Super-infection by Bacillus thuringiensis H34 or 3a3b can lead todeath in mice infected with the in£uenza A virus

Eric Hernandez a;*, Franc°oise Ramisse b, Philippe Gros c, Jean-Didier Cavallo a

a Service de Biologie Medicale, Hoªpital d'Instruction des Armees Begin, 69 Avenue de Paris, 94160 Saint-Mande, Franceb Centre d'Etudes du Bouchet, Laboratoire de Microbiologie, P.O. Box 3, 91710 Vert-le-Petit, France

c Service d'Anatomie Pathologique, Hoªpital d'Instruction des Armees Begin, 69 Avenue de Paris, 94160 Saint-Mande, France

Received 1 August 2000; accepted 12 August 2000

Abstract

Bacterial super-infections are the main cause of complication and mortality after influenza virus (IAV) infection. Since Bacillusthuringiensis (Bt) is considered non-pathogenic for humans and is widely sprayed in urban areas, the aim of this work was to evaluate thepotential pathogenicity of a combined infection Bt-IAV in a mouse model of pneumonia. Bacteria used for super-infections were Bt serotypeH34 isolated from human infection and the insecticidal strain 3a3b obtained from a commercial source. Virus strain was A/Scotland/20/74(H3N2) adapted to BALB/c mice by serial lung passage. Combined infection with 4% of the viral lethal dose 50% (LD50) and 102 spores ofBt H34 killed 40% of the mice. Mortality rates increased up to 55% and 100% when combined infections were done with respectively 104 and107 spores. The insecticidal strain Bt 3a3b was less pathogenic than Bt H34. A dose of 104 spores associated with 4% of IAV LD50 killed 50%of the mice. This inoculum must be compared with the doses usually sprayed in agriculture: 1011 spores m32. Total protection against super-infection was obtained when mice were treated with amantadine. Even if only a few cases of Bt human infection have been reported, theseresults suggest a possible risk for workers spraying Bt-based biopesticides during flu outbreaks. ß 2000 Federation of European Micro-biological Societies. Published by Elsevier Science B.V. All rights reserved.

Keywords: Bacillus thuringiensis ; Pulmonary infection; In£uenza virus; Combined infection

1. Introduction

The ubiquitous soil bacterium Bacillus thuringiensis (Bt)encodes a diverse array of pesticidal proteins and is widelyused around the world for insect pest control in agricul-ture and in forestry. Recently, DNA sequences of thisbacterium encoding the N-endotoxin have been includedin plants to promote resistance to insects. The main sero-types used as biopesticides are 3a3b and H14. Health im-plications of Bt have been suspected among populationsliving in intensive spray areas [1]. Infection in humans is,however, unusual, but many cases have been describedafter laboratory contamination, in gastro-intestinal tractinfection, but also in burn patients [2] and in periodontitis[3]. In 1995, we isolated a strain of B. thuringiensis var.konkukian (ser. H34) from soft tissue necrosis in a Frenchsoldier severely wounded by a land mine explosion in for-mer Yugoslavia. Experimental infection of BALB/c mice

demonstrated that this strain was able to induce myone-crosis [4]. It was also able to kill animals that had beeninfected by the pulmonary route [5]. In the same experi-mental conditions, serotype 3a3b and H14, both widelyused as aerosolized biopesticides, were also capable of in-ducing myonecrosis after intradermal injection and pneu-monia after inhalation. A very high inoculum (108 bacteriaper animal) was necessary to obtain this pathogenic e¡ect[5].

As Bt is often considered non-pathogenic and is widelysprayed at high doses (1011 spores m32) near human com-munities [1] and since bacterial super-infection has longbeen recognized to be a major complication of in£uenza[6^11], we evaluated the potential consequences of a com-bined pulmonary infection with in£uenza A virus and Btin mice.

2. Materials and methods

This research followed the European guidelines for ani-mal experimentation.

0928-8244 / 00 / $20.00 ß 2000 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.PII: S 0 9 2 8 - 8 2 4 4 ( 0 0 ) 0 0 2 0 2 - 9

* Corresponding author. Tel. : +33 (1) 43-98-59-21;Fax: +33 (1) 43-98-53-36; E-mail : hnz.eric@freesurf.fr

FEMSIM 1256 26-10-00

FEMS Immunology and Medical Microbiology 29 (2000) 177^181

www.fems-microbiology.org

2.1. Mice

Female BALB/c mice (Charles River, Saint-Aubin-les-Elbeuf, France) were maintained under speci¢c patho-gen-free conditions until used at the age of 5^6 weeks.

2.2. Bacteria

Bt H34 konkukian was isolated from a human patientand identi¢ed by biochemical tests. H-serotyping was per-formed at the WHO Collaborating Center for Entomo-pathogens (Institut Pasteur de Paris, France).

Bt 3a3b was obtained from a commercial source (Ab-bott Laboratories, Agricultural Division, Millas, France).

Spore suspensions were harvested from a 10-day-oldculture on minimal agar medium (yeast extract, 10 g l31 ;NaCl, 5 g l31 ; agar, 20 g l31), suspended in sterile waterand incubated for 1 h at 65³C to eliminate vegetativeforms. Inoculum concentrations were adjusted after countin agar plates by dilution in sterile water.

2.3. Hemolytic activity

Hemolytic activity of the supernatants was quanti¢ed byincubating a solution of human group O erythrocytes(50 Wl ; dilution 3% in phosphate-bu¡ered saline (PBS))(Sigma Laboratories, Saint-Quentin, France) with 50 Wlof a 18-h culture supernatant (two-fold dilutions inPBS). The combination was incubated for 60 min at37³C and left to precipitate for 60 min at room temper-ature [12]. The hemolytic activity was obtained by theultimate dilution yielding 100% hemolysis. Erythrocytesin sterile culture medium were used as a negative controland erythrocytes in distilled water as a positive one.

2.4. Virus stock

The in£uenza A virus (IAV) was strain A/Scotland/20/74 (H3N2). It was initially isolated from a patient andgrown in MDCK cells. Because mice are not natural hostsfor in£uenza virus, it was adapted to BALB/c mice byserial lung passage. Virus stock for respiratory infectionwas prepared from lung homogenates of infected micediluted in PBS (Sigma Laboratories, Saint-Quentin Falla-vier, France) and stored at 380³C until use. The dosesnecessary to achieve 100% mortality (LD100) and 50%mortality (LD50) 10 days after infection were determinedafter intranasal infection of 20 BALB/c mice per group.Infectious inocula were diluted extemporaneously in PBS,from 4% to 2% of LD50.

2.5. Monoclonal anti-thuringolysin antibody

The monoclonal anti-thuringolysin antibody was pro-vided by Professor Takeshi Honda (Research Institute

for Microbial Diseases, Osaka University, Japan) who de-scribed the identity of hemolysin of Bt and B. cereus [13].Ready to use preparation at 50 Wg Wl31 was stored at+4³C.

2.6. Amantadine

Amantadine (Symmetrel0) was obtained from Dupont-Merck-Pharma Laboratories, Quebec, Canada.

2.7. Infection protocol

2.7.1. Infection of the control groupsb In£uenza virus: for each inoculum (2% and 4% LD50 of

IAV) 20 mice were inoculated intranasally after slightether anesthesia.

b B. thuringiensis : three groups of 10 mice were inoculatedintranasally after ether anesthesia respectively with 102,104 or 107 spores of Bt H34 or Bt 3a3b.

Daily weight loss was used as a criterion of morbidity.In both cases, mortality and morbidity were recorded

during 15 days.

2.7.2. Combined infectionTwenty-four hours after infection with 2% or 4% of the

viral LD50, mice were super-infected with B. thuringiensisspores (20 mice per group).

Inocula used for super-infection with Bt H34 and Bt3a3b were respectively 102, 104 or 107 spores per mouse.

Results of infection with Bt H34 results are expressed inmortality. Results of infection with Bt 3a3b are expressedin mortality for the 107 inoculum. Weight loss was used asthe criterion of morbidity when the inocula were 102, 104

spores per mouse.In both cases, mortality and morbidity were recorded

during 15 days.

2.8. Histological examinations

Histological examinations were performed immediatelyafter death or sacri¢ce (by injection of a penthotal over-dose 15 days after infection). The lungs were perfused with500 Wl of a 10% formaldehyde solution by intra-trachealinjection [5]. The tissues were then embedded in para¤nblocks and sectioned at 4^5 Wm thickness. The sectionswere mounted onto slides and stained with Gram orMay^Grunwald^Giemsa stain.

2.9. Treatment with anti-thuringolysin antibodies

Mice infected with 104 or 107 spores of Bt H34 and 4%IAV LD50 were treated 6 h before super-infection with asingle dose of anti-thuringolysin antibody (1.5 Wg dilutedin 50 Wl of PBS).

FEMSIM 1256 26-10-00

E. Hernandez et al. / FEMS Immunology and Medical Microbiology 29 (2000) 177^181178

2.10. Treatment with amantadine

Mice infected with 107 spores of Bt H34 and 4% of IAVLD50 were treated 24 h before super-infection with aman-tadine administered orally at a regimen of 5 mg permouse, twice a day, during 5 days.

2.11. Statistical analyses

Mortality rates were compared by the Pearson or theYates M2 test.

3. Results

3.1. Hemolytic activity

Hemolytic activities were 1/256 for Bt H34 and 1/128 forBt 3a3b. These titers decreased to 1/8 and 1/4 after treat-ment of the supernatant with anti-thuringolysin antibod-ies.

3.2. Infection of the control group with IAV alone

Infection with 2% of IAV LD50 induced neither weightloss nor mortality.

All the animals survived after infection with 4% of theviral LD50. The maximum weight loss was 2.8 þ 0.27 g permouse. This maximum occurred at day 10 after infection.This represents 20% of their initial weight. Weight in-creased then from day 10 to day 15. No histological le-sions were detectable in the lungs examined at day 15 afterthe experimental challenge.

3.3. Infection of the control group with Bt H34 or Bt 3a3b

As previously described, a dose of 108 spores of Bt H34killed all the mice in less than 8 h. The same dose of Bt3a3b was lethal for 80% of the animals [5]. Infection withinocula from 102, 104 or 107 spores of Bt H34 or Bt3a3binduced a local in£ammatory reaction without mortality.

3.4. Combined infection with IAV and Bt H34

When mice were infected with 4% of IAV LD50, super-infection with 102, 104, or 107 spores of Bt H34 was lethalfor respectively 40, 55, and 100% of the animals. Deathoccurred between days 8 and 12.

Statistical analysis demonstrated that mortality was sig-ni¢cantly higher in the groups super-infected with Bt H34.The highest mortality rate was observed in the groupsuper-infected with 107 spores per mouse but there wasno signi¢cant di¡erence between the 102 and 104 inocula.Results are presented in Table 1.

In the same experimental conditions, lethal rates were

Table 1Mortality rates after infection with BtH34 and 3a3b alone or combined with 4% of the LD50 of A/Scotland/20/74 (H3N2); 20 mice per group, mortalityrates observed during 15 days and statistical analyses

Spores per mouse Mortality rates (%) after combined infection IAV(4% of LD50) (n = 20 mice per group)

Control groups Bt without IAV

Bt H34a Bt 3a3b Bt H34 Bt 3a3b

0 0(0/20) 0(0/20) 0 0102 40(8/20) 20(4/20) 0 0104 55(11/20) 40(8/20) 0 0107 100(20/20) 70(14/20) 0 0

Statistical analysis Bt H34: control^102 P = 0.005a, 102^104 P = 0.34 (NS), 104^107 P = 0.002a. Statistical analysis Bt 3a3b: control^102 P = 0.16a (NS),control^104 P = 0.005a, 104^107 P = 0.06.aResults obtained after Yates correction.

Fig. 1. Survival curves of mice after infection with 4% (A) of 2% (B) ofthe LD50 of A/Scotland/20/74 (H3N2) alone and associated with variousinocula of Bt H34: mortality observed during 15 days.

FEMSIM 1256 26-10-00

E. Hernandez et al. / FEMS Immunology and Medical Microbiology 29 (2000) 177^181 179

respectively 20, 40, and 70% when the viral infection wasdone with 2% of LD50 (Fig. 1A,B).

Statistical comparison between the lethality rates dem-onstrated that there was no signi¢cant di¡erence betweenthe control group and the 102 inoculum. The highest mor-tality rates were, however, observed in the groups infectedwith the highest inocula. These results are presented inTable 1.

3.5. Combined infection with IAV and Bt 3a3b

When mice were infected with 4% of the viral LD50, 104

or 107 bacterial spores were lethal for 50% of the animals.There was a signi¢cant di¡erence between these inoculaand the control group (P = 0.005). The 102 spores permouse inoculum was not lethal. There was no di¡erencewith the control group (Table 1).

When viral infection was done with 2% of LD50, all theanimal super-infected with Bt 3a3b survived. In this spe-cial case, weight loss as a measure of morbidity correlatedwell with the dose of Bt 3a3b used to infect the mice.

The greatest weight loss were respectively 24.6% and29.8% on day 10 after infection in the groups infectedwith 104 and 107 spores of Bt. There was no di¡erencebetween the control group and the mice super-infectedwith 102 spores (19%). The results are presented in Table2.

3.6. Protection by amantadine

Total protection against the co-infection was obtainedby using amantadine at a regimen of 5 mg per mouseadministered orally, every day, for 5 days.

3.7. Protection by the monoclonal anti-thuringolysinantibody

Mice treated 6 h before infection with a single dose ofmonoclonal antibody (250 Wg per mouse) and infectedwith 104 spore per mouse were protected against the lethale¡ect for 24 h. Antibodies did not protected the mice in-fected with 107 spores.

4. Discussion

Except for B. anthracis and B. cereus, the majority of

Bacillus species have apparently little pathogenic potential.Bt is considered highly pathogenic for insects, but is rarelyassociated with disease in humans.

Apart from gastro-intestinal tract infection or after lab-oratory contamination, only a few cases of infection dueto this bacterium have been described in the literature [2^4,14]. Many Bt diseases, however, could have been over-looked because identi¢cation is di¤cult in routine labora-tory procedures. It requires the observation of the toxincrystal which involves a specialized technique not usuallyused for medical purposes and H-serotyping. In the ab-sence of this technique, Bt might be identi¢ed as B. cereuswhich is often involved in human infections. Another pos-sible reason is that this bacterium is susceptible to manyantibiotics used in the community for the prophylaxis ofbacterial super-infection during £u outbreaks.

Despite these di¤culties, human isolates are not excep-tional. A large epidemiological study, conducted in Ore-gon during two seasons of intensive aerial spraying of Btvar. kurstaki for gypsy moth control, identi¢ed 55 humanBt-positive cultures in a local general hospital [1]. Amongthese 55 isolates, 95% were considered clinically irrelevant.The bacterium was, however, isolated in pure culture fromthree immunocompromised patients, speci¢cally in an an-tecubital abscess, in a blood culture specimen, and in gallbladder content. In these three cases, the implication of Btas responsible for infection must be considered.

There is no report of pulmonary super-infection by Btin the literature after in£uenza infection. It has, however,previously been described that workers exposed to aerialspraying could elicit pulmonary antibodies directedagainst this bacterium. The majority of nasal lavage cul-tures from these exposed workers were positive for thecommercial Bt sprayed, and most of them presented aller-gic skin sensitization [15]. These two clinical studies dem-onstrate that Bt is able to penetrate the human body andto elicit an immune response. If this penetration is possi-ble, super-infection cannot be excluded especially in im-munocompromised patients.

During in£uenza infection, bacterial super-infections arefavored by various alterations of respiratory epithelial cells[16], resident macrophages [17,18], and recruitable leuko-cytes [19] due to the direct cytopathic e¡ect of the virus[20]. The set of experiments presented in this study wasdesigned to examine the virulence of Bt in such circum-stances.

Experimental infections, caused by various inocula,

Table 2Morbidity of combined infection between 2% of LD50 of A/Scotland/20/74 (H3N2) and Bt 3a3b: weight loss

B. thuringiensis 3a3b spores per mouse Weight loss on post-infection day (%) Number of survivors

5 7 10

0 9.6 19.2 19 20/20102 9.7 19.3 19.1 20/20104 16.9 26 24.6 20/20107 15.25 25.4 29.8 20/20

FEMSIM 1256 26-10-00

E. Hernandez et al. / FEMS Immunology and Medical Microbiology 29 (2000) 177^181180

demonstrated that Bt H34 and Bt 3a3b were capable, indi¡erent degrees, of super-infecting and even killing micepresenting a subacute IAV infection.

Bt 3a3b seemed to be less virulent than Bt H34. Weightloss as a measure of morbidity, however, showed that alow inoculum of this bacterium was able to seriously com-plicate a IAV respiratory tract infection in mice.

The partial protection observed when animals weretreated with monoclonal antibodies directed against thethuringolysin suggests that this hemolysin was implicatedin the pathogenicity. This toxin was, however, probablynot involved alone. Many other Bt-produced toxins, suchas phospholipase C and sphingomyelinase could also beimplicated [5].

Protection obtained after treatment with amantadine,which inhibits IAV infection, demonstrated that Bt wasnot pathogenic alone. The virus transiently altered thedefense of the respiratory tract and rendered the host sus-ceptible to a pulmonary super-infection by a very low in-oculum of Bt.

Taken together, these results suggest that Bt sprayingaround human populations could be dangerous, especiallyfor immunocompromised patients. Even if no cases ofpulmonary infection have been described, they underlinethe necessity of worker protection during aerial sprayingby the use of such safety materials as security masks,gloves, work clothes and goggles.

Acknowledgements

We gratefully acknowledge Professor Takeshi Honda(Research Institute for Microbial diseases, Osaka Univer-sity, Japan) for providing the anti-thuringolysin monoclo-nal antibody. We also thank our friend Alan Talmon,Minister of the French Reformed Church, and Ian AlanHolder (Shriners Hospitals for Children, Cincinnati, OH,USA) for the help and the time provided to us. Financialsupport was provided by the Gouvernement Franc°ais,Ministe©re de la Defense, Service de Sante des Armees.

References

[1] Green, M., Heumann, M., Sokolow, R., Foster, L.R., Bryant, R. andSkeels, M. (1990) Public health implications of the microbial pesticideBacillus thuringiensis : an epidemiological study, Oregon, 1985^86.Am. J. Publ. Health 80, 848^852.

[2] Damgaard, P.H., Granum, P.E., Bresciani, J., Torregrossa, M.V.,Eilenberg, J. and Valentino, L. (1997) Characterization of Bacillus

thuringiensis isolated from infections in burn wounds. FEMS Immu-nol. Med. Microbiol. 18, 47^53.

[3] Helgason, E., Caugant, D.A., Olsen, I. and Kolsto, A.-B. (2000)Genetic structure of population of Bacillus cereus and B. thuringiensisisolates associated with periodontitis and other human infections.J. Clin. Microbiol. 38, 1615^1622.

[4] Hernandez, E., Ramisse, F., Cruel, T., Ducoureau, J.P., Alonso, J.M.and Cavallo, J.D. (1998) Bacillus thuringiensis serovar H34 konku-kian superinfection: report of one case and experimental evidence ofpathogenicity in immunosuppressed mice. J. Clin. Microbiol. 36,2138^2139.

[5] Hernandez, E., Ramisse, F., Cruel, T., le Vagueresse, R. and Cavallo,J.D. (1999) Bacillus thuringiensis serotype H34 isolated from humanand insecticidal strains serotypes 3a3b and H14 can lead to death ofimmunocompetent mice after pulmonary infection. FEMS Immunol.Med. Microbiol. 24, 43^47.

[6] Loosli, C.G. (1949) The pathogenesis and pathology of experimentalair-borne in£uenza A infections in mice. J. Infect. Dis. 84, 153^158.

[7] Loosli, C.G. (1968) Synergism between respiratory viruses and bac-teria. Yale J. Biol. Med. 40, 522^540.

[8] Jakab, G.J. (1982) Immune impairment of alveolar macrophagephagocytosis during in£uenza virus pneumonia. Am. Rev. Respir.Dis. 126, 778^782.

[9] Jakab, G.J., Astry, C.L. and Warr, G.A. (1983) Alveolitis induced byin£uenza virus. Am. Rev. Respir. Dis. 128, 730^739.

[10] Gardner, I.D. (1981) Suppression of antibacterial immunity by infec-tion with in£uenza virus. J. Infect. Dis. 144, 225^231.

[11] Nickerson, C.L. and Jakab, G.J. (1990) Pulmonary antibacterial de-fenses during mild and severe in£uenza virus infection. Infect. Im-mun. 58, 2809^2814.

[12] Coolbaugh, J.C., Wende, R.D. and Williams, R.P. (1972) Microtitra-tion of Bacillus cereus hemolysin. Appl. Microbiol. 24, 997^998.

[13] Honda, T., Shiba, A., Seo, S., Yamamoto, J., Matsuyama, J. andMiwatani, T. (1991) Identity of hemolysins produced by Bacillusthuringiensis and Bacillus cereus. FEMS Microbiol. Lett. 63, 205^209.

[14] Jackson, S.G., Goodbrand, R.B., Ahmed, R. and Kasatiya, S. (1995)Bacillus cereus and Bacillus thuringiensis isolated in a gastroenteritisoutbreak investigation. Lett. Appl. Microbiol. 21, 103^105.

[15] Bernstein, I., Bernstein, J., Miller, M., Tierzieva, S., Bernstein, D.,Lummus, Z., Selgrade, M., Doer£er, D. and Seligy, V. (1999) Im-mune responses in farm workers after exposure to Bacillus thuringien-sis pesticides. Environ. Health Perspect. 107, 575^582.

[16] Sanford, B.A. and Ramsay, M.A. (1987) Bacterial adherence to theupper respiratory tract of ferrets infected with in£uenza A virus.Proc. Soc. Exp. Biol. Med. 185, 120^128.

[17] Astry, C.L. and Jakab, G.J. (1984) In£uenza virus-induced immunecomplexes suppress alveolar macrophage phagocytosis. J. Virol. 50,287^292.

[18] Jakab, G.J., Warr, G.A. and Sannes, P.L. (1979) Alveolar macro-phage dysfunction associated with viral pneumonitis. Adv. Exp.Med. Biol. 1216, 81^90.

[19] Warr, G.A., Jakab, G.J., Chan, T.W. and Tsan, M.F. (1979) E¡ectsof viral pneumonia on lung macrophage lysosomal enzymes. Infect.Immun. 24, 577^579.

[20] Jakab, G.J. (1990) Sequential virus infections, bacterial superinfec-tions, and ¢brogenesis. Am. Rev. Respir. Dis. 142, 374^379.

FEMSIM 1256 26-10-00

E. Hernandez et al. / FEMS Immunology and Medical Microbiology 29 (2000) 177^181 181