Silkworm as an animal infection model for the screening of environmental , clinical and veterinary pathogens

Silkworm, Bombyx mori, has passive immunity and can be infected by pathogenic bacteria. Therefore, it can be used as a robust bacterial infection model for screening of pathogenic isolates from various sources. In this work, 11 environmental, clinical and veterinary isolates were screened for pathogenicity using silkworm larvae by injecting bacterial suspension through their dorsal surface and observing response. Experimental conditions were established by using Bacillus thuringiensis SW_R_F_1, Escherichia coli O157:H7, E. coli DH5α and 0.6% saline. Nine out of 11 isolates were detected pathogenic after screening. The biochemical and genomic analysis of the nine test isolates confirmed their pathogenicity. The LD50 of Pseudomonas aeruginosa 47D and Salmonella Typhimurium 77 were 4.63×10 at 12 hours was 8.02×10 cells/100μl/gram at 24 hours respectively. These results indicated that silkworm exhibits differential pathological response for pathogenic and nonpathogenic bacteria, and can be used as an alternative to animal model for screening diverse isolates.


Introduction
Animal model for toxicology testing to drug discovery, 1,2 is a prominent issue all over the world for various reasons, offers a biological system in which researchers can test their hypothesis in vivo and can investigate phenomenon under study. 2Animal model can also be employed for microbial pathogenicity testing.With the recent advent of isolation of novel microorganisms, 3 it is essential to challenge their ability of infect within biological system to understand host-pathogen interactions. 4Although use of mammalian models in bacterial infection study is well established, it is often a difficult job when a large number of bacterial strains need to be tested because this requires a large number of animals to be challenged.It seems costly, time consuming, requires complex experimental procedure and set-up as well as moral conflict from animal rights point of view.Silkworm, Bombyx mori, an invertebrate animal excels those points and can be used as an animal model to screen pathogenic bacteria from various sources.
Although being arthropod, insects have robust innate immunity and they show pathological response during infection. 5Along with short response time, another feasible feature of fifth instars silkworm larvae (SL) are large enough for ease handling in comparison to other model insects like Caenorhabditis elegance and Drosophila melanogaster. 6Silkworm does not require costly setting for rearing and easy to discard after experiments through simple autoclaving.Understanding of host-microbe's interaction and screening of microbial isolates from different sources by silkworm animal model may provide an economic tool for developing countries like Bangladesh.Using the model to survey the pathogenicity of isolated bacteria from different sources will give valuable pieces of information to understand the dynamics between microorganisms in different niches of host or environmental source and their pathogenic properties.In the present work, SL been used to establish as a robust animal model to screen pathogenic bacteria from environmental, clinical and veterinary sources.

Materials and Methods
Bacterial Strains: Eleven bacterial strains isolated in our lab from clinical, veterinary seven and environmental samples (Table I) were used in this study.Additionally, E. coli O157:H7 and E. coli DH5α were used as controls.A B. thuringiensis SW_R_F_ 1 strain that was isolated from flacherie diseased dead SL (our unpublished data) was also used for the present investigation.Phenotypic and genotypic properties of bacterial strains: Hemolysin and protease tests were carried out to assess the pathogenicity of the isolates by red blood cell lysis and proteolysis respectively.5% sheep blood agar was used for Hemolysin as described previously eight and protease test was done on skimmed milk agar plate nine for all 11 isolates.The isolates were initially purified on nutrient agar media and a single colony was stabbed in both medium for 24 hours at 37ºC.
Eight pathogenicity specific common genes were selected to detect in different isolates (Table II).DNA of all gram negative isolates were extracted by boiled DNA extraction procedure 10 whereas gram positive organism's DNA was extracted by using bacterial DNA extraction kit (Jena Bioscience GmbH, Germany) following manufacturer's instruction.PCR reactions were carried out using ProFlexTM 3×32-well PCR system (Applied Biosystems, USA) with G2 master-mix (GoTaq® G2 Hot Start Polymerase, Promega, USA).Reaction volume was 15μl in all cases.The PCR products were separated in 1% agarose gel electrophoresis and their molecular weights were detected by using 1kb and 100bp ladder (Bioneer, Korea).
Experimental protocol: Five different treatments, such as insect pathogen B. thuringiensis SW_R_F_1 and human pathogen E. coli O157:H7; and 0.6% saline, non-pathogenic E. coli DH5α strain along with no treatment control, were used to establish experimental protocol.After injecting each of the treatments using 1 ml 27G ¾˝ syringes into the hemolymph through dorsal surface of SL, mortality of SL was observed for 67 hours without feeding.There were 15 insects in each treatment groups.Dead SLs were detected by observing lack of their responsive movement.The data were plotted in a survival curve. 11urvival curve of different treatments were compared which illustrates the death pattern of SL that can be related with treatment injected in each group of larvae.Eleven environmental, veterinary and clinical isolates (Table I) were tested following above described method to validate the implication of SL model for the determination of pathogenicity of unknown isolates.Each treatment was done in triplicate, each comprising five larvae.A high dose of bacteria (≈ 10 9 cells per gram of larvae) was used.The data was observed for 44 hours and plotted in a survival curve 12 and each treatment was compared with control conditions.

Postmortem identification of pathogenic bacteria:
After collecting dead caterpillar from different treatment groups, their bodies were cut into pieces by sterile scissor, resuscitated in alkaline peptone water followed by culture in selective media such as mannitol salt agar, bismuth sulfite agar and ceramide which are used for the isolation of Staphylococcus, Salmonella and Pseudomonas respectively. 13Antibiogram of both bacterial culture from original treatment and culture after bacterial isolation from dead larvae of different treatments was done.The bacterial culture equivalent to 0.5 McFarland solution 14 was spread with cotton swab on Muller-Hinton agar.After 30 minutes of incubation in room temperature, different antibiotic disks (Table III) was placed on the solid medium and resistance pattern of all bacteria observed after 24-hour incubation at 37ºC.By plotting percentage of dead larvae against a range of doses in defined time and fitting the data in 2-degree polynomial regression model, LD50 was determined by calculating the dose which represents 50% reduction of live SL. 15 The generalized polynomial equation of two degree fitted with the experimental data is:

Determination of LD50
Here, β0, β1, β2 are parametric estimation of x, x is independent variable (log(dose)), y (% of dead SL) is dependent variable and ε is estimated error of the fit.The statistical analysis and LD50 determination were carried out on R (version 3.0.1)using ggplot2 package. 16,17

Results
Silkworm larvae act well as animal model for pathogen detection: After plotting percentage of survival data of silkworm from control treatments showed distinct survival pattern.B. thuringiensis SW_R_F_1 caused a sharp death and E. coli O157:H7 caused a continuous death with time which is indicated by a decline in number of live silkworm.In contrast, no treatment, 0.6% saline condition and E. coli DH5α caused an initially lagged reduction of live silkworm larvae number.By comparing survival curve of twelve challenged isolates with control treatment, their pathogenicity was deduced and categorized into three groups (Table I)highly pathogenic (sharp decline), pathogenic (continuous decline) and nonpathogenic (lagged decline) (Fig. 1).

Post-mortem identification of pathogens confirms infections:
Growth of bacteria which were isolated from dead silkworm of treatment group A (Table I

Discussion
The target of this study was to develop a rapid protocol to screen a large number of bacterial isolates from different sources.Advent of new technology has instrumented scientists in a way that they confront a large number microorganism of new species or strains, many of which are previously uncharacterized or poorly characterized.So, it is often required to characterize their pathogenicity.Although mice are a well establish animal model to test pathogenicity, it would be a hassle to use it for screening large number of isolates such as handling many mice, cost, rearing space and consideration ethical issues concerning killing of mammals in large number.In contrast, silkworm does not have those constrains and therefore can be a robust solution for screening problem.
E. coli DH5α, E. coli O157:H7, Bacillus thuringiensis SW_R_F_1 and 0.6% saline were used to establish experimental control along with a group of silkworm larvae with no treatment conditions.Their survival curve indicates that B. thuringiensis SW_R_F_1 and E. coli O157:H7, which are pathogenic to silkworm and human respectively, while used as positive control in this work caused a continuous fall of live silkworm larvae number after challenge dose.Negative treatment condition showed a decline after an initial plateau period.This distinctive trend between positive and negative control set the ground on which inference can be drawn about the pathogenicity of test isolates.Previously other studies also established experimental condition using human pathogen S. aureus five or silkworm pathogen nuclear polyhedrons virus 18 as positive control.A limitation of current work is that LD50 is remained undetermined for control treatment conditions.But the experimental condition established here remains feasible for screening experiments.
Eleven bacterial isolates from three different sources were screened in silkworm larvae insect model and their pathogenicity was inferred by comparing survival curve of each test treatment with control.Previously, silkworm was used to evaluate pathogenicity of bacteria attached to cedar pollen. 12In this experiment, silkworm has been used to screen out pathogenic bacterium from a wide variety of sources.The insect model described in this work clearly identifies Salmonella Typhimurium isolate 77, P. aeruginosa 47D, S. aureus 39/Diab/MH1, Bacillus cereus 44D Swab2, P. aeruginosa SN28, Salmonella Enteritidis isolate 56, Salmonella Heidelberg isolate 53, S. aureus SN7, Enterobacter sp.44D/MH1, as pathogens despite of their sources.Also, test bacteria under study were isolated from dead larvae, detected by growing in selective media and their antibiotic pattern was compared with original culture.In all cases, except S. aureus 39/Diab/MH1, the antibiotic resistance pattern was identical.This confirms that in most treatment conditions death of silkworm was caused by the infection by injected bacterium.The difference in antibiogram pattern for S. aureus 39/Diab/MH1 before and after post-mortem samples might be due to contamination during handling or preparation or the inoculated bacterium might have gained its' resistance gene after injection by horizontal gene transfer.Salmonella Typhimurium 77, a veterinary pathogen isolated from poultry and a clinical isolate P. aeruginosa 47D were selected for LD50 determination.The LD50 for P. aeruginosa 47D was 4.63×10 7 cells per 100 cells/100 μl/gram in 12 hours and for Salmonella Typhimurium 77, it was 8.02×10 7 cells/100μl/gram for 24 hours.For P. aeruginosa, similar LD50 (≈10 6 cells per dose in OEP vaccination and ≈ 10 8 -10 9 cells in three component vaccination) was observed in immunized rat, but not in non-immunized rat (≈10 3 -10 4 cells). 19The same scenario was observed in case of Salmonella Typhimurimu 77. 20The higher LD50 of gram negative organism in insect model previously was correlated with 60fold increase of expression of C-reactive protein (CRP). 21e putative 10 pathogens (from group A and B in Table I) were also characterized as pathogens by the presence of four specific genes responsible for pathogenicity and two biochemical properties.These biochemical tests were used to identify novel pathogenic genes five-inch silkworm infection model.Among the test isolates, Bacillus cereus 44D Swab2, P. aeruginosa 47D and S. aureus 39/Diab/MH1 showed protease positive test and produced zone of lysis in sheep blood agar.It should be noted that these two properties had been recognized as biochemical properties of a pathogen five.To detect pathogenicity associated genes, in P. aeruginosa 47D and P. aeruginosa N28 isolate, outer membrane protein gene oprL and exotoxin A gene toxA were detected by PCR.Hemolysin A gene, hlyA was detected in E. coli O157:H7 which was positive control in this work.Secretion system apparatus protein gene, ssaT, which is a pathogenicity island II specific gene for Salmonella, found in all Salmonella strains used in this study (e.g.Salmonella sp.Enteritidis isolate 56, S. Typhimurium isolate 77, S. Heidelberg isolate 53).The biochemical properties and molecular characterization of pathogenic genes detected in 10 putative pathogens identified by silkworm larvae model corroborated the authenticity and effectiveness of the insect larvae model as a method of detection of pathogens isolated from various sources.Only a small set of pathogenic gene and phenotypic characteristics was investigated, so negative results in these experiment does not necessarily conclude to lack of virulence of the isolates.Therefore, the informality, sensitivity and specificity of this model are yet to be determined, and further more scrutinized experiments are required.It could be concluded that the minimal experimental timelength requirement, cost effectiveness, less space required for large number pathogen handling and the robustness of the insect model potentially excursing the model as an effective alternative method to animal model for pathogenicity detection for bacteria of different sources.The findings revealed that (i) Silkworm, B. mori, larvae may be used as an effective and authentic model for pathogen detection; (ii) The method is cost-effective, less time consuming, needs less space and robust.

Conclusion:
This work presents a study of silkworm with conclusion that this insect can be used as a feasible animal model to screen pathogens of wide range of origin simultaneously.Silkworm insect model can also help to classify the level of pathogenicity of screened pathogens.LD50 of different pathogen in this model can provide a quantitative way to compare this model with another higher animal.

Fig. 1 :
Fig. 1: Establishment of experimental control.Bacillus thuringiensis SW_R_F_1 causing a steep death, E. coli O157:H7 causing a continuous death while E. coli DH5α, saline and no treatment condition caused lagged death.

Fig. 3 :
Fig. 3: LD50 determination for P. aeruginosa 47D (a) and Salmonella sp.77 (b).With increasing dose, percentage of dead SL increased.Four curves were fitted in polynomial regression model for each organism for four different time and statistically significant fitted curve was used to calculate LD50.

Table I :
Groupings of the different bacterial isolates and strains used in the study based on their pathogenicity screening test with their controls.
*Group A, rapid death; group B, gradual death; group C, lagged death.

Table II :
List of pathogenic gene specific primers

Table III :
The antibiogram pattern of original culture and culture isolated from dead SL *

Table IV :
Summary statistics of significant polynomial model fitted in LD50 plot.Here, β0 is estimate of intercept of fitted regression curve; β1, β2, β3 are parametric estimates of x, x 2 and x 3 respectively; and ε is the estimate of standard error in the model (equation 1).

Table V :
Calculation of LD50