Multi-modality imaging is a valuable approach for studying bacterial colonization in small animal models. This protocol outlines infection of mice with bioluminescent Citrobacter rodentium and the longitudinal monitoring of bacterial colonization using composite 3D diffuse light imaging tomography with μCT imaging to create a 4D movie of C. rodentium infection.
This protocol outlines the steps required to longitudinally monitor a bioluminescent bacterial infection using composite 3D diffuse light imaging tomography with integrated μCT (DLIT-μCT) and the subsequent use of this data to generate a four dimensional (4D) movie of the infection cycle. To develop the 4D infection movies and to validate the DLIT-μCT imaging for bacterial infection studies using an IVIS Spectrum CT, we used infection with bioluminescent C. rodentium, which causes self-limiting colitis in mice. In this protocol, we outline the infection of mice with bioluminescent C. rodentium and non-invasive monitoring of colonization by daily DLIT-μCT imaging and bacterial enumeration from feces for 8 days.
The use of the IVIS Spectrum CT facilitates seamless co-registration of optical and μCT scans using a single imaging platform. The low dose μCT modality enables the imaging of mice at multiple time points during infection, providing detailed anatomical localization of bioluminescent bacterial foci in 3D without causing artifacts from the cumulative radiation. Importantly, the 4D movies of infected mice provide a powerful analytical tool to monitor bacterial colonization dynamics in vivo.
Small animal models, in particular those utilizing mice, are routinely used to investigate bacterial pathogenesis or to test intervention strategies for infections, such as antibiotics, probiotics, prebiotics and vaccines1-7. The main experimental readouts from small animal infections are pathogen load, spatial and temporal localization of the infection, and changes to the immune response of the infected organism. In vivo optical imaging is a valuable tool for infectious disease research and can be used to monitor multiple experimental readouts through the use of reporter genes (luciferase, fluorescent proteins, beta-lactamase, etc), fluorescent dyes, nanoparticles or chemiluminescent probes targeted to a protein, biological process, or microorganism6.
Bioluminescence imaging (BLI) is an optical imaging modality used to monitor the colonization of small animals, such as mice and rats, by pathogenic bacteria3,6,8,9. Mice are infected with recombinant bacteria expressing a luciferase, such as the luxCDABE operon from Photorhabdus luminescens. These bacteria can then be detected through their light production using a CCD based in vivo imaging system3,6,9. Importantly, only metabolically active micro-organisms are bioluminescent (BL), meaning only viable bacterial cells are detected by this methodology 10,11. Using 2D BLI, the location of the BL source is inferred from the surface of the animal where the signal is emitted8. The exact anatomical localization of the BL foci in vivo has to be determined through the ex vivo analysis of organs3,6,9 In contrast, composite 3D diffuse light imaging tomography (DLIT) can be used to compile a quantitative 3D reconstruction of the BL source12. DLIT is performed by collecting BL images taken using defined narrow band-pass optical filters and subsequently inputting them into a diffuse optical tomography 3D reconstruction algorithm1,7,12,13.
Currently, multi-modality imaging is the only methodology available to get true non-invasive anatomical localization of bioluminescent foci in vivo without the need for ex vivo analysis. Recently, we used a combination of DLIT co-registered with μCT imaging to evaluate Citrobacter rodentium (C. rodentium) colonization dynamics following prophylactic treatment with a probiotic bacterium7. C. rodentium is a murine specific enteric pathogen used to model human infection with enteropathogenic and enterhemorrhagic Escherichia coli14. C. rodentium infection causes colitis, typically associated with mild weight loss, diarrhea, polarized Th1 immune response and distinct pathological changes, including colonic crypt hyperplasia and attaching and effacing lesion formation14. In addition to this, C. rodentium pathogenesis has been thoroughly studied using BLI and its colonization dynamics in C57BL/6J mice are well documented, making this bacterium an ideal model microorganism for use with multi-modality imaging3,4,7.
This protocol is the first to outline a methodology for integrated DLIT-μCT imaging of a bacterial infection using a single multimodality imaging platform, the IVIS Spectrum CT, and the generation of a 4D movie showing the true dynamics of this infection non-invasively.
The 4D movie of bacterial infection provides a useful tool to visualize and interpret large amounts of multi-modality imaging data quickly and easily. This technique facilitates the detailed analysis of how an infection spreads through an individual mouse and can be used to investigate how deletion of host or bacterial genes or particular intervention strategies effect bacterial load, distribution, and localization during a longitudinal study7. These videos also provide useful teaching aids and a means of disseminating information to the public.
There are several critical steps in this protocol that could affect the quality of the data obtained from DLIT-μCT imaging and the ability to compile a 4D video of infection. The most important part of this protocol is the successful and homogenous infection of mice with C. rodentium. It is essential that the mice used for the study are between 18-20 g and that the bacterial inoculums are freshly prepared and approximately 5 x 109 cfu, as described previously2,3. Prior to infection of the mice it is important to check that the inoculum is bioluminescent using the Spectrum CT and once the inoculum has been prepared, it must be continually homogenized before each mouse is gavaged to ensure that the mice receive similar infectious doses. The DLIT-μCT imaging of mice has been optimized so that the auto exposure function in Living Image 4.3.1 software automatically determines the optimized imaging parameters for the signal to be well above the noise. However, the auto exposure function relies on user defined settings and parameters which need to be modified as described in the procedure. Failure to do this will result in poor images with a low number of photons collected that do not result in an obvious progression in the infection, as the Spectrum CT’s factory settings for autoexposure are programmed for imaging tumors expressing firefly luciferase. Reconstructions performed using 560-620 nm give the best agreement between simulated and measured data and, therefore, are the more reliable data to include in the reconstruction.
A limitation to the use of DLIT-μCT is that ionizing radiation from the μCT scan causes sub-lethal radiation damage that is cumulative over a longitudinal study18. Sub-lethal radiation exposure can weaken the immune response, cause DNA damage, and apoptosis in internal organs19. Ultimately, cumulative sub-lethal radiation damage can cause death if the LD50/30 for ionizing radiation is exceeded, which is between 5 to 7 Gy depending on the mouse strain and age of the mice used18,20,21. Although some of the molecular damage from ionizing radiation can heal, since the overarching principle is to estimate dose conservatively, this is not typically accounted for in study planning. Instead, the aim is to stay as far below these limits as possible while still accomplishing the study goals. This is particularly important in this study because of the normal immune response to the infection, the frequency of imaging, and the fact that transgenic, immuno-comprised, or heavily infected animals may be more susceptible to ionizing radiation.
When planning the experiment to generate a 4D movie of infection, it is important to consider the length of the experiment, the number of μCT scans required during this period and the LD50/30 for ionizing radiation for the mouse strain being used. Another potential limitation to DLIT-μCT is the strength of the reporter expression within the bacterial strain being used, as this will affect bacterial detection limits and imaging times. It is highly recommended that researchers use validated bacterial strains that are fully virulent, but optimized for maximal lux operon expression, as demonstrated previously for BLI2,3.
One caveat to the current design of the 4D imaging is that each movie is comprised of individual DLIT-μCT scans which have different photon scaling. This can make the images difficult to interpret if the changes to the localization of the BL foci, or its intensity are subtle, or if there is one intense BL focus surrounded by multiple weak foci. Therefore, for longitudinal visualizations, it is important to keep the color bars consistent across the time points.
The concept of a 4D movie of infection can be applied to any suitably labeled bacterial pathogen. Future development of this technique will aim to use fluorescence imaging tomography (FLIT) as well as DLIT to facilitate the investigation of host responses to infection using a combination of bioluminescent bacterial pathogens and injectable fluorescent near infrared probes to investigate host responses to infection. In addition to this, in this protocol we only describe the use of bioluminescent bacteria to create 4D movies of infection. However, in some instances it may be necessary to use fluorescent labeled bacteria, for example tagged with iRFP, so that the bioluminescence reporter can be used for investigating host genetics during infection. Importantly, the use of multi-modality imaging combining DLIT/FLIT-μCT will allow us to non-invasively investigate multiple parameters during a bacterial infection, which will contribute significantly to the reduction, refinement, and replacement of the use of animals in scientific research as outlined in the NC3R’s initiative (http://www.nc3rs.org.uk/).
The authors have nothing to disclose.
The in vivo imaging facility at Imperial College was funded by the MRC.
Name of Reagent/Material | Company | Catalog Number | Comments |
Bioluminescent C. rodentium | Frankel lab | ICC180 | Wiles et al., 2004 |
Veet | Boots | Optimal depilation time is 7 min. Depilation works better if the cream is rubbed in well. | |
Isofluorane (100% v/v) | Abbott | B506 | |
Medical Oxygen | BOC Medical | Size F Cylinder. Note: an appropriate regulator is required. | |
Luria Bertani broth | Merck | 1.10285.0500 | 25 g in 1L Demineralised water. |
Luria Bertani agar | Merck | 1.10283.0500 | 37 g in 1L Demineralised water. |
Kanamycin sulphate | Sigma (Fluka) | 60615 | |
50 ml Polypropylene conical Falcon tubes | BD (Falcon) | 352070 | |
Universals | Corning (Gosselin) | E5633-063 | |
1 ml syringe | BD (Plastipak) | 300013 | |
Oral dosing needle (16G x 75 mm) curved | Vet Tech | DE005 | |
Microbanks (Cryovial) | Pro-Lab Diagnostics | PL.170/Y | |
IVIS Spectrum CT | Caliper- a PerkinElmer Company | 133577 Rev A/ Spectrum CT | |
6kVA UPS | Caliper- a PerkinElmer Company | ||
XGI-8 anesthesia system | Caliper- a PerkinElmer Company | 118918 | |
XAF-8 Anaesthesia filter charcoal | Caliper- a PerkinElmer Company | 118999/00 | |
Living Image v4.3.1 SP1 | Caliper- a PerkinElmer Company | ||
Benchtop shaking incubator | New Brunswick Scientific | Innova 44 |