6. Discussion

▼ 83 (continued)

Behavioural tests addressing whole organism-level effects, may provide primary information about the hazard potential of chemicals for organisms. Focusing on the implications of a behavioural approach for ecotoxicology as branch of stress ecology, in this study the behavioural analyses were combined with chronobiological procedures such as time series analysis and power spectral analysis.

Alterations in fish behaviour including biorhythmical aspects were analysed as indicators of the sublethal toxicity of the cyanotoxin MC-LR and the xenobiotic chemical PCB 28. It was shown that dissolved MC-LR between 0.5 and 50 µg l-1 and PCB 28 concentration at 100 and 150 µg l-1 acted as stressors and caused significant changes in the behaviour and circadian rhythms of activity of both fish species, Danio rerio and Leucaspius delineatus.

▼ 84 

Since sublethal effects of special stressors had to be investigated it was necessary to exclude the influence of varying environmental factors or other substances which normally occur under natural conditions. For that reason an artificial system was chosen, whereby the experimental design was characterised by a high level of invariant test conditions. Since the physico-chemical parameters of the used aquarium water were regarded to be at the normal levels of uncontaminated water for fish (see 4.2), only the added test substances (MC-LR and PCB 28) have had the potential to act as chemical stressors.

6.1 Toxicity of Microcystin-LR and Trichlorobiphenyl

6.1.1 Microcystin-LR

Several papers have shown the effects of MC-LR on fish through more or less invasive methods. In contrast to the above-mentioned effects of intraperitoneal injection and oral application (see, MC-LR in water produced no significant histopathological effects in fish (Phillips et al., 1985; Tencalla et al., 1994). Bury et al. (1995) registered an inhibited growth and disturbances of the ion balance in fish exposed to MC-LR diluted in water at concentrations of 41-57 µg l-1. Rodger et al. (1994) found gill and liver damage as well as mortality for brown trout (Salmo trutta) at concentrations between 16 and 19 µg l-1. However in the last case the authors were not able to attribute the effects to microcystin exposure alone, as also the pH was significantly elevated and the scum of algae cells led to physical irritations of the gills.

In contrast to these results the approach of using behavioural parameters as toxicity parameters was rather sensitive. This coincides with other studies suggesting that impairment of locomotor behaviour may be a more sensitive indication for fish health than more traditional toxicity endpoints (Little and Finger, 1990; Schreck, 1990; Siegmund and Biermann, 1992; Steinberg et al., 1995; Hopkins et al., 2003). Steinberg et al. (1995) described that the lowest observed effect concentration (LOEC) of atrazine detected by behavioural analyses was more than three orders of magnitude below the acute toxicity values. A review by Little and Finger (1990) revealed that the lowest behaviourally effective toxicant concentration that induced changes in swimming behaviour of fish ranged from 0.1% to 5.0% of the LC50.

▼ 85 

The lowest observed effect concentration (LOEC) of MC-LR for behavioural and chronobiological parameters appear to be ≤ 0.5 µg l-1 for both tested fish species Danio rerio and Leucaspius delineatus. For MC-LR a LOEC in the same range as found for behavioural changes was determined by Oberemm et al. (1997) and Wiegand et al. (1999) applying ontogenetic and enzymatic parameters, respectively. Larvae of Danio rerio reared under exposure to MC-LR at a concentration of 50 µg l-1 during the sensitive stages of embryonic development showed decreased survival, and larval growth and development were particularly retarded at a concentration of 0.5 µg l-1 (Oberemm et al., 1997, 1999). Wiegand et al. (1999) found that the activity of biotransformation enzymes (microsomal and cytosolic glutathione-S-transferases) was elevated in zebrafish embryos, even at a toxin concentration of 0.1 µg l-1.

These effects of MC-LR (including the findings of the present study) occur at environmental relevant concentrations of microcystins, associated with cyanobacterial blooms (see

6.1.2 Trichlorobiphenyl

In the present study the non-coplanar congener PCB 28 was tested and significant behavioural and chronobiological changes were found from a concentration of 100 µg l-1 upwards. In pre-tests the lower PCB 28 concentration of 50 µg l-1 did not affect the behaviour of Danio rerio as well as Leucaspius delineatus.

▼ 86 

Fingerman and Russel (1980) found effects of Aroclor 1242 on locomotor activity and on neurotransmitters (dopamine and norepinephrine) in the brain of the Gulf killifish, Fundulus grandis. Furthermore Aroclor 1254 led to significant changes of swimming behaviour of carp (Cyprinus carpio) with a LOEC of ≤ 22 µg l-1 (Schmidt et al., 2004). However it is not possible to judge which congeners of Aroclor are responsible for the behavioural and neurological alterations. Commercial PCB mixtures elicit a broad spectrum of toxic responses that are dependent on several factors including chlorine content, purity, dose, species and strain, age and sex of animal and route and duration of exposure (Giesy and Kannan, 2002).

Regarding behavioural parameters in fish the PCB mixture Aroclor 1254 shows a lower LOEC than the single congener PCB 28 and three causes may serve (alone or together) as an explanation: the proportion of higher chlorinated congeners, the proportion of coplanar (dioxin-like) congeners and the potential interactions between the single components of the mixture. The toxicity of a specific PCB congener is dependent upon both its degree of chlorination and the position of its chlorine atoms (Safe, 1984). Less chlorinated PCBs are lesser bioconcentrated and more readily metabolised and excreted (Giesy and Kannan, 2002). For PCB 28 no accumulation in the brain of fish was found (Qi et al., 1997).

Toxic effects due to coplanar PCBs occur at relatively smaller concentrations than those due to non-coplanar PCBs (Giesy and Kannan, 2002). For example in rainbow trout (Oncorhynchus mykiss), coplanar congeners affected early life stage mortality (Walker et al., 1996), in contrast to non-coplanar PCBs that did not induce early life stage mortality (Walker and Peterson, 1991; Hornung et al., 1996). However the main portion (81%) of Aroclor 1254 consists of non-coplanar PCBs (Frame et al., 1996), thus there is only a modest amount of coplanar PCBs in Aroclor 1254.

▼ 87 

Interactions among individual PCB congeners were found by Seegal et al. (1990), whereby a mixture of three non-coplanar PCBs (one of them was PCB 28) was more potent in reducing brain dopamine content than the equal amounts of each congener in tissue cultures of nonhuman primate brain. A further study found that three single non-coplanar congeners did not cause any toxic effects on embryos and newly hatched larvae of the Japanese medaka (Oryzias latipes), but all three together caused an inhibition of swim bladder inflation in the newly hatched larvae (Kim and Cooper, 1999).

Another uncertainty associated with the assessment of toxicity based on exposure to commercial PCB mixtures is related to the relative amounts of polychlorinated dibenzofurans (PCDFs) and polychlorinated naphthalenes (PCNs) identified as contaminants in technical PCB preparations (Giesy and Kannan, 2002), whereby dioxin-like potencies of polychlorinated naphthalenes were found by Blankenship et al. (2000).

Therefore, for analysing relations between structure and toxic effects of PCBs it is useful to separately determine the toxicity of single congeners. This is moreover important, because of the qualitative differences between technical mixtures of PCBs and the mixtures occurring in the environment (see Since the susceptibility of fish to Aroclor 1254 was clearly higher compared with effects of the single congener PCB 28 tested in the present study, PCB 28 cannot be the congener which led to the major toxic effects of PCBs. However, combined effects of this single congener with other PCB congeners can not be excluded.

6.2 Fish-species specific differences to chemical stressors

▼ 88 

Inherent sensitivity varies among species and chemicals because of the differences in the absorption (the amount of chemical entering the organism), translocation (movement of the chemical within the body), and biotransformation (metabolic activation, detoxification and excretion) of chemicals in relation to the presence of target and non-target receptors (Grue et al., 2002). However publications in which the susceptibilities of different fish species to water contaminants are compared are scarce. Physiological and behavioural responses to the same chemicals may differ markedly between species, for example the level of locomotor activity of rainbow trout (Oncorhynchus mykiss) and Atlantic salmon (Salmo salar) differed by one order of magnitude under the influence of zinc (Sprague, 1964; Sprague, 1968). If individuals or species differ in their responses to a particular gradient of an abiotic factor, changes in intra- or interspecific interactions can be expected. For example, the greater sensitivity in the avoidance response of rainbow trout (Oncorhynchus mykiss) to metal contamination compared with brown trout (Salmo trutta) may explain observed differences in the distributions of the two species in the Clark Fork River in Montana (Hansen et al., 1999).

Some publications are concerned with fish species specific differences of susceptibilities to cyanotoxins. In an acute toxicological study Bury et al. (1997) observed that brown trout (Salmo trutta) was less tolerant to MC-LR than rainbow trout (Oncorhynchus mykiss). From recent studies, it is known that some species from temperate lakes are highly sensitive to cyanotoxins e.g., whitefish (Cor e gonus lavaretus) (Ernst, 2001), whereas species from subtropical lakes are able to use it as food e.g., silver carp (Hypophthalmichthys molitrix), (Xie et al., 2004). Xie et al. (2004) found no detectable microcystin in the muscle and blood of silver carp (Hypophthalmichthys molitrix) in contrast to previous experimental results on rainbow trout (Oncorhynchus mykiss) and argue that silver carps have a mechanism to degrade or eliminate MC-LR actively and to inhibit its transportation across the intestines. Since both Danio rerio and Leucaspius delineatus showed highly sensitive reactions to MC-LR in the present study there is no hint for such an elimination mechanism.

For PCBs species specific differences were found in the extent of bioaccumulation for three estuarine fish species (red mullet Mullus barbatus, sea mullet Mugil cephalus and sea bass Dicentrarchus labrax) from the same area, whereby the extent of bioaccumulation was dependent on lipid content, habitat, dietary intake, growth rate and the metabolism rate of each species (Pastor et al., 1996). Significantly higher levels of PCBs were shown for slow growing Arctic char (Salvelinus alpinus) than for fast growing char (Hammar et al., 1993). Furthermore significantly higher mean tissue concentrations of Aroclor 1260 were found in predator species (0.23 ± 0.38 ppm) compared to bottom feeders (0.14 ± 0.24 ppm), whereas lower chlorinated Aroclors (1248 and 1254) did not lead to significant differences in residues in bottom feeding and predatory fish (EPA, 1999). Applying an individual-based model (IBM), the lower observed PCB concentrations in rainbow trout (Oncorhynchus mykiss) compared with lake trout (Salvelinus namaycush) from one lake were explained by the greater longevity in lake trout and the observed variation in PCB accumulation rate within the population of rainbow trout was explained by variation in prey PCB concentrations (Madenjian et al., 1994).

▼ 89 

In the present study the two fish species Danio rerio and Leucaspius delineatus were selected in order to analyse potential specific differences in effects from sublethal exposure to MC-LR and PCB 28. Because of its uncomplicated rearing and reproduction the species Danio rerio is often used in toxicological research. To compare the effects of Danio rerio with an endemic species the sensitive and partially endangered species Leucaspius delineatus was investigated. Since the observed differences in the reactions of both species to MC-LR and PCB 28 were rather small, the results of the species Danio rerio are comparable to those of the native European species Leucaspius delineatus. Some possible reasons could be the similar systematic origin from the family Cyprinidae, comparable size and weight and age (see 4.1).

6.2.1 Effects on locomotor activity of Danio rerio and Leucaspius d e lineatus

Within the context of different behavioural responds, locomotion plays an important role because it is an integrative indicator of the internal status of the animals. The usefulness of locomotor responses in behavioural ecotoxicology is based on the fact that they are objective and automatically quantifiable and can be measured effectively in a variety of fish species to characterise the consequences of sublethal exposures (Little et al., 1993).

The experimental system, that was used to observe and report the fish behaviour, was an artificial one. Beside its above-mentioned advantages there are restrictions, most obvious, the limited volume of the aquaria.

▼ 90 

Interestingly despite this relatively small volume the method allowed not only the description of the swimming velocity but also of the swimming mode of fishes. So the relationship between number of turns and motility clearly indicated species specific differences (see 5.1.5 and 5.2.5). This implied that the pattern of swimming activity were not generally determined by the limited volume of the aquaria. Microcystin-LR induced behavioural effects

Some changes in the spontaneous locomotor behaviour of both fish (Danio rerio and Leucaspius delineatus) suggest that bothspecies have a comparable susceptibility to MC-LR in this study. Whereas lower concentrations increase the motility, the highest exposure concentration, 50 µg l-1, caused a significant decrease in activity of both Danio rerio and Leucaspius delineatus. That corresponds well to the hormesis theory (see 6.3.1).

However, some differences between the two species were found in their reaction to MC-LR. So analysis of mean motility including exposure-time dependent aspects showed that Leucaspius delineatus reacted earlier (at MC-LR concentrations of 0.5, 5, and 15 µg l-1) and for a longer time (at 50 µg l-1) than did Danio rerio. Along with motility, turns characterise the swimming activity changes. The numbers of turns were significantly increased during the night phase at all MC-LR concentrations for Leucaspius delineatus and at elevated MC-LR concentrations for Danio rerio. Furthermore the regression analysis between motilities and turns was indicative for MC-LR induced effects on swimming performance of fish. The statistic relation (r2) between the increase in number of turns and increasing motility was weaker at elevated concentrations of MC-LR compared to the control for Danio rerio and, however, was stronger compared to the control for Leucaspius delineatus. Additionally for Danio rerio the significantly decreased slope of the regression at the highest MC-LR concentration of 50 µg l-1 indicated that the fish decreased the number of turns at a given motility and swam more smoothly. In contrast, for Leucaspius delineatus the slope of regression increased under the influence of MC-LR indicating the fish swam more jerkily under exposure. Trichlorobiphenyl induced behavioural effects

▼ 91 

Some reactions of Danio rerio as well as Leucaspius delineatus exposed to PCB 28 arevery similar, as it was shown for the MC-LR induced reactions. Elevated exposure concentrations of PCB 28 (150 µg l-1) caused a significant decrease of activity in Danio rerio as well as in Leucaspius delineatus. Analysis of mean motility including exposure-time dependent aspects showed that both Danio rerio and Leucaspius delineatus reacted to the highest PCB 28 concentration (150 µg l-1) from the first to the last exposure interval. Beside those similarities, PCB 28 led to some changes of spontaneous locomotor behaviour of both fish species which suggest some differences between Danio rerio and Leucaspius d e lineatus. At the lower PCB concentration of 100 µg l-1 the mean motility over the whole measuring time per day did not change significantly for Leucaspius d e lineatus, whereas the activity level of Danio rerio was significantly reduced. However differentiating the activity rhythms of Leucaspius delineatus according to daylight and nighttime activity significant effects of PCB 28 at the lower PCB concentration (100 µg l-1) were found, too.

Indicated by the analysis of exposure-time dependent aspects Danio rerio reacted more rapidly and over a longer time period than did Leucaspius deline a tus at PCB concentrations of 100 µg l-1. Significant changes of motility of Leucaspius d e lineatus were only recorded in the second exposure interval and afterwards disappeared from the third exposure interval up. The numbers of turns were significantly decreased during the daylight phase at PCB 28 concentrations of 100 and 150 µg l-1 for Danio rerio, and only at elevated concentrations (150 µg l-1) for Leucaspius delineatus.

The regression analysis between number of turns and motility indicated similar statistic relations under exposure of PCB 28 or control conditions for Danio rerio. In contrast, for Leucaspius delineatus the statistic relation between the increase of number of turns and increasing motility was stronger and the slopes of the regressions were significantly increased under PCB 28 exposure. Therefore, only for Leucaspius delineatus the variability of the pattern of swimming performance was decreased and the kind of swimming performance was changed in that kind that the fish swam more jerkily under the influence of PCB 28.

6.2.2 Effects on chronobiology of Danio rerio and Leucaspius d e lineatus

▼ 92 

Analysis of temporal patterns in biological data gives insight into acclimation processes. If the fish are well acclimated to the experimental conditions they will obtain stable rhythms of locomotor activity synchronised by external rhythms, e.g., by temporal variations of light and food (e.g., Schwassmann, 1980; Boujard and Leatherland, 1992; Siegmund and Biermann, 1992). In contrast, disturbance may lead to a desynchronisation between biological rhythms and their zeitgeber, whereby if the peaks of activity are occurring later (delay) or earlier (advance) in comparison to unexposed conditions the rhythm is phase shifted. Cyclic aspects of behaviour under unexposed conditions

Since the swimming activity of fish may show distinct day/night differences (e.g., Plaut, 2000; Campbell et al., 2002) biorhythmic aspects have to be necessarily considered in behavioural studies. Under unexposed conditions both fish species of this study Danio rerio and Leucaspius delineatus exhibited a significantly diurnal activity (indicated by the effects of zeitgeber) which coincides with other studies dealing with the circadian periodicity of Danio rerio (Baganz et al., 1998; Hurd et al., 1998; Plaut, 2000) and Leucaspius delineatus (Siegmund & Wolff, 1973). The daily activity of both fish species was synchronised with the artificial alternation of light and dark and the feeding times within the 24 hours period. The maximum of daily activity of Danio rerio as well as Leucaspius delineatus was registered shortly after the onset of light. This was the time period of the daily mating and spawning behaviour of Danio rerio.

Spectral analysis of activity in both fish species showed a time pattern which was characterised by a dominant 24 hours rhythmicity, but also by ultradian components with period length between 4.8 and 12 hours. Single cosinor analysis revealed significant circadian rhythms with equal periods of 24 h ± 1.14 h for both fish species, whereas the values of amplitude and MESOR were higher for Danio rerio than for Leucaspius delineatus. Microcystin-LR induced chronobiological changes

▼ 93 

There were some similar reactions of Danio rerio and Leucaspius delineatus in response to MC-LR exposure. In both fish species a degree of desynchronisation to the light/dark change was found which led to a phase shift in Danio rerio as well as in Leucaspius d e lineatus indicating that the influence of the zeitgeber light decreased under exposure to the toxin. This was expressed by the reduced value of the Degree of Functional Coupling (DFC) which is an objective parameter of the coordination of different organismal functions both with each other and with the external circadian zeitgeber (Scheibe et al., 1999). A reduced value of DFC for activity of alpacas (Lama guanicoe f. pacos) was for instance found for an accidentally hurt animal in comparison with healthy animals by Scheibe et al. (1999).

Further stress symptoms caused by MC-LR (at a concentration of 50 µg l-1) were the reduced amplitude which was significant only for Danio rerio and the significantly decreased MESOR for Danio rerio as well as Leucaspius delineatus.The latter one shows the drastically reduced activity level in both fish species. The significantly increased MESOR or amplitude at lower exposures reflect a significant increase in daytime motility.

However there were some differences between the two test species concerning the observed phase shift which is graphically represented in the polar plots. For Danio rerio a phase delay occurred, whereby at all concentrations the changes could only be registered during the light phase. Therefore, Danio rerio remained still significantly diurnally active (indicated by the effects of zeitgeber). In previous studies of the author it was shown that the phase delay was associated with circadian variations during the daily spawning time and an evidently reduced reproduction success at an MC-LR concentration of 50 µg l-1 (Baganz et al., 1998). In contrast, the phase of Leucaspius delineatus advanced, whereby this shift was so drastically that a phase reverse occurred and this species became significantly nocturnal as indicated by the effects of zeitgeber (for considerations of ecological effects see 6.3.3). Comparable changes in the circadian periodicity of fish activity (phase shift of activity, degree of desynchronisation to the light/dark change) induced by nitrogen compounds were found e.g., by Biermann (1992).

▼ 94 

Performing cosinor analysis on motility of Danio rerio resulted in distinctive single peaks and a stable period length of circadian rhythms for all tested MC-LR concentration as well as the control. However, for Leucaspius delineatus these single peaks of circadian motility rhythms where found only at the control and the lowest applied MC-LR concentration. Interestingly, at elevated MC-LR concentrations clear double peaks of circadian motility rhythms of Leucaspius delineatus occurred, indicating two oscillations of shorter periods. These two main oscillations were found by the power spectrum analysis of Leucaspius delineatus, too, whereby the dominance of the circadian rhythmic peak was clearly reduced, and simultaneously, the proportion of a harmonic oscillation with a 12 h rhythm increased under the influence of higher concentrations of MC-LR. Because of this increased amount of ultradian rhythms, the period length calculated by cosinor analysis as the mean value of all frequency parts significantly decreased for Le u caspius delineatus. In contrast to Leucaspius delineatus, for Danio rerio the ultradian rhythms were more affected than the circadian component by MC-LR exposure. The same effect was found by Scheibe et al. (1999) for alpacas (Lama gu a nicoe f. pacos) under stress conditions. These findings emphasize the importance of the circadian rhythmic component. Trichlorobiphenyl induced chronobiological changes

Similar reactions of Danio rerio and Leucaspius delineatus in response to PCB 28 (at a concentration of 150 µg l-1) were reflected by a significantly reduced amplitude as well as MESOR. The reduced MESOR indicates the reduced activity of both species (influenced by PCB 28) over the whole measuring time per day. The values of the effects of zeitgeber were significantly reduced for Danio rerio as well as Leucaspius delineatus,whereby both speciesremained diurnally active with a stable period length under exposure. In contrast, some effects on cyclic aspects of behaviour clearly indicate that Leucaspius delineatus reacted more sensitively to PCB 28 than Danio rerio did. A degree of desynchronisation to the light/dark change (indicated by the reduced values of DFC) which led to a phase advance, was only found for Leucaspius delineatus under exposure of PCB 28. Furthermore only for Leucaspius delineatus the dominance of the circadian rhythmic peak was reduced whereas for Danio rerio the complex of ultradian components did not change under PCB 28 exposure.

6.2.3 Stress potential of Microcystin-LR and Trichlorobiphenyl

Taken together, the analysed parameters gave clear evidence that both the cyanobacterial toxin MC-LR as well the xenobiotic chemical PCB 28 revealed stress symptoms on the behavioural and chronobiological level in fish occurring under sublethal conditions.

▼ 95 

Most of the changes in level and rhythms of swimming activity of both fish Danio rerio and Leucaspius delineatus suggest a comparable susceptibility to MC-LR as well as PCB 28. For MC-LR it could be shown that the reaction of Leucaspius d e lineatus was a bit more sensitive than of Danio rerio. For PCB 28 the analysis of the activity level indicated that Danio rerio tended to be more sensitive than Le u caspius delineatus, whereas the registered rhythmical parameters indicate that Leucaspius delineatus reacted in this respect more sensitively to PCB 28 than Danio rerio did.

Some reactions of the Danio rerio as well as the Leucaspius delineatus were independent from the kind of chemical stressors (MC-LR and PCB 28), despite of their different chemical and physical properties. That means that the behavioural parameter responds non-specifically (but very sensitive) to any toxic chemicals. So, for both species elevated concentrations of the stressors led to a reduction of the activity.

Furthermore the rhythmical parameters (MESOR, amplitude) changed significantly under exposure of both substances (MC-LR or PCB 28), as indicated by the cosinor analysis. Analysis of the degree of synchronisation between activity rhythms and their zeitgeber (by power spectral analysis, DFC values and effects of zeitgeber) as well as the quantification of the harmonic frequency structure of activity rhythms (by power spectral analysis) proved to be good indicators for environmental changes, corresponding to findings of Siegmund and Biermann (1989, 1990) and Scheibe et al. (1999) (see 6.2.2).

▼ 96 

For Leucaspius delineatus both MC-LR as well as PCB 28 led to an increase of the slope of regression between number of turns and motility indicating that the swimming patterns were influenced in the same direction independently from the kind of stressor. However this could not be verified for Danio rerio, since PCB 28 did not affect the swimming patterns in contrast to MC-LR. The phenomenon of stressor-induced change of swimming mode has to be examined for its relevance in further studies.

6.3 Applications of behavioural approaches in ecotoxicology

This thesis is engaged in the recently still emerging field of behavioural ecotoxicology which integrates the three different disciplines: ethology, toxicology and ecology (Dell´Omo, 2002). Because the behaviour is the cumulative manifestation of genetic, biochemical, physiological and environmental cues, behavioural data may provide a link between individual response and population change, especially for those behaviours that manifest ecologically as changes in structure and function of the community (Little et al., 1985).

Toxicant-induced changes in behaviour may indicate toxicity (the failure of adaptive mechanisms) or conversely may be the adaptive response of an animal to mitigate or obviate the potential effects (Dell´Omo, 2002). Since fish have coevolved with cyanotoxins over phylogenetical long periods it could be hypothesized that protective adaptive mechanisms have developed. The fast and sensitive behavioural reactions found in this study support this theory. On the other hand it is rather unlikely that appropriate adaptive responses on the behavioural level to anthropogenic contaminants that are recent additions to the environment, have developed.

▼ 97 

Basic knowledge of exposure related behavioural alterations relevant for ecotoxicological assays remain scarce, and systems that have the ability to link toxicology data with swimming behaviours are still needed (Vogl et al., 1999).

The research on fish behaviour as an indicator of toxic effects is currently getting more and more attention (e.g., Chon 2002) and the findings in this study indicate that the non-invasive automatic registration of activity data is a suitable approach on the way to more sensitive ecotoxicological research methods and practicable for a range of applications.

Locomotor activity as a main component of behaviour is generated and controlled by different physiological processes and motivational states of an organism. Since it is important for such activities as feeding, predator avoidance, competitive interactions, migration behaviour (Reidy et al., 2000) it is also relevant in the ecological context. However one should be careful not to overinterprete effects found in the laboratory (Zala and Penn, 2004) since in natural ecological setting, individuals might be able to avoid the source of exposure or develop tolerance to pollutants (Barron, 2002).

6.3.1 Hormesis theory

▼ 98 

It has long been assumed that chemicals have a threshold level of safe exposure, and that dosage effects are linear; however, these assumptions have turned out to be obviously incorrect. Among others, Calabrese and Baldwin (2003) established the hormesis theory as a fundamental new concept in toxicology used to determine risks and risk regulations. According to these authors hormesis is a dose-response relationship phenomenon characterized by a stimulation of response at low doses and an inhibition of response at high doses.

Chemical stressors that elicit hormesis in organisms may induce therefore more than one mode of action, resulting in either a J-shaped or an inverted U-shaped response. Hormetic responses have been reported in hundreds of studies for a broad range of species (protozoa, bacteria, fungi, plants, invertebrates and vertebrates including humans), biological endpoints (e.g., survival, growth, reproduction), and both inorganic and organic chemicals (Calabrese and Baldwin, 1997). Therefore, such non-monotonic effects (or hormesis) are the rule rather than the exception in toxicology studies. Hormesis was also found on the behavioural level, e.g., male scent-marking behaviour increased when mice during foetal life were exposed to low doses of estrogenic pesticides (methoxychlor, DDT and a synthetic oestrogen (DES)), but marking behaviour declined again at the highest dose of DES (vom Saal et al., 1995). A stimulatory effect of a sublethal, light anaesthetic dose of tricane methane sulfonate (MS222) on velocity of mummichog (Fundulus heteroclitus) was observed by Kane et al. (2004).

Some of the dose-responses registered in the present study correspond to the hormesis theory: there was an increase of daytime activity level at lower MC-LR concentrations and a decrease of these effects at elevated concentrations of MC-LR for both Danio rerio and Leucaspius delineatus. Another hint of hormesis in the present study was the increase of the proportion of the 24 hours periodicity related to the whole harmonic frequency structure at lower MC-LR concentrations whereas this proportion decreased at elevated concentrations for Danio rerio. That means the synchronisation between activity rhythms and their zeitgeber was stronger at lower MC-LR concentrations and weaker at higher MC-LR concentrations compared to the control.

▼ 99 

However, at moderate concentrations no significant effects of MC-LR on the daytime motility were observed for both fish species. Furthermore the power spectrum of MC-LR concentrations of 15 µg l-1 isequal to that of the control for Danio rerio. Such an absence of a biological effect at moderate doses of a chemical (imazalil sulphate) on growing populations of the green algae Scenedesmus qua d ricauda was as well observed by Prokhotskaya et al. (2000), whereas lower concentration resulted in an increase of cell number and higher concentrations resulted in cell death.

Alternatively, the fact that PCB 28 did not elicit hormetic responses may be indicative for a single mode of action caused by higher levels of stressor exposure.

According to the hormesis theory the inhibition follows often an initial stimulatory response, appearing to represent a modest overcompensation of a disruption in homeostasis. However, higher exposures may exceed the capacity to compensate and lead to an adverse response compared with the controls.

▼ 100 

The following explanations for the phenomenon of hormesis were summarized by Prokhotskaya et al. (2000): Weak treatments impair the regulatory mechanisms, what leads to an activation of regulated processes. The further increase in the factor strength disturbs the functioning of regulated systems and suppresses metabolic reactions, which can result in cell death. Another explanation is based on the assumption that weak treatments induce detrimental changes in the cell qualitatively similar to those induced by stronger treatments, although to a lesser extent. Cell response involves protective mechanisms which, in addition to compensating for impairments, can result in hypercompensation. Stimulation of membrane receptors is believed to be a cause for the activation of cell functions. Under stronger actions, the activity of membrane receptors is suppressed.

From the ecotoxicological point of view it is important to determine whether hormesis has a positive, neutral or adverse effect on the overall health of organisms. Positive effects of hormesis e.g., on the longevity are demonstrated by some authors (Neafsey, 1990). So hormesis is an active survival strategy of organisms under altered conditions (Prokhotskaya et al., 2000).

6.3.2 Behavioural strategies

The strategies used by organisms to resist chemical exposure can be classified into several general categories: avoidance or escape reactions, exclusion (for example, many aquatic animals exposed to toxic chemical secrete mucus onto exposed surfaces), removal (in-coming toxicants might be actively pumped out), detoxification, possibly followed by excretion (e.g., by sequestration in granules or via metabolic transformation), and repair of damage caused by toxicants (Forbes and Calow, 1996). Therefore, the increase in daytime motility of Danio rerio and Le u caspius delineatus at lower MC-LR exposures in the present study can be interpreted as an escape reaction and/or as an increased spatial orientation behaviour. Tembrock (1984) described spatial orientation as an active behavioural adaptation to cope with changing ecological conditions. However increases of activity are costly for the organisms in terms of metabolic resources and especially energy. This energy needed for faster swimming is not available for other living processes, e.g., growth or reproduction. Thus, the determination of whether the low dose stimulation (according to the hormesis theory) is beneficial or harmful is not always obvious and must be judged in every single case.

▼ 101 

Increased MC-LR concentration caused significant decreases in motility of both species over the whole measured time per day. Hence an escape and/or orientation behaviour was not any more the dominant strategy under increased exposure. It is hypothesized that physiological disruptions (connected with an increased energy demand) are responsible for the observed effects, whereby at the point of maximum achievable metabolic scope under these conditions the amount of surplus energy available for swimming was supposed to be significantly decreased. This assumption was supported by the findings of some studies which found significantly lowered swimming speeds in fish exposed to sublethal concentrations of toxicants (McGeer et al., 2000; Campbell et al., 2002; Hopkins et al., 2003). Certainly at least a part of the saved energy is needed for the biotransformation process of the toxins.

Organisms are able to metabolise toxins by oxidation, reduction and hydrolysis (phase 1; catalysed e.g., by the cytochrome enzyme P450 monooxygenase (CYP 1A)), and by conjugation (phase 2, e.g., via glutathione-S-transferase).

The biotransformation process of MC-LR by elevation of microsomal and cytosolic glutathione-S-transferase activity, was described by Pflugmacher et al. (1998), who identified in various aquatic organisms including fish an enzymatically formed glutathione conjugate of MC-LR as the first step of biotransformation and its degradation to a cysteine conjugate. An elevated activity of biotransformation enzymes: microsomal and cytosolic glutathione-S-transferase; glutathioneperoxidase in zebrafish embryos was found by Wiegand et al. (1999).

▼ 102 

Exposure to PCBs led to a significant elevation of the cytochrome enzyme P450 monooxygenase (see chapter as well as of the conjugation enzyme system (cytosolic glutathione-S-transferase) in fish (Koponen et al., 2000; Schmidt 2004).

PCB biotransformation has been shown to lead to two classes of PCB metabolites that are present as contaminant residues in the tissues of selected biota: hydroxylated (HO) and methylsulfone (MeSO2) PCBs (Letcher et al., 2000). Hydroxylated PCBs in lake trout (Salvelinus namaycush) blood plasma were found by Campbell et al. (2003). Both hydroxylated and methylsulphonyl metabolites of PCBs are reported to be toxic (Brouwer et al., 1997).

The significance of energy saving behaviour under exposure is also expressed by compensation processes between daylight and night activity. So it was shown that for Danio rerio increases in daylight activity from MC-LR exposure (at 0.5 µg l-1) were compensated for, at least partly, by decreases in activity at night. The same applied to Leucaspius delineatus, whosedecrease in daylight motility (at MC-LR concentrations of 15 µg l-1 and 50 µg l-1; at PCB 28 concentration of 100 µg l-1) corresponded with a motility increase in the dark. Obviously, this activity shift compensates the elevated energy expenditure during the metabolic more active phase. Therefore, the data show the importance of evaluating activity rhythms differentiated according to daylight and nighttime activity, while analysing only the average activity over the days span would minimize the effects of the stressor by levelling out the phase related effects.

▼ 103 

Reducing the activity (and the costs that go along with it) can be an effective strategy if exposure is of temporary nature and the energy made available for stress decreasing processes (e.g., enzyme production) is sufficient to overcome this period. However a lowered activity over longer-term exposures is likely to indicate an impaired performance e.g., in the form of reduced feeding and mating activity (Forbes and Calow, 1996) and may lead to a metabolic collapse in the organism. In previous studies the reduced activity level and the phase shift of activity of Danio rerio at the MC-LR concentration of 50 µg l-1 coincided with a reduced spawning activity and success (Baganz et al., 1998). The species Le u caspius delineatus reproduces in April-May near shorelines and among vegetation where in meso- to eutrophic lake blooms of cyanobacteria may occur because of, for example, wind and current, and therefore an impact on its spawning behaviour is possible. However it is difficult to synchronise the reproduction of Leucaspius delineatus under laboratory conditions since the reproduction of this species is limited to a special period of the year.

Because swimming performance is of central importance for many aspects of fish biology, their reduction following exposure to contaminants could ultimately diminish the fitness in affected individuals and have consequential implications for inter- and intraspecific interactions (Hopkins et al., 2003). Plaut (2000) pointed out that a reduction in swimming capability, resulting in a reduction in the rate of activity, may decrease the ability to gather food, making the fish vulnerable to predation. Impairments in foraging behaviour of mummichogs (Fundulus heter o clitus) from contaminated sites appeared to have accounted for their reduced growth and longevity in comparison to those from uncontaminated sites, whereby grass shrimps, which are an important prey species of the mummichogs, had a greater population density and a larger size-frequency distribution at the polluted site, apparently because of reduced predation pressure (Weiss et al., 1999).

6.3.3 Chronobiological analysis

Quantifying stress induced behavioural changes of continuous recordings facilitated the registration of rhythmical changes and provided therefore novel information on the stress potential of the investigated contaminants. Time patterns in swimming activity showed an at least as sensitive response to the chemical stressors, as it was shown for the basic behavioural parameters. So the LOEC-value for rhythmical parameter was in the same range as for basic behavioural parameter for both MC-LR as PCB 28.

▼ 104 

With finding a mathematical equation for modelling the circadian activity rhythms of fish over a given time period by using the cosinor analysis it was possible to explain the substantial proportion of data which inhered a 24 hours periodicity.

The power spectral analysis using the program "Zeit" (Scheibe et al., 1999, 2002) with their amplitude coefficients gives a measure of how well the activity rhythms of fish fit infradian, circadian and ultradian sinusoidal frequencies. The remaining proportion of data was only the white noise, that means the part of the raw data which inhered no more significant rhythmic components (StatSoft, Inc., 2004).

Deferment of phase relations, changes in frequency structure, loss of rhythmicity or reduction of amplitude are regarded as signs of adaptation, disease or pre-mortal state (Scheibe et al., 1999). For analysing changes of circadian rhythms of humans caused by diseases (e.g., Alzheimer’s Disease) some studies deal with an evaluation of various rhythmic parameters like MESOR, amplitude, acrophase and period (e.g. Volicer et al., 2001). For some species of hoofed animals the ultradian rhythms of activity were more affected than the circadian component by external disturbances using the power spectral analysis (Berger et al., 2002, 2003). However these utilised analytical methods of biorhythmic research have rarely been applied for estimating the risks of aquatic contaminants, and it would be useful to bring them more into focus of ecotoxicology. Methods of biorhythm research require continuous activity records over extended periods of time.

▼ 105 

Since the response of rhythmic changes seems to be a characteristic reaction of fish exposed to waterborne toxicants, it is a good indicator of sublethal stress (Siegmund & Biermann, 1990). This has been verified by investigations on fish in the present study, whereby the results proved that the extended methods of time series investigation like cosinor analysis and power spectral analysis can be valuable tools for the study of harmful environmental factors. In the present study the amount of ultradian rhythms of Leucaspius delineatus were clearly increased under MC-LR exposure.

The DFC which is an objective parameter of coordination of different organismal functions both with each other and with the external circadian zeitgeber, enables an assessment of the organismal state (Scheibe et al., 1999). Analysis of DFC in the present study indicates that the synchronisation between circadian rhythms of activity and their external zeitgeber was weaker under the influence of the chemical stressor (see 5.1.8 and 5.2.8). The reduction of the synchronising effects of light on the locomotor activity rhythm in both Danio rerio and Leucaspius delineatus was partly accompanied by a shift of the most active period of the day to another time period of the day.

Disturbances of that equilibrium (the homeostasis) can be answered by a variety of integrative reactions (biochemical, physiological, behavioural) with different priorities. The high flexibility of biological rhythms (Aschoff 1984) results in a rather early alteration of circadian periodicity to deal with stress conditions.

▼ 106 

According to the above mentioned enhanced spatial orientation behaviour under exposure, the observed phase shift of behavioural rhythms in both Danio rerio and Leucaspius delineatus species could be interpreted in a similar way as an enhanced or new temporal orientation. Rhythmicity has a great adaptive value for the precise temporal fit of the organism into the ecosystem (Tembrock, 1992), the ecological niche has therefore not only a spatial but also a temporal component.

Because various environmental conditions oscillate cyclically, the ability to anticipate temporal changes in the environment would enable an organism to be prepared, both physiologically and behaviourally, to perform specific activities when the environmental conditions are most favourable to the species (Hoenen & Gnaspini, 1999). Therefore, the phase shift of behavioural rhythms under exposure may lead on the one hand to physiological disturbances of the temporal coordination of internal processes (internal desynchronisation; Aschoff, 1969). On the other hand it may lead to interspecies disadvantages under natural conditions, e.g., in the context of efficiently searching for prey and avoiding predators.

As mentioned above biorhythmic changes can also come along with a reduced reproduction success. In fishes it was demonstrated that feeding time affects growth performance (Boujard and Leatherland, 1992; Bolliet et al., 2001) which indicated the importance of synchronisation between external rhythmical events and internal processes. It can be assumed that a variety of behavioural and physiological rhythms which are usually coupled become dissociated under stress conditions.

▼ 107 

Such a dissociation between circadian rhythms of swimming activity and heart rate of fish indicated harmful effects on fish caused by temperature stress (Siegmund, 1981). Disturbances of coordination between respiratory-cardiovascular processes are used for identifying acute toxicity syndromes of rainbow trout (O n corhynchus mykiss) (McKim et al., 1987). Bolliet et al. (2004) suggests that physiological rhythms involved in nutrient utilisation might not be as flexible as feeding rhythm thereby leading to a desynchronisation between rhythms and possibly a decrease in growth and feeding efficiency.

The ability to predict a regularly occurring environmental change accurately and consistently and anticipate the necessary behavioural adjustments is critical to the survival of a population (Adgins-Regan and Weber, 2002). Johnson et al. (2003) addressed the adaptive significance of circadian rhythmicity by testing the relative fitness under competition between various strains of cyanobacteria expressing different circadian periods, whereby strains that had a circadian period similar to that of the light/dark cycle were favoured under competition.

The effects of MC-LR and PCB 28 were more drastically at the time of switching the light on or off as it was shown by the smoothed curve of average motility. Daily variations in sensitivity of fishes to harmful stimuli were also shown by Spieler et al. (1977), whereby fathead minnows (Pimephales promelas) and golden shiners (Notemigonus crysoleucas) exhibited differences in sensitivity to potentially lethal levels of chlorine, formalin, or heat, depending on the time of day. Therefore, the meaning sense and the reliability of behaviour measurements to recognise stressful impacts can be enhanced by a chronobiological analysis of data, as confirmed by several authors (Siegmund & Biermann, 1992; Steinberg et al., 1995; Grillitsch et al., 1999).

6.4 Digression: Biomonitoring

▼ 108 

Some of the research done within the scope of this study may potentially result in an enhancement of recently used biomonitoring methods. Although biomonitoring was not the subject of this study, behaviour was used as a biomarker for indicating the presence of a stressor and that is the aim of biomonitoring, too.

Automated biomonitoring or biological early warning systems are defined as systems that detect toxic conditions on a continuous basis in whole organisms (Butterworth et al., 2000). In contrast to physico-chemical analyses biomonitors facilitate an unspecific indication of pollutants including synergistical and antagonistical effects, in water monitoring. Using behavioural endpoints for biomonitoring has the advantages a) that their high sensitivity is comparable with other toxicological tests, e.g., enzymatic tests (see 6.1), b) that they have the capability of an online monitoring process without disturbing the test organisms and c) that organisms respond with behavioural changes within short time periods (Blübaum-Gronau et al., 2000).

Quantifying fish behaviour in the present study by using the online video-processing system BehavioQuant® (Spieser et al., 2000) proved to be useful for the detection of discontinuities and slight alterations in the normal behaviour. The experimental design of this study was developed under the explicit heeding of different standards (see 4.2 and Baganz et al., 2000). The tests organisms were sufficiently acclimated to the standard test conditions to enhance the reliability of the measurement by reducing higher variability of behaviour; recommended are 2-4 weeks (Siegmund and Biermann, 1990; Baganz et al., 2000). Even a sound experimental design combined with proper statistical analyses is essential for approaches using fish movement for the bioindication of stressors (Vogl et al., 1999).

▼ 109 

The present study showed irrefutably that in behavioural experiments different distinctive patterns of animals’ reactions need to be necessarily considered, depending on the species, the stressor and the specific time of the day. Since critical concentrations at a very low level can be measured and potential dose-effect relationships can be registered, biomonitoring using fish behaviour is about to become a standard. On this way some research has to be done to simplify the modelling of answer reactions detection to make it more reliable and more independent of the kind of reaction. So the results of this study indicate that in some cases it is necessary to register the absolute deviation from standard values independent on the direction of the reaction (e.g., if there is an increase or a decrease of activity and/or of the amount of rhythmical parameter). Furthermore for a successful use of behavioural tests in biomonitoring circadian periodic changes of activity should be regarded to prevent false alarms. That means that the alarm algorithms of biomonitoring should be calculated by defining a range of valid values following the temporal pattern of swimming behaviour under standard test conditions.

Some current biomonitoring systems use constant light conditions (LL) to reduce the circadian deviations of activity by eliminating the external zeitgeber. However it is probable that this procedure generates a stress potential for the test organisms. In addition the absence of zeitgeber reduces the synchronisation between all single fish of the school and may lead therefore to a higher variability of the behavioural pattern which is not useful in the context of biomonitoring.

The experiments of the present study were performed over rather long time intervals and so the focus was on circadian rhythms. During the shorter time observations used for biomonitoring and pre-warning especially ultradian rhythms have to be taken into account, because they can indicate disturbances even in time periods in the range of minutes. Ultradian rhythms (which have a period length of oscillation smaller than 20 hours; τ< 20 h) are commonly based on oscillations at the subcellular, cellular or supra-cellular level (Peters and Veeneklaas, 1992). The procedures of time series analysis that proved to be useful for analysis of circadian rhythm in this study can easily be exploited for the analysis of ultradian rhythms in biomonitoring applications.

© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
XDiML DTD Version 4.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated: