Intravital imaging is becoming more popular and is being used to visualize cellular motility and functions. In contrast to in vitro analysis, which resembles in vivo analysis, intravital imaging can be used to observe and analyze cells directly in vivo. In this review, I will summarize recent imaging studies of autoreactive T cell infiltration into the central nervous system (CNS) and provide technical background. During their in vivo journey, autoreactive T cells interact with many different cells. At first, autoreactive T cells interact with endothelial cells in the airways of the lung or with splenocytes, where they acquire a migratory phenotype to infiltrate into the CNS. After arriving at the CNS, they interact with endothelial cells of the leptomeningeal vessels or the choroid plexus before passing through the blood–brain barrier. CNS-infiltrating T cells become activated by recognizing endogenous autoantigens presented by local antigen-presenting cells (APCs). This activation was visualized in vivo by using protein-based sensors. One such sensor detects changes in intracellular calcium concentration as an early marker of T cell activation. Another sensor detects translocation of Nuclear factor of activated T-cells (NFAT) from cytosol to nucleus as a definitive sign of T cell activation. Importantly, intravital imaging is not just used to visualize cellular behavior. Together with precise analysis, intravital imaging deepens our knowledge of cellular functions in living organs and also provides a platform for developing therapeutic treatments.
Antigen presentation; Autoimmunity; Central nervous system; Intravital imaging; T cells
APCs, antigen presenting cells; BBB, blood–brain barrier; CFP, cyan fluorescent protein; CFSE, 5-(and-6)-carboxyfluorescein diacetate succinimidyl ester; CMTMR, 5-(and-6)-(((4-chloromethyl)benzoyl)amino)tetramethylrhodamine; CNS, central nervous system; CSF, cerebrospinal fluid; EAE, experimental autoimmune encephalomyelitis; FRET, fluorescence resonance energy transfer; GFP, green fluorescent protein; GRIN, gradient index; MBP, myelin basic protein; MIP, macrophage migration inhibitory factor; MRI, magnetic resonance imaging; N.A., numerical aperture; NFAT, nuclear factor of activated T cells; OVA, ovalbumin; p.t., post transfer; YFP, yellow fluorescent protein
During inflammation, many different types of immune cells migrate to and accumulate in the lesion. Those cells interact with each other and work together. The functions of each cell population have often been studied in vitro after isolation from animals, which has provided valuable information. Under these experimental conditions, cells are cultured and/or stored. However, an in vitro system lacks blood flow and often lacks three-dimensional structure, which resembles environment in the organ. In addition, purified cytokines and growth factors are often added to the culture, which might create different conditions from in vivo, where cells are exposed to a mixture of those factors. Therefore, to understand cellular functions, in vivo analysis should be considered.
In vivo experiments have certain disadvantages, as it is much more complicated and difficult to perform than in vitro experiments. The experimental animals must be kept under physiological conditions in order for proper experiments to be performed. If this is not the case, the results obtained are not accurate. In addition, researchers need to consider how to identify and analyze cells in living animals. In contrast to in vitro experiments in which antibody staining and cell labeling can be easily done, it is more challenging to mark target cells in vivo. Furthermore, the number of cells that can be analyzed in vivo, especially by intravital imaging, is usually much fewer than that in vitro. This means that there is always a risk that intravital imaging detect only special event at special location. Therefore, the best approach is to combine intravital imaging and conventional methods. Intravital imaging provides information regarding cellular functions under physiological conditions, whereas conventional methods provide larger quantities of data regarding cellular status.
Our group is focusing on the infiltration of encephalitogenic T cells into the central nervous system (CNS). We use experimental autoimmune encephalomyelitis (EAE), a widely used animal model for multiple sclerosis,1 which is considered an autoimmune disease in humans. EAE can be induced by active immunization of CNS-specific antigens emulsified in complete Freunds adjuvant (active EAE). Immunized antigen is taken up by dendritic cells and macrophages and presented to CD4+ helper T cells. Thereafter, CD4+ T cells migrate through the body. Alternatively, EAE can be induced by adoptive transfer of myelin antigen-specific T cells (transfer EAE)2 or by using transgenic mice that harbor myelin antigen-specific T cells in high numbers (spontaneous EAE).3 In any case, CNS-infiltrating CD4+ T cells recognize specific antigens presented by local antigen-presenting cells (APCs), and they become activated, produce inflammatory cytokines, and initiate the inflammatory reaction. During inflammation, both innate immune cells (such as macrophages) and adoptive immune cells (such as T and B cells) infiltrate into the CNS and contribute to CNS inflammation. Macrophages have both pro- and anti-inflammatory roles during inflammation.4 and 5 Infiltrated B cells produce antibodies in the cerebrospinal fluid,6 which may either enhance or control inflammation. It was shown that regulatory T cells (Treg) infiltrate into the CNS, although their function there is still largely unknown. Our ultimate goal is to illustrate the functions of and interactions among infiltrating immune cells during CNS inflammation. In this review, we will focus on cellular interactions in EAE, especially by using intravital imaging.
Many different methods of intravital imaging are available. For example, magnetic resonance imaging (MRI) is used for diagnosis of MS patients to detect inflammation. MRI is non-invasive and provides valuable information. However, a conventional MRI machine with a 3T magnetic field does not provide sufficient resolution for single cell imaging.7 Recently, a higher-powered MRI with a 7T magnetic field was introduced that can visualize CNS inflammation with surprisingly high resolution.7 Still, it is not sufficient to see single cells in the CNS. The same holds true for computed tomography. The above methods are very useful for diagnostic use, but not for single-cell imaging.
To achieve single-cell imaging, microscopic techniques are commonly used. In the earliest phase of intravital imaging, leukocytes were imaged in the blood vessels of frogs by using bright-field microscopy (reviewed in8). This opened up new methodologies for allowing the observation of cellular motility directly in vivo. However, the target tissue must be thin and relatively transparent because bright-field techniques are used. Furthermore, cell types are hard to distinguish. The use of fluorescent microscopy allows one to focus on specific cell types after proper labeling (for discussion of labeling, please refer to the next section.). Now researchers can analyze the cells of interest in the living animal. However, fluorescent microscopy can only achieve a relatively small penetration depth. Imaging is thus possible only close to the surface.
It is possible to increase the penetration depth of fluorescent imaging, either by using stronger labeling, objectives with higher numerical aperture (N.A.), or stronger excitation power. The development of confocal microscopy equipped with stronger lasers increased the penetration depth dramatically. Confocal microscopy has better spatial resolution and provides clearer images than does fluorescent microscopy. One disadvantage of confocal microscopy is slow image acquisition because of the need to do line scanning. This can be improved by using spinning disk confocal9 or light sheet microscopy,10 which can perform faster acquisition. Another disadvantage of confocal microscopy is phototoxicity, which is difficult to prevent because fluorochromes are excited by strong laser light. Excitation laser power can be reduced, but the emitted signal becomes weaker.
To increase the penetration depth and reduce phototoxicity, two-photon microscopy was developed.11 Two-photon microscopy can share most equipment parts with confocal microscopy, except the excitation laser. The difference between one-photon microscopy (confocal microscopy) and two-photon microscopy is the mechanism of excitation. One photon excites one fluorescent molecule in confocal microscopy, whereas two photons excite one fluorescent molecule in two-photon microscopy. To achieve this two-photon excitation, high photon density is absolutely necessary.12 Therefore, instead of a continuous confocal laser that emits photons spontaneously, a pulsed two-photon laser can accumulate generated photons and emit them in time intervals.13 As a result, without changing the average laser power, a two-photon laser increases the peak power dramatically. Commonly used commercial two-photon lasers pulse at the frequency of 80 MHz (80 million pulses per second), which can provide a sufficient pulse even during very fast scanning. Two-photon excitation occurs only at the focal point due to the requirement of high photon density. To some extent, excitation of fluorochromes produces oxygen radicals, which induce cellular toxicity. Since fluorochromes are excited only at the focal point in two-photon excitation, two-photon microscopy minimizes phototoxicity. Another advantage of two-photon microscopy is penetration depth. Because two photons excite one fluorescent molecule, each photon contributes only half the amount of energy compared with conventional one-photon excitation. This indicates that two-photon microscopy is equipped with a laser of twice the wavelength than that of confocal microscopy. Because longer-wavelength light has less of a scattering effect in tissues, two-photon microscopy shows higher penetration depth. All of these factors result in two-photon microscopy being an indispensable method for intravital imaging.
It is necessary to use anesthesia to stabilize animal movement. At the same time, animal conditions must be kept as close to physiological as possible during intravital imaging. We use a fentanyl mixture for induction and isoflurane during intravital imaging. Animals are intubated via tracheostomy and connected to a small animal ventilation machine. Then, isoflurane is continuously delivered during intravital imaging. As an alternative, anesthesia injection of a ketamine/xylazine mixture can be used. Injection anesthesia is relatively easy to perform because no additional equipment is necessary. However, additional injections to keep animals anesthetized are absolutely required for longer imaging times, which might be not be feasible without stopping image acquisition.
We use additional equipment as follows to monitor and control animal conditions. An anesthesia monitor is used to monitor O2 and CO2 concentration in the inspiratory and expiratory gas. The machine can also monitor airway pressure and isoflurane concentration. Because an animals body temperature decreases during anesthesia, we install a heat pad under the animal. The heat pad is connected to a temperature sensor, and it keeps the body temperature stable during intravital imaging. An electrocardiogram is monitored continuously. For intravenous injection during imaging, an intravenous cannula is inserted into the tail vein and saline solution is injected slowly to prevent blood clotting. These machines are monitored and controlled by custom-made software with adjustable alarms. All of this equipment is not strictly necessary, but it helps for stable imaging.
To acquire stable images, it is very important to stabilize the animal mechanically. To accomplish this, we used custom-made devices. For spinal cord imaging, we used a forceps-like device and clamp the spinal cord from both sides of the imaging area (similar to14). For spleen imaging, the spleen is isolated from the body without damaging the blood vessels and placed onto a heated stage.15 Other researchers have published schemes for stabilization of lymph nodes,16 ear skin,17 and liver.18
To identify target cells in vivo, it is necessary to stably label them. Commonly used labeling methods are listed in Table 1. Early studies used synthesized dyes, such as 5-(and-6)-carboxyfluorescein diacetate succinimidyl ester (CFSE) and 5-(and-6)-(((4-chloromethyl)benzoyl)amino)tetramethylrhodamine (CMTMR).19 In addition to these simple dyes, functional dyes that can monitor cellular function, such as intracellular calcium levels, are widely used. Although these dyes stain cells very strongly, cells lose fluorescence and become undetectable if they are proliferating.
|57 and 58
|59 and 60
|59 and 61
|61 and 62
|43, 63 and 64
|27 and 65
|66 and 67
|61 and 68
|Kaede before conversion
|Kaede after conversion
|66 and 67
|ERK FRET sensor
|43 and 44
|PKA FRET sensor
Genetic modification induces expression of fluorescent proteins. In addition to global expression of green fluorescent protein (GFP),20 a variety of transgenic mice has become available as listed in Table 2. By using knock-in technology, a fluorescent protein can be inserted under the control of a specific promoter. For example, yellow fluorescent protein (YFP) was inserted under control of the CD11c promoter to cause YFP expression only in dendritic cells.21 Alternatively, a fluorescent protein can be fused with another protein that is expressed only in special cell lineage. For example, transgenic mice expressing the FoxP3-GFP fusion protein can be useful for the study of regulatory T cells.22 These gene-modified animals continuously produce fluorescent protein within the cells, even in proliferating cells. The gene that encodes the fluorescent protein can be delivered by a viral vector, such as a retro-, lenti- or adeno-virus. Usually viral transduction can be quickly done to compare the generation of gene-modified animals. In general, protein labeling is weaker than chemical labeling, which influences penetration depth.
|NK T cells
|IFNg producing cells
We primarily used retroviral gene transfer to label our derived encephalitogenic T cells.23 To achieve retroviral transduction, the target cells must be proliferating. Therefore, we stimulate T cells in vitro with a specific antigen; in our case, CNS-specific autoantigen together with APCs. After inducing proliferation of autoreactive T cells, they are co-cultured with retroviral vector-producing cells.23 Transduced T cells can be selected by using antibiotics because the retroviral vector contains an antibiotic resistance gene.
When EAE is induced by adoptive transfer of encephalitogenic T cells, these T cells do not infiltrate into the CNS directly, but spend some days in peripheral organs. Odoardi et al. showed that T cells accumulate in the lung immediately after transfer. 24 Intravital imaging showed that encephalitogenic T cells actively migrate within the airways. Interestingly, encephalitogenic T cells acquired a migratory phenotype in the lung and penetrated into the CNS more efficiently than did in vitro activated T cells. This result indicates that encephalitogenic T cells become mature in the lung. It is not yet clear whether this maturation reflects intrinsic T cell changes or influences from external factors. Because lung airways are exposed to the external environment and stimulation can be delivered from numerous sources, it may well be possible that some endogenous factors induce T cell maturation. Indeed, it has been shown that gut microbiota, which are similarly a mixture of many antigens, activate encephalitogenic T cells in the spontaneous EAE model.25
The lungs are not the only organ where T cells maturate. Flügel et al. showed that encephalitogenic T cells accumulated in the spleen and acquired a migratory phenotype there. 24 and 26 Gene profiling demonstrated that migratory T cells showed phenotypic changes, especially up-regulated cell adhesion molecules and chemokines that are important for migration.24 The up-regulation of chemokine receptors is confirmed by cell surface staining of T cells prepared from the spleen.26 More directly, retransfer of encephalitogenic T cells prepared from the spleen showed faster infiltration into the CNS.
Intravital imaging of encephalitogenic T cells in the spleen showed that T cells are continuously moving within the organ.15 In our set-up, penetration depth was limited to approximately 100 μm from the surface, which is less than in lymph nodes or spinal cord. According to our analysis, encephalitogenic T cells did not show any directed movement. We sought to arrest T cells in an antigen-dependent manner by applying soluble antigen intravenously. Our analysis showed that T cells became slower immediately after soluble antigen injection and were arrested 20 min after antigen injection. This was surprisingly fast, yet the following results suggested that this is within reason and that the entire process happens in 20 min. First, injection of peptide, which does not need to be processed to be presented to T cells, arrested the T cells more rapidly than did whole protein. Second, MHC class II blocking diminished the soluble antigen effect dramatically. More directly, DQ-OVA, which is non-fluorescent but becomes fluorescent after protein digestion, caused fluorescent signals within 15 min after intravenous injection. Importantly, this quick effect was further supported by conventional analysis. Both mRNA and protein level data show that inflammatory cytokines are produced as early as 30 min after soluble antigen treatment. Because soluble antigen trapped encephalitogenic T cells in the spleen and prevented CNS infiltration, clinical EAE was ameliorated. This experiment elegantly showed the benefit of intravital imaging. Intravital imaging clearly showed the behavioral changes of T cells before and after injection of soluble antigen in the same animal, which was difficult to analyze using conventional methods.
After encephalitogenic T cells acquire a migratory phenotype, they leave the peripheral organs and arrive at the CNS. We asked the question, how do T cells enter the CNS? Because conventional histological studies suggested that “early bird” T cells are detected in the spinal cord leptomeninges, we performed intravital imaging there. We used myelin basic protein (MBP)-specific GFP-labeled T cells (TMBP-GFP cells) to induce clinical EAE. Animals showed the first clinical sign of disease on day 3 post transfer (p.t.). Therefore, we performed imaging at the spinal cord leptomeninges between day 1 and day 3 p.t.27
Intravital imaging showed that a few TMBP-GFP cells arrived at the leptomeninges on day 1 p.t., which is long before disease onset. Those pioneer cells adhered to the intraluminal surface and moved along the vessels. Similar intraluminal crawling was reported for other cell types, such as monocytes28 and neutrophils.29 Within the next 24 h, the number of intraluminal cells increased. Because two-photon microscopy detects signals by scanning, it is hard to detect flowing cells and rolling cells, indicating that the cells that we detected were crawling. According to our analysis, intraluminally crawling T cells prefer to migrate against the direction of flow in blood vessels. However, velocities are similar regardless of the direction of movement. The infusion of anti-integrin α4 antibody diminished intraluminal crawling, indicating that intraluminal crawling is VLA-4-dependent. Although the precise significance of intraluminal crawling is still unknown, we can speculate that those cells are looking for extravasation sites. Interestingly, TMBP-GFP cell crawling was observed only in the leptomeningeal vessels, but not in other blood vessels, such as those in ear connective tissues and near peripheral nerves.
Intraluminal crawling was followed by extravasation. Intravital imaging recorded that crawling TMBP-GFP cells arrested and then extravasated.27 During extravasation, we often observed leakage of fluorescent dextran, which was injected intravenously to fill the blood plasma. This leakage indicates that the blood–brain barrier (BBB) had opened. However, the leakage was observed for only a short time, indicating that the BBB closed again after T cells crossed it. According to our observations, multiple TMBP-GFP cells extravasated, one after the other, at the same place. This suggests that there are special locations where lymphocytes prefer to extravasate. In accordance with our observations, Abtin et al. showed that neutrophils extravasated adjacent to perivascular macrophages in inflamed skin. 30 In addition, they showed that this localization is due to chemokines produced by perivascular macrophages. Another group suggested that there was influence from pericytes.31 They showed that pericytes attracted myeloid leukocytes by producing macrophage migration inhibitory factor (MIF). Although these studies focused on innate immune cells in peripheral organs, similar mechanisms may exist in the spinal cord leptomeninges.
Why do T cells infiltrate into the immune-privileged CNS? The CNS is protected by the BBB, and the infiltration of immune cells is tightly controlled but not prohibited.32 After ovalbumin (OVA)-specific GFP-labeled T cells (TOVA-GFP cells) were transferred into naïve animals, a very small number of cells were found in the CNS,27 supporting the idea of immune surveillance in the CNS. However, Davalos et al. used intravital imaging to suggest that fibrinogen leakage from blood vessels induced clustering of microglia, which further induced neuronal damage. 33 This small amount of damage may change the permeability of the BBB and recruit immune cells to the CNS.
The spinal cord leptomeninges is not the only location where encephalitogenic T cells begin infiltration. It was shown that small numbers of T cells enter the CNS and are distributed in the parenchyma within 3 h after adoptive transfer, suggesting direct infiltration into the CNS parenchyma.34 In addition, it was shown that T cells enter the CNS via dorsal blood vessels at the 5th lumbar spinal cord.35 This is due to CCL20 production caused by activation of sensory neurons by the soleus muscles. T cells also seem to infiltrate via the cerebrospinal fluid (CSF). Reboldi et al. showed that CCR6-deficient mice are resistant to EAE and, interestingly, T cells were stacked at the choroid plexus, where CCL20 is constitutively expressed.
Once encephalitogenic T cells enter the CNS, it was shown that T cells recognized antigen presented by bone marrow-derived perivascular macrophages.36 To visualize the interaction between encephalitogenic TMBP-GFP cells and APCs, we performed intravital imaging at the spinal cord leptomeninges. We visualized APCs by intrathecal injection of fluorescent dextran (size: 70 kDa) into the cisterna magna. Intravital imaging found that TMBP-GFP cells interact with APCs for a relatively long time, whereas TOVA-GFP cells showed only a short period of contact. This result suggests that TMBP-GFP cells recognize endogenous antigen presented by local APCs. Indeed, TMBP-GFP cells in the spinal cord meninges and parenchyma, but not in the spleen, produced inflammatory cytokines, indicating activation. In addition, TOVA-GFP cells in the spinal cord leptomeninges became arrested after administering OVA-pulsed APCs intrathecally, which again suggests antigen-dependent interaction and subsequent activation. However, due to lack of proper methods, it was not possible to visualize T cell activation in vivo.
There were several remarkable attempts to detect cellular activation in vivo. One of them involved detecting an immunological synapse, which is the special structure formed when TCR recognizes its specific antigen in the context of MHC.37 For this purpose, lck or CD3ζ is fused to GFP and expressed in cell lines using retroviral gene transfer.38 After TCR-mediated stimulation, the fusion protein was recruited into an immunological synapse, and it was imaged in vitro. However, it is not easy to apply these fusion proteins to in vivo experiments. In in vitro experiments, one can predict where immunological synapses will appear by using a monolayer of APCs. In contrast, an immunological synapse can appear at any place on the cell surface in vivo, which indicates that one must scan entire cells with precise z-stacks to detect it. Such precise scanning requires more time and loses temporal resolution. Another attempt to detect cellular activation in vivo involved using GFP knock-in mice under the control of the immediate early gene Nr4a1 (Nur77) was performed.39 By quantifying GFP expression, activation status and signal strength can be analyzed. However, this was not suitable for intravital imaging because there is an unavoidable time gap between T cell stimulation and GFP expression. A simpler approach is to use calcium sensing dyes to analyze neuronal activities.40 Unfortunately, this approach cannot be used for T cells because T cells lose their staining within a short time due to proliferation and actively pumping out the dyes.41 To overcome these problems, we decided to use protein-based sensors that detect T cell activation immediately after TCR stimulation. More specifically, we attempted to detect increasing intracellular calcium and translocation of nuclear factor of activated T cells (NFAT) from cytosol to nucleus (Fig. 1). These sensors are functionally distinguishable. Calcium signaling can be induced by relatively weak stimulation, whereas NFAT translocation can occur only after absolute T cell activation. It makes sense to use two sensors to obtain a more precise picture of the status of T cell activation.
Scheme of activation sensors. (A) TCR stimulation induces calcium release from endoplasmic reticulum (ER). Emptying calcium in the ER opens Calcium Release-Activated Channels (CRAC) on cell surface, which induce influx of extracellular calcium. Increased intracellular calcium induced DE phosphorylation of NFAT, followed by relocation of NFAT from cytosol to nucleus. (B) Structure of Twitch calcium sensing protein. CFP and YFP are connected with troponin C calcium sensing protein. Twitch changes its confirmation according to calcium concentration. Excitation of CFP results emission of blue and yellow at low and high calcium environment, respectively. Right pictures show representative cells of both activated (line) and not activated (dotted lines). (C) Protein structure of NFAT-based activation sensor Picture shows representative cells of both activated (line) and not activated (dotted lines).
We used the calcium sensor Twitch, which consists of cyan- and yellow-fluorescent protein (CFP and YFP, respectively) connected with a troponin C domain. In a low calcium environment, excitation of CFP induces emission from CFP. In contrast, in a higher calcium environment, calcium binding to the troponin C domain changes the protein conformation.42 Subsequently, excitation of CFP results in emission from YFP due to a fluorescence resonance energy transfer (FRET) effect. To improve expression levels in mouse T cells, we developed a codon-diversified Twitch and used it for intravital imaging in the mouse EAE model.41 Twitch was expressed in MOG-specific T cells (TMOG-Twitch cells) by using retroviral gene transfer and imaged in peripheral lymph nodes and spinal cord leptomeninges. In peripheral lymph nodes, TMOG-Twitch cells showed occasional short-duration calcium spikes, often coincident with lower motility. Because it is not likely that MOG antigen is presented in the peripheral lymph node, we considered those short-duration calcium spikes to be antigen-independent. The application of antigen stimulation arrested T cells within a short time, as we observed in the spleen.15 At the same time, TMOG-Twitch cells showed saturated long-duration calcium elevation, indicating that the Twitch sensor detected T cell activation. TMOG-Twitch cells were also imaged in the spinal cord leptomeninges at the time of EAE onset. Substantial numbers of TMOG-Twitch cells showed elevated intracellular calcium, and the duration of calcium spikes was approximately 6 min, which is considerably longer than those observed in lymph nodes. Importantly, those activations were often observed in perivascular areas or near APCs. Because Twitch-labeled OVA-specific T cells rarely showed long-duration calcium spikes, we concluded that TMOG-Twitch cells are activated by endogenous autoantigens.
To detect the subcellular location of NFAT, truncated NFAT was fused to GFP and expressed in MBP-specific T cells (TMBP-NFAT-GFP cells).43 TMBP-NFAT-GFP cells were first imaged when T cells were within the leptomeningeal vessels. Intravital imaging showed that both rolling and crawling cells had cytosolic NFAT, indicating that they were not activated. In contrast, substantial numbers of extravasated TMBP-NFAT-GFP cells showed NFAT-GFP in their nuclei. Intravital imaging clearly showed that a non-activated T cell, which has cytosolic NFAT, interacted with local APC and the interaction quickly induced translocation of NFAT to the nucleus. Interestingly, some, but not all, APCs stimulated T cells efficiently. Similar observations were also reported by another group.44
In summary, by using protein-based activation sensors, we could visualize T cell activation in the CNS after contact with local APCs. Of course, these activation sensors can be applied in other cells, as shown.45
Although CD4+ T cells are considered the key player in initiating CNS inflammation, other types of cells, which can be either brain-resident cells or infiltrating cells, also contribute. One of them is Treg, which can suppress the function of encephalitogenic T cells. Because depletion of Treg at the acute phase of EAE enhances clinical severity dramatically, we aimed to image Treg in the spinal cord leptomeninges to analyze interaction with encephalitogenic T cells.46 We crossed T-Red mice, in which T cells express RFP,47 and DEREG mice, in which Treg express GFP and diphtheria toxin receptor under FoxP3 promoter.48 Intravital imaging at the spinal cord leptomeninges was performed at the peak of EAE with or without Treg depletion. We found that the encephalitogenic T cells moved slower and stopped more often in the absence of Treg. This suggests that Treg can influence inflammation in the CNS. During intravital imaging, we observed that Treg interacted with both effector T cells and APCs, indicating that suppression of disease can be via direct effect on effector T cells or indirect effect via APCs.
There are other players in CNS inflammation. B cells are known to produce antibodies in the CSF6 and contribute significantly to CNS inflammation. The depletion of B cells is beneficial for both EAE49 and MS.50 Mononuclear phagocytes, such as microglia33 and macrophages51 induce neuronal damage. In addition, oligodendrocytes have a critical role in myelination and neurons are targeted to be destroyed. However, in vivo imaging to study these cells, with the exception of neurons, has rarely been performed, and their roles are largely unknown.
The results obtained from intravital imaging can be used for developing therapeutic treatment. For example, we showed that the infusion of anti-integrin α4 antibody diminished intraluminal crawling within minutes. As a consequence, infiltration of encephalitogenic T cells into the CNS is also blocked, resulting in prevention of clinical EAE.27 Indeed, anti-integrin α4 antibody is approved as an MS treatment and shows beneficial effects. Our intravital imaging clearly showed the mechanism of this antibody treatment. In addition, we have shown that the calcium inhibitor, BZ194, ameliorated clinical EAE in both preventive and therapeutic treatments.52 Intravital imaging showed that BZ194 treatment increased T cell motility in the CNS. We speculate that BZ194 prevented T cell arrest by blocking intracellular calcium signaling; therefore, T cells do not get sufficient stimulation to induce inflammation. Furthermore, we have shown the effect of soluble antigen treatment in EAE. When soluble antigen was given before the onset of EAE, the treatment ameliorated clinical severity dramatically.15 In contrast, soluble antigen worsened EAE when it was applied after the onset of disease.53 In both cases, soluble antigen activates encephalitogenic T cells. The difference lies in where the T cell activation occurs. Before the onset of EAE, the majority of encephalitogenic T cells are in the periphery, and activation of them does not result in deleterious effects. However, after the onset of EAE, many encephalitogenic T cells are in the CNS, and their activation results in a lethal level of inflammation. One always needs to keep in mind that results from rodent models cannot be applied directly to humans. However, intravital imaging holds great potential for understanding the cellular mechanisms of disease pathogenesis and for developing and evaluating therapeutic treatments.
Intravital imaging in the immunology field started in the early 2000s to study cellular motility in the explanted organ.54 Currently, multicolor imaging and functional imaging have become popular. There are interesting, and potentially very robust, new methods that have been introduced recently. One of them involves gradient index (GRIN) lenses.55 This method uses an endoscope that can perform imaging within the tissue. Because the penetration depth of two-photon microscopy is superior, but still limited, such an endoscope is the method of choice. Recently, an interesting study using photoconvertible dyes was published.56 This study explored the functional difference between migratory and resident dendritic cells in the lymph nodes. This kind of study has very high potential because cells are migrating in the body, and the consequences of a particular event may not happen in the same place. For example, a cell receives stimulation in one organ and shows effector function in another organ. Lastly, it is extremely important to analyze data and obtain fruitful results. Two-photon microscopy has become user-friendly and it is now easier to acquire excellent images. However, this is only one component of intravital imaging and researchers must translate imaging data to fruitful messages.
The author received the research funding from Genzayme.
Our work is supported by Deutsche Forschungsgemeinschaft (KA2951/2-1 and KA2951/3-1), Hertie Foundation, Novartis Foundation for Therapeutic Research, Genzyme, and German Academic Exchange service (DAAD), Ludwig-Maximilians University Munich, and Max-Planck Society.