We show a technique for in vivo live bioluminescence and near-infrared imaging of optic neuritis and encephalitis in the experimental autoimmune encephalomyelitis (EAE) model for multiple sclerosis in SJL/J mice.
Experimental autoimmune encephalomyelitis (EAE) in SJL/J mice is a model for relapsing-remitting multiple sclerosis (RRMS). Clinical EAE scores describing motor function deficits are basic readouts of the immune-mediated inflammation of the spinal cord. However, scores and body weight do not allow for an in vivo assessment of brain inflammation and optic neuritis. The latter is an early and frequent manifestation in about 2/3 of MS patients. Here, we show methods for bioluminescence and near-infrared live imaging to assess EAE evoked optic neuritis, brain inflammation, and blood-brain barrier (BBB) disruption in living mice using an in vivo imaging system. A bioluminescent substrate activated by oxidases primarily showed optic neuritis. The signal was specific and allowed the visualization of medication effects and disease time courses, which paralleled the clinical scores. Pegylated fluorescent nanoparticles that remained within the vasculature for extended periods of time were used to assess the BBB integrity. Near-infrared imaging revealed a BBB leak at the peak of the disease. The signal was the strongest around the eyes. A near-infrared substrate for matrix metalloproteinases was used to assess EAE-evoked inflammation. Auto-fluorescence interfered with the signal, requiring spectral unmixing for quantification. Overall, bioluminescence imaging was a reliable method to assess EAE-associated optic neuritis and medication effects and was superior to the near-infrared techniques in terms of signal specificity, robustness, ease of quantification, and cost.
Multiple sclerosis is caused by the autoimmune-mediated attack and destruction of the myelin sheath in the brain and the spinal cord1. With an overall incidence of about 3.6 cases per 100,000 people a year in women and about 2.0 in men, MS is the second most common cause of neurological disability in young adults, after traumatic injuries2,3. The disease pathology is contributed to by genetic and environmental factors4 but is still not completely understood. Autoreactive T lymphocytes enter the central nervous system and trigger an inflammatory cascade that causes focal infiltrates in the white matter of the brain, spinal cord, and optic nerve. In most cases, these infiltrates are initially reversible, but persistence increases with the number of relapses. A number of rodent models have been developed to study the pathology of the disease. The relapsing-remitting EAE in SJL/J mice and the primary-progressive EAE in C57BL6 mice are the most popular models.
The clinical EAE scores, which describe the extent of the motor function deficits, and body weight are the gold standards to assess EAE severity. These clinical signs agree with the extent of immune cell infiltration and myelin destruction in the spinal cord and moderately predict drug treatment efficacy in humans5. However, these signs mainly reflect the destruction of the ventral fiber tracts in the spinal cord. Presently, there is no easy, non-invasive, reliable, and reproducible method to assess in vivo brain infiltration and optic neuritis in living mice.
The in vivo imaging agrees with the 3 “R” principles of Russel and Burch (1959), which claim a Replacement, Reduction, and Refinement of animal experiments6, because imaging increases the readouts of one animal at several time points and allows for a reduction of the overall numbers. Presently, inflammation or myelin status is mainly assessed ex vivo via immunohistochemistry, FACS-analysis, or different molecular biological methods7, all requiring euthanized mice at specific time points.
A number of in vivo imaging system probes have been developed to assess inflammation in the skin, joints, and vascular system. The techniques rely on the activation of bioluminescent or near-infrared fluorescent substrates by tissue peroxidases, including myeloperoxidase (MPO), matrix metalloproteinases (MMPs)8, and cathepsins9 or cyclooxygenase2. These probes have been mainly validated in models of arthritis or atherosclerosis9,10. A cathepsin-sensitive probe has also been used for fluorescence molecular tomographic imaging of EAE11. MMPs, particularly MMP2 and MMP9, contribute to the protease-mediated BBB disruption in EAE and are upregulated at sites of immune cell infiltration12, suggesting that these probes may be useful for EAE imaging. The same holds true for peroxidase or cathepsin-based probes. Technically, imaging of inflammation in the brain or spinal cord is substantially more challenging because the skull or spine absorb bioluminescent and near-infrared signals.
In addition to inflammation indicators, fluorescent chemicals have been described, which specifically bind to myelin and may allow for quantification of myelination13. A near-infrared fluorescent probe, 3,3′-diethylthiatricarbocyanine iodide (DBT), was found to specifically bind to myelinated fibers and was validated as a quantitative tool in mouse models of primary myelination defects and in cuprizone-evoked demyelination14. In EAE, the DBT signal was rather increased, reflecting the inflammation of the myelin fibers5.
An additional hallmark of EAE and MS is the BBB breakdown, resulting in increased vascular permeability and the extravasation of blood cells, extracellular fluid, and macromolecules into the CNS parenchyma. This can lead to edema, inflammation, oligodendrocyte damage, and, eventually, demyelination15,16. Hence, visualization of the BBB leak using fluorescent probes, such as fluorochrome-labeled bovine serum albumin5, which normally distribute very slowly from blood to tissue, may be useful to assess EAE.
In the present study, we have assessed the usefulness of different probes in EAE and show the procedure for the most reliable and robust bioluminescent technique. In addition, we discuss the pros and cons of near-infrared probes for MMP activity and BBB integrity.
1. EAE Induction in SJL/J Mice
2. Bioluminescent and Near-infrared Imaging of Optic Neuritis and Brain Inflammation
3. Image Analyses
Time Course of Bioluminescence of Optic Neuritis
The bioluminescence signal of the inflammation probe was the strongest around the eyes and occurred exclusively in EAE mice with optic neuritis. A signal occurred in neither the non-EAE mice nor the mice not injected with the inflammation probe. The signal disappeared when the mice recovered. Hence, the signal is specific for optic neuritis, and the peak of the signal parallels the peak of the clinical EAE scores. Figure 1 shows two examples of SJL/J mice imaged at day 10 and 14 after EAE induction. The bioluminescent signal was the highest on day 10 and disappeared when the mice started to recover. The time courses of the clinical scores matched the disappearance of the bioluminescent signal (example-1) or preceded (example-2) the decline of the scores.
Figure 1. Time Course of Optic Neuritis and Clinical Scores in EAS-SJL/J Mice. Bioluminescent images of optic neuritis were captured 10 min after the injection of the inflammation probe (100 µL i.p.) during the 1st flare of the disease in two SJL/J mice at different time points after immunization. The bioluminescent images (left panel) were captured 10 and 14 days after immunization and are presented as rainbow pseudocolor photograph overlays. Lut bars range from blue (low-) to red (high-bioluminescence). Scale bar: 1 cm. The right panel shows bar charts of the individual total bioluminescence counts in ROIs (mean ± sd of 3 images each) and the time courses of the clinical EAE scores. The red arrows mark the imaging days. Please click here to view a larger version of this figure.
Assessment of Treatment Efficacy
Inflammation probe
The effects of the medication (R-flurbiprofen 5 mg/kg/d)5 are shown in Figure 2. Five examples of the bioluminescent images in each group are presented. The scores in the vehicle and treatment group were heterogeneous, but in all mice, the bioluminescent signal in the eye showing inflammation in the optic nerve was lower in the medication group (Figure 2A). The quantification of the total bioluminescent counts in ROIs confirmed significant treatment efficacy (Figure 2B, with box plots of the total bioluminescent counts, unpaired 2-tailed t-test, P < 0.05). The imaging results agreed with the therapeutic effects of the medication in terms of the clinical scores and histopathologic manifestations of EAE in the spinal cord and optic nerve5.
Figure 2. Efficacy of Medication in EAE-SJL/J Mice Using Bioluminescent Imaging. SJL/J mice received vehicle or medication (R-flurbiprofen 5 mg/kg/d) from day 5 after immunization. Images were captured after the injection of the inflammation probe (100 µL i.p.) at the 1st EAE peak, n = 10 per group. EAE developed in 7/10 in both groups, and EAE non-responders were without signal. A) Bioluminescence images in living mice captured 5 – 15 min after the i.p. injection of 100 µL of the inflammation probe are presented as rainbow pseudocolor photograph overlays. LUT bars range from blue (low-) to red (high-intensity). Scale bar: 1 cm. The individual peak of the total bioluminescent counts in ROIs was used for quantification. ROIs were restricted to the head. The PLP injection sites were shielded with black cloth. B) Box plots showing the quantification of the total counts in ROIs. The box represents the interquartile range, the line is the median, and the whiskers show the minimum to maximum. The total counts differed significantly between groups (2-sided unpaired t-test, P < 0.05), showing a reduction of optic neuritis and encephalitis in mice receiving medication. Please click here to view a larger version of this figure.
Pegylated fluorescent nanoparticles
Epifluorescence images of near-infrared nanoparticles revealed vascular leaks around the eyes and in the brain in both treatment groups (Figure 3A, 2 examples). The particles distribute very slowly from the blood to the interstitial space, except for sites of inflammation, where the dye accumulates, allowing for an assessment of BBB integrity. The leakage was evident at 3 h and 24 h after the injection of the nanoparticles, but it was stronger at the later time point. There was no specific signal in non-EAE mice or mice not injected with nanoparticles (right panel). Hence, the signal was specific. The quantitative analysis of radiant efficiency in ROIs (Figure 3B) did not reveal significant differences between treatment groups (n = 3 per group, unpaired 2-tailed t-test, P < 0.05).
Figure 3. Assessment of Blood Brain Barrier Disruption with Pegylated Fluorescent Nanoparticles. SJL/J mice received vehicle or medication (R-flurbiprofen 5 mg/kg/d) from day 5 after immunization. Images were captured 3 h and 24 h after the injection of near-infrared-labeled nanoparticles (70 µL i.v.) during the 1st EAE peak, n = 3 per group. EAE non-responders were without signal. The auto-ROI tool was used for the quantification of the radiant efficiency. The ROIs were restricted to the head. The PLP injection sites were shielded. A) Exemplary epifluorescence images of living mice, captured 3 h and 24 h after the injection of nanoparticles. Images of mice without EAE or without the injection of nanoparticles were used as imaging controls. Scale bar: 1 cm. LUT bars range from dark red (low-) to yellow (high-intensity). B) Bar charts showing the quantification of the radiant efficiency in ROIs (mean ± SD). Treatment groups did not significantly differ (2-sided unpaired t-test). Please click here to view a larger version of this figure.
Matrix metalloproteinase-sensitive probe
After subtracting the auto-fluorescence, images of the protease-activatable MMP probe revealed inflammation in the brain and the spinal cord in EAE mice (Figure 4A, examples of 4 mice per group). There was no signal in mice not injected with the probe, and a weak signal occurred in an EAE mouse without clinical symptoms (non-responder). Images in Figure 4 show the signal at Ex/Em 640/700 subtracted by the image at Ex/Em 675/720. Differences between the treatments were revealed only after the quantitative analysis of the radiant efficiency in auto-ROIs after spectral unmixing (Figure 4B, unpaired 2-tailed t-test, n = 6 and 4, P < 0.05).
Figure 4. Assessment of Metalloproteinase Activity with Near-infrared MMP-sensitive Imaging Probe. SJL/J mice received vehicle or medication (R-flurbiprofen 5 mg/kg/d) from day 5 after immunization. Images were captured 24 h after the injection of the probe (150 µL i.v.) at the 1st EAE peak, n = 6 and 4. The PLP injection sites were shielded. EAE non-responders and mice without MMP probe injections were used as controls. Each 2 images were captured at Ex/Em 640/700 nm (specific signal) and 675/720 nm (auto-fluorescence). Using the spectral unmixing tool, the auto-fluorescence was subtracted and subsequently, the auto-ROI tool was used to identify the sites of specific MMP activity. The total radiant efficiency was used for quantification. A) Exemplary epifluorescence images in living mice captured 24 h after probe injection. The images are the result of spectral unmixing (UMX). Scale bar: 1 cm. LUT bars range from dark red (low-) to yellow (high-intensity). B) Box plots showing the quantification of the total radiant efficiency in ROIs. The box represents the interquartile range, the line is the median, and the whiskers show the minimum to maximum. The asterisk indicates a significant difference between groups (2-sided unpaired t-test, P < 0.05). Please click here to view a larger version of this figure.
The present video shows techniques for bioluminescence and near-infrared fluorescence in vivo imaging of EAE in SJL/J mice. We show that bioluminescence imaging using an inflammation-sensitive probe mainly shows optic neuritis, and the quantification agrees with the clinical evaluation of EAE severity and the effects of medication. However, the bioluminescence imaging method was not able to detect inflammation of the lumbar spinal cord, which is a primary site of EAE manifestation17, likely because the signal is absorbed by the spine.
The near-infrared imaging is more sensitive, but at the expense of high interference with auto-fluorescence, which does not occur with bioluminescence. The exposure time of NIR is much shorter (1 – 2 s) as compared to BLI (2 – 5 min), which makes NIR the preferred method for time series with rapidly changing signal intensities.
Critical Steps within the Protocol and Limitations of the Technique
Signals of the immunization site interfere with spinal cord imaging in EAE, but this can be circumvented by locating both injection sites at the base of the tail, which was not inferior to the recommended injection sites (neck and base of the tail). Nevertheless, it is necessary to shield the injection sites with black cloth.
For bioluminescence imaging, it is crucial to capture a time series of images, because the kinetics differ between animals. To avoid biases caused by kinetics, imaging of pairs of control (e.g., vehicle or wild type) and verum (e.g., drug or transgenic) mice is advantageous. According to the manufacturer, the signal should be stable for 30 min. However, in EAE, we observed faster and more transient kinetics, with peaks occurring at 5 – 10 min and a substantial decline at 15 min.
For near-infrared imaging, the manufacturer recommends the Ex/Em setting for a specific probe. Nevertheless, it was useful to run a filter series initially and to always capture images at least at two excitation/emission combinations, which closely match the reported Ex/Em maxima and can later be used for spectral unmixing of unspecific signals.
It is important to use white mice, like SJL/J, because light and fluorescence are highly absorbed by black fur. Black mice need to be cautiously shaved on the head and back one day before imaging. Skin lesions must be avoided, because they will appear as inflammatory spots and interfere with the imaging of the brain or spinal cord. For NIR imaging, the manufacturer recommends the depilation of all mice, even white mice. Even after shaving, the head and back results in black mice were less convincing than with white mice (not shown). For the primary-progressive EAE model in C57BL6 mice, white C57BL6 mice, which are commercially available, may be an alternative. Hairless, immunocompetent SKH1 mice are useful for near-infrared imaging, but not EAE, because these mice have an albino genetic background and do not reliably develop EAE (maximum score: 0.5 – 1). Bioluminescent imaging using the inflammation probe in these mice revealed several inflammatory spots in the skin (not shown) where hair follicles are lost.
NIR imaging of protease activity revealed brain and spinal cord inflammation, but signals were superimposed by auto-fluorescence, requiring spectral unmixing before quantitative analysis. Hence, NIR imaging was less robust than bioluminescence imaging and was more expensive. However, the use of protease activatable probes may be useful for the assessment of drugs that specifically target matrix metalloproteinases.
Advantages and Disadvantages
The different imaging techniques are complimentary and address specific questions. The advantages of bioluminescent imaging are affordability (about 20 Euro/mouse); a lack of auto-bioluminescence, which would interfere with the signal; high specificity; convenient i.p. injection and image analysis; and robustness and reliability. Disadvantages are long exposure times and signal absorption by black fur and bone.
Advantages of NIR are the broader availability of NIR-labeled probes, ease of custom NIR labeling, short exposure time, and high sensitivity. Disadvantages are high costs (50 – 100 Euro/mouse), absorption by fur, strong interference with auto-fluorescence, and the necessity for spectral unmixing and image processing before analysis.
A number of probes are available that detect sites of inflammation because they are activated by pro-inflammatory enzymes that are upregulated at sites of inflammation (e.g., peroxidases or metalloproteinases) due to the infiltration of immune cells. Some of these probes will also detect cancers demonstrating immune cell infiltration into the tumor microenvironment or the release of enzymes by the tumor itself (e.g., MMPs). Probes that accumulate in tissue due to capillary leakage will detect disruptions in the BBB, but also at other sites of inflammation and cancer.
Significance of the Technique with Respect to Existing/Alternative Methods
The combination of "imaging plus clinical sores" was superior to "scores only" for the assessment of the disease status and the detection of medication effects. The bioluminescence signal around the eyes also agrees with previous histopathological studies showing myelin destruction and immune cell infiltrates in the optic nerve5. About 2/3 of MS patients develop episodes of optic neuritis. So far, there are no reliable, non-invasive methods to quantify optic neuritis in living mice except diffusion magnetic resonance imaging (MRI) and optical coherence tomography (OCT), which are technically demanding18. OCT has been introduced into EAE as a method showing retinal changes and atrophy during EAE, suggesting an autoimmune reaction in the optic nerve19.
Compared to the method described in this manuscript, OCT is a higher stressor for the mice, because it requires deep anesthesia. The readout is not a direct visualization of the optic nerve19.
Near-infrared imaging of fluorescent nanoparticles was useful to visualize the disruption of the BBB, which is another hallmark of EAE and MS. Besides MRI20,21, there is no non-invasive method for the in vivo monitoring of BBB integrity. This would be quite useful, because experimental drugs and phyto-medications specifically act by tightening the barrier22,23, which normally hinders excessive lymphocyte recruitment into the CNS24. Preventing leukocyte attachment or transmigration through the BBB and reducing its leakiness is an effective strategy in MS treatment25, and nanoparticle imaging may help to assess efficacy of candidate drugs. So far, the particles are expensive. Intravital microscopy is another technique used to visualize BBB integrity26, but it normally requires long-lasting, deep anesthesia (e.g., with ketamine and xylazine) because of the craniotomy and disallows re-awakening the mice, hence preventing time course analyses. However, intravital microscopy provides high-resolution images at the cellular to subcellular levels, which is not achievable with in vivo imaging.
Future Applications or Directions after Mastering the Technique
In summary, the imaging techniques presented in the present video help to assess the individual courses of the disease and to monitor the effects of medication, which were in part not revealed by clinical scores alone. The techniques agree with the 3 "R" principles of Replacement, Reduction and Refinement in animal experiments and are useful add-on tools in drug research.
The authors have nothing to disclose.
This research was supported by the Deutsche Forschungsgemeinschaft (CRC1039 A3) and the research funding program “Landesoffensive zur Entwicklung wissenschaftlich-ökonomischer Exzellenz” (LOEWE) of the State of Hessen, Research Center for Translational Medicine and Pharmacology TMP and the Else Kröner-Fresenius Foundation (EKFS), Research Training Group Translational Research Innovation – Pharma (TRIP).
AngioSpark-680 | Perkin Elmer, Inc., Waltham, USA | NEV10149 | Imaging probe, pegylated nanoparticles, useful for imaging of blood brain barrier integrity |
MMP-sense 680 | Perkin Elmer, Inc., Waltham, USA | NEV10126 | Imaging probe, activatable by matrix metalloproteinases, useful for imaging of inflammation |
XenoLight RediJect Inflammation Probe | Perkin Elmer, Inc., Waltham, USA | 760535 | Imaging probe, activatable by oxidases, useful for imaging of inflammation |
PLP139-151/CFA emulsion | Hooke Labs, St Lawrence, MA | EK-0123 | EAE induction kit |
Pertussis Toxin | Hooke Labs, St Lawrence, MA | EK-0123 | EAE induction kit |
IVIS Lumina Spectrum | Perkin Elmer, Inc., Waltham, USA | Bioluminescence and Infrared Imaging System | |
LivingImage 4.5 software | Perkin Elmer, Inc., Waltham, USA | CLS136334 | IVIS analysis software |
Isoflurane | Abbott Labs, Illinois, USA | 26675-46-7 | Anaesthetic |