Environmental Toxicology and Chemistry, Vol. 28, No. 1, pp.

113122, 2009
2009 SETAC
Printed in the USA
0730-7268/09 $12.00 .00

SUBCHRONIC EXPOSURE OF HONEYBEES TO SUBLETHAL DOSES OF PESTICIDES:

EFFECTS ON BEHAVIOR

YASSINE ALIOUANE, ADESSALAM K. EL HASSANI, VINCENT GARY, CATHERINE ARMENGAUD, MICHEL LAMBIN,
and MONIQUE GAUTHIER*
Centre de Recherches sur la Cognition AnimaleUMR CNRS 5169, Universite Paul Sabatier, 118 route de Narbonne,
31062 Toulouse cedex France

( Received 7 March 2008; Accepted 7 July 2008)

AbstractLaboratory bioassays were conducted to evaluate the effects on honeybee behavior of sublethal doses of insecticides
chronically administered orally or by contact. Emergent honeybees received a daily dose of insecticide ranging from one-fifth to
one-five-hundredth of the median lethal dose (LD50) during 11 d. After exposure to fipronil (0.1 and 0.01 ng/bee), acetamiprid (1
and 0.1 g/bee), or thiamethoxam (1 and 0.1 ng/bee), behavioral functions of honeybees were tested on day 12. Fipronil, used at
the dose of 0.1 ng/bee, induced mortality of all honeybees after one week of treatment. As a result of contact treatment at 0.01
ng/bee, honeybees spent significantly more time immobile in an open-field apparatus and ingested significantly more water. In the
olfactory conditioning paradigm, fipronil-treated honeybees failed to discriminate between a known and an unknown odorant.
Thiamethoxam by contact induced either a significant decrease of olfactory memory 24 h after learning at 0.1 ng/bee or a significant
impairment of learning performance with no effect on memory at 1 ng/bee. Responsiveness to antennal sucrose stimulation was
significantly decreased for high sucrose concentrations in honeybees treated orally with thiamethoxam (1 ng/bee). The only significant
effect of acetamiprid (administered orally, 0.1 g/bee) was an increase in responsiveness to water. The neonicotinoids acetamiprid
and thiamethoxam tested at the highest dose (one-tenth and one-fifth of their oral LD50, respectively) and fipronil at one-five-
hundredth of LD50 have limited effects on the motor, sensory, and cognitive functions of the honeybee. Our data on the intrinsic
toxicity of the compounds after chronic exposure have to be taken into account for evaluation of risk to honeybees in field conditions.

KeywordsPesticides Chronic Sublethal Honeybee Behavior

INTRODUCTION TS (BASF, Ludwigshafen, Germany) with insecticidal prop-

The honeybee Apis mellifera is valuable for the economy erties similar to those of imidacloprid contained in Gaucho
due to the products of the hive (honey, pollen, royal jelly), (Bayer AG, Leverkusen, Germany). Fipronil is now a major
which generate considerable income for beekeepers, as well pesticide for use on crops but also as an antiparasitic, with an
as to its contribution to crop pollination, which is valued at estimated world market of $150 million [2]. It is suspected of
more than $15 billion a year in the United States alone (U.S. having harmful effects on honeybees and has been forbidden
Department of Agriculture, http://ars.usda.gov/main/main. in France because of its potential involvement in bee declines.
htm). Honeybees also contribute to plant biodiversity by pol- The previously mentioned insecticides are neurotoxic com-
linating wild plants. Honeybees and their products are poten- pounds that act on ion channels within the insect nervous
tially exposed to several contaminants present in the environ- system. The neonicotinoids imidacloprid, acetamiprid, and
ment, such as chemical products released into the hive to fight thiamethoxam have the same target at the cellular level, acting
against diseases and parasites and pesticides used in agriculture mostly as agonists of insect nicotinic acetylcholine receptors
against pests. (nAChRs) [3]. The and subunit compositions of these
The continuing need for novel and selective insecticides receptors define different nicotinic receptor subtypes differing
acting on pests has led to the development of new groups of in their pharmacological properties. In honeybees, as in insects
compounds. The newest major group of insecticides are the in general, the subunit composition of nAChRs is unknown.
neonicotinoids, which include imidacloprid, acetamiprid, and Patch-clamp experiments performed on honeybee brain neu-
thiamethoxam [1]. Worldwide annual sales of neonicotinoids rons in culture have shown that imidacloprid is a partial agonist
total $1 billion, and they are used against piercingsucking of nAChRs [47]. At least two types of nAChRS have been
pests (aphids, leafhoppers, and whiteflies) of major crops. In described in the honeybee brain: -bungarotoxin (-BGT)-
France, the use of imidacloprid has been suspended because sensitive and -BGT-insensitive nAChRs [8]. These receptors
of concerns that it may have a drastic effect on bee populations, are involved in tactile and olfactory learning and memory [9
causing loss of honeybees and weakening hives. Acetamiprid 12], which are essential functions for foraging behavior. Fi-
and thiamethoxam are presented as potential alternatives to pronil disrupts inhibitory neurotransmission by blocking the
imidacloprid subject to proof that they are harmless to non- -amino-butyric-acid (GABA) gated-chloride channels as well
target species. Fipronil belongs to the phenylpyrazole group as the glutamate gated-chloride channels (GluCl) [4,13]. Since
and is the first product of this group to be introduced for pest mammals are devoid of this type of chloride channel, the action
control. It is the active molecule of the insecticide Regent of fipronil in blocking the glutamate-activated chloride channel
is considered to be responsible, at least partially, for its higher
* To whom correspondence may be addressed (gauthiem@cict.fr). selective toxicity to insects over mammals [14]. Gabaergic
Published on the Web 8/13/2008. interneurons and neurons bearing GluCl receptors have been
113
114 Environ. Toxicol. Chem. 28, 2009 Y. Aliouane et al.

found in several honeybee brain neuropiles, where they act as experiments, acetonitrile for thiamethoxam) was 0.3% (v/v)
modulators of excitatory synapses [4,15]. In olfactory path- for oral and 10% (v/v) for topical applications in final solu-
ways, gabaergic interneurons contribute to shaping the neural tions.
representation of odors [16,17]. Contrary to nAChRs, GABA
and GluCl receptors are also found outside the central nervous Animals
system on the muscle membrane, where they regulate the ex-
The tests were performed all year-round in Toulouse, in the
citatory glutamate neurotransmission at the neuromuscular
south of France, on emergent honeybees (A. mellifera). In
junction [18,19].
September the hives received a single treatment against varroa
We have previously performed experiments on the biolog-
(Apivar; Laboratoires Biove, Arques, France), and for the
ical effects of acute sublethal oral or contact exposure of hon-
first month after treatment no sampling took place. Since 2004
eybees to acetamiprid and thiamethoxam. Thiamethoxam in-
the pesticides Gaucho and Regent TS have been forbidden in
duced no effect on behavioral functions, whatever the dose
France, thus minimizing exposure of honeybees to these pes-
and the delivery mode. Acetamiprid had an activating effect
ticides. Honeybees were collected from hives placed in con-
on behavior, which appeared as an increase in sucrose re-
trolled room temperature (23C). Working on emergent bees
sponsiveness and in locomotor displacements, but also induced
makes it possible to control their age and keep them alive
long-term memory impairment after oral absorption [20]. The
longer in laboratory conditions. The bees were caught on a
experiments conducted with fipronil in similar conditions
brood frame when emerging from the cells. They were caged
showed a slight decrease in sucrose responsiveness and olfac-
in groups of 40 individuals and maintained in darkness under
tory memory impairment after topically applied sublethal dos-
controlled conditions (40% relative humidity, temperature
es [21].
33C). Pollen and sucrose solution (50% w/v) were provided
As neonicotinoids are strongly suggested to be systemic
ad libitum for the first week. The bees were then allowed to
[22], the xylem transport in the plant could result in the pres-
make a purging flight before returning to their cages, where
ence of tiny quantities of the molecules in nectar and pollen.
they were subjected to an 11-d exposure period. During this
Uptake of fipronil has also been demonstrated in the root of
period, the bees were fed with sucrose solution (50% w/v) and
sunflowers, leading to transport into leaves [23]. Pollen and
water. The feeders were changed daily with fresh solutions.
nectar foraging on plants treated with systemic insecticides
can lead to an accumulation of these products in the hive, and Exposure protocols
young honeybees can be exposed to repeated sublethal doses
of pesticides during their early life. The present study examines Two modalities of exposure were used: oral and contact
whether the rather limited behavioral effects we observed after exposure. For oral treatment, the sucrose solution used for
acute exposure to the three pesticides would be amplified by feeding the bees contained the test compound or contained the
repeated exposure. In the laboratory, we reproduced subchron- solvent (control). The volume of the test compound sugar so-
ic intoxication of young honeybees with sublethal doses of lution was adjusted daily to the number of survivors on the
acetamiprid, thiamethoxam, and fipronil and are now able to basis of a consumption of syrup estimated at 33 l/bee/d [25].
report the effects of oral and contact exposure on sensory Control groups ingested a sugar solution containing the ap-
perception of water and sugar, locomotor displacements, and propriate solvent. Individual contact exposure consisted of ap-
olfactory learning and memory capabilities. plying the solution containing the compound under investi-
gation or the solvent alone (control) to the thorax of the hon-
MATERIALS AND METHODS eybee. To do so, each honeybee was caught in the cage daily
The following experiments have been performed by two and maintained with an insect forceps while 1 l of the solution
examiners but not in blind tests. was applied to the thorax using a micropipette with a tip. After
the drop disappeared, which took several seconds, the hon-
Drugs eybee was released into a new cage where the treated bees
were gradually collected. Throughout the exposure period, the
Fipronil (98.5% purity), thiamethoxam (97% purity), and
mortality per day was evaluated in control and treated groups
acetamiprid (99% purity) all were purchased from Cluzeau
and has been presented as cumulative-death curves in the re-
Info Labo (Sainte-Foy-La-Grande, France). The three com-
sults section.
pounds were used at doses of one-fifth to one-five-hundredth
of the median lethal dose (LD50) that elicit sublethal effects
Behavioral assays
[20,21,24]. Based on previous studies [25] and our own ex-
periments, the doses used per bee were 0.1 and 0.01 ng (LD50 For each pesticide and for each dose, a control group re-
oral 48 h: 46 ng) for fipronil, 1 and 0.1 ng (LD50 oral 48 h: ceiving the solvent was tested. The behavioral assays were
5 ng; LD50 contact 24 h: 29 ng) for thiamethoxam, and 1 and conducted in parallel in the treated and in the control groups.
0.1 g (LD50 oral 72 h: 14.5 g; LD50 contact 24 h: 7 g) The numbers of animals for each group and for each experi-
for acetamiprid. Stock solutions of fipronil and acetamiprid ment are indicated in Table 1. At the end of the exposure period
were prepared in acetone (Sigma, Saint Quentin Fallavier, (11 d), honeybees were individually tested for locomotor ac-
France) in accordance with the European and Mediterranean tivity, water and sucrose responsiveness, and learning abilities.
Plant Protection Organism guidelines [26]. Thiamethoxam was Locomotor activity was evaluated in free-moving animals. Ex-
dissolved in acetonitrile (Sigma) in accordance with the man- periments to test water and sugar responsiveness and learning
ufacturers recommendations. For topical application, the stock were carried out in restrained honeybees. Honeybees were cold
solution was diluted with distilled water to obtain the specific anesthetized and individually fixed in plastic tubes and main-
concentration. For oral delivery, the specific concentration was tained with a drop of wax/resin mixture deposited on the back-
obtained after a final dilution in sucrose solution (50% w/v). side of the thorax. The antennae and the mouthparts were left
The proportion of solvent (acetone for acetamiprid and fipronil free.
Chronic pesticide exposure and honeybee behavioral functions Environ. Toxicol. Chem. 28, 2009 115

Table 1. The numbers presented in this table represent the number of honeybees used for each experiment. For controls, the values are in italics. Asterisks
indicate in the corresponding behavioral assay that at least one value or one parameter shows a significant difference compared to control

Fipronil Thiamethoxam Acetamiprid

0.1 ng 0.01 ng 1 ng 0.1 ng 1 g 0.1 g

Size of groups Oral Topic Oral Topic Oral Topic Oral Topic Oral Topic Oral Topic

Mortality
Treated 17* 16* 36 23 56 32 65 49 58 44 34 29
Controls 17 16 37 21 56 33 59 49 55 40 35 31
Locomotion
Treated 27 27* 56 32 65 49 30 30 25 22
Controls 29 29 56 33 59 49 30 30 27 24
Water experiments
Treated 34 32* 38 35 40 36 26 26 27* 26
Controls 33 32 38 35 40 36 29 33 28 28
Sucrose experiments
Treated 27* 30 32* 25 28 32 27 30 25 22
Controls 29 30 28 27 24 33 30 30 27 24
Learning experiments
Treated 25* 30* 30 23* 33 29* 29 30 25 25
Controls 28 30 28 22 32 29 30 30 27 24

Locomotor activity Water responsiveness and sucrose responsiveness experi-

The locomotion of naive honeybees was tested in an open- ments were conducted in independent groups.
field apparatus consisting of a white Plexiglas box (30 30
Olfactory learning
4 cm) with a glass front allowing observation. The box
stood vertically and was illuminated from above to induce We used the olfactory conditioning of the proboscis exten-
upward displacements. The back vertical side was divided into sion reflex (PER) to assess cognitive functions in honeybees
squares of 5-cm sides, defining six vertical levels, allowing exposed to pesticides. The proboscis extension induced by
the bee to be localized into the apparatus. A hole made in the antennal contact with sucrose can be conditioned to an odorant
bottom right-hand side allowed individual honeybees to be as long as the odorant precedes the sucrose stimulation. After
introduced into the box for a 3-min observation period. The several paired associations of odor and sugar stimulations
position of the honeybee was recorded every 3 s on the screen (sometimes one association can be sufficient), the odorant be-
of a computer using specially adapted software [27]. Two suc- comes a conditioned stimulus (CS), and the honeybee responds
cessive recordings of the honeybee in the same square were to the olfactory stimulation by a proboscis extension (condi-
reported as a 3-s period of immobility. Relevant parameters tioned response). The learned olfactory stimulus can be re-
were the distance covered in the box, the time spent in each tained for a short (1-h) or long (24- and 48-h) period, and
level, and the displacements of the honeybee from lower to olfactory memory is evaluated through retrieval tests. After a
upper levels. Observations of the honeybees behavior such as 3-h fast period, honeybees were trained along five trials with
trembling or abnormal movements of legs or wings were also 1-min intertrial intervals. The olfactory stimulation (or CS)
reported. was directed toward the antennae for 6 s and diffused through
an air puff by means of a 5-ml syringe. The CS was an odor
Water responsiveness of coffee (5 mg of coffee powder introduced in the syringe)
Proboscis extension induced by stimulating the antennae for thiamethoxam experiments. We used one of the two odor-
with water is used to test exposed honeybees responsiveness ants, 1-hexanol and 1-nonanol, (Sigma) as CS for fipronil and
to water. One hour after being fed with a drop of water and a acetamiprid experiments. The odorant 1-hexanol or 1-nonanol
drop of 50% sucrose solution, restrained honeybees were stim- was diffused from a piece of filter paper placed in the syringe
ulated on the antennae with a drop of water. The number of and soaked with 2 l of the solutions. The odorants were
animals responding with a proboscis extension was evaluated renewed daily. Three seconds after the onset of the olfactory
in control and treated groups. The response to water was tested stimulus, the antennae were touched with a drop of sugar water
again, 3 h after the first test. Water consumption was controlled (40% w/v) for 3 s, and this stimulation induced proboscis
during the exposure period. The volume of consumed water extension. The honeybee was then allowed to feed on a sucrose
was measured in control and treated groups daily. solution for 3 s. Memory tests were performed 1, 24, and 48
h after the training phase, presenting the CS alone (coffee
Sucrose responsiveness flavor for thiamethoxam experiment) or in random presentation
Stimulating the antennae with a sucrose solution can also with a new odorant (1-hexanol for bees conditioned with
induce proboscis extension. The reflex is used to evaluate the 1-nonanol and, vice versa, for acetamiprid and fipronil exper-
sucrose responsiveness in honeybees displaying equal moti- iments). A bee that has learned the predictive significance of
vation to sugar. One hour after being fed with 50% (w/v) the odorant for food will extend the proboscis to the olfactory
sucrose solution, the honeybees antennae were stimulated at stimulation. When the honeybee did not respond to the sugar
3-min intervals with sucrose solutions of ascending concen- stimulation during the training phase, it was excluded from
trations of 0.03, 0.1, 0.3, 1, 3, 10, and 30% (w/v). The per- the experiment. A honeybee that did not respond to the odorant
centage of bees responding to sucrose stimulation by a pro- during the memory test was tested after 10 min for the sugar-
boscis extension was evaluated in treated and control groups. elicited proboscis extension. If no response was observed to
116 Environ. Toxicol. Chem. 28, 2009 Y. Aliouane et al.

Fig. 1. Percentages of honeybee mortality recorded during 11 d of exposure to fipronil, 0.1 ng (A) and 0.01 ng (B), after oral and topical
treatments. Controls received the solvent. Number of subjects in each group is indicated in parentheses. Statistical comparisons of the curves:
For fipronil, 0.1 ng/bee, oral versus control p 0.003; topic versus control p 0. 011; for fipronil, 0.01 ng/bee, oral versus control p 0.758,
topic versus control p 0.893, as determined by the KaplanMeier test.

sugar, the honeybee was discarded. The number of animals the values were compared between control and treated groups
that were excluded from the statistical analysis for the pre- for acquisition (from the second to the fifth trial) and for each
viously mentioned reasons was equivalent in control and treat- retention test (at 1, 24, and 48 h) using Fishers exact test.
ed groups and represented less than 5% of the total number. Within-group comparison for level response to conditioned
odorant versus new odorant was performed using McNemars
STATISTICAL METHODS test [29]. For each of these tests, a p value of less than 0.050
Comparison of the mortality curves between the control was considered significant. All the statistical tests were per-
and treated groups was performed with the KaplanMeier test. formed with SPSS12 software (SPSS Science, Chicago, IL,
Locomotor activity was evaluated through analysis of three USA).
relevant parameters: the path length during the 3-min obser-
RESULTS
vation period, the duration of immobility, and the time spent
in each level of the box. Students t tests were performed for All honeybees in the treated and control groups consumed
mean comparison between treated and control group values the syrup distributed daily. The consumption of the total
after variance comparison with Levenes test [28]. The mean amount ensures that honeybees were exposed orally to pesti-
water consumption per day per bee was calculated as the total cides. Attacks between animals were visible in fipronil-treated
amount of water consumed each day in each group divided by honeybees at the dose of 0.1 ng/bee and at the end of the 11
the number of surviving honeybees, and this calculation does d of exposure at the dose of 0.01 ng/bee. These attacks were
not allow error bars to be represented on the figure. The daily associated with frequent behavior of wing fluttering accom-
values of consumed water were compared between the treated panied by the emission of alarm pheromone.
and the control groups with a Students t test. Figures repre-
senting the proportion of bees releasing a PER in response to Mortality
water, sugar solution, or olfactory stimulation do not include Cumulative mortality curves were established for each pes-
error bars, which is the general practice for this type of data, ticide, for each dose and for each modality for the 11-d ex-
as standard errors or standard deviations do not accurately posure period. Fipronil (0.1 ng/bee, oral and contact) induced
reflect the variability for proportions. Responsiveness to water the death of all the animals one week after the beginning of
was compared between control and treated groups at 1 and 3 treatment. Mortality increased significantly compared to con-
h with a chi-square test. The comparisons between the groups trol, from day 3 of fipronil oral exposure (Fig. 1A, p 0.003),
for sucrose responsiveness were conducted using Fishers exact and from day 5 when fipronil was topically applied (Fig 1A,
test, which directly yields a p value. For olfactory learning, p 0.011). At the dose of 0.01 ng/bee, fipronil induced 25
Chronic pesticide exposure and honeybee behavioral functions Environ. Toxicol. Chem. 28, 2009 117

and 20% mortality, respectively, in orally and topically treated

animals at the end of the treatment (Fig. 1B). These percent-
ages were not statistically different from mortality in control
individuals (21.6 and 17.5%, p 0.758 and 0.893, respec-
tively). Because fipronil at the dose of 0.1 ng induced complete
mortality, the dose of 0.01 ng/bee was retained to study the
behavioral effects of chronic intoxication. At the end of the
exposure period with thiamethoxam (11th day), the percentage
of mortality in bees treated orally and topically with 1 ng was
10% (data not shown). The mortality in these groups was not
different from that of controls (4 and 10% respectively). A
maximum of 20% of dead bees was observed in the groups
exposed orally and topically to 0.1 ng thiamethoxam. The same
mortality level was observed in control animals (15% for oral
and 20% for topic exposure). Acetamiprid (1 g/bee) orally
and topically administered induced, respectively, 29.3 and
31.8% mortality after 11 d, but these values were not different
from the control (21.8 and 22.5%, respectively). At the dose
of 0.1 g/bee, mortality of orally and topically treated animals
was equivalent to their respective controls. Mortality ranged
from 20 to 26% for all the groups (data not shown).

Locomotor activity
Topical treatment of fipronil at the dose of 0.01 ng/bee
produced only the effect of spending more time in immobility
(Students test, t 2.631, p 0.011, df 58). The path
length and vertical displacements in the box decreased in treat-
ed animals, but the difference between the treated animals and
the controls was not significant in either case. No effect of
fipronil was observed at the same dose after oral treatment.
Thiamethoxam and acetamiprid had no significant effect on
the three parameters of locomotor activity compared to con-
Fig. 2. Honeybees behavioral response to water after fipronil exposure
trols, regardless of dose (0.1 and 1 ng/bee for thiamethoxam (A) or acetamiprid exposure (B). (A) Water consumption measured
and 0.1 and 1 g/bee for acetamiprid) or exposure route (oral daily in fipronil-treated and control groups. The total water volume
delivery or topical application). In all cases, honeybees seemed per day was divided by the number of survivors in each group, giving
less active than controls in the box and spent less time in the the mean water consumption per day per bee. See text for statistical
analysis. (B) Water responsiveness evaluated after chronic oral ex-
sixth upper level of the box and more time in levels 1 and 2. posure with acetamiprid (0.1 g/bee). *** p 0.001 (2 test). PER
The time spent in immobility was increased, and the distance proboscis extension reflex.
covered was reduced in treated honeybees. None of these dif-
ferences were significant. We never observed honeybees trem-
bling on the floor, fallen backward, or displaying abnormal respectively). Oral exposure to a dose of 0.1 ng/bee and contact
movements of wings, legs, or body. exposure to thiamethoxam (0.1 and 1 ng/bee) had no effect
on sucrose responsiveness (see Fig. 3B for contact 1 ng/bee).
Water consumption and water responsiveness Acetamiprid 0.1 g had no effect on the response rates of
A significant increase of the volume of water consumed by honeybees to the ascending concentrations of sucrose solutions
honeybees topically treated with fipronil was observed (Stu- for oral and topical treatment (data not shown). At the dose
dents test, p 0.0001), and this response was obvious during of 1 g/bee, topical treatment induced a nonsignificant increase
the first week of treatment (Fig. 2A). No effect was observed of responses to sucrose concentrations (Fig. 3C). We note that
on water responsiveness after fipronil oral treatment. Thia- this is the only case where the response curve of treated an-
methoxam induced no effect on water consumption and re- imals exceeded that of control bees.
sponsiveness. Acetamiprid given orally induced an increase in
water responsiveness (0.1 g/bee) at 1 h (2 test, p 0.0002) Olfactory learning
and at 3 h (2 test, p 0.0006) (Fig. 2B) but had no effect Fipronil (0.01 ng/bee) absorbed orally or applied topically
on water consumption. for 11 d had no effect on learning performance. An example
of the result for oral intoxication is given in Figure 4A. The
Sucrose responsiveness conditioned PER level was approximately 60% in each group.
Oral exposure to fipronil (0.01 ng/bee) induced a decrease Comparison of performance in the control group between the
in sucrose responsiveness with a significant difference for the 1-, 24-, and 48-h memory tests revealed a decay of memory
0.3 % sucrose concentration (Fig. 3A; Fishers exact test, p with time. This observation was valid for the control groups
0.042). Topically applied fipronil had no significant effect (Fig. of thiamethoxam and acetamiprid experiments as well (see Fig.
3A). Oral thiamethoxam (1 ng/bee) induced a decrease of hon- 4B and C). No effect of fipronil on the response to CS com-
eybees sucrose responsiveness to 3 and 10% sucrose concen- pared to controls was observed, regardless of the time of the
trations (Fig. 3B; Fishers exact test, p 0.001 and 0.008, test. This indicated no memory impairment induced by fipronil
118 Environ. Toxicol. Chem. 28, 2009 Y. Aliouane et al.

Fig. 3. Sucrose responsiveness of honeybees after oral or contact exposure to insecticide compared to controls. The nature and the dose of the
insecticide are indicated in the figure. The performance is expressed as the percentage of proboscis extension reflex (PER) to antennal stimulation
with sucrose solutions of increasing concentrations recorded in each group. Number of subjects in each group is indicated in parentheses. * p
0.050 (Fishers exact test).

1, 24, and 48 h after learning. At these times the honeybees tioned response to a new odorant for long periods. The same
were also tested for a new odorant. Control honeybees re- effect was observed with fipronil 0.01 ng/bee after topical
sponded significantly more to the CS than to the new odorant exposure (data not shown). Oral treatment of thiamethoxam
(McNemars tests, p 0.002 and 0.031 for tests performed at (0.1 and 1 ng) induced a slight and nonsignificant decrease of
1 and 24 h, respectively). This was also the case for treated performance during learning and in retrieval tests (data not
honeybees at the 1-h test (McNemars test, p 0.031). For shown). Only with topical application did we observe a sig-
later retrieval tests (at 24 and 48 h), the response levels to the nificant decrease of learning performance or retention level.
CS and the new odorant were not different in fipronil-treated For a dose of 0.1 ng/bee, the learning curve of topically treated
animals, indicating a problem of generalization of the condi- animals was not different from the control curve, and the
Chronic pesticide exposure and honeybee behavioral functions Environ. Toxicol. Chem. 28, 2009 119

Fig. 4. Olfactory learning and memory performances of honeybees exposed to insecticides compared to controls. Performances during training
(left) and memory tests (right) are expressed as percentage of proboscis extension reflex (PER) to conditioned stimulus (CS) recorded in each
group. (A) Honeybees treated orally with fipronil and control honeybees were conditioned to one odorant (CS) during training and were tested
for CS and a new olfactory stimulus (NS) at 1, 24, and 48 h. * p 0.050 (MacNemars test for within-group comparison of response to CS
and NS). (B and C) Honeybees exposed topically to thiamethoxam (0.1 and 1 ng) and control honeybees were conditioned to respond by a PER
to a coffee flavor during training and memory tests. Number of subjects in each group is indicated in parentheses. * p 0.050 (Fishers exact
test for between-group comparison).

1-h retention level was equivalent in the two groups, with a more differences between the two groups (Fig. 4B). At the
performance approximately of 50%. The memory test per- dose of 1 ng/bee, topical application of thiamethoxam induced
formed 24 h after learning showed a significant decrease in a significant decrease in learning performance for the third and
performance in the treated group compared to controls fourth trials (Fishers exact tests, p 0.025 and 0.033, re-
(Fishers exact test, p 0.020), but at 48 h, there were no spectively). At the end of the learning session, control bees
120 Environ. Toxicol. Chem. 28, 2009 Y. Aliouane et al.

reached 70% response rate, whereas thiamethoxam-treated the foraging profile of the honeybee. Inside the honeybee col-
bees reached only 50% response rate. Consequently, memory ony, division of foraging labor correlates directly with sucrose
performance of the latter was lower than that of controls at 1, responsiveness. Pollen foragers respond to lower concentra-
24, and 48 h, but the difference was not significant (Fig. 4C). tions of sucrose than do nectar foragers [31]; they perform
Acetamiprid induced no effect on learning and memory (data better on associative learning assays [32], and they display
not shown). At the end of the learning session (fifth condi- increased walking activity compared to nectar foragers [33].
tioning trial), all the animals gained a conditioning level of Therefore, induced modifications of one (or all) of these func-
60%. During the retention tests, the response level to the CS tions by pesticide intoxication may have repercussions on hon-
was lower in treated honeybees than in controls at each point eybee foraging, leading to a perturbation of foragers activity
in time, but the difference was not significant. In control and and, as a consequence, a disruption of nectar and pollen hoard-
treated bees, the response rate to the new odorant was low and ing. Contrary to what was observed after acute treatment [21],
significantly different from the response rate to the conditioned fipronil chronically applied to the thorax affected locomotor
odor. activity. The exposed honeybees stayed in the lower part of
the apparatus, and the time spent in immobility was signifi-
DISCUSSION cantly increased. None of the relevant parameters (time spent
Repeated exposure of honeybees to fipronil at the dose of in each level, duration of immobility, and number of displace-
0.1 ng/bee (LD50/50), defined as sublethal in our previous ments) were different from their counterpart control values for
work [21], induced complete mortality in individuals exposed other treatments regardless of dose and delivery method. We
for one week. This effect on mortality was not observed after previously observed an activating effect of neonicotinoids at
neonicotinoid exposure. Sustained exposure to fipronil (LD50/ low doses on locomotor activity. Acetamiprid in acute topical
500) or to neonicotinoids (1/5 LD50 1/100) induced treatment at doses of 0.1 and 0.5 ng/bee increased locomotor
limited behavioral modifications. Chronic sublethal doses of activity [20], but in the present experiment this effect was not
acetamiprid induced no greater effect than at acute doses [20] observed. Imidacloprid (LD50: 1020 ng/bee) in acute topical
on water responsiveness and induced less effect than at acute treatment at doses of 2.5, 5, 10, and 20 ng/bee induced an
doses on locomotor activity, sucrose responsiveness, and ol- inability of the honeybees to move in the apparatus. However,
factory memory. The experiments with thiamethoxam show an increase of locomotor activity was induced by the lowest
that repeated exposure to a dose that has no behavioral effect dose of 1.25 ng/bee [27]. These results are in agreement with
when applied in acute conditions [20] results in the appearance the fact that at low doses these compounds act as agonists of
of some behavioral deficits. We may conclude from these ob- the cholinergic system and induce excitation, whereas at higher
servations that acetamiprid seems to be the least toxic of the doses they evoke a toxic effect. However, repeated exposure
three molecules for honeybees after repeated exposure to sub- to low doses of acetamiprid did not transform motor excitation
lethal doses. into significant immobility and decreased displacements. Thia-
We reported a mean mortality level of 21% for acetone and methoxam, which, along with imidacloprid, belongs to the
12% for acetonitrile in control animals orally or topically treat- nitroguanidine neonicotinoid group, did not induce this acti-
ed with the solvent. Comparison to the mortality level (10%) vating effect after acute [20] or chronic exposure (present
reported in nontreated animals for a 10-d observation period study).
[30] indicates that acetone enhances mortality of individuals We observed modifications of behavioral response to water
in our experiments and can be considered partly responsible after treatment with pesticides. Fipronil induced an increase
for the mortality of fipronil-treated (0.01 ng/bee) and aceta- in water consumption during the exposure period. Oral ace-
miprid-treated (1 g/bee and 0.1 g/bee) animals. tamiprid treatment (0.1 g/bee) induced the enhancement of
Oral thiamethoxam delivered at the highest dose (one-fifth water responsiveness, and a nonsignificant increase of sucrose
of the LD50 corresponding to 30 g/L) had no significant responsiveness was induced by topical acetamiprid (1 g/bee).
effect on mortality. Similarly, chronic oral exposure of hon- These modifications were previously observed after acute ex-
eybees to either imidacloprid or its plant metabolites induced posure to acetamiprid [20]. Acetamiprid exposure of honey-
no lethal effect at concentrations of 20 and 40 g/L [26]. bees could modify the hive equilibrium, shifting nectar for-
Acetamiprid 1 g/bee (one-tenth of the LD50) induced the agers with a high sucrose threshold to pollen foragers with a
highest observed mortality level (30%), but this level was not low sucrose threshold. Only oral exposure with thiamethoxam
statistically different from that of the control group. Oral fi- at the highest repeated dose (1 ng/bee) induced a partial de-
pronil at a dose of one-fiftieth of the LD50 (0.1 ng /bee, 3 crease of sucrose responsiveness. The same dose in acute treat-
g/L) induced complete mortality after one week of treatment. ment had no significant effect [20]. Fipronil had no obvious
This result is in agreement with the 40% mortality obtained effect on sucrose responsiveness, only a tendency to a decrease
with 2.2 g/L fipronil in the subchronic exposure study con- after oral exposure, an observation reminiscent of the one ob-
ducted by Decourtye et al. [25]. By contrast, acute contact served after acute treatment [21]. We observed that sucrose
with the dose of 0.1 ng/bee induced no additional mortality responsiveness was decreased in emergent control bees com-
compared to controls in the 24 h following the treatment [21], pared to foraging adult bees [20,21], a difference that can be
indicating that fipronil at a sublethal dose becomes lethal on related to age and foraging experience [34]. The difference is
repeated exposure. It is noteworthy that no significant mortality clear for weak sugar concentrations up to 3% and can explain
was obtained in conditions of chronic exposure when the dose the fact that sucrose responsiveness of young bees was less
was decreased to 0.01 ng/bee, which corresponds to a dose of affected by acetamiprid and fipronil than in foragers.
one-five-hundredth of the LD50. Contact thiamethoxam (0.1 ng/bee) induced a decrease of
We report limited effects of the three pesticides on the memory 24 h after learning followed by a recovery at 48 h
motor, sensory, and cognitive functions of the honeybee. The that rules out long-term memory impairment. Learning per-
behavioral functions we have taken into account are linked to formance was decreased in bees treated with thiamethoxam at
Chronic pesticide exposure and honeybee behavioral functions Environ. Toxicol. Chem. 28, 2009 121

the dose of 1 ng/bee, but there were no significant repercus- americana neurons, thiamethoxam failed to activate the neu-
sions on olfactory memory. No effect of fipronil was observed ronal nAChRs, a result already found by Nauen et al. [40] on
on learning performance and olfactory memory, a result al- Heliotis virescens neurons. Both studies report as a certainty
ready found in honeybees chronically consuming a higher dose that thiamethoxam does not interact with the nicotinic recep-
(0.075 ng/bee) [25]. However, fipronil induced an impairment tors. Nevertheless, the biological effect of thiamethoxam is in
of odorant response specificity; this is shown by the fact that, most cases comparable to other neonicotinoid insecticides
after long postlearning periods, treated honeybees responded [41]. The toxic action of thiamethoxam is then related to its
indiscriminately to the learned odorant and the new one, a rapid conversion to clothianidin, a metabolite compound that
phenomenon known as odor generalization. Data in the lit- binds to insect nAChRs with high affinity and is considered
erature have shown that each of the odorants we used as CS a full agonist of these receptors [40]. However, we failed to
(1-hexanol or 1-nonanol) induces a low response level to the find any relevant biological effect of thiamethoxam on the
other one [35]; thus, in our experiments, the two odorants honeybee after acute sublethal treatment [20], and we observed
should be perceived as dissimilar by bees. The lack of odorant- only a limited impairment of sucrose sensitivity and olfactory
specific memory suggests that fipronil favors increased odor learning after chronic treatment (present study). No explana-
generalization. This effect was not observed after acetamiprid tions can be put forward for these results, as bioassays in
treatment, as honeybees still discriminated between known and honeybees have shown comparable toxic effects of imidaclo-
unknown odorants. These results suggest that cholinergic and prid, thiamethoxam, and clothianidin, with LD50 values in-
gabaergic pathways do not support the same role in olfactory duced by contact treatment ranging from 18 to 30 ng/bee [41].
processes. Anatomical and physiological evidence indicates In our evaluation of the sublethal toxicity, we cannot es-
that the olfactory message is conveyed from the antennae to timate the no-observed-effect-concentration for fipronil. At a
higher brain centers through activation of cholinergic neurons. concentration of 0.3 g/L (corresponding to the oral dose of
Modifying the excitatory level of the cholinergic pathways 0.01 ng/bee), at least one behavioral parameter was statistically
with acetamiprid treatment seems to have no major effects on different from the control value (Table 1). By contrast, ace-
olfactory perception. Inhibitory local circuits within the an- tamiprid and thiamethoxam have no or limited effect when
tennal lobe are necessary for building up odorant-specific sig- applied chronically at sublethal doses. Therefore, the no-ob-
nals. It has been suggested that GABA and GluCl receptors served-effect-concentration for the behavioral assays are 3
play this role [4,16]. Fipronil blocking of these receptors can mg/L and 3 g/L for acetamiprid and thiamethoxam, respec-
be responsible for the olfactory generalization. tively. Hence, fipronil appears to be the most toxic of the three
The effects of the low doses of fipronil after subchronic molecules tested. The evaluation of the pesticide risk to hon-
exposure can be linked to sensitization of receptors following eybees (the hazard ratio) will be a combination of the intrinsic
prolonged stimulation. In cockroach thoracic ganglion neu- toxicity of the molecules as we have attempted to define it
rons, the repetitive activation of GluCl receptors by the major and the exposure of bees to the compounds in natural con-
metabolite of fipronil, fipronil sulfone, decreases the 50% in- ditions, which will depend on their status (larvae, workers,
hibitory concentration values, indicating that the receptors are foragers) in the hive.
more sensitive to the inhibitor [36]. Receptor sensitization can
be responsible for the death of animals that have undergone AcknowledgementThis research benefited from financial support of
repeated exposure at the dose of 0.1 ng/bee and for the be- the European Community in the framework of the Apiculture Program
havioral effects we reported at the dose of 0.01 ng/bee (in- 2007, Agreement 07-09, between Viniflhor, the Centre National de la
Recherche Scientifique, and the University Paul Sabatier. The authors
creased water consumption, decreased mobility, odor gener- are grateful to J.C. Sandoz and E. Lebourg and to V. Raymond-
alization). Although all neonicotinoid insecticides act selec- Delpech and P. Haas-Hammel.
tively on insect nAChRs, their agonist actions vary from partial
to full efficacy. A recent study conducted on cockroaches has REFERENCES
shown that the agonist efficacy, defined as the maximum in- 1. Tomizawa M, Casida JE. 2005. Neonicotinoid insecticide toxi-
cology: Mechanisms of selective action. Annu Rev Pharmacol
ward current induced by neonicotinoid insecticides on isolated Toxicol 45:247268.
neurons, is positively correlated with insecticidal activity [37]. 2. Buckingham SD, Biggin PC, Sattelle BM, Brown LA, Sattelle
Injected at toxic doses, low-efficacy compounds like imida- DB. 2005. Insect GABA receptors: Splicing, editing, and targeting
cloprid cause excitatory symptoms, whereas high-efficacy by antiparasitic and insecticides. Mol Pharmacol 68:942951.
3. Matsuda K, Buckingham SD, Kleier D, Rauhh JJ, Grauso M,
compounds (acetamiprid) cause depressive/paralytic symp- Sattelle DB. 2001. Neonicotinoid insecticides acting on insect
toms. Using acute intoxication with sublethal doses, we failed nicotinic acetylcholine receptors. Trends Pharmacol Sci 22:573
to find such a distinction between imidacloprid [27] and ace- 580.
tamiprid [20], the two compounds showing an activating effect 4. Barbara GS, Zube C, Rybak J, Gauthier M, Grunewald B. 2005.
on honeybees biological functions. In the vertebrate, a 3-d Acetylcholine, GABA and glutamate induce ionic currents in cul-
tured antennal lobe neurons of the honeybee, Apis mellifera. J
exposure to imidacloprid, thiacloprid, or nicotine up-regulates Comp Physiol A 191:823836.
the neuronal 42 nAChRs that are insensitive to -BGT [38]. 5. Barbara GS, Grunewald B, Paute S, Gauthier M, Raymond-Del-
Patch-clamp recordings performed on cultured rat cortical neu- pech V. 2007. Cholinergic currents of cultured antennal lobe neu-
rons show 42 nAChR desensitization induced by 30-min rones from adult honeybees brains. Invertebr Neurosci 8:1929.
6. Deglise P, Grunewald B, Gauthier M. 2002. The insecticide im-
exposure to nicotine [39]. It was originally proposed that nic- idacloprid is a partial agonist of the nicotinic receptor of honeybee
otine-induced up-regulation is related to desensitization of the Kenyon cells. Neurosci Lett 321:1316.
receptor, but the relations between the two phenomena are not 7. Nauen R, Ebbinghaus-Kintscher U, Schmuck R. 2001. Toxicity
clearly defined [38]. Such up-regulation of nAChRs has not and nicotinic acetylcholine receptor interaction of imidacloprid
and its metabolites in Apis mellifera (Hymenoptera: Apidae). Pest
been described in insects and cannot be counted for the be- Manag Sci 57:577586.
havioral modifications observed in honeybees after neonico- 8. Gauthier M, Dacher M, Thany SH, Niggebrugge C, Deglise P,
tinoid exposure. In the study of Tan et al. [37] on Periplaneta Kljucevic P, Armengaud C, Grunewald B. 2006. Involvement of
122 Environ. Toxicol. Chem. 28, 2009 Y. Aliouane et al.

alpha-bungarotoxin-sensitive nicotinic receptors in long-term affected by imidacloprid according to the season. Pest Manag
memory formation in the honeybee (Apis mellifera). Neurobiol Sci 59:269278.
Learn Mem 86:164174. 25. Decourtye A, Devillers J, Genecque E, Le Menach K, Budzinski
9. Cano Lozano V, Bonnard E, Gauthier M, Richard D. 1996. Mec- H, Cluzeau S, Pham-Delegue MH. 2005. Comparative sublethal
amylamine-induced impairment of acquisition and retrieval of toxicity of nine pesticides on olfactory learning performances of
olfactory conditioning in the honeybee. Behav Brain Res 81:215 the honeybee Apis mellifera. Arch Environ Contam Toxicol 48:
222. 242250.
10. Cano Lozano V, Armengaud C, Gauthier M. 2001. Memory im- 26. European and Mediterranean Plant Protection Organism. 1992.
pairment induced by cholinergic antagonists injected into the Guideline on test methods for evaluating the side-effects of plant
mushroom bodies of the honeybee. J Comp Physiol A 187:249 protection products on honeybees. EPPO Bull 22:203215.
254. 27. Lambin M, Armengaud C, Raymond S, Gauthier M. 2001. Imi-
dacloprid-induced facilitation of the proboscis extension reflex
11. Dacher M, Lagarrigue A, Gauthier M. 2005. Antennal tactile
habituation in the honeybee. Arch Insect Biochem Physiol 4:129
learning in the honeybee: Effect of nicotinic antagonists on mem- 134.
ory dynamics. Neuroscience 130:3750. 28. Levene H. 1960. Robust tests for equality of variances. In Olkin
12. Thany SH, Gauthier M. 2005. Nicotine injected into the antennal I, Ghurye SG, Hoeffding W, Madow WG, Mann HB, eds, Con-
lobes induces a rapid modulation of sucrose threshold and im- tributions to Probability and Statistics: Essays in Honor of Har-
proves short-term memory in the honeybee Apis mellifera. Brain old Hotelling. Stanford University Press, Menlo Park, CA, USA,
Res 1039:216219. pp 278292.
13. Janssen D, Derst C, Buckinx R, Van den Eynden J, Rigo JM, 29. Eliasziv M, Donner A. 1991. Application of the McNemar test
VanKerkhove E. 2007. Dorsal unpaired median neurons of Lo- to non-independent matched pair-data. Stat Med 10:19811991.
custa migratoria express ivermectin- and fipronil-sensitive glu- 30. Schmuck R. 2004. Effects of a chronic dietary exposure of the
tamate-gated chloride channels. J Neurophysiol 97:26422650. honeybee Apis mellifera (Hymenoptera: Apidae) to imidacloprid.
14. Narahashi T, Zhao X, Ikeda T, Nagat K, Yeh JZ. 2007. Differential Arch Environ Contam Toxicol 47:471478.
actions of insecticides on target sites: Basis for selective toxicity. 31. Page RE, Erber J, Fondrk MK. 1998. The effects of genotype on
Hum Exp Toxicol 26:361366. response thresholds to sucrose and foraging behavior of honey-
15. Bicker G, Schafer S, Kingan T. 1985. Mushroom body feedback bees (Apis mellifera). J Comp Physiol A 182:489500.
interneurons in the honeybee show GABA-like immunoreactivity. 32. Scheiner R, Page RE, Erber J. 2001. The effects of genotype,
Brain Res 360:394397. foraging role, and sucrose responsiveness on the tactile learning
16. Stopfer M, Bhagavan S, Smith B, Laurent G. 1997. Impaired performance of honeybees (Apis mellifera L.). Neurobiol Learn
odour discrimination on desynchronization of odour-encoding Mem 76:138150.
neural assemblies. Nature 990:7074. 33. Humphries MA, Fondrk MK, Page RE. 2005. Locomotion and
17. Sachse S, Galizia CG. 2002. Role of inhibition for temporal and the pollen hoarding behavioral syndrome of the honeybee (Apis
spatial odor representation in olfactory output neurons: A calcium mellifera L.). J Comp Physiol A 191:669674.
34. Pankiw T, Page RE. 1999. The effect of genotype, age, sex, and
imaging study. J Neurophysiol 87:11061117.
caste on response thresholds to sucrose and foraging behavior of
18. Collet C, Belzunces L. 2007. Excitable properties of adult skeletal honey bees (Apis mellifera L.). J Comp Physiol A 185:207213.
muscle fibers from the honeybee Apis mellifera. J Exp Biol 210: 35. Guerrieri F, Schubert M, Sandoz JC, Giurfa M. 2005. Perceptual
454464. and neural olfactory similarity in honeybees. PLoS Biology 3:
19. Marrus SB, Portman SL, Allen MJ, Moffat KG, DiAntonnio A. e60.
2004. Differential localization of glutamate receptor subunits at 36. Zhao X, Yeh JZ, Salgado VL, Narahashi T. 2005. Sulfone me-
the Drosophila neuromuscular junction. J Neurosci 24:1406 tabolite of fipronil blocks -amino acid- and glutamate-activated
1415. chloride channels in mammalian and insect neurons. J Pharmacol
20. El Hassani AK, Dacher M, Gary V, Lambin M, Gauthier M, Exp Ther 314:363373.
Armengaud C. 2008. Effects of sublethal doses of acetamiprid 37. Tan J, Galligan JJ, Hollingworth RM. 2007. Agonists actions of
and thiamethoxam on the behavior of the honeybee (Apis mel- neonicotinoids on nicotinic acetylcholine receptors expressed by
lifera). Arch Environ Contam Toxicol 54:653661. cockroach neurons. Neurotoxicology 28:829842.
21. El Hassani AK, Dacher M, Gauthier M, Armengaud C. 2005. 38. Tomiza M, Casida JE. 2000. Imidacloprid, Thiacloprid, and their
Effects of sublethal doses of fipronil on the behavior of the hon- imine derivatives up-regulate the 42 nicotinic receptor in M10
eybee (Apis mellifera). Pharmacol Biochem Behav 82:3039. cells. Toxicol Applied Pharmacol 169:114120.
22. Halm MP, Rortais A, Arnold G, Tasei JN, Rault S. 2006. New 39. Marszalec W, Yeh JZ, Narahashi T. 2005. Desensitization of nic-
otine acetylcholine receptors: Modulation by kinase activation
risk assessment approach for systemic insecticides: The case of
and phosphatase inhibition. Eur J Pharmacol 514:8390.
the honey bees and imidacloprid (Gaucho). Environ Sci Technol 40. Nauen R, Ebbinghaus-Kintscher U, Salgado V, Kaussman M.
40:24482454. 2003. Thiamethoxam is a neonicotinoid precursor converted to
23. Aajoud A, Raveton M, Aouadi H, Tissut M, Ravanel P. 2006. clothianidin in insects and plants. Pest Biochem Physiol 76:55
Uptake and xylem transport of fipronil in sun flowers. J Agric 69.
Food Chem 54:50555060. 41. Iwasa T, Motoyama N, Ambrose JT, Roe MR. 2004. Mechanism
24. Decourtye A, Lacassie E, Pham-Delegue MH. 2003. Learning for the differential toxicity of neonicotinoid insecticides in the
performances of honeybees (Apis mellifera L.) are differentially honeybee, Apis mellifera. Crop Prot 23:371378.