Einstein Contained Aerosol Pulmonizer (ECAP): Improved Biosafety for Multi-Drug Resistant (MDR) and Extensively Drug Resistant (XDR) Mycobacterium tuberculosis Aerosol Infection Studies

Abstract

A new apparatus enhances the biosafety of containment (biosafety level 3 [BSL-3]) and provides experimental reproducibility for aerosol infection experiments with MDR and XDR Mycobacterium tuberculosis. The methods are generally applicable to the study of airborne pathogens.

Introduction

The study of emergent pathogens requires special care as antibiotic resistance and new mechanisms of pathogenesis may be present. While aerosol infection studies are necessary, they carry the greatest potential risk of any laboratory procedure. In the course of our work with Multi-Drug Resistant (MDR-) and Extensively Drug Resistant (XDR-) Mycobacterium tuberculosis, we have developed a new apparatus that enhances the biosafety of aerosolization experiments. This apparatus and the associated protocols were developed and validated for TB; however, they are likely to be of wider use.

Several different systems have been developed to generate aerosol infections in mice, guinea pigs, and rabbits (Druett, 1969; Hartings & Roy, 2004; Henderson, 1952; Louveau et al., 2005; Rosebury, 1947; Schwebach et al., 2002; Smith, 1970; Tsenova et al., 2006). The systems differ in their level of safety, ease of operation, level of required maintenance, adaptability to different animal models, number of animals that can be accommodated simultaneously, and quantitative repeatability of delivered dosage. Smith and colleagues developed the Madison Aerosol Exposure Chamber (MAEC), an excellent stand-alone system that is easy to use, has low maintenance requirements, is quantitatively reproducible, and can be used to infect large numbers of mice, guinea pigs, or rabbits in a single run (Smith, 1970). However, the Institutional Biosafety Committees of the Albert Einstein College of Medicine and the Howard Hughes Medical Institute (HHMI) have raised three serious concerns about MAEC safety: 1) when loading and unloading the inoculum culture to/from the stand-alone MAEC, the culture is open to the lab air—a potential source of contamination to the operators; 2) whole body aerosol-exposed mice, which can harbor spreadable pathogens on their external surfaces, are transiently open in the laboratory air during transfer from the exposure chamber to their housing cages; and 3) the stand-alone MAEC lacks back-up protection against operator exposure due to failure of the door gasket, exhaust air tubing system, or HEPA filter leak. These redundancies are important in BSL-3 safety. Biosafety concerns became most critical in the context of proposed aerosol infection studies with MDR- and XDR- M. tuberculosis. A new apparatus, the Einstein Contained Aerosol Pulmonizer (ECAP), was designed and constructed. The ECAP and its integration into a series of biosafety cabinets (BSCs) connected by airtight tunnels add redundant layers of safety to the stand-alone MAEC. With this new apparatus, the proposed studies have, in fact, been safely completed and others are in progress (Ashiru et al., 2010; Ioerger et al., 2009).

The ECAP joins the MAEC to a Baker Class III ventilated and negative pressure glovebox (Figure 1). The ECAP incorporates the excellent aerosolization protocols developed by Dr. Don Smith and Dr. David McMurray (Smith, 1970). This combined apparatus and its associated protocols for enhancing safe operation and maintenance were integrated into our BSL-3 laboratory (Figure 2A). The glovebox exposure chamber is connected to a Class II Type A2 biosafety cabinet (BSC) by an airlocked tunnel. Mice never leave a BSC as they are moved from pre-exposure housing to the aerosol exposure chamber to post-exposure housing. At appropriate times, mice are sacrificed in a Class II Type A2 BSC, then transferred through an airlock into another Class II Type A2 BSC where organs are dissected, homogenized, and spread on media for bacterial counts. Safety designs, implementations, and animal protocols were comprehensively reviewed and approved by both the Einstein and HHMI safety offices.

Figure 1.

A. Madison Aerosol Exposure Chamber (AEC) opened showing carousel with cylindrical mouse-cages.

B. Einstein Contained Aerosol Pulmonizer (ECAP), integrates a Madison Aerosol Exposure Chamber (MAEC) with a Baker Class III ventilated glovebox, and a control box.

Figure 2A.

Coupled hood series in the BSL-3 suite. The arrow indicates the direction of airflow between the rooms. The BSL-3/4 aerosol room has the most negative air pressure in the suite. A. Madison AEC; B. Baker Class III glovebox; C. animal changing hood; D. airtight wall tunnel; E. doors between rooms; F. BSL-3 suite room partition walls.

The traditional stand-alone MAEC diameter is 2′ wide and 3′ long. During operation, an 18-cage carousel is placed inside of the MAEC and the unit’s front door is locked. Each of the individual cages in the carousel is a perforated cylinder, 3″ in diameter and 8.5″ long; up to five mice can be exposed in each cage (Figure 1A). To create the new integrated assembly, the MAEC was joined to a Baker Class III ventilated glovebox (Figures 1B, 2A), so that the inside of the chamber could be accessed only through the glovebox. This results in approximately a 3.5′ working distance inside the ventilated and negative pressure glovebox. Loading and unloading of caged experimental animals, aerating the cages and the animals, and setting up the nebulizer are all performed within the ventilated glovebox. The opposite side of the ventilated glovebox is joined, via an airtight wall tunnel, to a Class II Type A2 BSC (Figure 2A) in the animal holding area. A knife-edge gasket was substituted for the flat gasket of the MAEC chamber to better seal its door. The Collison nebulizer (BGI, Inc., Waltham, MA) mounted on the airtight door of the MAEC introduces the aerosolized infectious agent. The nebulizer is activated by HEPA-filtered compressed air. Secondary airflow from the ambient air inlet tube is drawn from the HEPA-filtered air already in the glovebox. All MAEC air is exhausted by a vacuum pump via two HEPA filters in series. This MAEC exhaust is HEPA-filtered as it vents into the glovebox exhaust system that is also HEPA-filtered. This system of two separate, successive filters provides an additional layer of safety and operator protection. As is normal inside BSL-3 facilities, air pressure is lower than in the outside corridor. Inside the BSL-3 facility, the aerosol delivery system is inside a smaller room (section 4) that has lower air pressure than the rest of the BSL-3 area (BSL-3/4). Air pressure is lowest of all inside the aerosol exposure chamber.

The internal MAEC and the glovebox were individually tested and each passed a pressure decay test, holding an internal overpressure against a 3″ water column for 30′. The apparatus was then tested biologically both for containment and for the reproducibility of mouse exposures. The non-pathogenic species M. smegmatis mc²155 (Jacobs et al., 1987; Snapper et al., 1990) as well as three different strains of fully drug-sensitive and virulent M. tuberculosis—H37Rv, CDC1551, and Erdman—were used for tests. In brief, mycobacterial strains were prepared as previously described (Schwebach et al., 2002). For biocontainment tests, four sentinel plates were placed on top of the ventilated glovebox during operation, four around the tunnel connecting the Class III glovebox to the Class II Type A2 cabinet, and four inside the ventilated glovebox immediately after mice were removed post-aerosol exposure. In each set of four plates, two contained only Middlebrook 7H10 (Difco, Detroit, MI), 0.5% glycerol, 10% oleic acid-albumin-dextrose-catalase (OADC) (Becton Dickinson, Sparks, MD), and the other two were supplemented with cycloheximide at 100 ug/ml. Plates were left open overnight, then closed, incubated at 37°C for 6 weeks, and carefully examined for M. tuberculosis colonies. For M. smegmatis experiments, plates were read after 7 days.

For infection studies, female C57BL/6 mice were exposed to aerosolized mycobacteria for 20 minutes, followed by a clean air purge of 30 minutes inside the ECAP and then an additional 30 minutes outside the exposure chamber but inside the glovebox. Infected mice were sacrificed 3–4 hours for M. smegmatis and 24 hours post-aerosol infection for M. tuberculosis. Lungs were dissected, removed, placed in a stomacher bag containing 4.5 mls of PBS-0.05% Tween, and then homogenized using the stomacher apparatus (Seward Lab Systems, Northampton, UK) at high speed for 4 minutes. Serial dilutions were plated on Middlebrook 7H10 plates (Difco, Detroit, MI), supplemented with 0.5% glycerol, 10% oleic acid-albumin-dextrose-catalase (OADC) (Becton Dickinson, Sparks, MD), and incubated for 3 weeks at 37°C before M. tuberculosis colonies were counted. Figure 2B shows the results of M. smegmatis mc²155 at four different infection dosages, with four groups of five mice for each dosage level. The consistency of dosage in each of the four groups is shown by the (small) standard deviation at each input dosage level. Table 1 compiles the results of experiments with M. tuberculosis CDC1551, Erdman, and H37Rv at different input dosages with groups of five mice for each dosage level; four dosage levels were used for Erdman and H37Rv and three levels for CDC1551. Again, the consistency of dosage is manifest in the (small) standard deviation at each input dosage level. Day-to-day variation was also small (data not shown). Mice were sacrificed and dissected. The lungs were homogenized and the cells diluted and plated. Plates were incubated at 37°C for 3–7 days for M. smegmatis or 3–4 weeks for M. tuberculosis and colonies counted in Class II Type A2 biosafety cabinets within the BSL-3 facility.

Figure 2B.

M. smegmatis was prepared at 4 different concentrations based on Optical Density (OD₆₀₀) readings before aerosol delivery, using C57BL/6 mice from NCI. We used 5 mice for each of 4 aerosol delivery runs, with input dosages at the level of n = 5, 6, 7, 8 (log increase). This figure shows a direct correlation between the input dosage and total bacterial load in lungs (measured in CFUs) measured at 24 hours post exposure. Error bars show standard deviation from CFUs recovered from 5 mice in each group.

Table 1.

Dose estimate at time of infection as determined by Optical Density (OD₆₀₀), input dosage as determined by Colony Forming Units (CFU; because of the slow growth of M. tuberculosis, colony counts become available only one month after exposure), and bacterial loads in mice infected with 3 different strains of virulent M. tuberculosis.

CDC1551	OD*/ml	CFU**/ml	avg/lung	std dev
5	***	***	***	***
6	8×105	2×106	107	7
7	8×106	1×107	588	58
8	8×107	1×108	3146	381

Erdman	OD/ml	CFU/ml	avg/lung	std dev
5	1×105	1×104	28	40
6	1×106	4×104	60	36
7	1×107	2×106	650	84
8	1×108	3×108	2383	801

H37Rv	OD/ml	CFU/ml	avg/lung	std dev
5	3×105	3×104	8	6
6	3×106	2×105	13	19
7	3×107	7×106	203	56
8	6×107	1×107	690	257

Note: For each of the four strains tested, and for each of the four dosage levels (5, 6, 7, 8), absolute values are shown for OD dose, CFU dose, average bacterial load recovered from lungs, and standard deviation of that load. For each strain at each dosage level, 5 mice were used, for a total of 75 mice in all.

^*The input concentration of M. tuberculosis for each experiment was set via OD₆₀₀; CFUs were estimated using the heuristic that 1 OD₆₀₀ represents 3 × 10⁸ CFU/ml.

^**At the same time, a sample of the input culture was also diluted and plated for actual CFUs that were counted 3–4 weeks later.

^***This dosage level was not used for the CDC1551 strain.

Over 100 additional experimental studies, consisting of more than 250 separate aerosol runs with over 7,000 mice, further confirmed the ability of the ECAP to deliver the intended dosage levels both accurately and consistently. More than 10 different strains of mice and over 15 different mycobacterial strains were used, including MDR- (Figure 3) and XDR-M. tuberculosis (data not shown). Sentinel plates showed no untoward microbes on any part of or around the integrated apparatus and no releases of contaminated aerosol or exposure of laboratory personnel has been detected throughout these experiments, which were conducted over a period of five years at the Albert Einstein College of Medicine. The ECAP allows up to 90 mice per run, with the same consistent dosage as the stand-alone MAEC. The ECAP design has been retrofitted into existing BSL-3 facilities and incorporated into new ones, notably the Human Vaccine Institute at Duke University and the K-RITH facility, a collaboration between the Howard Hughes Medical Institute and the University of Kwazulu-Natal in South Africa. Since the design and manufacture of this prototype, The Baker Company has sold several units.

Figure 3.

Course of infection of C57Bl/6 mice by drug-susceptible (V4207) and multi-drug-resistant (V2475) M. tuberculosis KZN strains. C57Bl/6 mice were infected with a low-dose (LD, 10 ± 0 CFU/lung bacterial load at 24 hr post-aerosol infection) and a high dose (HD, 580 ± 150 CFU/lung bacterial load at 24 hr post-aerosol infection) of V4207, as well as with a low-dose (22 ± 3 CFU/lung bacterial load at 24 hr post-aerosol infection) of V2475. Bacterial loads in the lungs (4 mice per group) were assessed at days 1, 7, 14, 21, and 28, and plotted with standard deviations.