Cocaine affects foraging behaviour and biogenic amine modulated behavioural reflexes in honey bees

Subjects

All experiments were performed at Macquarie University, Sydney, Australia. Bees used were of the standard commercially available strains in Australia, and reared according to standard bee keeping practices. For foraging experiments, a colony containing approximately 5,000 bees was housed in a 400 m² flight enclosure.

Pharmacological treatments

For topical application, 3 µg freebase cocaine dissolved in 1 µL dimethylformamide (DMF) was applied to the dorsal thorax of bees using a glass microcapillary. This was the same non-toxic dose that increased dance rate in the study by Barron et al. (2009). DMF is a solvent that can penetrate bees’ cuticle and allows cocaine to pass into the haemocoel (Barron et al., 2007). This method has previously been used for administering cocaine to honey bees (Barron et al., 2009; Søvik, Cornish & Barron, 2013). As a control, bees were treated with DMF alone in the same manner.

For volatilised treatments, freebase cocaine was dissolved in ethanol, and carefully pipetted onto a nichrome wire filament connected to a power source (McClung & Hirsh, 1998). Ethanol was evaporated from the filament at room temperature. To treat bees, a single bee was kept in a 50 cm³ airtight container encapsulating the filament. The filament was heated for 10 s and bees were kept in the container, exposed to volatilised cocaine, for one minute. Unlike vertebrates, insects have an open gas exchange system that transports oxygen directly to tissues where it is needed in the gaseous phase, bypassing the haemolymph. Air is taken in through spiracles in the thorax and abdomen, passed through trachea, before gas exchange takes place via tracheoles (Chapman, 2013). This system allows volatilised cocaine to be delivered directly to cells throughout the bee nervous system. As a control, pure ethanol was applied to the filament, allowed to evaporate, and the clean filament was used for treatments using the method outlined above (for details see Søvik, Cornish & Barron, 2013). All reagents were supplied by Sigma-Aldrich (St. Louis, MO, USA).

Previously we have shown that the pharmacokinetics of these two methods are markedly different (Barron et al., 2009; Søvik, Cornish & Barron, 2013), but without measuring the rate cocaine enters and is cleared from the brain following administration it is not possibly to conclusively state how different.

Effects of cocaine on honey bee foraging preferences

To examine if bees developed a preference for a feeder associated with cocaine treatment, 60 individually paint-marked bees were trained to two ad libitum 1.5 M sucrose feeders placed at the closed ends of two 2 m long tunnels that intersected at a 45° angle (Fig. 1). The walls and floor of the tunnels were solid opaque plastic; the ceiling was covered with mesh. From the perspective of approaching from the hive, the entrance to the left tunnel and the walls surrounding the feeder in the left tunnel were marked with horizontal green and white stripes, the entrance to the right tunnel and the walls surrounding the feeder in the right tunnel were marked with vertical blue and white stripes. The tunnels created two visually distinct and spatially separated environments in which feeders were located. The bee’s choice of feeder could easily be assayed visually by observing which tunnel they entered and which feeder they alighted on. The colours blue and green were chosen because bees have distinct photoreceptors for these two colours (Chittka & Menzel, 1992), further, the 90° difference in orientation of the striped patterns is easily differentiated by honey bees (Frisch, 1971) and was added to make the tunnels even more distinctive. This design was chosen in order to increase the distinctiveness of the two tunnels (i.e., in order to make it as easy as possible for the bees to tell the two tunnels apart). This allowed detecting changes in preference rather than discriminatory abilities.

Bees were trained and tested in a five-day protocol. On day one of a trial, bees were trained to use both tunnels by alternating the availability of tunnels every 15 min while progressively stepping a 1.5 M sucrose feeder deeper into each tunnel over a 4 h period. Bees were released from the tunnel after feeding by lifting the mesh.

On day two, bees were further trained to use the tunnels by alternating the availability of the tunnels every 30 min for 3 h, then simultaneously opening both tunnels to provide bees with a free choice of feeders for 1 h. During this time the number of visits of each bee to each feeder was recorded. These were converted to a preference index as follows: ${\displaystyle \text{Preference index}=\frac{\left(\text{number of visits to green tunnel}-\text{number of visits to blue tunnel}\right)}{\text{total number of visits}}}$ This preference index is similar to that used for aversive conditioning by Vergoz et al. (2007), but because individual bees varied in the total number of visits made, we divided difference in visits made by the total number of visits to allow for comparison between bees. At this stage the median preference index was not significantly different from zero (Wilcoxon signed rank test. W = 456, p = 0.166, n = 75) indicating there was no preference toward either tunnel.

On days three and four of a trial, bees had access to the green tunnel only for 2 h a day, which offered a 1 M sucrose feeder. Bees were randomly assigned to cocaine or control treatment groups. We used the slower topical treatment method so that cocaine would persist in bees’ systems for the majority of their time interacting with the tunnels (previous work suggested topical cocaine treatment influenced bee behaviour for approximately 1.5 h following treatment (Barron et al., 2009), whereas the effects of volatilised treatment appeared to be shorter in duration). Bees were treated with either 1 µl DMF containing 3 µg cocaine or 1 µl DMF alone on their first visit to the feeder each day.

With this assay design bees had more opportunities to visit the green tunnel than the blue tunnel, and therefore had more reinforcing experiences in the green tunnel than the blue tunnel. Thus, we expected all bees to develop a weak preference for the green tunnel. However, the aim of this experiment was to test whether cocaine treatment affected the magnitude of the preference for the green feeder.

On day five of a trial, all bees were given simultaneous access to both tunnels for 1 h to test the preference of bees for the different tunnels. The number of visits by each bee to each tunnel was recorded. During the test both tunnels contained empty feeders, and once bees had reached the end of a tunnel they were released. The number of visits to each tunnel by each bee was converted to a preference index as described. Five replicate trials of this experiment were performed. For analysis data from all trials were pooled.

Effect of volatilised cocaine on foraging rate

Previously, Barron et al. (2009) did not find a difference in foraging rate between bees treated with cocaine and controls, using the topical treatment method. As topical treatment is rather slow (Barron et al., 2007) and rate of cocaine delivery to the central nervous system affects the magnitude of behavioural responses (Samaha & Robinson, 2005), we decided to test if the number of foraging trips was affected following the more rapid volatilised treatment method (Søvik, Cornish & Barron, 2013). In a flight cage bees were trained to visit an ad libitum sucrose feeder where they were given individually distinctive paint marks. Bees that returned five times after being marked were caught and treated with 5 µg volatilised freebase cocaine or control. We chose 5 µg as this was the highest volatilised dose previously tested that did cause deleterious motor effects (Søvik, 2013). Bees were assigned to treatment groups randomly. The number of visits treated bees made to the feeder in the 40 min following treatment were recorded. Sucrose concentration has previously been shown to affect foraging rate in bees (Seeley, 1995), studied responses of bees to both low (0.5 M) and high (2.0 M) sucrose solutions.

Effects of volatilised cocaine on sucrose responsiveness

To test if volatilised cocaine affected sucrose responsiveness we used cage-reared bees of known age and social history. Upon emergence, bees were placed in mesh cages (20 ×16 ×3 cm) with ad libitum access to honey. The cages contained eighty bees each and were kept at 34 °C for 6 days. When the bees were 7 days old, they were fastened individually in an 8 mm tube in a way that prevented the bees from escaping but allowed the proboscis and antenna to move freely (Bitterman et al., 1983). This method is most commonly used for proboscis extension learning experiments (Felsenberg et al., 2011) but has also been used to measure bees’ responsiveness to sucrose (Scheiner, Page Jr & Erber, 2004). Once harnessed, bees were treated with 0 or 10 µg volatilised cocaine and tested for sucrose responsiveness. The 10 µg was chosen based on an initial pilot experiment suggesting that this dose was sufficient to elicit increased responsiveness to sucrose (E Søvik, 2012, unpublished data). We repeated this experiment with 0, 5, 10, 20, or 50 µg volatilised cocaine to examine if the effect seen with 10 µg was dependent on the cocaine dose used.

The sucrose responsiveness test consisted of touching a drop of 10% sucrose solution to the antennae of bees 3 min after drug exposure, and recording whether or not the proboscis was extended. After the test, bees were tested for their response to water and honey. Bees responding to water, or failing to respond to honey were excluded from the analysis.

Effects of volatilised cocaine on responsiveness to electric shocks

To examine effects of cocaine on responsiveness to electric shock, bees were fastened between two conducting brass plates with a piece of electrical tape (for details see Vergoz et al., 2007). After treatment with 0, 5, 10, 20 or 50 µg volatilised cocaine, brass plates were connected to an electrical supply, and bees were shocked with gradually increasing voltage (0.5 V every 5 min) from 0 to 10 V. The first voltage at which a bee extended its stinger (a reflexive response) was recorded for each bee. Testing occurred in front of an extraction fan so no alarm pheromone would linger in the testing room and affect bees yet to be tested (Vergoz et al., 2007). Comparisons between groups were based on EV₅₀ (half maximal effective voltage): the point at which half of all bees in the treatment group extended their stingers.

Effects of cocaine on honey bee foraging preferences

Repeatedly treating bees with 3 µg cocaine in DMF at a sucrose feeder enhanced bees’ preference for that feeder in a choice assay when compared to bees treated with DMF as a control (Mann–Witney test: U = 2, 185, p = 0.0038; Effect size: r = −0.25). Treating bees with cocaine at a feeder while they were foraging resulted in a greater preference for that feeder in a free-choice test when compared to bees treated with DMF (vehicle control) while foraging at the feeder (Fig. 2A).

Effect of volatilised cocaine on foraging rate

Bees treated with 5 µg volatilised cocaine once at a 0.5 M feeder made significantly more return visits to the feeder in the 40 min following treatment, than controls (t₇₀ = 5.0710, p = 0.00003; Effect size: d = 0.9905; Fig. 2B). Bees treated with cocaine at a 2 M feeder showed no increase in visitations after cocaine treatment (t₇₀ = −0.2087, p = 0.8353; Effect size: d = 0.0399; Fig. 2B). This demonstrated that bees altered the rate at which they returned to a low quality feeder following volatilised cocaine treatment, but not to a high quality feeder (Fig. 2B).

Effects of volatilised cocaine on sucrose responsiveness

Treatment with 10 µg of volatilised cocaine increased bees responsiveness to sucrose (χ² = 6.0268, df = 1, p = 0.0141; Effect size: d = 0.6331; Fig. 3A). The effect was dependent on the cocaine dose. Bees treated with 5 and 10 µg of cocaine were significantly more responsive to sucrose than controls (χ² = 14.089, df = 4, p = 0.0070; Fig. 3B), while bees treated with 20 or 50 µg of cocaine did not differ from controls. The control treatment differed quite markedly between two experiments; however, this is likely because the two experiments were performed at different times of the year. Sucrose responsiveness varies with season and environmental conditions. The important aspect is the difference between the cocaine treated bees and the control treated bees in a given experiment.

Effects of volatilised cocaine on responsiveness to electric shocks

Cocaine affected bees’ responsiveness to shock in a dose dependent manner (Fig. 3C). We used the EV₅₀ for statistical comparisons. All bees treated with cocaine were significantly more sensitive to electric shock than control treated bees (F_4,40 = 5.4, p = 0.0015; Fig. 3C). There were no differences between the cocaine treatment groups with the exception of bees treated with 50 µg cocaine. The bees treated with 50 µg were significantly more sensitive than all other cocaine treated groups. The EV₅₀ of cocaine treated bees (50 µM EV₅₀ = 2.1; 20 µM EV₅₀ = 3.5; 10 µM, EV₅₀ = 2.6; 5 µM EV₅₀ = 3.1) was lower than in control treated bees (EV₅₀ = 5.3).