Reporter enzyme fluorescence (REF) imaging and quantification of tuberculosis in live animals

Abstract

The slow growth rate of Mycobacterium tuberculosis hinders research progress, since estimating the bacterial numbers present in all experiments normally relies on determination of colony forming units on agar plates. M. tuberculosis colonies can take as long as four to six weeks to become visible. Whole animal imaging is an emerging technology that has broad applications in all areas of biological sciences, including monitoring infections. Imaging allows bacterial numbers to be determined in real-time for each infected animal, individually, which allows inter-animal variability to be observed and controlled for. Reporter enzyme fluorescence (REF) utilizes custom substrates that allow production of a fluorescent product after cleavage by a bacterial enzyme. In our recently published studies, we demonstrate that the enzyme β-lactamase, a naturally occurring enzyme expressed by M. tuberculosis, can be used for REF. The resulting imaging system is the first that allows non-invasive detection of natural M. tuberculosis strains directly in pulmonary infected living animals. Use of REF for M. tuberculosis infected mice allows detection of ~10⁴ CFU in the lungs, which is very sensitive. This system also displays promise for allowing rapid evaluation of differences in virulence strains and efficacy of therapeutics and vaccines. This system could be developed into a diagnostic tool for tuberculosis through the use of REF to identify infected tissues or other diagnostic specimens.

This article refers to:

Tuberculosis remains a major global public health problem with one third of the world's population infected and 1.8 million deaths in 2008.1 More effective vaccines, anti-tuberculosis drugs, and more efficient diagnostic technologies are urgently needed. Mycobacterium tuberculosis, the causative agent of tuberculosis, grows very slowly, taking several weeks to over a month to obtain visible colonies on agar plates. This problem impacts the pace of virulence studies, evaluation of therapeutics and development of vaccines, making rapid methods for quantification of bacteria under laboratory conditions very valuable.

Non-invasive real-time imaging technologies can facilitate tuberculosis research by providing the ability to monitor the disease process in live animals. Bacterial numbers present and the tissues infected can be determined through imaging, since the signal level correlates well with bacterial numbers. As a result, quantification of imaging signal can serve as a surrogate for determination of colony forming units on agar plates. Clearly, before imaging alone can be used for this purpose, the approach must be well validated and the correlation between signal and bacterial numbers must be accurate and consistent. Ultimately, however, we expect that imaging will be relied upon more and more within the infectious diseases field to provide quantitative data regarding infections in animals. This is particularly likely in slow-growing microorganisms, such as mycobacteria, where the need to determine bacterial numbers delays progress. Imaging systems based on bioluminescence and fluorescence have been successfully used for real-time evaluation of vaccines,2,3 monitoring of tumorigenesis4 and to follow infection for Escherichia coli,5 Salmonella typhimurium,6 Listeria monocytogenes,7 pulmonary and systemic Streptococcus pneumoniae8,9 and Staphylococcus aureus.10 In addition, our own studies applied fluorescence imaging to quantification and tissue localization of tuberculosis in mice.

Imaging with Reporter Enzyme Fluorescence

A number of recombinant mycobacterial reporter strains have been developed for detection of mycobacteria using luminescence and fluorescence.11–15 These approaches require specific laboratory strains where the appropriate reporter gene has been introduced. We wished to develop an imaging system that did not require recombinant strains, since expression of a foreign gene can impact bacterial fitness in unexpected ways, particularly when expressed from plasmids.16–18 Although use of reporter systems that allow fluorescence or bioluminescence imaging have greatly facilitated progress, one of the main advantages of REF for imaging is that it does not require recombinant strains. The ability to detect all M. tuberculosis strains either under laboratory conditions or during infections in animals is likely to have a profound impact on the field. This characteristic also makes it possible that ultimately REF could be used to image infections directly in humans.

Our recent studies describe development of a novel fluorescence based in vivo imaging system, designated Reporter Enzyme Fluorescence (REF), to track and quantify M. tuberculosis infection in live mice.19 This system uses a naturally produced M. tuberculosis enzyme, β-lactamase and custom-designed fluorogenic substrates that generate fluorescent products when cleaved by β-lactamase. The substrates are composed of a fluorochrome and quencher connected by a β-lactam ring. The custom substrates for REF also carry acetylated D-glucosamine connected to the carboxylate of cysteine through γ-amino-butyric acid producing a charged molecule, which facilitates uptake into eukaryotic cells.20 Once this substrate is within cells, β-lactamase secreted by intracellular bacteria cleaves the β-lactam ring. Once the β-lactam ring is hydrolyzed by β-lactamase, the fluorochrome moiety is released from the quencher and generates a near infrared fluorescent molecule (Fig. 1A). Since the custom made fluorogenic molecule uses Cy5.5 as its dye, the resulting fluorescence displays similar characteristics to those of the dye. The fact that mycobacteria persist and replicate intracellularly during infections makes it necessary for the substrate to penetrate mammalian cells, but the fluorescent product does not need to be membrane permeable and it may actually be advantageous for the dye to be retained within the cell. These characteristics result in REF displaying very high sensitivity because the Cy5.5 dye builds up within infected cells over time (Figs. 1C and 2). Using this imaging system, we successfully detected ∼10⁴ CFU of M. tuberculosis in lungs of living mice, and we were also able to examine the efficacy of anti-tuberculosis therapy in real-time. These studies allowed us to better understand the characteristics necessary to produce an optimal substrate for imaging in live animals and provide the foundation for design of improved substrates for REF.

New Applications for REF in Tuberculosis Research

The high sensitivity of REF and its ability to be used directly on clinical strains of M. tuberculosis makes it a useful tool for a number of applications, in addition to whole body imaging of disease processes. Since REF results in fluorescence within host cells that are infected by mycobacteria, it allows specific identification of host cell specificity in situ. This may be important in tuberculosis research because it is well known that M. tuberculosis can infect nearly any cell type but is mostly observed within macrophages where a great deal of its replication is thought to occur during infections. However, it is possible that transient infection of other cell types occurs and may play important roles in different aspects of pathogenesis, such as dissemination or latency or may be involved in sequestration of therapeutic survivors in a privileged niche that is less accessible to antimicrobials.

Quantification of M. tuberculosis bacilli in vitro can be carried out using REF, whether the bacteria are intracellular or extracellular. This fact allows use of REF to study efficacy of therapeutics in laboratory culture or when the bacteria are growing intracellularly. In combination with fluorescence-activated cell sorting (FACS) or confocal microscopy, both infected and uninfected host cells can be quantified using the REF fluorophore as the marker for infected cells and a secondary marker for all cells. This same approach can also be applied to homogenized tissues that are obtained from infected animals that have been given the REF substrate prior to sacrifice to identify each of the cell subsets that are infected at different times. In combination with cryosectioning, histopathology staining and fluorescent confocal microscopy, the infected cells and or tissues can be visualized directly to provide the dynamic relationship between the infected cells and the tissues or organ that they are found within during infections. The ability to follow the bacteria as well as the cell type infected throughout the course of tuberculosis pathogenesis could provide the temporal and spatial characteristics of host tissue invasion during dissemination, a poorly understood aspect of the disease. In addition, the impact of vaccination or treatment with antimicrobials on the location and number of bacilli in different tissues and organs can be monitored in real-time. This information can, therefore, provide a rapid method to screen vaccine and therapeutic candidates, facilitating development of new strategies to combat tuberculosis.

One of the major problems in animal studies for infectious diseases in general, but also in the tuberculosis field, is the variability observed between laboratory animals, even using inbred strains. Our experience imaging different bacterial species in animals suggests that part of this variability is due to variability in actual results from inoculation via different methods. However, methods such as the use of aerosol chambers where multiple animals are infected under the same conditions at the same time control many of the potential sources of variability and we have confirmed this fact by imaging subsequent to infection using this method. However, even in this case, despite the homogeneity of inoculation, differences of as much as two- to five-fold can be observed in colonization, cytokine response, clearance and growth within different tissues and organs between animals. One advantage of imaging with REF is that individual animals can be followed over time, allowing outlier animals to be identified and followed separately, with growth within organs/tissues followed internally to each animal, rather than as the mean for a group that may not be homogeneous. Following differences in virulence in this manner is likely to provide new information regarding tissue distribution and virulence that were not previously possible or anticipated and may well be more sensitive than handling animals as groups.

Since REF allows sensitive detection of M. tuberculosis under both in vitro and in vivo conditions, it offers potential for development as a diagnostic tool for sputum or other diagnostic specimens from patients suspected to have tuberculosis. Current acid-fast smear technology allows detection of 5,000–10,000 bacilli per milliliter in sputum,21–23 a number of bacteria that is only found in patients that can transmit infection to another person. REF allows detection of as few as 100 bacilli under laboratory conditions within only a few hours after incubation in the presence of the substrate, suggesting that it could be more sensitive than acid-fast smear and similar in sensitivity to culture, but offering a result in a shorter time period than culture.23 The high sensitivity of the REF system is most likely due to the high catalytic activity of the β-lactamase enzyme,24,25 in combination with the low background of the custom substrates used. The list of applications of REF are likely to grow as more laboratories validate the approach using different in vitro and in vivo assays.

Future Strategies to Improve REF

The search continues for appropriate enzymes with high catalytic activity where appropriate substrates can be designed for development of additional REF imaging systems for other infectious agents. In the case of the current REF system in tuberculosis, efforts are ongoing to develop substrates that are specific to the M. tuberculosis β-lactamase. This is necessary because of the potential that other bacterial species that produce β-lactamase could interfere with detection of tuberculosis in diagnostic specimens or patients. We have the advantage of a crystal structure for the tuberculosis β-lactamase, BlaC, which has allowed us to identify differences between the M. tuberculosis enzyme and other common β-lactamases26 that can be utilized to design specific substrates. It should also be possible to utilize the available enzyme structures and kinetic information for β-lactams to design substrates with improved enzyme kinetics. These approaches offer promise for development of improved REF systems for tuberculosis that will be more sensitive and specific. We hope that studies with the tuberculosis REF system will serve as a model for development of similar imaging and detection systems for other pathogenic organisms.

We have demonstrated that REF is effective for small animal imaging, but even greater sensitivity is needed to apply this type of imaging to humans due to the greater tissue depth that will cause increased light scattering and signal loss. It is possible that PET or SPECT probes could be developed using similar substrates to those used for REF, but optical imaging is much less expensive, though more challenging to implement. Pulmonary tissue may offer advantages for optical imaging in larger mammals due to the reduced signal loss of one log for every 3 cm of lung as compared to one log for every 1 cm of muscle.27 The current substrates can be readily improved through the use of probes that use longer wavelengths close to or within the near infrared window (650–900 nm), where hemoglobin and water will absorb the smallest portion of the total signal.28 Improving the catalytic activity of the substrates used as well as the tissue distribution should dramatically impact the maximal signal intensity for this system. In addition, the use of microwatt laser sources where the excitation light can penetrate from 4–7 cm of deep muscle29 offers the potential of exciting these sensitive signal sources. The combination of more sensitive substrates with improved excitation and detection methods may allow fluorescence molecular tomography to accurately quantify and detect NIR signal in deep tissue.30 REF offers an opportunity to continue improving technologies for optical imaging of infectious diseases that are inexpensive, noninvasive and extremely sensitive.

Figures and Tables

Figure 1 Description of the reporter enzyme fluorescence (REF) approach to imaging bacterial infections. (A) The reaction that generates RE F signal is catalyzed by a bacterial enzyme, preferably naturally produced. In this case, the bacteria is Mycobacterium tuberculosis (Mtb) and the enzyme is β-lactamase. A custom substrate is used that either is based on fluorescence resonance energy transfer (FRET) quenching of the fluorophore or uses a structure that is not fluorescent until cleavage. (B) The structure of CNIR5, one of the primary substrates used successfully for REF. The β-lactam ring indicated is cleaved by β-lactamase and separates the fluorescent dye, in this case Cy5.5, from the quencher (QSY22) that absorbs light at the Cy5.5 wavelength of emission (∼690 nm). Modifications that impact catalytic activity and specificity can include changes to any component of this structure, but those substitutions surrounding the β-lactam ring have the greatest impact. (C) Proposed model that outlines the high sensitivity of REF. The substrate (CNIR5) is cleaved to a product by intracellular Mtb. The product (Cy5.5) is retained within the host cell and builds up to very high levels until substrate is no longer available.

Figure 2 Observed fluorescent product from REF accumulating in host macrophages over time post-infection with mycobacteria. Host cell (J774A.1 murine macrophages) nuclei were stained with DAPI (blue), mycobacteria were expressing GFP (green) and the CNIR5 fluorescent dye is Cy5.5 (red). Note the increase in product (red) fluorescence between four and 24 h post-infection with mycobacteria. This confocal microscopy supports the model in Figure 1. A pool of fluorescent product is visible within infected cells that is sufficient for fluorescence-activated cell sorting (FACS) and confocal microscopy, explaining the high levels of signal observed over time, even when animals are infected with low bacterial numbers.

Acknowledgements

This work was funded by grant 48523 from the Bill & Melinda Gates Foundation and grant AI47866 from the National Institutes of Health.

Addendum to: