The BALB/c Mouse Model for the Evaluation of Therapies to Treat Infections with Aerosolized Burkholderia pseudomallei

Burkholderia pseudomallei, the causative agent of the disease melioidosis, has been isolated from the environment in 45 countries. The treatment of melioidosis is complex, requiring lengthy antibiotic regimens, which can result in the relapse of the disease following treatment cessation. It is important that novel therapies to treat infections with B. pseudomallei be assessed in appropriate animal models, and discussions regarding the different protocols used between laboratories are critical. A ‘deep dive’ was held in October 2020 focusing on the use of the BALB/c mouse model and the inhalational route of infection to evaluate new antibiotic therapies.


Melioidosis
Burkholderia pseudomallei, the causative agent of the disease melioidosis, is a saprophytic bacterial pathogen living in soil and stagnant waters and is endemic to many parts of the world. It is naturally resistant to multiple antibiotics and the current treatment regime is lengthy, comprising two phases totaling 6-9 months of antibiotic intervention [1][2][3]. In addition, strains resistant to standard-of-care therapeutics are being reported [4][5][6][7]. The recent identification of B. pseudomallei in a commercially available aromatherapy spray, in an aquarium and in the environment in the US highlights the organism's ability to survive in disparate environments and the potential for further global spread [8][9][10].
Estimated at 165,000 cases worldwide with 89,000 deaths per year [11], melioidosis is considered to be a neglected disease by many, further supporting the argument for more funding to be made available for the development and evaluation of new therapeutics and treatment strategies to reduce the global health burden of this disease [12]. To enable the effective evaluation of these therapies, well-characterized animal models that recapitulate human disease are required. Models in several animal species have been developed, but this manuscript focuses on the commonly utilized BALB/c mouse infected with B. pseudomallei by the inhalational route.

The Participants
A 'deep dive' was held in October 2020, coordinated by the US Defense Threat Reduction Agency (DTRA), to discuss the BALB/c mouse as a model to evaluate therapeutics for the treatment of infections with B. pseudomallei. The participants included scientists from

Current Therapies for the Treatment of Melioidosis
The treatment of melioidosis is complex and lengthy, and success is dependent on the severity of the infection when therapy is initiated. It typically comprises a minimum of two weeks of intravenously delivered ceftazidime or meropenem, followed by 3-6 months of orally delivered co-trimoxazole or co-amoxiclav [1][2][3]. Despite the completion of the treatment regimen, B. pseudomallei infection can relapse, typically following a period of immunosuppression [13]. This relapse of the disease accounts for approximately 10% of cases but is dependent on the country [14]. The development of natural resistance to these antibiotics, in combination with the occurrence of adverse effects and the potential relapse of the disease, warrants the determination of the efficacy of new therapeutics in treating melioidosis. To evaluate these products effectively, well-characterized and well-developed animal models are required, and, in the case of melioidosis, this includes the use of the BALB/c mouse.

Mouse Models
The two most common mouse strains used in B. pseudomallei research are BALB/c and C57BL/6. The Steering Group on Melioidosis Vaccine Development (SGMVD) previously communicated the role of both mouse strains in vaccine development due to their different sensitivities to infection, and that when taken together, they may be more representative of human disease [15]. BALB/c mice are more sensitive to infection and represent a more acute disease, typically used in therapeutic evaluation, whereas C57BL/6 mice (which are more resistant and represent a more chronic infection) are more commonly used in vaccine studies [16][17][18]. These classifications (in the context of acute or chronic infection models) are limited by a consensus of what defines a chronic infection in a mouse model of melioidosis; therefore, studies can be conducted for lengthy periods of time.
Several models delivering B. pseudomallei by multiple routes (including intraperitoneal, subcutaneous, intravenous, oral, intratracheal and intranasal) have been developed. In the context of both public health and particularly biodefense, however, a well-characterized inhalational model of melioidosis is most appropriate and necessary for the development and evaluation of novel medical countermeasures [19][20][21][22][23].
With respect to therapeutic evaluation, the BALB/c mouse model is a robust model of B. pseudomallei infection, typically using females 6-8 weeks old (Table 1). A target dose of approximately 20-50 LD 50 equivalents achieves an acute, lethal infection. Lower LD 50 challenges can lead to a chronic infection with a prolonged time to death (weeks) [24]. Variables including sex, age, vendor, the route of challenge and a geotemporally diverse set of B. pseudomallei strains have been previously investigated in the BALB/c mouse strain. The USAMRIID has previously reported that there were no statistically significant differences between age-matched male and female BALB/c mice following exposure to aerosolized B. pseudomallei strain K96243 [26]. This work also demonstrated that older mice were more susceptible to infection (when comparing 8-week-old mice with 28-week-old mice). Retrospective analyses of limited data sets indicated that BALB/c mice from different vendors responded similarly to infection with B. pseudomallei, although documented differences between source vendors can result in differences in mouse phenotypes that could potentially impact experimental reproducibility [27][28][29].
BALB/c mice exhibit a heterologous presentation of the disease, which is generally representative of that documented in humans [19,30]. The median lethal dose (LD 50 ) values for BALB/c mice exposed to aerosolized B. pseudomallei have been determined for many isolates. One study demonstrated that the LD 50 ranged from less than 1 CFU to approximately 4300 CFU when aerosolized B. pseudomallei strains were delivered to BALB/c mice (compared to C57BL/6, which demonstrated values from approximately 43 CFU to approximately 10,000 CFU) [31]. The two B. pseudomallei isolates routinely used for therapeutic evaluation are strains K96243 and 1026b, which have LD 50 values of approximately 4-25 and 4 CFU, respectively, in the BALB/c mouse model [20,26,32].

Preparation of Challenge Material
As discussed above, the typical bacterial strains used within the four laboratories are K96243 and 1026b, both isolated from melioidosis patients in Thailand in the 1990s and having comparable virulence. K96243, probably the most widely used strain, was isolated at Mahidol University Hospital in 1996 and passaged four times before arriving at the Central Public Health Laboratory, UK. Dstl received the isolate from the Central Public Health Laboratory, UK, in 1999 following a further 3 passages, and the strain was eventually deposited in the Biodefense and Emerging Infections Research Resources Repository (BEI Resources) by Battelle Biomedical Research Centre following another 17 passages [31,33].
While there is value in standardizing strains used for therapeutic evaluation, the heterogeneity in B. pseudomallei supports the use of multiple strains. It is important that the genetic instability of B. pseudomallei is also taken into consideration with the production of seed banks with minimal passages. Genetic differences have been identified between B. pseudomallei strain K96243 in different laboratories; therefore, having a good understanding of the provenance of strains and access to genome sequence information is imperative [34]. In addition, as endemic strains naturally become antibiotic-resistant, it is important that different strains are characterized in animal models. The same is true of B. pseudomallei strains isolated from different geographical locations. One example is the Australian isolates, which have been shown to cause more neurological problems. Some of these have been evaluated in animal models of melioidosis, including the Balb/c mouse [31].
Differences in bacterial challenge material preparation were observed between the four institutions ( Table 2). ITI cultures B. pseudomallei in Brain Heart Infusion (BHI) broth with shaking for 12-14 h at 37 • C, inoculated from a fresh stock grown on Trypticase Soy agar (TSA). The following day, the concentration is adjusted, serially diluted and plated on TSA. The USAMRIID inoculates bacteria from a frozen stock into 4% glycerol (Sigma Aldrich, St. Louis, MO, USA), 1% tryptone (Difco, Becton Dickinson, Sparks, MD, USA), and 0.5% NaCl (Sigma Aldrich) broth (GTB), which is placed into a shaking incubator for 24 ± 4 h at 37 • C. The following day, an OD 660 is measured and diluted with GTB to the starting concentration for challenge, which is then confirmed by bacterial enumeration on sheep blood agar (SBA) plates. Dstl inoculates bacteria onto Luria agar (LA) and then into Luria-Bertani broth (LBB) for 16 ± 4 h at 37 • C with shaking at 200 rpm. Following challenge, the concentration of viable bacteria in the starting inoculum is determined by performing 10-fold bacterial dilutions plated on LA. Battelle inoculates thawed bacteria into Luria-Bertani broth with 4% glycerol (LBG) and incubates at 37 • C (shaking at 250 rpm) for 18-24 h. Following incubation, the starter culture is diluted in LBG and incubated at 37 • C (shaking at 250 rpm) for 18-20 h. Following incubation, a portion of the culture is pelleted, washed with Dulbecco's phosphate-buffered saline (PBS) with 0.01% gelatin and 9.7% α-α trehalose (BSGT) and adjusted to the required concentration. All cultures are enumerated on LBG agar (LBGA).

Aerosolization of Challenge Material
The aerosol system at Dstl, USAMRIID and ITI consists of an exposure chamber attached to an automated bioaerosol exposure (Biaera Technologies L.L.C., Hagerstown, MD, USA) system for the control of chamber airflow, chamber pressure, aerosol generation and aerosol sampling (Table 3) [35]. Aerosols are generated by a three-jet Collison nebulizer (BGI, Inc., Waltham, MA, USA) and spray tube in an AeroMP apparatus (Biaera Technologies L.L.C., Hagerstown, MD, USA). The temperature within the chamber during the spray is ambient, 18-25 • C (dependent on the laboratory). Relative humidity (RH) can be controlled and is usually set between 40 and 75%, with a recent shift to using a higher RH (~70%) due to better reproducibility between runs. The flow rates detailed in Table 3 are based on the inherent flow of the specific AGI impingers and aerosol systems used by the different laboratories. PBS-Phosphate-buffered saline; GTB-4% glycerol, 1% tryptone and 0.5% NaCl broth; BSG-(PBS) with 0.01% gelatin and 9.7% α-α trehalose; BHI-Brain Heart Infusion broth; LBG-Luria-Bertani broth with 4% glycerol; SBA-sheep blood agar; LBGA-Luria-Bertani broth with 4% glycerol; CBA-Columbia blood agar; TSA-Trypticase Soy agar.
The generated aerosol is sampled either for the duration of the exposure or for a defined period of time at the midpoint of the challenge by using an all-glass impinger (AGI) (Ace Glass Inc., Vineland, NJ, USA). The concentration of B. pseudomallei in the aerosol is determined by recovering samples from the exposure chamber using an AGI operating at 12 L/min containing 10 mL of PBS (Dstl) or BHI broth with antifoam (ITI). The USAMRIID runs their system at 19.5 L/min with the AGI collection run at 6 L/min containing 10 mL of GTB and 0.001% antifoam A. An Aerodynamic Particle Sizer (APS) Spectrometer (TSI, Inc., Shoreview, MN, USA) in conjunction with an Aerosol Diluter (TSI, Inc., Shoreview, MN) is used (although not routinely in all laboratories) to determine the particle size distribution. The concentrations of viable bacteria in the AGI samples are analyzed by performing 10-fold bacterial dilutions, as discussed above. The aerosol system at Battelle is used as previously described [36]. No significant differences in the spray factors have been determined for the K96243 or 1026b strains of B. pseudomallei.
The inhaled or retained dose of bacteria delivered as an aerosol is then calculated. The aerosol concentration (C e ) is calculated by (C sampler × V sampler )/(Q sampler × t), where C sampler = the titrated concentration of the sampler, V sampler = the volume of the collection media in the sampler, Q sampler = the flow rate through the sampler, and t = the total time the sample was taken. The total volume (V t ) of the experimental atmosphere inhaled is determiend (V t = V m × length of exposure), where V m applies Guyton's formula, which determines the volume of air breathed per unit time as a function of body weight (W) in V m /W (mL g −1 ) = 2.1 × W 0.75 [37]. V t is then multiplied by the aerosol concentration to give the presented dose (D p = C e × V t ) [3]. The retained dose calculation then assumes that each mouse retains 40% of the organisms that were inhaled in the lungs (Retained dose = Presented dose × 0.4) [38]. These calculations are a useful indication of the dose; however, without sacrificing the animals to determine the actual bacterial count delivered and present in the lung, it is not possible to determine the actual dose. It is also worth noting that there are some assumptions associated with the calculations, including a constant and equal minute volume between experimental animals and a constant aerosol concentration over time. However, the calculated dose is an important comparator between experiments to enable comparisons of therapeutic efficacy.
At Battelle, aerosol concentration (CFU/L) is calculated from the impinger CFU divided by the product of the average impinger sample rate and sample time. Rodent respiration rates and minute volumes are calculated using Guyton's formula as above. The total accumulated tidal volume (TATV) is calculated as the product of the minute volume and total exposure time. The total inhaled dose (CFU) is calculated from the product of the aerosol concentration and TATV.

Route of Infection and Challenge Dose
The aerosol systems described above deliver the bacteria as a whole-body or nose-only exposure. There should not be any differences in the inhaled dose between the two systems, as the respiratory rates will be the same (i.e., the volume of air the animal breathes in). The differences in terms of the actual dose would relate to the aerosol concentration and the volume of the chamber (the larger the volume, the larger the dilution).
In this mouse model, the median time to death is observed at day 3-4 post-challenge with K96243 or 1026b if challenge doses are within the range of 20-135 LD 50 equivalents [32,39,40]. Time to 100% lethality in infected and untreated mice ranged from day 2 to 6 post-challenge [23]. One possibility with the whole-body system is oral exposure following the grooming of contaminated fur; however, it has been demonstrated that high doses (>1 × 10 6 CFU) of B. pseudomallei are required to establish an infection in mice following ingestion, and therefore, it is unlikely to be a contributing factor for this disease [41].

Pathogenesis of Disease
The participants acknowledged that the best way to define the initiation time point for therapeutic administration is to characterize the natural history of the disease. The pathogenesis of infection may be slightly different depending on the bacterial strain, the in vitro culture conditions and the infecting dose.
Following exposure to aerosolized B. pseudomallei, BALB/c mice are typically transiently bacteremic between day 2 and day 4 post-exposure. The concentration of bacteria in the blood is low (less than 100 CFU/mL) and only observed in a subset of animals [19,31,42]. The lungs, as the primary site of infection following exposure to aerosolized bacteria, are the most heavily colonized, typically followed by the liver and spleen.
Following an aerosol exposure with 20 LD 50 equivalents of strain K96243, Dstl observes high levels of bacteria in the lungs, liver and spleen and some kidney, brain and blood samples at 24 h post-challenge. By 36 h post-challenge, all organs are colonized, with untreated animals succumbing to infection by day 3-4 post-challenge [32,39,43].
Following a targeted 50 × LD 50 exposure, Battelle observes bacteremia and bacterial colonization in the liver, lung and spleen from 24-60 h post-exposure. Between 24 and 60 h post-challenge, the levels in the liver and spleen increase as the level in the lungs decreases. Bacteria are also present in the kidney from 36 to 60 h post-challenge.
The USAMRIID observes a peak bacterial load in the lungs at approximately day 7 post-challenge, reaching 10 8 CFU per gram of lung following exposure to 5 CFU of strain K96243 (less than 1 LD 50 equivalent). A reduction in the level of bacteria colonizing the lungs between days 7 and 10 post-challenge is then observed, with few survivors beyond 14-21 days post-challenge. The bacterial burden in the spleen aligns with that observed in the lungs, peaking at approximately day 7 post-challenge, with a reduction in bacterial concentration observed at later time points. A normal healthy BALB/c spleen is approximately 0.1-0.2 g, and a mouse with melioidosis can have a spleen that is enlarged 10-fold or more. In some cases, based on clinical signs, mice would not reach euthanasia criteria, but upon necropsy, the enlarged spleen is clearly indicative of localized disease. Multiple pyogranulomatous lesions are observed in the spleen, other organs and anatomical locations, which is characteristic of melioidosis. The bacterial load enumerated in the liver is similar to those described for spleens and lungs. BALB/c mice have also been shown to have inflammation in the liver, lung, nasal cavity, ear canal and brain [19,20,44].
One study stated that BALB/c mice challenged with 5 CFU of strain K96243 presented with higher temperatures relative to C57BL/6 mice challenged with 18 CFU (approximately 0.05 LD 50 equivalents) during the first 5 days post-challenge [19]. Furthermore, BALB/c mice lost weight starting at 24 h post-challenge, reaching their nadir at day 5-6 postchallenge and regaining thereafter.
In addition, when enumerating bacteria on agar, it is also important to consider colony morphology, particularly when isolating bacteria from animals. B. pseudomallei colonies are known to have different morphologies, and colony switching has been reported [45]. It is important that this is considered with confirmation via a secondary method, e.g., PCR, if required.

Time of Treatment Initiation and Antibiotic Comparators
While there are small differences in the course of the disease within laboratories (based on the natural history studies detailed above), the group determined that initiating therapies 24-36 h post-challenge represented an appropriate time point for routine studies (Table 4). These initiation times are based on previously generated experimental data and the development of clinical signs of disease. It was agreed that both ceftazidime and co-trimoxazole are valid positive controls to be included in therapy studies, as (1) both are standard-of-care treatments for human melioidosis, (2) there is a significant body of historical data in the BALB/c mouse model and (3) each represents a different route of administration. Therefore, the selection should be based on the laboratory and the scientific purpose of the study.
Dstl defines post-exposure prophylaxis (PEP) as therapy delivered at 6 h post-challenge, when bacteria can be detected in the lung, liver and spleen [39]. When the time to the initiation of therapy is delayed to 24 h post-challenge (when bacteria have disseminated to the brain and are detected in the blood), the survival of mice treated with comparator antibiotics is reduced. Co-trimoxazole is typically included at Dstl, 78 mg/kg delivered orally twice a day for 14 days (Dstl, unpublished data). When administered at 24 h postchallenge, approximately 80% of animals treated with this antibiotic will be colonized by day 42 post-challenge and start succumbing to the disease [46]. At Dstl, the concentration of bacteria is typically determined in tissues at the time of treatment initiation, at the end of treatment and at the end of the study.
ITI defines the initiation of therapy at 24 h post-challenge as PEP and treatment as 42-48 h post-challenge [40]. Irrespective of the treatment initiation time, the concentration of colonizing bacteria present in infected and untreated animals is determined. ITI typically includes ceftazidime, administering 150 mg/kg by the intraperitoneal (IP) route every 6 h for 14 days. This results in 90-100% survival when evaluated as PEP and 40-50% when evaluated as treatment (at day 45 post-challenge) [40]. A dose of 150 mg/kg delivered every 6 h achieves the therapeutic 'time above MIC' parameter, and therefore, higher doses do not increase efficacy (at the end of the study). Livers and spleens harvested at day 45 post-challenge are colonized with B. pseudomallei, with less than 50% of surviving animals shown to be clear of bacteria. If these studies are extended, deaths are typically observed at day 60 post-challenge, correlating with the percentage of animals colonized at day 45. Standard therapeutic intervention (PEP) at the USAMRIID is 24 h post-challenge, with treatment typically defined as the intervention at 48 h post-challenge. Delaying the initiation of therapy to 72 h post-challenge has resulted in very low levels of survival. The USAMRIID delivers 150-300 mg/kg of ceftazidime every 6 h for 21 days via IP injection following an aerosol infection with strain 1026b. Typically, the dose used is 150 mg/kg delivered every 6 h, which offers 30-80% protection at day 60 post-challenge. No dose of ceftazidime evaluated has consistently resulted in sterile spleens at day 60 post-challenge [47,48]. In experiments involving therapeutic evaluation, tissue burden (e.g., lung and spleen) is only routinely investigated at the end of the study. No strong correlation has been observed between the challenge dose and percentage survival at day 60 post-challenge in infected animals receiving ceftazidime (150-300 mg/kg). In addition, no correlation between the challenge dose and splenic bacterial burden has been observed in ceftazidime-treated animals at day 60 post-challenge.
Battelle initiates ceftazidime as treatment at 24 h post-challenge, delivering 150 mg/kg or 300 mg/kg by the IP route every 6 h for 14 days. Ninety percent survival is observed at the end of treatment when 300 mg/kg is administered, compared to 50-60% survival at the end of treatment when 150 mg/kg is administered [49]. At day 45 post-challenge, 70% and 40% protection has been afforded for 300 mg/kg and 150 mg/kg doses, respectively. By 75 days post-challenge, 40% and 13-20% survival rates have been observed for 300 mg/kg and 150 mg/kg doses, respectively.
As unsuccessful clearance of B. pseudomallei is believed to result in the relapse of the disease, murine studies have been extended to day 70 post-challenge (following treatment for 14 days) to determine whether a sterilizing antibiotic could be identified. However, it has been noted that this is an arbitrary end point, as human treatment includes oral eradication therapy, which can be up to 6 months. Furthermore, there are additional variables to consider, including the presence of bacterial persisters and differentially culturable cells that can be produced following exposure to antibiotics or other stresses [50,51]. The group agreed that, in some instances, longer studies are warranted, but an end point of 45 days post-challenge (assuming 14 days of treatment) is reasonable in this mouse model. Extending studies to 75 days post-challenge may be useful to interrogate the additional benefit of innate defense regulator molecules and other adjunctive therapies, which may not be obvious in shorter studies. Although not utilized by the four laboratories, an alternative to observing the natural relapse of the disease is to artificially stimulate this using an immunosuppressant. As the BALB/c mouse discussed in this manuscript is a screening model focusing on survival and a reduction in bacterial burden in tissues at the end of the study, this adequately enables comparisons between standard-of-care antibiotics and test article groups to be made without the requirement for immunosuppression.

Minimum Inhibitory Concentration (MIC) or Minimum Bactericidal Concentration (MBC)?
Generally, the in vitro measurement of antibiotic activity is the MIC, defined as the lowest concentration of an antibiotic that prevents the growth of the bacteria in the inoculum [52]. The MIC forms the in vitro basis for modeling using pharmacokinetic parameters to predict the pharmacodynamics of the antibiotic. However, discussions are warranted regarding the importance of the MBC (the lowest concentration of the antibiotic that kills 99.9% of the bacteria in the original inoculum) as an alternative and more informative value for bactericidal drugs, evaluating the killing and clearance of pathogens, rather than just the inhibition of growth. For pathogens that can result in high tissue and blood concentrations (such as B. pseudomallei), the inhibition of growth may not be adequate (as demonstrated by the very short treatment window for most of the biothreat pathogens and the incidence of relapse in melioidosis).

Pharmacokinetics in Infected Animals
The importance of the pharmacokinetic profiles of novel antibiotics, in both naïve and infected animals, and the penetration of a drug into the lung epithelial lining fluid (ELF) under both conditions was discussed. Where possible, human exposure data should be used to model human-simulated dosing regimens to provide reasonable estimates of efficacy for candidate drugs, as is conducted with comparator antibiotics. ITI provided an overview of the PK profile of ceftazidime in the context of lung penetration in the BALB/c mouse (both infected with B. pseudomallei and naïve). The ceftazidime levels in the lung (ELF) were unchanged following infection when compared to naïve animals; however, the plasma levels in infected animals were reduced compared to the naïve controls, potentially accounting for the residual colonization observed in tissues (Heine, unpublished data). A particular emphasis on the importance of PK in antibacterial development was highlighted, in addition to the need to maintain adequate exposure and prevent the emergence of resistance.
The group discussed the mechanisms by which samples are sterilized to enable PK analysis outside of high containment. Following the successful validation of the process, Battelle and ITI filter-sterilizes samples. Battelle has validated the use of 0.22 µm PES filters to enable the detection of ceftazidime in plasma samples. During method development, the impact of filtering on the loss of drug should be examined, as some hydrophobic drugs can bind to the filter. Once outside of high containment, the analysis utilized to detect the antimicrobial (e.g., HPLC, mass spectrometry) is dependent upon the drug.

Determination of the Host Response to Infection
Several laboratories have characterized the host response following a B. pseudomallei infection by carrying out in vivo infection studies or by examining the immune response in primary cells harvested from mice [53]. The immunological response is similar for each of the bacterial strains, although the magnitude and duration of the response do vary. The host response is generally characterized by chemokine or cytokine responses [19,20,42], with the cellular response with clinical chemistries occasionally assessed [19,20,54].
Following inhalational exposure, physiological and immunological changes are rapidly seen in BALB/c mice [19,20,31,42,55]. Following challenge with K96243, an initial reduction in the levels of serum alanine aminotransferase is demonstrated, followed by an increase at 48 and 60 h post-challenge [54]. Levels of aspartate aminotransferase were also increased at 48 and 60 h post-challenge, and a reduction in blood glucose was observed. In another study, reductions in the liver enzyme alkaline phosphatase and serum albumin were observed by day 3 post-challenge, while serum globulin levels increased [20].
The chemokine and cytokine responses in the lungs and spleens of mice challenged by K96243 were similar to those in the sera, with a rapid and significant induction of IL-6, MCP-1, IFN-γ and TNF-α by 24 h post-challenge [55]. A follow-on study showed that the induction of cytokine and chemokine responses was similar in the lung, liver and spleen for some cytokines (IL-1α, IL-6, IL-12p40, IL-12p70, IL-17, G-CSF), chemokine (C-C motif) ligand (CCL)4 and TNF-α but not for others (IL-1β, IL-4, IL-10, IFN-γ, chemokine, (C-X-C motif) ligand (CXCL)1, CCL2, CCL3 and CCL5) [49]. In the spleens of mice that received a low challenge dose of K96243, there was a peak of low-level expression of IFN-γ and TNF-α at 2 days post-challenge and IL-6 at day 15 post-challenge [19].
For mice challenged with 1026b, there was an induction of IFN-γ, IL-10, IL-1β, IL-6, KC and TNF-α in the lungs of mice at 24 h post-challenge. This had reduced to negligible levels by day 4 post-challenge, except for IL-10, which maintained a reduced, but significant, level [22].

1.
Comprehensive natural history studies are the only mechanism by which an appropriate and consistent intervention time may be defined.

2.
Bacterial strains K96243 and 1026b are both considered type strains of B. pseudomallei and are routinely utilized; however, the inclusion of additional strains may be warranted during the development of novel therapeutics and following the natural evolution of strains with different antibiotic susceptibility profiles.

3.
Both ceftazidime and co-trimoxazole are appropriate antibiotics to include in studies evaluating novel therapies for melioidosis.

4.
The bacterial load in a panel of tissues should be taken from a cohort of mice at the time of treatment initiation to determine the extent of pathogen dissemination (and to understand the stage of the disease). This could also be applied to other experimental time points (e.g., at the end of the treatment period).

5.
Initiating treatment at 24-36 h post-challenge appears to be appropriate for therapeutic intervention following inhalational infection with the K96243 strain of B. pseudomallei. For 1026b, this may be extended to 48 h post-challenge. 6.
Although there are variations and limitations in using this animal model, we are recapitulating key aspects of human disease, predominantly the acute nature of the disease, following exposure by the aerosol route of infection. This article has identified differences between laboratories and concludes that some differences are warranted (different laboratory setups, equipment, scientific questions, practices), and some differences can be aligned, e.g., limitations of the dosing length.

1.
Is the bacteriological medium used to grow the bacterial challenge important to standardize?

2.
What is the target challenge dose for the aerosol challenge? Should this be more stringent to down-select candidates earlier? Currently, the targeted dose is one that reproducibly results in lethality in control animals, and there is an acceptance that variation is inherent due to inhalation exposure.