Zebrafish: a multifaceted approach for drug safety assessment

Many compounds fail in preclinical development because of safety-related problems, but identifying 'predictable' safety or toxicity liabilities earlier in the process could lead to improved design and selection of compounds that are more likely to be approved. It is estimated that a 10% improvement in predicting failure before the initiation of clinical trials could save $100 million (€68.4 million) in development costs per drug.¹ One area where this can be applied is in preclinical in vivo models, which are currently limited to rodents, dogs and monkeys. These models, although effective, are costly and use large amounts of compounds, often precluding early detection of liabilities associated with the compounds.

A model gaining increasing recognition from the pharmaceutical industry is the zebrafish, which offers a novel approach to toxicity and safety screening. This approach is both cost effective and high throughput.

Figure 1: An adult zebrafish. Scalebar = 0.89cm

What are zebrafish?

Zebrafish (Danio rerio; Figure 1) is a small freshwater fish originating from the rivers of northern India. The zebrafish model system possesses numerous advantages for medium- to high-throughput screening of compounds, such as the relative ease of maintaining large stocks of fish, its high fecundity and its rapid embryonic development ex utero, whereby all the major organs are represented at 3 days post-fertilization (dpf), facilitating experimental manipulation. As the larvae are transparent, it allows direct in vivo observation of tissue by noninvasive methods (Figure 2).² Furthermore, the organization of the zebrafish genome and the genetic pathways controlling signal transduction and development are highly conserved between zebrafish and man.³ At 7 dpf, the larvae are approximately 4 mm long. Because of this small size, assays can be undertaken in 96-well plates and, as the larvae can live in as little as 200 μL of fluid, only a few milligrams of compound are required for screening (Figure 3). Thus, in vivo analysis of the effects of compounds can be undertaken much earlier in the drug discovery process than has previously been possible, which is facilitated by the fact that zebrafish are dimethylsulphoxide (DMSO) tolerant up to 1% v/v and readily absorb compounds from the water. The high fecundity of zebrafish and the relative ease of maintaining large stocks also provide investigators with ample numbers of larvae to analyse in high-throughput screens compared with conventional in vivo models. These properties have established zebrafish as excellent model systems.⁴

Figure 2: A zebrafish larva at 3 dpf. Organs such as the heart are clearly visible because of the optical clarity of the larvae at this age. Scalebar = 0.24mm

Safety pharmacology

In the drug development process, lead candidate compounds are assessed for adverse effects in safety pharmacology studies. These are usually performed before first exposure to man according to a hierarchy, with a core battery of tests conducted that assess the cardiac, central nervous and respiratory systems. These assessments are usually followed by supplemental studies on other organ systems such as the gastrointestinal (GI) tract and kidneys in more detailed studies if required.

Figure 3: Zebrafish larvae at 7 dpf in an eppendorf tube containing 1% DMSO solution and 96-well plate format.

In vivo safety pharmacology is often a bottleneck for advancing compounds into development, but zebrafish could accelerate the process by providing new ways of identifying early compounds that show adverse effects on physiological systems.

Cardiac assessment

One of the principal causes for compounds failing the drug development process is adverse effects on the heart.⁵ A major concern for compound development is the possibility of evoking QT prolongation (associated with Torsade de Pointes), a serious and often fatal heart arrhythmia. Drug-induced QT prolongation is purported to result from blockade of the human ether-a-go-go gene product (hERG) potassium channel. The zebrafish zERG gene is very closely related to the human gene and its knockdown in zebrafish larvae results in a very specific atrioventricular block.⁶ Furthermore, drugs that are known to prolong QT interval by hERG channel blockade, such as cisapride, pimozide and terfenadine, along with those that do not act on the hERG, such as YS035, have been identified in this assay with an excellent correlation with humans.^7,8 In addition to QT interval, effects on cardiac behaviour, such as bradycardia, tachycardia and force of contractions, can be assessed. These findings have been taken further by the production of a transgenic zebrafish line that expresses green fluorescent protein in the myocardium, which can be used to screen small molecules.⁹

Central nervous system assessment

Adverse effects of compounds on the central nervous system (CNS) account for 10% of drug withdrawals during the last 30 years.¹⁰ One assessment generally used for investigating effects of compounds on CNS is locomotor activity.¹¹ Zebrafish have recently become the focus of neurobehavioural studies because larvae display learning, sleep, drug addiction and neurobehavioural phenotypes, which are quantifiable and relate to those seen in humans.^12–16 It is also known that the zebrafish brain structure and function is similar to other vertebrates, lending further support to its use as an effective model for evaluating behaviour.¹⁷ Zebrafish behaviour can be easily observed by tracking their movements in 96-well plates, which has enabled the throughput of compounds being screened to increase immensely.¹⁸ Using this approach, it has been possible to confirm the effects of sedatives, such as diazepam and melatonin, as well stimulants, such as caffeine.¹⁸ The potential of compounds to induce seizure can also be assessed in zebrafish larvae as they display seizure activity when exposed to convulsants such as pentylenetetrazole.¹⁹ Furthermore, studies have shown learning and memory in zebrafish, with cognitive impairment also being demonstrated.²⁰ Drug dependence/abuse liability is another area included in follow-up studies for novel CNS drugs before approval.¹¹ Dependency and addiction have been demonstrated in adult zebrafish in a number of studies,^14,21,22 and assays for this are now being developed in the larvae.^23,24

GI assessment

Although not a regulatory requirement for assessment prior to Phase I trials, adverse drug reactions in the GI system account for approximately 18% of reported clinical adverse events.²⁵ Zebrafish do not have stomachs, but spontaneous intestinal motility can be observed in larvae as early as 4 dpf, with powerful anterograde and retrograde contractions present.²⁶ Zebrafish also have a rich receptor pharmacology influenced by the enteric nervous system, and a very good correlation has been demonstrated between zebrafish and humans.^27,8 These retrograde contractions can be detected in response to agents such as ementine and, thus, zebrafish may be suitable models for detecting the possible emetic potential of agents. These features, along with the transparency of the larvae, enable drug-induced alterations of the GI system to be studied using imaging of the live intact animal.

Figure 4: The embryonic zebrafish retina (left) is almost identical to the adult human retina (right). The relative positions of the various cell types are the same in both tissues.

Visual function assessment

A number of reports to the national registry of drug-induced ocular side-effects has led to identification of many ocular adverse reactions during drug treatment; for example, bisphosphonates have been associated with ocular inflammation, topiramate causes angle-closure glaucoma and cetirizine has been linked to oculogyric crisis.^28,29 Current methods using measurements of electroretinograms in dogs are intensive and technically difficult, but zebrafish have a cone-dense retina giving them rich colour vision that is similar to a human's. This offers a distinct advantage in preclinical studies over commonly used nocturnal rodents that do not possess specialized colour vision.³⁰ In zebrafish, the visual system is well developed by 5 dpf (Figure 4).³¹ An optokinetic assay has been developed to measure the visual function in larvae, which allows adverse ophthalmic events to be assessed with smaller quantities of compound than has previously been possible.^32,33 In the optokinetic assay, larvae are immobilized in methylcellulose and arranged within a striped drum with the dorsal side up, tails towards the centre of a dish and heads facing outwards towards the drum (Figure 5). The drum is rotated, eliciting a series of smooth pursuits followed by saccadic eye movements as the eyes flick onto the next stripe. Changing parameters, such as the velocity of rotation, stripe width and contrast, allows quantification of visual acuity, light adaptation and contrast sensitivity.³⁴

Figure 5 a): Larvae placed in the centre of the drum with black and white strips rotating on the inner wall. b) 10 larvae immobilized with 4% methylcellulose in petri dish with head facing outwards.

Toxicology

Although toxicology is not the subject of this article, it is worth mentioning that zebrafish have long been used for toxicity testing of environmental and agrochemical agents³⁵ and, more recently, for evaluating the toxicity of pharmaceutical agents.³⁶ An in vivo toxicology assessment can be made in 1 week using larval zebrafish, which is a much shorter timeframe than in corresponding mammalian studies. Studies have already demonstrated that zebrafish can be used to investigate hepatotoxicity, ototoxicity and neurotoxicity of novel compounds.^37,38 Toxicology assays in zebrafish have been validated with a number of known embryotoxic and teratogenic compounds, and the good correlation with mammals suggests that the pathways mediating these effects are highly conserved; for example, the exposure of developing larvae to ethanol results in craniofacial abnormalities and learning deficits, which are known to be associated with foetal alcohol syndrome.²⁰

The advantages of zebrafish

Conclusion

The ability to identify failures earlier in the drug development process is becoming increasingly important. Zebrafish offer an efficient and cost-effective in vivo model in which small amounts of compound can be used to assess the putative side-effect profile of compounds, allowing earlier decisions to be made in the successful optimization of compound series or development of particular early preclinical candidates. Zebrafish larvae are also DMSO tolerant and transparent, allowing assays to be conducted noninvasively with some of the more insoluble compounds. Additionally, the predictability of the zebrafish model to humans has shown a very high correlation and further work is ongoing.³⁹

The issue of pharmacokinetic–pharmacodynamic relationships of compounds in zebrafish is being explored by a number of laboratories. The wealth of information that can be garnered from larval zebrafish to evaluate the safety pharmacology and toxicology of compounds makes this an exciting new approach to aid and advance drug discovery, permitting earlier, informed decisions on the prioritization of compounds for development, saving both time and money and, consequently, reducing attrition rate.

Jonathan Best is Principal Research Scientist at Summit plc (UK).

Wendy Alderton is Managing Director of the Zebrafish Business Unit at Summit plc (UK).

References

1. J.A. Kramer et al., Nat. Rev. Drug Discov., 6(8), 636–649 (2007).

2. M. Westerfield, The Zebrafish Book: A Guide for the Laboratory use of Zebrafish (Danio rerio) (University of Oregon Press, Eugene, OR, USA, 2000).

3. J.H. Postlethwait et al., Genome Res., 10(12), 1890–1902 (2000).

4. D.J. Grunwald et al., Nat. Rev. Genet., 3(9), 717–724 (2002).

5. W. Suter, Curr. Opin. Chem. Biol., 10(4), 362–366 (2006).

6. U. Langheinrich et al., Toxicol. Appl. Pharmacol., 193(3), 370–382 (2003).

7. S. Berghmans, IIR's 4th Annual QT Prolongation and Safety Pharmacology: Predictive Safety Assessment in Non-Clinical and Clinical Studies, (Amsterdam, The Netherlands, 28 February–3 March, 2006).

8. S. Berghmans et al., J. Pharmacol. Toxicol. Methods, 58(1), 59–68 (2008).

9. C.G. Burns et al., Nat. Chem. Biol., 1(5), 263–264 (2005).

10. M. Fung et al., Drug Information Journal, 35(1), 293–317 (2001).

11. R.D. Porsolt et al., Fundam. Clin. Pharmacol., 16(3), 197–207 (2002).

12. G.M. Cahill, Cell Tissue Res., 309(1), 27–34 (2002).

13. S. Guo, Genes Brain Behav., 3(2), 63–74 (2004).

14. J. Ninkovic et al., J. Neurobiol., 66(5), 463–475 (2006).

15. M.B. Orger et al., Methods Cell Biol., 77(1), 53–68 (2004).

16. I.V. Zhdanova et al., Brain Res., 903(1–2), 263–268 (2001).

17. V. Tropepe et al., Genes Brain Behav., 2(5), 268–281 (2003).

18. S. Berghmans et al., 5th International Conference on Methods and Techniques in Behavioural Research (Wageningen, The Netherlands, 30 August–2 September, 2005).

19. S.C. Baraban et al., Neuroscience, 131(3), 759–768 (2005).

20. M.J. Carvan et al., Neurotoxicol. Teratol., 26(6), 757–768 (2004).

21. R. Gerlai et al., Pharmacol. Biochem. Behav., 67(4), 773–782 (2000).

22. T. Darland et al., Proc. Natl. Acad. Sci. USA, 98(20), 11691–11696 (2001).

23. S. Bretaud et al., Neuroscience, 146(3), 1109–1116 (2007).

24. B. Lockwood et al., Pharmacol. Biochem. Behav., 77(3), 647–654 (2004).

25. Safer Pharmacology Working Group Report (Academy of Medical Sciences, November, 2005).

26. A. Holmberg et al., J. Fish. Biol., 63(2), 318–331 (2003).

27. A. Holmberg et al., J. Exp. Biol., 207(Pt 23), 4085–4094 (2004).

28. F.W. Fraunfelder et al., Ophthalmology, 111(7), 1275–1279 (2004).

29. R.M. Santaella et al., Drugs, 67(1), 75–93 (2007).

30. G.H. Jacobs, Biol. Rev. Camb. Philos. Soc., 68(3), 413–471 (1993).

31. J. Bilotta et al., Int. J. Dev. Neurosci., 19(7), 621–629 (2001).

32. S.S. Easter, Jr. et al., Dev. Biol., 180(2), 646–663 (1996).

33. Y.Y. Huang and S.C.F. Neuhass, Front. Biosci., 13(5), 1899–1916 (2008).

34. O. Rinner et al., Invest Ophthalmol. Vis. Sci., 46(1), 137–142 (2005).

35. S. Bretaud et al., Neurotoxicol Teratol., 26(6), 857–864 (2004).

36. A.L. Rubinstein, Expert Opin. Drug Metab. Toxicol., 2(2), 231–240 (2006).

37. C. Ton et al., Hear Res., 208(1–2), 79–88 (2005).

38. C. Parng et al., J. Pharmacol. Toxicol. Methods, 55(1), 103–112 (2007).

39. Barros et al., Br. J. Pharmacol., 154(7), 1400–1413 (2008).