Behavioral response of Zebrafish Danio rerio Hamilton 1822to sublethal stress by sodium hypochlorite: ecotoxicologicalassay using an image analysis biomonitoring system

Danielly de Paiva Magalhaes Æ Rodolfo Armando da Cunha ÆJose Augusto Albuquerque dos Santos ÆDaniel Forsin Buss Æ Darcılio Fernandes Baptista

Accepted: 28 March 2007 / Published online: 15 May 2007

� Springer Science+Business Media, LLC 2007

Abstract We evaluated behavioral responses of zebrafish

Danio rerio exposed to sublethal concentrations of sodium

hypochlorite using an image analysis biomonitoring system

(IABS). First, the limits of normal variation in swimming

activity of zebrafish were determined by monitoring trav-

eled distance of 40 control fishes using the IABS. An acute

toxicity test was performed to determine the LC50(24 h) for

D. rerio to NaOCl. To evaluate the toxic effects in swim-

ming activity, 32 fishes were exposed to 40%, 30%, 20%,

10% of the LC50 and 32 were used as control using the

IABS. We considered toxic concentrations where more

than 10% intervals of the treated group were below the

limits of normal variation and were significantly different

from the control group. Two main responses were ob-

served: an escape response (increased swimming activity)

at 10% treated group, a gradual decrease in swimming

activity from the 20% of the LC50 on, and an avoidance

response at higher concentrations. The response of the 20%

treated group were considered as a NOAEL and responses

of the 30% and 40% treated groups indicated significant

hypoactivity (adverse effect). This behavioral biomonitor-

ing system has proven to be a useful tool to detect sublethal

toxicity that could be incorporated in biomonitoring pro-

tocols in Brazil.

Keywords Zebrafish � Danio rerio � Image analysis �Swimming activity � Sodium hypochlorite � Sublethal

toxicity

Introduction

Identifying and measuring environmental effects caused

by a chemical requires, in most cases, multidisciplinary

efforts. According to the triad approach for stream ecosys-

tem integrity assessment (Monda et al. 1995), chemical

analyses should be combined with biological responses (e.g.,

community structure and functioning) and with toxicological

assessments.

Toxicological assessments are commonly performed

using acute toxicity tests because they are relatively simple

to perform and produce fast results, usually evaluating the

lethal concentration for 50% of the test organism. How-

ever, environmental contamination in natural ecosystems

often occurs at concentrations well below the lethal con-

centration, which may cause sublethal effects (Cabrera

et al. 1998). In this way, acute toxicity tests ignore the so

called ‘ecological death’––where although the chemical

effect of low toxicant exposures are not sufficient to kill the

organism it unables them to function in an ecological

context because their normal behavior is altered (Scott and

Sloman 2003). Such modifications on behavior may affect

predator avoidance, olfactory capacity, schooling, migra-

tion, feeding, diurnal rhythmic behavior, among other

effects (Little and Finger 1990; Scott and Sloman 2003).

The behavior of an animal is a link between physio-

logical and ecological processes (Gerhardt 1998). Thus, it

is an ideal indicator for studies on the effect of pollutants in

the environment and in the evaluation of potential adverse

effects on the biota.

D. de PaivaMagalhaes (&) � R. A. da Cunha �J. A. A. dos Santos � D. F. Buss � D. F. Baptista

Laboratorio de Avaliacao e Promocao da Saude Ambiental,

Departamento de Biologia, Instituto Oswaldo Cruz, FIOCRUZ,

Av. Brasil 4.365, Manguinhos, Rio de Janeiro,

RJ CEP 21045-900, Brazil

e-mail: daniellypaiva@uol.com.br

123

Ecotoxicology (2007) 16:417–422

DOI 10.1007/s10646-007-0144-2

In the last decade, several biosensor systems were

developed using behavioral responses of aquatic organ-

isms (fish, water-fleas, protozoans and mussels) for in-

field or laboratory biomonitoring (Charoy et al. 1995;

Wolf et al. 1998; Tahedl and Hader 1999, 2001).

Behavioral response biomonitoring are efficient for early

warning responses to pollutants in aquatic systems and

may be used for quality control of industrial treatment

plants, to detect toxic waves, and to assess surface water

quality for public supply.

Textile plants have been reported as highly pollutant

industrial activities (Dorn et al. 1993; Lin and Peng 1994).

Villegas-Navarro et al. (2001) tested the toxicity of each

process of the textile industry and found that none of the

processes were atoxic and the dyeing and Bleaching stages

were the most toxic to Daphnia magna. Also, a wide

variety of auxiliary chemicals may be used in the fabric

processing. One of these chemicals is the sodium hypo-

chlorite (NaOCl), used in the bleaching of fibers. It is also

used in other industrial activities, agriculture, pharmaceu-

ticals, hospitals, and as disinfectant in domestic use

(Santamarta 2001). However, although studies (Crebelli

et al. 2005; Lacorte et al. 2003) have alerted about the high

toxic potential of sodium hypochlorite in the environment,

mainly due to residual chlorine in the water, this substance

continue to be used indiscriminately and its discharge is

scarcely controlled.

Among many organisms that may be used to test textile

effluents effects, fishes are good sentinels for behavioral

assays because their swimming behavior is easily observed

and quantified in controlled settings (Scott et al. 2003).

Swimming activity is a general measures of swimming

behavior, is commonly used to assess contaminant-related

changes in locomotion (Little and Finger 1990). Swimming

activity includes variables such as the frequency, duration,

speed and traveled distance, as well as the frequency and

angle of turns, position in the water column and form and

pattern of swimming (Little and Finger 1990). Evaluation

of swimming activity is useful to determine sublethal

effects of chemicals in the laboratory or under controlled

conditions in field. Detection of abnormal activity is based

on comparisons of response of exposed fishes either with

activity measured during baseline, pre-exposure period or

observation of fishes under a control treatment (Ellgaard

et al. 1978; Thomas et al. 1996; Wolf et al. 1998; Kane

et al. 2004; Swain et al. 2004).

The two main objectives of this study were: i) to

determine the limits of normal variation in swimming

activity of D. rerio Hamilton 1822 in laboratory; and ii)

whether sodium hypochlorite affected swimming acti-

vity of D. rerio using an image analysis biomonitoring

system.

Material and methods

Test organism

Adult male and female Zebrafish weighing between 0.2 and

0.5 g, body length varying from 2.4 to 3.2 cm, and more

than 120 days old, were maintained in laboratory condi-

tions for one week for acclimatation prior the experiments.

Fishes were fed with Spirulina Flakes 200 (Alcon Gold)

once a day and kept in filtered dechlorinated tap water

(water temperature between 23 and 26�C, room tempera-

ture between 22 and 24�C). Fishes were not fed for 24 h

before the tests.

Chemical

The sodium hypochlorite (NaOCl) P.A. 5–6% ISOFAR

brand was used as toxic agent in this study. Test concen-

trations were extracted from a stock solution of 1 mL of

sodium hypochlorite diluted in 1 L of filtered dechlorinated

tap water.

Image Analysis Biomonitoring System (IABS)

A schematic diagram of the Image Analysis Biomonitoring

System (IABS) is shown in Fig. 1. We used an opaque

glass aquarium, 35 · 35 · 25 cm, 30 L capacity, divided

in two independent compartments, one for the control and

one for the treated group. Inside each compartment, an

opaque acrylate with 3 mm hole was divided into four

holding boxes (9.5 · 5· 2 cm) where fishes were kept

individually. Both compartments were equipped with

submerged water pumps for water mixing. An illuminating

cabin provided the holding boxes a shadowless diffuse soft

lightning. A recording cabin made of acrylate held the

analogic video camera. Images were sent to a Videomex-

V� (Columbus Instruments), which analyzed images as

white pixels over a black background using the software

Traveled Distance of Multiple Objects (TDMO). Traveled

distance data were sent to a computer and recorded in

Excel database for analyses.

Limits of normal variation in swimming activity

of Danio rerio

The limits of normal variation in swimming activity of

D. rerio were determined by monitoring traveled distance

of 40 fishes kept in filtered dechlorinated tap water using

the IABS during 5 h. Fishes had a 1 h acclimation period in

the system prior the monitoring. The limits of normal

variation were determined by linear regressions of mean

418 D. P. Magalhaes et al.

123

values of the 40 fishes at each time interval ± 1 standard

deviation.

Evaluation of the toxic effects in swimming activity

of Danio rerio

Tests followed Brazilian regulations for acute toxicity tests

using D. rerio (L5.019-I, CETESB 1990). A 24 h acute

toxicity test was performed to determine the LC50 for

D. rerio to NaOCl, using the following exposure concen-

trations: 22.8, 28.5, 34.2, 39.0, 45.6, 51.3, 57.0 mg/L. A

control group was kept in filtered dechlorinated tap water.

The Probit method (Goldstein 1964) was used to determine

the LC50 for 24 h exposition to the NaOCl.

To evaluate the toxic effects in swimming activity, 32

fishes were exposed to 40%, 30%, 20%, 10% of the LC50

concentration and 32 were used as control. Two replicates

of four fishes of a control and a treated group were used

concurrently for each concentration. Fishes were placed

carefully in the holding boxes and kept in static condition

with no water change during the experiment. Fishes had a

1 h acclimation period in the system prior the experiment

and then exposed for 5 h to the chemical, a sufficient

period for detecting the effects of a sublethal concentration

without volatilization of the chemical. The exposure con-

centration was introduced 15 min prior the monitoring and

homogenized using a water pump. The same procedure was

used in the control group by adding water. Exposure period

was divided in 60 recording intervals of 5 min and mean

values of the traveled distance were recorded.

Mean values of treated and corresponding control group

were compared to the limits of normal variation. Behav-

ioral tests were considered valid if traveled distance of the

control group were in accordance to the established limits

of normal variation (i.e., 10% intervals or less were below

or above the limits). This procedure was adopted to guar-

antee that no interference occurred during the experiment.

For the treated groups, we considered toxic concentrations

where more than 10% intervals were below the limits of

normal variation. In such cases, control vs. treated results

were tested for significance using a Mann-Whitney U-test.

Water samples were taken at the beginning and at the

end of the experiment and pH, dissolved oxygen and

conductivity were analyzed from both control and treated

holding tanks using a SensION 378 (Hack).

Results

During the establishment of the limits of normal variation,

D. rerio had decreasing swimming activity during the

monitoring period, with significant differences according to

the Wilcoxon test (p < 0.05) between initial and final

traveled distances. Therefore, the limits (±1 standard

deviation) were established respecting this pattern (Fig. 2).

Acute toxicity tests indicated a LC50 (24 h exposure to

NaOCl) of 48 mg/L (Fig. 3). Based on this result, we

established sublethal concentrations to evaluate the effects

in swimming activity of D. rerio: 10% (4.8 mg/L), 20%

(9.6 mg/L), 30% (14.4 mg/L) and 40% (19.2 mg/L) of the

LC50.

Fig. 1 Schematic diagram of

the Image Analysis System

(IAS): 1––microcomputer

containing a software for data

storage in Excel database; 2––

monitor; 3––Videomex-V�

system; 4––record cabin in

acrylate; 5––video camera; 6––

two independent opaque glass

aquarium, 10L capacity each;

7––exposure chambers, each

one with four holding boxes;

8––water mini-pumps; 9––

silicon rubber for chemical

introduction; 10––drain; 11––

illumination cabin

Fig. 2 Mean traveled distance (•) and limits of normal variation in

swimming activity of Danio rerio (linear regressions of ±1 standard

deviations)

Behavioral Response of Zebrafish Danio Rerio 419

123

Behavioral tests were valid, since traveled distance

variation of all four control groups were within the estab-

lished limits of normal variation (i.e., less than 10%

intervals were below or above the limits).

Danio rerio showed a sigmoid dose-response curve

when exposed to tested sublethal concentrations, decreas-

ing the traveled distance with increasing concentrations

(Fig. 4). Mean values of the 10% treated group were over

the established upper limits of normal behavior in 15% of

the intervals (Fig. 5) and the Mann-Whitney U-test indi-

cated significant differences between control and treated

groups (p < 0.05). However, this behavior was considered

a state of hyperactivity––a stressor effect but not an

adverse effect. Little effect was observed with the 20%

LC50 treated group: values of 94% of the intervals were

within the established normal limits. Mean values of the

30% and 40% treated groups were adversely affected by

the chemical: traveled distance was below the lower limits

of normal variation in 58% and 87% of the intervals,

respectively. Effects were observed in the 30% treated

group after 70 min of exposure and after 40 min in the

40% treated group. Mann-Whitney U-tests indicated sig-

nificant differences between treated and respective control

groups in both cases (p < 0.05).

Little difference in physical–chemical parameters was

found between water samples taken at the beginning and

the end of the experiment from both control and treated

Fig. 3 Dose-response curve for 24 h acute toxicity test of Daniorerio exposed to NaOCl

Fig. 4 Sigmoid dose-response curve of mean traveled distance of

Danio rerio exposed to 10, 20, 30 and 40% concentrations of the LC50

of NaOCl and Control group. Bars indicate standard errors

Fig. 5 Mean traveled distance of Danio rerio exposed to sublethal

concentrations of NaOCl: continuous line represents the limits of

normal variation, (*) treated groups means and (•) corresponding

control groups means. a) concentration of 10% of the LC50; b) 20%;

c) 30% and d) 40%

420 D. P. Magalhaes et al.

123

groups (Table 1). Even the pH, that could have lowered

due to the formation of hypochlorous acid from the reac-

tion of sodium hypochlorite and water, showed little

change. This assures that observed behavioral changes

resulted only due to the effects of sodium hypochlorite.

Discussion

Similarly to our findings, many studies showed that

behavioral responses are not necessarily correlated to toxic

concentrations: lower concentrations may induce an in-

crease in swimming activity, while higher concentrations

may induce a decrease in activity. Little et al. (1989)

observed that swimming capacity was heightened when

fishes were exposed to the organophosphorus defoliant

DEF or 2,4–DMA at 0.5% of LC50, while higher concen-

trations (5–50% of the LC50) reduced their swimming

capacity. Same results were observed by Finger et al.

(1985): a 30-days exposure of bluegill fish to low sublethal

concentration of fluorene (0.12–0.25 mg/L) significantly

increased their swimming capacity, while higher concen-

trations (1.0 mg/L) decreased their swimming capacity.

This increase in swimming activity typically characterizes

an escape response, where the organism tries to avoid the

area impacted by the chemical (Smith and Bailey 1988).

Ellgard et al. (1978) stated that this could signal the

approach to the tolerance limits for a pollutant, i.e., from

this point, higher concentrations would probably be toxic to

the test organism.

In our study, the lowest concentration used in the

behavioral test (10% of the LC50) induced an increase in

swimming activity (hyperactivity) of D. rerio. Based on

the findings of Ellgard et al. (1978), we did not consider

this response as an adverse effect. Such concentrations,

below the NOAEL (no observed adverse effect level),

would not be considered as the cause of adverse effects, but

could be used to determine the sensitivity of the biosensor

species exposed to a chemical during a determined period.

We observed that from the 20% of the LC50 on, swimming

activity decreased gradually, characterizing a dose-

response effect. Mean values of the 20% treated group

were within the normal limits of variation, therefore con-

sidered as a NOAEL, and the 30% and 40% of the LC50

concentrations induced significant hypoactivity. This

decrease in swimming activity may be an avoidance

behavior, designed to lessen the probability of death or the

metabolic costs incurred in maintaining physiological

homeostasis (Olla et al. 1980; Schreck et al. 1997).

In this study, no significant alteration in the pH, dis-

solved oxygen and conductivity were found when Sodium

Hypochlorite was added. Chen et al. (2001) have found

similar results, and indicated that although no significant

changes in physical and chemical parameters occurred, the

addition of the NaOCl to the effluent increased its toxicity.

This information is particularly useful in Brazil because

many textile or paper mill industries throw their effluents in

rivers and streams. Since routine monitoring often do not

consider such parameter, the toxicity effects of these

effluents are probably underestimated. Although sodium

hypochlorite is widely used and highly toxic, we found

very few studies relating the effects of sodium hypochlorite

in the environment and/or in biological systems. A possible

explanation is the difficulty to isolate the effect of sodium

hypochlorite from other substances used in the textile and

paper mill, or domestic use. We faced such difficulties in

Table 1 Water physical and

chemical parameters for initial

and final samples for each

exposure concentrations and

corresponding controls

Treatments Samples Dissolved oxygen (mg/L) pH Conductivity (lS/cm) Temperature (�C)

10% LC50(24 h) Initial 5.3 6.8 91.8 24.3

Final 5.6 6.5 90.2 23.3

Control Initial 6.1 6.8 88.0 24.4

Final 5.3 6.6 88.5 23.3

20% LC50(24 h) Initial 6.0 7.1 103.2 24.1

Final 5.8 6.7 99.7 23.3

Control Initial 6.4 7.0 96.5 24.0

Final 5.8 6.7 96.3 23.4

30% LC50(24 h) Initial 5.4 6.7 103.0 23.3

Final 5.5 6.7 101.5 23.3

Control Initial 6.0 6.8 95.8 24.3

Final 5.6 6.6 96.2 23.2

40% LC50(24 h) Initial 6.3 6.7 105.4 22.0

Final 6.1 6.6 104.9 21.2

Control Initial 6.3 6.6 97.6 21.8

Final 6.0 6.7 98.3 21.4

Behavioral Response of Zebrafish Danio Rerio 421

123

the field while conducting studies on the effect of the

textile industry on fishes and macroinvertebrates assem-

blages in streams in south-east Brazil (Baptista et al. un-

publ. data): most effects were probably synergic, and the

present study is an attempt to isolate the effect of Sodium

Hypochlorite on biological systems.

In resume, we registered two main responses of D. rerio

to the NaOCl: an escape response at low concentrations

and an avoidance response at higher concentrations. We

found out that the swimming activity was an easy-to-

measure efficient way to detect toxic effects of NaOCl,

since exposure concentrations necessary to cause such

biological responses were well below the lowest concen-

tration necessary to cause mortality according to acute tests

(22.8 mg/L––Fig. 3). In this way, a behavioral biomoni-

toring using an image analysis system may be a useful tool

to detect sublethal toxicity and could be incorporated in

biomonitoring protocols in Brazil.

Acknowledgements We are grateful to Valdinei Valin for the water

chemical analysis and to Jorge Vieira for building some components

of the IABS. This research was partially funded by Fundacao Os-

waldo Cruz – PAPES III and PDTSP/Aguas.

References

Cabrera C, Ortega E, Lorenzo ML, Lopez MDC (1998) Cadmium

contamination of vegetable crops, farmlands, and irrigation

waters. Rev Environ Contam T 154:55–81

CETESB (Company of Technology and Environmental Sanitation)

(1990) Agua – Teste de toxicidade aguda com peixe – Parte I –

Sistema Estatico, CETESB-L5.019-I, Sao Paulo

Charoy CP, Janssen CR, Persoone G, Clement P (1995) The

swimming behaviour of Brachionus calyciflorus (rotifer) under

toxic stress. I. The use of automated trajectometry for determin-

ing sublethal effects of chemicals. Aquat Toxicol 32:271–282

Chen CM, Shih ML, Lee SZ (2001) Increased toxicity of textile

effluents by a chlorination process using Sodium Hypochlorite.

Water Sci Tech 43:1–89

Crebelli R, Conti L, Monarca S, Feretti D, Zerbini I, Zani C, Veschetti

E, Cutilli D, Ottaviani M (2005) Genotoxicity of the desinfection

by-products resulting from peracetic acid- or hypochlorite-

desinfected sewage wastewater. Water Res 39:1105–1113

Dorn PB, Salanitro JP, Evans SH, Kravetz L (1993) Assessing the

aquatic hazard of some branched and linear nonionic surfactants

by biodegradation and toxicity. Environ Toxicol Chem 12:1751–

1762

Ellgaard EG, Tusa JE, Malizia AA Jr. (1978) Locomotor activity of

bluegill Lepomis macrochirus: hyperactivity induced by suble-

thal concentrations of cadmium, chromium and zinc. J Fish Biol

1:19–23

Finger SE, Little EE, Henry MG, Fairchild JF, Boyle TP (1985)

Comparison of laboratory and field assessment of fluorine, part I:

Effects of fluorine on the survival, growth, reproduction, and

behavior of aquatic organisms in laboratory tests. In: Boyle TP

(ed) Validation and Predictability of laboratory methods for

assessment the fate and effects of contaminants in aquatic

ecosystems. STP 865. American Society for Testing and

materials, Philadelphia, PA, pp 120–133

Gerhardt A (1998) Whole effuent toxicity testing with Oncorhynchusmykiss (Walbaum 1792) survival and behavioral response to a

dilution series of a mining effluent in South Africa. Arch Environ

Con Tox 35:309–316

Goldstein A (1964) Biostatistics: an introductory text. Collier-

Macmillan Canada, Toronto, pp 154–159

Kane AS, Salierno JD, Gipson GT, Molteno TCA, Hunter C (2004) A

video-based movement analysis system to quantify behavioral

stress response of fish. Water Res 38:3993–4001

Lacorte S, Latorre A, Barcelo D, Rigol A, Malmqvist A, Welander T

(2003) Organic compounds in paper-mill process waters and

effluents. Trends Anal Chem 22:725–737

Lin SH, Peng CF (1994) Treatment of textile wastewater by

electrochemical method. Water Res 28:277–282

Little EE, Archeski RD, Flerox BA, Kozlovskaya VI (1989)

Behavioral indicators of sublethal toxicity in rainbow trout.

Arch Environ Con Tox 19(3):380–385

Little EE, Finger SE (1990) Swimming behavior as an indicator of

sublethal toxicity in fish. Environ Toxicol Chem 9:3–19

Monda DP, Galat DL, Finger SE (1995) Evaluation ammonia toxicity

in sewage effluent to stream macroinvertebrates: 1. A multi-level

approach. Arch Environ Con Tox 28:378–384

Olla BL, Pearson WH, Studholme AL (1980) Applicability of

behavioral measures in environmental stress assessment. Rapp

P-V Reun-Cons Int Explor Mer 179:162–173

Santamarta J (2001) Por um futuro sem contaminantes organicos

persistentes. Agroecol e Desenv Rur Sust 2:46–56

Scott GR, Sloman KA, Rouleau C, Wood CM (2003) Cadmium

disrupts behavioural and physiological responses to alarm

substances in juvenile rainbow trout (Oncorhynchus mykiss).

J Exp Biol 206:1779–1790

Schreck CB, Olla BL, Davis MW (1997) Behavioral response to

stress. In: Iwama GK, Pickering AAD, Sumpter JP, Schreck CB

(eds) Fish stress and health in aquaculture. Cambridge Univ

Press, Cambridge, pp 145–170

Smith EH, Bailey HC (1988) Development of a system for continuous

biomonitoring of a domestic water source for early warning of

contaminants. In: Gruber DS, Diamonds JM (eds) Automated

biomonitoring: living sensors as environmental monitors. Ellis

Horwood, Chichester, UK, pp 182–205

Swain HA, Sigstad C, Scalzo FM (2004) Effects of dizocilpine

(MK-801) on circling behavior, swimming activity, and place

preference in zebrafish (Danio rerio). Neurotoxicol Teratol

26:725–729

Tahedl H, Hader DP (1999) Fast examination of water quality using

the automatic biotest ecotox based on the movement behavior of

freshwater flagellate. Water Res 33(2):426–432

Tahedl H, Hader DP (2001) Automated biomonitoring using real time

movement analysis of Euglena gracilis. Ecotoxicol Environ Saf

48:161–169

Thomas M, Cheretien D, Florion A, Trever D (1996) Real-time

detection of potassium cyanide pollution in surface waters using

eletric organ discharge wave emitted by the tropical fish,

apteronotus albifrons. Environ Technol 17:561–574

Villegas-Navarro A, Ramırez MY, Salvador MS, Gallardo JM (2001)

Determination of wastewater LC50 of the different process stages

of the textile industry. Ecotoxicol Environ Saf 48:56–61

Wolf G, Scheunders P, Selens M (1998) Evaluation of the swimming

activity of Daphnia magna by image analysis after administra-

tion of sublethal cadmium concentrations. Comp Biochem

Physiol A 120:99–105

422 D. P. Magalhaes et al.

123