efficacy and welfare impact of different euthanasia methods for zebrafish. Not only is a humane death central to welfare and the 3Rs, but stress during euthanasia can change scientific outcomes. However, the most frequently used methods of euthanasia have multiple shortcomings with regard to animal welfare and human safety. In this study, we propose the use of propofol for immersion euthanasia of adult zebrafish. Propofol has been known to rapidly induce anesthesia in many species, including zebrafish, but its efficacy as a euthanasia agent for zebrafish has not fully been explored. In this study, adult zebrafish were euthanized by immersion on one of 5 different preparations: ice bath, 250 ppm MS222, 600 ppm lidocaine hydrochloride, 100 ppm propofol, or 150 ppm propofol for 20 or 30 min. Display of aversive behaviors, time to loss of righting reflex, time to cessation of opercular movement, and time to recovery after transfer to clean tank water were assessed and recorded. Propofol at both concentrations induced loss of righting reflex and loss of opercular movement more quickly than did MS222 or lidocaine hydrochloride and caused no display of aversive behaviors as seen with ice bath or lidocaine exposure. However, fish exposed to propofol at either concentration for 20 min sometimes recovered, whereas a 30-min exposure was sufficient for euthanasia of all fish tested. These findings suggest that exposure to propofol for a duration of at least 30 min quickly and effectively euthanizes adult zebrafish without compromising end-of-life welfare.

Introduction

Since emerging as a model organism of interest in the 1980s, use of the laboratory zebrafish (Danio rerio) in biomedical research has exponentially increased. Today the zebrafish is considered a species of primary importance by the American College of Laboratory Animal Medicine and is used by over 1,500 laboratories worldwide for studying development, neurobehavior, genetic disorders, novel therapeutic drugs, host-microbe interactions, and more.^10,39 Despite their popularity in research, little is known about how finfish such as zebrafish process and experience aversive events. However, strong and ample evidence indicates that they predictably react and respond to noxious stimuli,^9,26,31-33 and that nociception leads to a subjective experience of pain similar to that of humans and other mammals.²⁵ Additional evidence indicates that commonly used anesthesia and euthanasia drugs may elicit aversive behaviors and avoidance, suggesting compromised welfare.^30,38 Furthermore, stress from husbandry and experimental procedures often alters experimental results.¹² For zebrafish, as for other animals, refinement of procedures such as euthanasia are essential to ensuring both good welfare and good science.

Accordingly, the American Veterinary Medical Association (AVMA) calls for a conservative approach to the euthanasia of fish that recognizes their potential to experience and perceive pain in a way analogous to other vertebrates and emphasizes reducing potential distress as much as possible.¹ Although these guidelines are primarily welfare-focused, other factors are also important when selecting a euthanasia method. These include speed, efficacy, tissue toxicity, ease of use, and potential occupational risks. AVMA-recommended methods of euthanasia appropriate for zebrafish include a variety of anesthetic chemicals acceptable for immersion and single-step physical methods. Of these, immersion in buffered tricaine methanesulfonate (MS222) has historically been the most frequently used method of euthanasia for adult laboratory zebrafish worldwide despite potential hazards to laboratory personnel and reports of both aversive behaviors and variable efficacy.^7,22,34,37 Lidocaine hydrochloride immersion has recently been recommended as a safer, faster, and less aversive chemical means of euthanasia.^6,7 Of the acceptable physical methods, rapid chilling to 2 °C to 4 °C (35.6°F to 39.2°F) in an ice bath is rapid, effective, and safe for personnel. However, zebrafish display a brief period of erratic hyperlocomotion after exposure to an ice bath that is suggestive of aversion and can be distressing for observers.²² In addition, this method is not approved for use in all countries.^19,36

Although not currently labeled for use in aquatic species, propofol immersion has been suggested as a potential method of anesthesia and euthanasia of finfish.^3,4,27 Propofol is a commonly used anesthetic induction agent in both veterinary and human medicine. It quickly exerts sedative-hypnotic effects by binding to γ-aminobutyric acid receptors. High doses of propofol result in pronounced vasodilation, respiratory depression, and apnea in mammalian patients after loss of consciousness; if not corrected, these effects can be fatal.^11,35 Prior studies in our department have suggested that immersion in high doses of propofol will effectively euthanize adult zebrafish,⁴ but this finding has not been confirmed in a rigorous comparison to other accepted euthanasia methods.

Therefore, the goal of this study was to assess propofol immersion as a method of euthanasia for adult zebrafish as compared with other AVMA recommended chemical (buffered MS222, lidocaine hydrochloride) and physical (rapid chilling) methods. An ideal method of euthanasia for zebrafish would be pain- and distress-free for the animals, cause fast, irreversible loss of consciousness with minimal histologic change, and be safe and easy to use for laboratory personnel. Because previous work has focused on handling safety and tissue toxicity associated with these methods, the current study compared their efficacy, speed, and displays of aversive behavior.^6,15,37 We hypothesized that immersion of adult zebrafish in propofol at the tested concentrations would result in irreversible loss of consciousness faster and with fewer displays of aversive behavior than other tested means.

Materials and Methods

Ethics statement.

This study was designed with the 3Rs in mind to test a potential refinement over current methods of zebrafish euthanasia. All procedures were approved by the Administrative Panel on Laboratory Animal Care at Stanford University, which is an AAALAC-accredited institution.

Animals and husbandry.

Fifty-eight mixed-sex 5-mo-old wild-type AB zebrafish from a single batch of embryos (Sinnhuber Aquatic Research Laboratory, Corvallis, OR) were used for this study. All fish used were surplus sentinels and had not previously been used for invasive research; no fish were purchased for this study. Active zebrafish racks in this facility are regularly screened by PCR for infectious agents (Aeromonas hydrophila, Edwardsiella ictaluri, Flavobacterium columnare, Mycobacterium spp., Ichthyophthirius multifiliis, Pleistophora hyphessobryconis, Pseudocapillaria tomentosa, Piscinodinium pillulare, Pseudoloma neurophilia, and Saprolegnia brachydanis) and no potentially pathogenic organisms had been identified on this rack at the time of this experiment. All fish used were apparently healthy and behaving appropriately for the species.

Zebrafish were group housed at a stocking density of 5 to 10 fish per liter on a single recirculating aquaculture system rack (Aquaneering, San Diego, CA) in a housing room with a 14h:10h-light:dark cycle. The system rack was fitted with a heater, 50-µm prefilter, 25-µm mechanical filter, fluidized bed biologic filter, carbon filter, and UV lamp providing a minimum of 100,000 mW/s/cm2. A 10% water volume change was performed daily with calcite-filtered reverse osmosis water. Water temperature and chemistry were maintained within standard ranges established as appropriate for this species (25 to 29 °C, pH 6.5 to 8.5, dissolved oxygen > 6 mg/L, conductivity 500 to 2,000 µS, ammonia < 0.02 ppm, nitrite < 0.5 ppm, and nitrate ≤ 50 ppm). Fish were fed a diet of artemia (E-Z Egg, Brine Shrimp Direct, Odgen, UT), hatched in-house, and commercially available powdered feed (GemmaMicro 300, Skretting USA, Tooele, UT) twice daily.

Experimental design.

Zebrafish were netted arbitrarily from their home tanks and assigned randomly to one of 7 euthanasia groups as described in Figure 1. Randomization was performed using the =RAND() function in Microsoft Excel with a numerical descriptor assigned to each euthanasia preparation. A total of 8 or 9 fish were assigned to each group.

Figure 1.

Euthanasia preparations, manufacturers, exposure times, temperatures, and concentrations.

Euthanasia stock preparations were established by mixing selected euthanasia agents with tank water to achieve the desired temperature (2 to 4 °C in the case of ice water) or concentration (for all other preparations). The pH of each preparation was measured before use and confirmed to be within tolerable ranges for zebrafish (6.0 to 8.5).2 Each fish was netted and placed individually in a 1-L clear, polycarbonate tank (Tecniplast USA, Inc., West Chester, PA) containing 0.5 L of the assigned euthanasia preparation. Fish in the ice water group were protected from direct contact with ice by placing them above a slotted spawning barrier that allowed chilled water, but not ice, to pass into the area containing the fish.

All fish were observed throughout the procedure by one of two observers who were experienced with zebrafish euthanasia. Inter-rater reliability was established via duplicate scoring of a training set of animals euthanized via MS-222 prior to initiation of this study; values were consistent between observers without additional training. Due to obvious visual differences between contexts (opacity, visible ice, and spawning barrier) observers were aware of the treatment. Displays of aversive behaviors (erratic movement, hyperlocomotion, piping at the surface, twitching) were noted, as were time to onset of anesthesia (loss of righting reflex, LORR) and respiratory arrest (cessation of operculation, COO).^5,18,21 The minimum time for LORR and COO was set to 5 s for all groups to account for human delay in tracking fish in the opaque propofol preparations.

Each animal was immersed in its assigned preparation for a predetermined period and then removed by net, rinsed with clean tank water, and transferred to a recovery tank on the original rack. Fish were monitored for signs of life (righting, operculation, body movement, response to tank tap) once every 15 min for 90 min after placement in the recovery tanks. Successful euthanasia was defined as reaching stage IV anesthesia that was irreversible for at least 90 min after transfer from the euthanasia preparation to fresh system water.⁵ Any animal that demonstrated signs of life in the recovery tank was noted and immediately euthanized by rapid chilling.

Statistical analysis.

All initial tests were performed in JMP Pro 14 for Windows (JMP, Cary, NC) with further post hoc tests performed in SAS 9.4 for Windows (SAS, Cary, NC) as needed. No data points were excluded.

To test whether the concentration and exposure time of the propofol treatments influenced whether fish would recover signs of life, a logistic regression was used with recovery after removal to the recovery tank (yes or no) as outcome variables, and concentration and exposure time as predictors.

To test for the effect of euthanasia method on time to LORR and time to COO we used a general linear model with euthanasia method as the single predictor variable. Significant results were tested posthoc using the Tukey adjustment for multiple comparisons. The assumptions of general linear models (normality of error, homogeneity of variance, and linearity) were tested posthoc after analyses.¹⁴ As one would expect for latencies, these variables were log-normal distributed and were logged for analysis to meet these assumptions.

To test for an effect of euthanasia method on aversive behavior (which was scored yes or no), logistic regression and likelihood ratio tests were used. These tests were also used to determine whether time to LORR and time to COO were predictors of aversive behavior.

Results

Efficacy of euthanasia method.

All fish that were exposed to ice water, lidocaine, or MS222 for 20 min were effectively euthanized. However, a 20-min exposure to propofol was associated with a 18.75% failure rate, with 3 fish reviving (1 of 8 at 100 ppm, and 2 of 8 at 150 ppm); the remainder of the animals in propofol groups were effectively euthanized. A 30-min exposure to propofol at either concentration effectively euthanized all fish tested. Statistically, for all fish submerged in propofol, recovery was not predicted by dose (LR-ChiSq = 0.4168; P = 0.5185) but was predicted by exposure duration (LR-ChiSq = 4.896; P = 0.0269).

Onset of anesthesia.

Time to LORR was significantly affected by euthanasia method (GLM: F6,51 = 108.5; P < 0.0001; Figure 2) with propofol at both tested concentrations and ice water resulting in rapid LORR in all fish within 10 s. Posthoc Tukey tests showed that lidocaine took significantly longer to induce LORR than did the other treatments. MS222 similarly failed to induce LORR as quickly as either propofol or ice water. There was no significant difference in time to LORR between propofol at either concentration and ice water.

Figure 2.

Rapid chilling or exposure to propofol at both tested concentrations induces loss of righting reflex significantly more quickly than do either MS222 or lidocaine hydrochloride. Time to Loss of Righting Reflex differed significantly between treatments (P < 0.0001). Data were log transformed for analysis and are plotted on log scales as LSM ± SE. Treatments with the same superscript do not differ significantly (posthoc Tukey test).

Respiratory arrest.

Time to COO was significantly affected by drug treatment (GLM: F6,51 = 189.8; P < 0.0001; Figure 3). Posthoc Tukey tests showed that MS222 and lidocaine took significantly longer to induce COO than did the other treatments, followed by the propofol, and with ice being significantly faster than the other treatments.

Figure 3.

Exposure to propofol at both tested concentrations results in respiratory arrest more quickly than do MS222 or lidocaine, but less quickly than rapid chilling. Time to Cessation of Operculation differed significantly between treatments (P < 0.0001). Data were log transformed for analysis and are plotted on log scales as LSM ± SE. Treatments with the same superscript do not differ significantly (posthoc Tukey test).

Aversive behavior.

Potential for aversive behavior was significantly affected by Drug Treatment (Logistic Regression: LR-ChiSq = 33.51; P < 0.0001). Neither time to LORR (LR-ChiSq = 2.408; P = 0.1207) nor time to COO (LR-ChiSq = 0.4983; P = 0.4803) predicted aversion. All (8 of 8) fish treated with ice showed aversive behavior, some (2 of 8) fish treated with lidocaine showed aversive behavior, and no (0 of 42) fish in the remaining treatments showed aversive behavior. Further posthoc tests could not be calculated due to the extreme nature of these differences.

Discussion

Although our hypothesis was incorrect in terms of irreversible loss of consciousness at the 20-min time point, our data suggest that both 100-ppm and 150-ppm propofol preparations are sufficiently potent to quickly induce an unconscious and apneic state in adult zebrafish, with a minimum 30-min exposure time necessary to achieve death in all fish tested. Shorter exposure times risk inducing an anesthetic state that is deep but reversible, as evidenced by recovery of some fish after a 20-min exposure to both tested propofol concentrations. Our data do not support a previous study that reported death of all adult zebrafish that were exposed to a 120-ppm propofol preparation for 10 min.⁴ Although not seen in the current study, zebrafish were also reported to recover after exposure to MS222 for less than 30 min after cessation of operculation.³⁷ Consequently, current AVMA euthanasia guidelines recommend that when performing euthanasia of finfish using chemical immersion as a sole method, fish should remain immersed for at least 30-min after cessation of operculation. Our current work supports this recommendation. Alternatively, propofol immersion may be used to rapidly achieve deep anesthesia prior to physical methods of euthanasia such as maceration, decapitation, or freezing.

The anesthetic effects of propofol have been established in a variety of mammalian and nonmammalian species including aquatics, but the mechanism by which it causes apnea and death in finfish is unclear.^3,27,35 Propofol exerts a profound anesthetic effect via increasing GABA-mediated inhibitory tone within the CNS, thus provoking rapid loss of consciousness and depression of somatosensory responses.^11,35 In addition, in healthy mammals, respiration is primarily driven by elevated concentrations of carbon dioxide in arterial blood. In these species, propofol inhibits this drive and leads to dramatic dose-dependent respiratory depression and a risk of prolonged apnea.^11,24 Conversely, respiration in teleost fishes is thought to be primarily driven by response to the oxygen concentration in the water; however, some evidence indicates that both pH-responsive and carbon dioxide-responsive ventilatory mechanisms may also exist in these species.^13,17 Therefore, the respiratory depression seen in finfish exposed to propofol may be due to either the suppression of carbon dioxide driven mechanisms as in mammals or to other neurologic effects.

Our study found that zebrafish exposed to propofol develop LORR and COO significantly faster than those exposed to either lidocaine or MS222. Achieving deep anesthesia quickly limits the amount of time that the conscious zebrafish is exposed to the euthanasia preparation and is at risk of experiencing pain or distress. In addition, it reduces the necessary active participation time of staff, especially when propofol immersion is used to produce deep anesthesia before using a secondary physical method. In our experience, lidocaine took significantly longer to induce LORR and COO than was previously reported at this concentration.⁶ However, others have reported that at higher doses (1,000 ppm lidocaine and 500 ppm MS222) and with supplemental buffering, lidocaine and MS222 can both produce stage IV anesthesia in adult zebrafish relatively quickly.²⁰ The concentration of MS222 used in the current study (250 ppm) is the lowest concentration generally recommended for zebrafish euthanasia; had a higher concentration been used, a more rapid onset of anesthesia may have occurred.

Our data supported our hypothesis that propofol immersion would elicit fewer displays of aversive behavior than other methods of euthanasia. Due to the opaque nature of propofol, we were limited in this study to assessing only gross behaviors associated with aversion and were unable to screen for finer measures of distress such as an increased operculation rate. No overt behavioral indicators of distress were seen in zebrafish exposed to propofol prior to loss of consciousness. Conversely, all fish exposed to ice water and one quarter of the fish exposed to lidocaine demonstrated evidence of distress including erratic fast movements and piping at the surface of the water.^18,21 These reactions are well-documented for both rapid chilling and lidocaine.^6,30,37 None of the fish exposed to MS222 demonstrated evidence of distress despite previous publications suggesting that MS222 exposure is aversive to zebrafish.^30,37,38 This incongruity may be due to differences in buffering practices between studies or to individual animal variations. Given that evidence of animal distress during euthanasia is a major welfare and scientific concern and can contribute to compassion fatigue in animal facility personnel, our results indicate that propofol offers a significant advantage over other tested methods.²³

In comparison with MS222, propofol has additional advantages in terms of safety and ease of preparation. Propofol preparations may be prepared without use of a hood or supplemental PPE as the liquid drug does not pose a significant inhalation or contact hazard for laboratory personnel.¹⁶ Because propofol is manufactured at a neutral pH for intravenous use, propofol dilutions do not need buffering. The preparation of lidocaine hydrochloride and ice preparations are similarly uncomplicated, although some lidocaine preparations may require buffering to a neutral pH.²⁰ MS222 is a hazardous powder that must be weighed and diluted within a biosafety hood to reduce personnel exposure, and stock solutions must always be buffered to an acceptable pH range prior to use.³⁷

Because propofol is a white, lipophilic drug designed for injection, its use in an aqueous context requires some considerations of practicality. When diluted and used for immersion, care must be taken to mix well and avoid allowing the mixture to settle, as this could lead to uneven drug distribution throughout the resulting emulsion. In addition, 100-ppm and 150-ppm propofol dilutions are essentially opaque. This opacity can obscure visualization of fish in the water column, especially in larger tanks or if multiple fish are being euthanized at once. An additional practical consideration is one of cost. Propofol is significantly more expensive than the other drugs used in this study, and most propofol formulations have shelf lives ranging from only 6 h to just 28 d after vial puncture.^8,28,29 However, many institutions that perform large animal anesthesia use propofol on a regular basis and likely dispose of a significant quantity of leftover drug to avoid expiry once vials are punctured. At these institutions, propofol that remains in the original vial after large animal procedures may be used for zebrafish euthanasia up to the expiry time without need for refrigeration, special handling, or additional cost. Zebrafish euthanasia may thus provide a reasonable and cost-effective use for propofol that would otherwise be wasted.

Our results demonstrate that at the high concentrations tested and with a 30-min exposure period, propofol is an effective euthanasia agent for adult zebrafish, similar to the AVMA recommended methods of MS222 immersion, lidocaine hydrochloride immersion, and rapid chilling. Propofol immersion also results in loss of consciousness and cessation of respiration more quickly than the other tested chemical methods of euthanasia. Moreover, fish do not demonstrate gross behavioral evidence of pain or distress when immersed in propofol as they do when exposed to ice water. The use of propofol immersion for euthanasia of adult zebrafish, or, alternatively, as a method of anesthesia before use of a physical method of euthanasia, can thus significantly improve end-of-life welfare for adult zebrafish.