|Home | About | Journals | Submit | Contact Us | Français|
Language: | French
This study was designed to determine the susceptibility in vitro and infectivity of 1 field isolate of Mycobacterium avium sbsp paratuberculosis after exposure to monensin sodium and tilmicosin phosphate. Minimum inhibitory concentrations (0.39 μg monensin sodium/mL; 1.60 μg tilmicosin phosphate/mL) were determined in quintuplicate. Organisms were then incubated with 3 different concentrations of each medication for 3 different lengths of time, then washed and resuspended in sterile physiologic saline and injected intraperitoneally into mice that were genetically susceptible to infection. Mice were euthanatized 50 d later and the number of hepatic granulomas was used as the indicator of infectivity. Neither time of incubation nor concentration of medication had any effect on the infectivity of the organisms. Monensin sodium significantly reduced the number of hepatic granulomas in genetically susceptible mice while tilmicosin phosphate did not. Antimycobacterial activity of monensin sodium suggests that the role of monensin in the control of bovine paratuberculosis should be evaluated further.
Cette étude a été faite afin de déterminer la sensibilité in vitro et l’infectivité d’un isolat clinique de Mycobacterium avium ssp. Paratuberculosis après exposition au monensin et au tilmicosin. Les concentrations minimales inhibitrices ont été déterminées après 5 répétitions (0,39 μg/mL pour le monensin et 1,60 μg/mL pour le tilmicosin). Par la suite, les microorganismes ont été incubés avec 3 concentrations différentes de chacun des antibiotiques pendant 3 périodes de temps différentes, lavés et remis en suspension dans de la saline physiologique et administrés par voie intrapéritonéale à des souris génétiquement susceptible à l’infection. Les souris ont été euthanasiées 50 j plus tard et le nombre de granulomes hépatiques utilisés comme indicateur d’infectivité. Ni le temps d’incubation, ni la concentration du medicament n’avaient d’effet sur l’infectivité des organismes. Le monension a réduit de manière significative le nombre de granulomes hépatiques chez les souris génétiquement sensible mais pas le tilmicosin. L’activité anti-mycobactérienne observée du monensin devrait faire que son role dans la maîtrise de la paratuberculose soit évalué plus à fond.
(Traduit par Docteur Serge Messier)
Monensin is a monovalent, polyether antibiotic marketed as a feed additive for use in the prevention and control of coccidiosis, and to improve feed-efficiency and weight-gain of food-producing animals (1–5). The first isolated polyether antibiotics had antimicrobial activity in vitro against certain Gram-positive bacteria and mycobacteria (2–5). The minimum inhibitory concentration (MIC) of monensin against Mycobacterium phlei was 3.1 μg/mL (5). Tilmicosin phosphate is a semi-synthetic, 16-membered macrolide that is indicated for the treatment of cattle with respiratory tract infection caused by Mannheimia haemolytica; its antimycobacterial activity was not known prior to this study.
Mycobacterium avium sbsp paratuberculosis, the causative organism of Johne’s Disease or paratuberculosis, in several ungulate species, is a facultative, intracellular, acid-fast bacterium that is fastidious and slow-growing in vitro (6,7). Evaluation of the antimicrobial susceptibility of this organism in vitro is complicated by its requirements for growth and protracted incubation time (8). A murine model of infection (9) was used as a biological method for evaluation of the effect of monensin sodium against this organism (10). Results of one study demonstrated a prophylactic effect of monensin (15 or 30 mg monensin/kg of complete feed) when fed to mice for 50 d (10). In another study, 450 mg of monensin/head per day was fed to cattle with naturally occurring Johne’s Disease for 120 d (11). Lesions in several tissues from the treated cattle were improved at the end of the trial, but lesions in tissues from all but one of the cows in the control group worsened.
The study reported here was designed to evaluate the susceptibility of M. avium sbsp. paratuberculosis to monensin sodium and to tilmicosin phosphate in vitro and the infectivity of M. avium sbsp. Paratuberculosis with a murine model, after the organisms were incubated for 3 different lengths of time with 3 different concentrations of monensin sodium or tilmicosin phosphate. The working hypotheses were that the minimum inhibitory concentration (MIC) of monensin sodium and of tilmicosin phosphate could be determined in vitro and that infectivity of M. avium sbsp. paratuberculosis, demonstrable by the number of hepatic granulomas in the mice of the murine model, would decrease as a function of the time of incubation and the concentration of monensin sodium or of tilmicosin phosphate.
The identity of tilmicosin phosphate was coded and unknown to the investigators until after completion of this study. Results (unpublished data) of pilot studies allowed the authors of this article (RBS, DRA) to develop a protocol for determination of the MIC of monensin sodium and tilmicosin phosphate against Mycobacterium avium sbsp. paratuberculosis. The standard tube dilution procedure (12) for determining the MIC of antimicrobial drugs was modified by replacing the Mueller Hinton broth, which is routinely used for aerobic antimicrobial susceptibility procedures, with Middlebrook 7H9 broth containing oleic acid, albumin, dextrose, and catalase (OADC) (100 mL/L of medium); Tween 80 (0.05%); and Mycobactin-J (2 μg/mL of medium; Remel Laboratories, Lenexa, Kansas, USA), and using a larger initial inoculum than that prescribed for rapidly growing organisms. A field isolate of Mycobacterium avium sbsp. Paratuberculosis was obtained from the Texas Veterinary Medical Diagnostic Laboratory, College Station, Texas, USA, and its identity was confirmed by polymerase chain reaction (PCR) using primers specific for the IS 900 sequence that is unique to Mycobacterium avium sbsp. paratuberculosis. Two-fold dilutions of the Middlebrook 7H9 broth (1 mL each), as described above, with monensin sodium or tilmicosin phosphate were prepared and inoculated with 1 mL of suspension of organism (1.5 × 108 organisms/mL or 0.5 McFarland). Final volume of each tube was 2 mL with 7.5 × 107 organisms/mL and the final concentrations of the respective drugs were as indicated in Table I.
The inoculated tubes with dilutions of the respective medication were prepared in quintuplicate; placed in a non-airtight, nonhumidified, covered plastic container (Hi-Top, Clear Storage Container 7020 [41.6 × 28.6 × 18.7 cm; 15.5L]; Newell Rubbermaid, Freeport, Illinois, USA); and were incubated for 30 d at 37°C in aerobic conditions routinely used in our laboratory (RBS, DRA). During the 30-day incubation, the tubes were periodically inspected for visible signs of bacterial growth. On day 30, tubes that contained the lowest concentration of the respective medication with no visible bacterial growth were considered to contain the MIC. Separate smears of samples of broth from tubes containing visible growth were prepared and stained with Ziehl-Neelsen and with Gram’s stains. Samples of broth from the tubes containing visible growth were streaked onto blood agar plates and incubated in standard aerobic conditions for 72 h to detect the presence of bacterial contaminants.
The same field isolate used for the evaluation of susceptibility in vitro was used for this portion of the study. Organisms (1 mL of 1.5 ×108 organisms/mL) were incubated at 37°C, in medium (9 mL; final volume was 10 mL) as described above with vancomycin (0.05 mg/mL of final volume as in Herrold’s selective medium), and contained monensin sodium or tilmicosin phosphate at 0.1, 1.0, or 10.0 μg/mL for 0.5, 3, or 7 d. Organisms were washed twice with sterile physiologic saline solution (PSS) (0.9% NaCl) and resuspended in sterile PSS at a concentration of 109 organisms/mL. Mice assigned to receive the organism, were inoculated intraperitoneally with 0.20 mL (2 ×108 organisms) of the suspension.
Mice were purchased at 3 weeks of age, housed in groups of 5, and allowed a 1-week adaptation period before the study was initiated. Two hundred and twenty female, 4-week-old, genetically susceptible mice (strain C57BL6/J) (The Jackson Laboratory, Bar Harbor, Maine, USA) were used as principals. Principal mice were randomly assigned to receive one of the treatments that consisted of intraperitoneal injection of M. avium sbsp paratuberculosis after incubation with monensin sodium or tilmicosin phosphate at 0.1, 1.0, or 10.0 μg/mL for 0.5, 3, or 7 d, or one of the control treatments. Mice that received the organism that was not exposed to either of the 2 drugs served as a positive control. Mice that received 0.20 mL of sterile PSS served as a negative control. Two mice were lost to further study, leaving only 8 mice in the sentinel group that did not receive an injection. The sentinels were housed in the same room with the principal mice and provided evidence of factors that might have breached the experimental system and caused unforeseen influence on the results. Four-week-old, female, genetically resistant mice (strain C3H/HeJ) (The Jackson Laboratory, Bar Harbor, Maine, USA) received an organism that was not exposed to the drugs in order to validate reproduction of the model. All mice were fed a standard murine diet and received fresh water ad libitum for 50 d.
On day 50 postinoculation, mice were euthanatized by asphyxiation with CO2, livers were harvested from all of the mice in each treatment group, placed in separate containers with neutral buffered formalin (10% formaldehyde), and labeled with the coded identity of the respective treatment group. The following day, a sample, approximately 2-mm wide, of liver from each mouse was taken, traversing the widest and thickest portion of the left hepatic lobe. Each sample was returned to the neutral buffered formalin to complete the fixation process. The same procedure was used for all sections examined. Two slides of each hepatic sample were prepared for routine histologic evaluation. One section was stained using the hematoxylin and eosin technique, and the 2nd section was stained using the Ziehl-Neelsen technique. Both slides of liver from each mouse were examined histologically for granulomas and acid-fast bacteria, respectively. The area of one section of liver was measured with a digital, computerized system (NIH Image, version 1.6; written by W. Rasband, public domain software, http://rsb.info.nih.gov/nih-image/National Institutes of Health, Bethesda, Maryland, USA). All slides were coded with a number so that the treatment received by the individual mouse was unknown to the pathologist (JFE) who examined the sections histologically for the presence of acid-fast bacteria, and counted and recorded the number of granulomas in each of the hepatic sections. The number of granulomas was then divided by the area of that specific section to provide the value for the response variable (granulomas/mm2 of hepatic section).
Validation of the murine model followed a completely randomized design (Table II). Independent variables (treatments) required to validate the murine model were the following treatments: positive control, negative control, genetically resistant mice, and sentinels. Effects of the independent variables (2 medications [monensin sodium, tilmicosin phosphate], incubation 3 times [0.5, 3, 7 d], and 3 concentrations [0.1, 1.0, 10.0 μg/mL]) required 2 × 3 × 3 factorial analysis (Table II). The dependent variable was the number of granulomas/mm2 of hepatic section. General linear regression (SYSTAT-10; SPSS Science Marketing Department, SPSS, Chicago, Illinois, USA), appropriate for the design (Table II) of each portion of the study, was used to evaluate the effect of treatment on the number of granulomas/mm2 in the hepatic section. The appropriate equation for the 2 × 3 × 3 factorial analysis was: Yijkl = μ + Mi +Cj +Ik +M*Cij +M*Iik +C*Ijk +M*C*Iijk +εijkl; where Yijkl is the response variable for the ith medication, the jth concentration, and the kth time of incubation; μ is effect common to all response variables observed; Mi is the ith medication; Cj is the jth concentration; Ik is the kth time of incubation; M*Cij is the interaction of the ith medication with the jth concentration; M*Iik is the interaction of the ith medication with the kth time of incubation; C*Ijk is the interaction of the jth concentration of the kth time of incubation; M*C*Iijk is interaction of the ith medication with the jth concentration and the kth time of incubation; and εijkl is the random error, approximately normally distributed as (0, σ2). The appropriate equation for the completely randomized analysis was: Yij = μ + Ti + εij; where Yij is the response variable at the ith treatment; μ is effect common to all response variables; Ti is the ith treatment; and εij is the random error, approximately distributed as (0, σ2).
On day 30 of incubation, the MIC of monensin sodium was 0.39 μg/mL (tube 5) and that of tilmicosin phosphate was 1.60 μg/mL (tube 6) (Table I). All 5 sets of tubes yielded the same results, thereby, eliminating the requirement for statistical analysis. Smears of samples of broth with visible growth, stained with Ziehl-Neelsen and with Gram’s stains, contained many acid-fast organisms but no non-acid-fast organisms. Samples of broth with visible growth produced no aerobic bacterial growth during the 72 h of incubation on blood agar plates confirming that contamination of the samples did not occur. Because the broth was inoculated with M. avium sbsp paratuberculosis at the beginning of the study and no contamination occurred, no other analysis was performed for identification of the acid-fast organisms observed.
Five principal mice had to be excluded from final analysis. One mouse died during intraperitoneal injection, 2 died due to effects of malocclusion, another mouse with malocclusion survived the duration of the study but was so emaciated that data for that mouse was not included in the analysis. Two mice had hydrocephalus, but only one survived for the duration of the study and was included in the analysis. The incidence of malocclusion (1.3%) and of hydrocephalus (0.9%) were within those anticipated by the supplier (3% and 1%, respectively). Therefore, those conditions in the experimental mice were not attributed to the experimental procedure.
Actual concentrations of the targeted values for the medications were as follows: monensin sodium −10 μg/mL = 10.064 μg/mL, 1.0 μg/mL = 1.1436 μg/mL, and 0.1μg/mL = 0.1048 μg/mL; tilmicosin phosphate −10 μg/mL = 10.24 μg/mL, 1.0 μg/mL = 1.024 μg/mL, and 0.1 μg/mL = 0.1024 μg/mL.
No acid-fast organisms were observed in any hepatic section. The model for creating hepatic granulomas functioned, as described originally (9,10), and was confirmed (Table IV) by the control groups of this study. The factorial analysis (Table IV), revealed that the significant difference (P = 0.024) was due to the effect of medication (monensin sodium or tilmicosin phosphate). There was no significant effect (Table III) of time of incubation (P = 0.244) or concentration of either medication (P = 0.344).
Because of that latter finding, values for all incubation times and all concentrations of the respective medication were combined for subsequent statistical analyses of effects of all treatments, appropriate for a completely randomized design. Significant effect (P = 0.001) of medication on the response variable was found. Genetically resistant mice (Table IV) had the fewest granulomas (0.041 ± 0.058). No significant difference was detected in the mean number of hepatic granulomas/mm2 in mice that served as negative controls (0.147 ± 0.076), for sentinel mice (0.127 ± 0.038), or for mice that received organisms exposed in vitro to monensin sodium (0.115 ± 0.092). Mean number of hepatic granulomas/mm2 in mice that served as positive controls (0.298 ± 0.225) was not significantly different from that for mice that received organisms exposed in vitro to tilmicosin phosphate (0.242 ± 0.180). The mean number of hepatic granulomas/mm2 in mice that served as positive controls and mice that received organisms exposed in vitro to tilmicosin phosphate were significantly greater than those of the other groups. Results confirmed that no unrecognized, extraneous entity or the PSS influenced the response variable; the model was appropriately reproduced; monensin sodium decreased the infectivity of this organism in vivo; but tilmicosin phosphate did not decrease the infectivity of this organism in vivo.
To the authors’ knowledge, this study was the first to successfully evaluate the antimicrobial susceptibility of M. avium sbsp. Paratuberculosis in vitro using monensin sodium or tilmicosin phosphate. The MIC of monensin sodium against this isolate of M. avium sbsp. paratuberculosis was considerably less than that reported for the same medication against M. phlei (5). An explanation for that difference was not determined by the results of this study and is the subject of another investigation. The protocol developed for evaluating susceptibility in vitro in this study was based on results of pilot studies (unpublished data) in our (RBS, DRA) laboratory and has been refined so that it can be performed consistently and with confidence. The medium, prepared as described above, contained Tween 80 (13). The same medium was used throughout the study. The only difference among treatments was the medication added to the medium. Similar to other studies, the effect noted in our study was attributed to the medication (monensin sodium or tilmicosin phosphate) and was not attributed to the Tween 80 (14,15).
Results of this study support the 1st hypothesis because the MIC of monensin sodium and tilmicosin phosphate were determined in vitro using the 30-day duration of incubation and modified medium described. Results further support the conclusion that monensin sodium and tilmicosin phosphate had antimycobacterial activity against the organisms used and can be included among the few medications that have been shown to have such activity in vitro (5,13–15). Studies are underway to further prove the method of testing in vitro and to evaluate the extent of susceptibility to monensin sodium with other isolates.
Hepatic granulomas, characterized as foci of macrophages and epithelioid cells surrounded by lymphocytes, were associated with the ability of the organism to proliferate within macrophages (9). Those granulomas remain the principal response variable for this model. The murine model, as used for this study, involved younger mice, a smaller volume, and fewer organisms injected than those used in the model as originally described and those used in previous studies (9,10). The importance of the series of controls included in this study was substantiated by results that confirmed appropriate reproduction of the model and that the response variable was not compromised by those modifications. The young age of these mice created some logistical problems related to handling and recognition of the congenital conditions. Therefore, use of older mice, as in the original model, is recommended for future studies using this model.
Results with the murine model in this study revealed that our 2nd hypothesis was correct. There was no significant effect of length of time of incubation, concentration of monensin sodium, or concentration of tilmicosin phosphate on the number of hepatic granulomas in the mice. However, incubation of the organism with monensin sodium significantly reduced the number of granulomas. Tilmicosin phosphate did not significantly affect the number of hepatic granulomas caused by this organism. This evidence supports the fact that results in vitro do not always predict results in vivo. The controlled study reported here for evaluation of these medications in vivo using a murine model is one step closer to the naturally-occurring disease than are studies in vitro. There is less encouragement to pursue other studies in vivo with tilmicosin phosphate than with monensin sodium against this organism. Although susceptibility of the organism to tilmicosin phosphate in vitro was encouraging, infectivity, as determined with the murine model, was not encouraging because the number of hepatic granulomas/mm2 within the hepatic section from mice that received the organism exposed to tilmicosin phosphate was no different than that for mice in the positive control group.
Results of studies with the murine model and with cattle suggest that the effect of monensin is not a laboratory oddity. Our first study using this murine model provided evidence of a prophylactic effect of monensin sodium against M. avium sbsp. paratuberculosis (10). We subsequently showed a beneficial effect of monensin on lesions in cattle with naturally-occurring Johne’s Disease (11). Susceptibility in vitro of the organism used in this study was quantitated; the MIC was 0.39 μg of monensin/mL. The beneficial effect of monensin, at a concentration below the MIC, was demonstrated in results with the murine model in this study. Exposure to monensin sodium at a concentration of at least 0.1 μg/mL of medium for at least 12 h reduced the infectivity of M. avium sbsp. paratuberculosis. Concentrations of monensin and tilmicosin phosphate used in this portion of the study encompassed the MIC of each medication determined in this study. Logarithmic differences in concentrations were chosen because the pharmacodynamic response to a medication is normally distributed as a function of the logarithm of the dose or concentration of that medication (16,17). Results were then evaluated with statistical methods applicable to normally distributed response variables, without transformation.
The mechanism of antimycobacterial action of monensin was not the subject of this investigation but possible mechanisms, proposed elsewhere (4,5), include direct or indirect ionophoric activity that may alter the bacterial intracellular utilization of metals and electrolytes. Proper utilization of those elements is vital to intermediary metabolism of the organisms or may render organisms susceptible to other substances in the medium, such as the vancomycin that was used routinely in selective media, that are otherwise not effective against the organism (4,8,18).
Further investigations should be performed to evaluate and define the role for monensin sodium in the control and management of bovine paratuberculosis. The ability to diagnose cattle infected with this organism has greatly improved (6). While, to date, no treatment for cattle with Johne’s Disease is considered practical, medications currently suggested for treatment may be used in an extra-label manner (6,19). Efficacy of those compounds has only been evaluated with, and considered acceptable for, individual clinical patients. The role of those compounds for successful, large-scale control of the disease is doubtful. Efforts continue for development of an effective vaccine against bovine paratuberculosis, but to date, vaccination has not been universally effective (20–22).
Clinical use of monensin sodium as an aid in the control of Johne’s Disease has been largely ignored but may be a rational, economical, and valuable tool. Although laws in the United States (US) do not currently allow extra-label use of feed additives, (AMDUCA — 21 CFR Part 530; Docket number 96N — 0081, RIN 0910-AA47: http:/www.fda.gov/cvm/index/amduca/amducafr.htm) regulations in other countries may permit such use. Monensin is a fermentation product of the organism Streptomyces cinnamonensis (4). A large portion of monensin administered orally to cattle is eliminated in feces (23–26). Concentrations of monensin that were used in our murine model encompassed the MIC of the drug for the test organism and were within those present in feces (1.3 to 8.5 ppm — depending on the amount fed) of cattle receiving monensin orally (23–26). Organisms shed in feces of infected ungulates are considered to be the source of infection for other susceptible animals (6,7). Depending on size of particles ingested, the intestinal transit time for cattle is usually longer than 12 h (27–29). Exposure of the organism to monensin during intestinal transit may, therefore, be sufficient to reduce the infectivity of the organism by the time it is excreted with the feces. Based on results of this study, it could be hypothesized that fecal concentrations of monensin may be sufficient to reduce the infectivity of M. avium sbsp. paratuberculosis in that feces if exposed for at least 12 h. It could be further hypothesized that prolonged exposure to monensin in the fecal patty or soil may continue to decrease the infectivity of the organism, rendering it less likely to infect another animal that would ingest it. Monensin decays within a few weeks, depending on conditions in feces and soil, and it presents no known environmental risk (26). We suggest that studies be performed to evaluate the effect of monensin on recovery of M. avium sbsp. paratuberculosis shed in feces from infected cattle and the natural infectivity of those organisms in susceptible cattle.
Results of this and previous studies do not suggest that monensin is a dramatically effective therapeutic agent for infected cattle that demonstrate advanced clinical signs of Johne’s Disease (11). In those animals, monensin may assist to prolong the usefulness of the animal, as has been attributed to other medications (6). We hypothesize that the most probable roles of monensin in the control of the disease may be to reduce fecal shedding, infectivity or both of the organism shed from infected animals, and as an aid in prevention of infection in susceptible, young replacement animals (replacement brood-stock), or both. Should those hypotheses prove to be true, prolonged, diligent use of monensin, in time, could reduce contamination of the environment with infective organism, thereby reducing exposure and incidence of the disease in susceptible animals such as replacement brood-stock. Appropriate use of monensin, concurrent with the practice of diagnostic testing and culling of infected animals from the herd, could assist other measures for the eventual elimination of the disease from a herd or reduce the influence of the disease (U.S. Voluntary Johne’s Disease Herd Status Program for Cattle; USAHA; 1998 http://www.aphis.usda.gov/vs/nahps/Johnes/vjdhspusaha1.htm).
Supported in part by Eli Lilly and Company, Greenfield, Indiana, USA.