In vitro and in vivo susceptibility of the honey bee bacterial pathogen Paenibacillus larvae subsp larvae to the antibiotic tylosin.

vitro and in vivo susceptibility of the honey bee bacterial pathogen


Introduction
American Foulbrood (AFB) disease caused by the spore-forming bacterium Paenibacillus larvae subsp.larvae (P.l.larvae) (Heyndrickx et al., 1996) (formerly Bacillus larvae) is a cosmopolitan disease of bacterial origin affecting the larval and pupal stages of honeybees (Apis mellifera L.) (White, 1920;Shimanuki, 1990).Diseased individuals turn brown, then black, and the resultant mass becoming a hard scale of material deposited on the side of the cell.AFB is one of the few bee diseases capable of killing a colony, and has unique problems for prevention and control because the spores can remain viable for long periods of time and survive environmental adversities (Matheson and Reid, 1992).The disease is highly contagious and if undetected can kill a colony and spread to others within the same apiary or to another apiary nearby by robbing and exchanging of brood combs, the main sources of contamination.
In areas where disease incidence is high, antibiotic treatments appear as an alternative to the burning of diseased bee colonies.Currently, the only antibiotic approved for prevention and control of AFB is oxytetracycline, however, there is evidence of oxytetracycline-resistant isolates of P.l.larvae in certain areas of the USA, Canada and Argentina (Alippi, 2000;Colter, 2000;Evans, 2003;Miyagi et al., 2000).
The purposes of the present work were to evaluate the susceptibility of 67 strains of P.l.larvae from diverse geographical areas to tylosin by determining their minimal inhibitory concentrations (MIC), since the information about their efficacy in vitro is quite limited (Alippi, 1994;Kochanski et al., 2001;Okayama et al., 1996) and to determine the response of colonies with clinical signs of AFB to tylosin at different doses and forms of application in order to determine the most effective dose for disease control and lack of recurrence of the disease.

Bacterial strains
The 67 P.l.larvae strains from diverse geographical origins used in this study are listed in Table 1.For the isolation of P.l.larvae strains from larvae affected by AFB and from honey samples, previously described techniques were employed (Alippi, 1995;Alippi and Aguilar, 1998;Alippi et al., 2004).In addition, reference strains from culture collections were included, and also strains of Pseudomonas aeruginosa, Escherichia coli and Staphylococcus aureus used as reference standards for quality control ranges of MICs were provided by Instituto Malbra ´n Collection, Instituto Malbra ´n, Buenos Aires, Argentina (Table 1).

Determination of minimal inhibitory concentrations of tylosin
As there is no NCCLS recommendation for the determination of MICs of P.l.larvae, a method developed for this species was used based on the NCCLS standard for Bacteria isolated from animals (NCCLS, 1999).Minimal inhibitory concentrations of tylosin tartrate (Sigma 1 ) were determined by the agar dilution method using MYPGP (Dingman and Stahly, 1983) as basal medium because P.l.larvae is not able to growth on Muller-Hinton agar.Tylosin concentrations tested were made using a stock solution of 5000 mg/ml in distilled water and sterilized using a syringe-driven filter unit with a 0.22 mm pore size.Appropriate dilutions were made in sterile distilled water and stored at À20 8C until used.The medium was maintained at 45 8C until the antibiotic solutions were incorporated and 25 ml of culture medium were poured onto each Petri plate of 90 mm in diameter.Increasing concentrations of tylosin tested were: 0.0039, 0.0078, 0.015, 0.03125, 0.0625, 0.125, 0.5, 1, 2, 4, 8, 16 and 32 mg/ml of medium.For the controls, MYPGP agar without antibiotic was employed.
Vegetative cells of each P.l.larvae strain grown on MYPGP agar for 48 h of incubation at 37 8C were suspended in sterile distilled water and adjusted to approximately 2.87 Â 10 8 cells/ml (A 620 nm = 0.388, equivalent to a Mc Farland standard of 1).The different bacterial strains were layered by using an automatic micropipette to place drops of 10 ml each over the surface of the solidified culture medium containing each antibiotic concentration.Ten replications were tested for each strain.The inoculated plates were examined for growth after 48 h of incubation at 37 8C.The lowest concentration of antibiotic preventing growth was defined as the MIC (complete inhibition of bacterial growth on the test plates, disregarding a single colony of faint haze caused by inoculum).
In the case of P. aeruginosa, E. coli and S. aureus used as reference standards, the technique outlined by the NCCLS for each species (NCCLS, 1999) was followed, with the only difference that MYPGP agar was used as basal medium.

Field experiments
The efficiency of tylosin tartrate for the control of AFB on diseased bee colonies was evaluated on three field experiments conducted at the Faculty of Agricultural Science Experimental Field, UNLP, La Plata 35 8S latitude, 57 8W longitude.In all experiments, colonies of honeybees derived from Apis mellifera ligustica L. were used.Queens were marked and their wings were clipped to avoid swarming.Colonies were distributed in a completely randomized design.The experimental procedures for inoculation were conducted as described by Alippi et al. (1999).
The amount of healthy brood, adult bees and larvae with AFB-clinical symptoms were measured in the same way in all of the experiments.Healthy brood cells (larvae and pupae) were quantified according to the following scale: grade 0, absence of brood cells; grade 0.5, between 1 and 499 brood cells; grade 1, between 500 and 3499 brood cells; grade 2, between 4000 and 7999 brood cells; grade 3, between 8000 and 11,499 brood cells; grade 4, between 12,000 and 15,999 brood cells and grade 5, between 16,000 and more than 16,000.The number of adult bees were quantified in a similar manner, with a minimum of grade 0 and a maximum grade 5, being grade 0, The first experiment started in March and ended in October 2000.The 10 colonies used were from packs graded 4 for bees, 3 for brood and 0 for AFB-infection.Two treatments with five repetitions each were applied: five colonies were treated with 1.5 g of tylosin tartrate (Tylan Esanco 1 ) and five colonies were used as AFB-inoculated controls.The total dose (1.5 g per colony) was divided in six parts preparing 70 g candies (55 g powdered sugar + 15 g 55% sugar syrup + 0.25 g active ingredient (a.i. of tylosin tartrate); cherry jelly was used as an attractant to ensure consumption.For controls a mixture of powdered sugar, sugar syrup and cherry jelly was used.The first candy was supplied as preventive 15 days before the inoculations of colonies with AFBcontaminated combs containing 20 AE 5 scales (Alippi et al., 1999).Inoculated colonies were evaluated once a month starting 30 days after inoculation.Each candy was replaced with a new one every 2 weeks.At the end of the experiment disease recurrence was also evaluated.
The second experiment was carried out between April and August 2001.The 10 colonies used were from packs graded 4 for bees, 3 for brood and 0 for AFB disease signs.Inoculations and evaluation of colonies were managed like those in the first experiment, but with modifications in doses and form of administration of the antibiotic.Two treatments with five repetitions each were applied: five colonies were treated with 0.750 g tylosin tartrate in a 70 g candy 15 days before inoculation and five colonies used as AFB-inoculated controls were supplied with a 70 g candy.Four monthly inspections were made starting 30 days after infection.
The third experiment took place between May and October 2002.Twelve colonies were used and were graded 4 for bees, 3 for brood and 0 for AFB-infection.The differences with the former experiment were: the treatments were applied by means of monthly supplied syrup.Six colonies were treated with 0.750 g tylosin tartrate and six colonies were used as AFB-inoculated controls (50% sacarose syrup).The total dose of tylosin for the first group was provided at the first application with 750 cc of syrup 30 days after infection.Four monthly applications of 50% syrup without antibiotic were made afterwards.Controls were supplied in five monthly applications consisting of 750 cc 50% sacarose syrup.Taste was improved by adding 0.5 ml raspberry essence to each jar.

Statistical analysis
Wilcoxon two-sample test was used for analyzed significant differences ( p < 0.05) between colonies treated with tylosin and untreated controls in each experiment as described by Spivak and Reuter (2001).Mantel Haenszel Chi-square test and also Fisher exact test (two tails), specially suited for small samples, were run in order to estimate the infection among treatments ( p < 0.05) in all three experiments.In the same way, the Fisher exact test was used to estimate significant differences in the number of brood cells and that of adult bees between tylosin treatments and AFB-inoculated controls.Results from all three experiments were combined for that purpose (Spivak and Reuter, 2001).

Minimum inhibitory concentrations of tylosin
MYPGP probed adequate for growth and interpretation of MIC values of P.l.larvae and also results for P. aeruginosa, E. coli and S. aureus are within the acceptable limits for quality control strains (NCCLS, 1999).All P.l.larvae strains were highly susceptible to tylosin with MIC values ranging from 0.0078 to 0.5 mg/ml depending upon the tested strain (Fig. 1).Results are summarized in Table 1.These values indicate that very low concentrations of tylosin are required to inhibit the growth of P.l.larvae.The appearance of oxytetracycline-resistant P.l.larvae strains in many countries has given a high priority to the search of alternatives being tylosin highly effective for the control of the Foulbrood pathogen in vitro when testing 67 strains from different geographical origins.These results are in accordance with previous studies on strains from Japan, where MIC values were between 0.025 and 0.1 mg/ml (Okayama et al., 1996).
The National Committee for Clinical Laboratory Standards do not provide a standard method for determining MIC values for P.l.larvae (NCCLS, 1999), nor have breakpoints for antibiotic resistance been established.The values of MIC obtained using MYPGP agar for reference strain of P. aeruginosa ATCC 27853 MIC > 32, S. aureus ATCC 29213 MIC = 2 and E. coli ATCC 29922 MIC > 32 are within the acceptable limits for quality control strains, being >32, between 0.5 and 4 mg/ml and >32, respectively.Isolates could be considered as susceptible when their MICs were 8, intermediate for MICs of 16 and resistant for MICs !32, as suggested for tylmicosin (NCCLS, 1999) or resistant when their MICs were 4 as considered for many antibiotics (Gales et al., 2001).Based upon the results of the present study, there are not any P.l.larvae resistant or intermediate strains for tylosin.Further studies that include resistant and/or intermediate P.l.larvae strains are needed to corroborate this hypothesis, but until present there is not tylosin resistance reported for this species.

Field experiments
In the first field experiment, three out of five AFBinoculated control colonies showed symptoms 60 days after inoculation, and 150 days after inoculation, all colonies showed clinical signs ranging from level 1 to level 6.On the other hand, the colonies treated with tylosin showed no symptoms at all.The results from the first field experiment showed that the disease was controlled in all of the treated colonies and no recurrence of the disease was observed after the end of the experiment by using six candies of 250 mg tylosin (Table 2 and Fig. 2).
During the second field experiment two colonies treated with tylosin showed symptoms 90 days after inoculation (level 1).One colony recovered 120 days after inoculation (level 0) and the other remained with less than three larvae affected (level 1) (Table 3).One of the AFB-inoculated controls showed symptoms after 30 days.After 90 days, all AFB-inoculated controls showed clinical signs of the disease and the infection increased throughout the end of the experiment (ranging from level 1 to level 6) (Table 3).The results from the second field experiment showed that the disease was controlled in 80% of the treated colonies and no recurrence of the disease was observed after the end of the experiment by using one candy of 750 mg of tylosin (Table 3 and Fig. 2).
In the third field experiment, after 90 and 120 days of the infection, in the tylosin treatments, no AFB diseased larvae were found, and after 150 days one colony exhibited two larvae with clinical symptoms of AFB (level 1).The AFB-inoculated controls showed clinical signs in three of them 60 days after infection (level 1).Two of them died after 90 and 120 days, respectively (level 6) (Table 4 and Fig. 2).Overall the results were consistent over 3 years of field experiments, in relation to infection level variable, both the Wilcoxon and the Fisher tests showed similar results.The tylosin treated colonies showed significant differences respect compared to the AFB-inoculated controls ( p < 0.05).Likewise, the analysis of the infection variable in all three experiments together with the Fisher and Mantel Haenszel Chi-square tests gave significant differences for tylosin compared to those of the AFB-inoculated controls (Fisher p < 0.05; Mantel Haenszel Chisquare = 17,364, p < 0.05) (Fig. 2).The result of the analysis of the brood variable in all three experiments under the Fisher test showed significant differences for tylosin compared to the AFBinoculated controls ( p < 0.05), which confirms the high effectiveness of tylosin.The analysis for the adult bee variable in all three experiments under the Fisher test showed no significant differences between tylosin and the AFB-inoculated controls ( p > 0.05).The reason for this could lay in the fact that during the autumn-winter season workers show an increase in lifespan associated with the decrease in the queen's egg laying capabilities.Thus, the disease had no noticeable effects on adults throughout the experiments.
Our results agree with publication data from other researchers about the effectiveness of tylosin for controlling clinical symptoms of AFB.Our previous studies (Alippi et al., 1999) demonstrated the efficacy of tylosin to control AFB by using a formulation of 1500 mg a.i. of tylosin tartrate applied in extender patties or in paper pack in suppressing AFB clinical signs after 1 year after treatment.Peng et al. (1996) found that dosages of 100 and 200 mg of tylosin protected the colonies for 3 and 4 weeks, respectively, but an additional feeding of tylosin at 100, 200, 400 and 800 mg doses can eliminate the signs of infection for an additional 3-week period.Peng et al. (1996) also noticed that bees were reluctant to accept 800 mg tylosin/7 g sugar dose, taking more than 4 weeks for its consumption.In our study bees took less than 21 days to consume 750 mg tylosin tartrate supplied with syrup, probably due to the addition of raspberry essence to improve the taste.Elzen et al. (2002a,b) working with colonies with different levels of infection with oxytetracycline-resistant AFB from a commercial apiary managed to control the infection in 15 and 45 days, respectively, by sprinkling 200 and 400 mg powdered tylosin tartrate.No re-infection was   (2002a,b) demonstrated that a total dose of 600 mg tylosin over 3 weeks applied as a dust in a powered sugar mixture effectively controls AFB during 3 months, the authors also tested a greasy patty method of application with an equivalent weekly tylosin dosage and found that the method was also effective.However, in all colonies treated with patties bee populations were significantly reduced due to the invasion and proliferation of the small hive beetle Aethina tumida that consume the food sources of a colony, including a greasy patty.
Regarding the form of application and the doses of tylosin used in our study, no significant differences were found among our three experiments ( p < 0.005), but taking into account that tylosin is converted into desmycosin that could remain stable in honey over 9 months (Kochanski, 2004a,b) we suggest to use a single dose of 750 mg and also in autumn treatment to prevent persistent residues in honey.Applications in syrup are not convenient because it is stored directly by the bees, and any applications during the nectar flow when honey is being stored should be avoided.
Nevertheless, the results obtained from Kochanski (2004b) were from honeys prepared by adding tylosin under laboratory conditions.Further studies are needed to determine the fate of tylosin and desmycosin residues in honey and royal jelly under field conditions at the recommended doses for field trials, and also the disposition profile of tylosin and desmycosin among honeybees, larvae and pupae for understanding its pharmacokinetics in bee colonies.

Fig. 2 .
Fig. 2. Severity scores of larvae with AFB-clinical symptoms in tylosin treated and control colonies at the last inspection date during 2000, 2001 and 2002 field experiments.

Table 1
Minimal inhibitory concentration values (MICs) (mg/ml) for selected isolates and designation and origin of the bacterial strains used in this study

Table 2
Clinical symptoms of AFB and amount of healthy brood and adult bees per colony and inspection date during the first field experiment starting inMarch 29, 2000

Table 3
Clinical symptoms of AFB and amount of healthy brood and adult bees per colony and inspection date during the second field experiment starting inApril 13, 2001

Table 4
Clinical symptoms of AFB and amount of healthy brood and adult bees per colony and inspection date during the first field experiment starting inMay 22, 2002