The live lung slice method consists of partitioning the lung in thick slices, which are then carefully maintained to assure tissue viability. To do so, after the exsanguination of the animal, a cannula is placed intratracheally and lungs are inflated with 1.5–2 % liquid low melting agarose to maintain lung structure. The lungs are extracted and agarose is solidified in cold physiological medium. With the use of a micro-slicer or a vibratome, 150–300 μm thick serial slices are obtained. The vibration offers the possibility of cutting living tissue with a minimal injury. The slices are maintained in cell culture media at 37 °C and 5 % CO₂. Lung slices are normally used in the first 8 h although they have been used up to 2–3 days after sampling [9]. The lung slices have normal cell activity allowing the study of cell interactions and at the same time the structure of the alveoli is maintained to allow the 3D imaging of cell movements. For complete tissue preparation methods review, please see Thornton et al. [10].

LLS preparations have several advantages in comparison to fixed tissue imaging. First, fixed tissue samples are usually much thinner than live lung slices. The increased thickness of the live slices allows the study of three-dimensional cell movements that would be lost in thinner preparations or cellular monolayers. Second, and more importantly, unlike the static fixed images, LLS allow the study of time-dependent processes ranging from fast cell–cell interactions to the study of complex and slow pathogenic immune responses in the lung. Third, the LLS technique is a useful tool to study cellular interactions deep in the lung where intravital imaging or the isolated, perfused lung preparation cannot reach due to the excessive thickness. Because only a section of lung is studied with LLS preparations, it has important limitations such as the loss of blood and lymph circulation and the lack of neural input. These considerations are very important in studies requiring cell egress from other organs or other parts of the lung. Leukocyte activation induced by the isolation procedure may also be a concern.

In a work aimed to establish LLS as a suitable ex vivo technique to investigate the immunomodulatory effects and the characterization of cytokine production after LPS and other challenges, cell death was imaged by live/death fluorescence using confocal microscopy [9]. To distinguish living from dead cells, different dyes are used, such as Acridine Orange and Propidium Iodide. These markers allowed a better characterization of rat type II pneumocytes in an in situ patch clamp study [11]. TUNEL fluorescence analysis is also a very useful staining approach to study cell death. Jacobs et al. used this technique to study ROS-induced cell apoptosis [12].

Pulmonary neuroephithelial bodies (NEB) are organized as clusters of pulmonary neuroendocrine cells during the development of the lung, and 4-Di-2-ASP fluorescent staining can be used to identify them in sliced lung preparations. Ca²⁺ signaling in the NEB microenvironment has been the focus of several studies. To facilitate imaging, Ca²⁺-specific fluorescent dyes have been developed, from the commonly used Fluo-3 to its brighter analog Fluo-4. The use of these dyes allows real-time imaging of cellular Ca²⁺ flux using confocal microscopy [13, 14].

The use of fluorescent antibodies in LLS preparations has improved protein and cellular identification due to the greater antigen accessibility compared with whole lung approaches. The use of Surfactant A and alveolar type II cell-specific antibodies allowed the identification of alveolar type II cells and the lamellar bodies contained within the cells [15].

The introduction of fluorescent markers opened the possibility to easily track and image target cell types or even cell subpopulations. Using a transgenic mouse where a yellow fluorescent protein expression is under the control of a promoter specific to antigen-presenting cells (CD11c-EYFP), Major Histocompatibility Complex (MHC) class II expression was studied. The lung slices were incubated with LPS and other pro-and anti-inflammatory molecules and the confocal fluorescence images showed that MHC-II expression was increased with pro-inflammatory challenges and repressed with dexamethasone [9].

In the isolated, perfused lung preparation, the left pulmonary artery is cannulated and perfused with autologous blood or a physiologic salt solution allowing passive drainage through the left ventricle. The lungs are ventilated with normal air or humidified gas containing different concentrations of oxygen and positive end-expiratory pressure is maintained. Using a perfusion pump, experimental agents can be added to the perfusate and introduced in the lung circulation. The lungs are suspended from a force displacement transducer to measure lung weight changes, while the whole system is maintained in a closed chamber with controlled humidity and temperature.

The main advantage of IPL technique over the live intravital approach is the unmatched stability of the sample. This important feature has permitted the use of directed microinjections of fluorescent dyes into endothelial or epithelial cell layers in the lungs [16]. Perfused whole lung preparations are a better approach to study lung physiology than LLS preparations. IPL preparations do not have the injured cut surfaces at the top and bottom of the preparation, which may induce undesirable effects when studying organ functionality. Vascular leakage is also better studied in isolated lungs than in LLS [17]. A clear disadvantage of isolated lung preparations is that it lacks the effects of the normal (or induced) physiological changes that occur systemically. This effect diminishes the possibility of leukocyte chemoattraction to the lungs from the bone marrow and circulation, which is an important limitation in inflammation-related studies. A possible leukocyte activation effect induced by the pump-assisted circulation used in the preparation of IPL is another disadvantage when compared with lung intravital microscopy [18].

The labeling of mitochondria with fluorescent tags has allowed for the study of mitochondria transference in lung injury models. Islam et al. showed that mouse bone marrow-derived stromal cells, containing red fluorescent mitochondria, were located adjacent to epithelial cells minutes after their instillation into LPS-challenged isolated perfused lungs. In a further experiment, they showed by confocal microscopy that the fluorescent mitochondria were transferred to alveolar type II cells in a gap junction dependent process [19]. The use of other mitochondria markers (Mitotracker green, Ca²⁺ binding Rhod 2AM) allows their identification in endothelial cells [20].

Leukocytes are critical mediators of lung injury induced by endotoxemia, and the time that leukocytes spend in lung microvessels (venules and capillaries) depends on their activation state as well as the activation state of the endothelium. The use of fluorescence microscopy and R6G-rhodamine revealed that LPS-challenged microvessels retained leukocytes for a longer time than unchallenged lungs. The passage time is easily quantified in the surface vessels of isolated and perfused lungs without introducing any mechanical injury to the lungs [21].

Reactive oxygen species (ROS) production can be measured using fluorescence microscopy and the 2′–7′-dichlorodihydrofluoroscein acetate dye [20]. This technique was employed by Ichimura et al. to show that physical pressure-induced stress is able to promote ROS-triggered endothelial cell expression of P-selectin.

Nitric oxide (NO) is a potent controller of the vascular tone in systemic and pulmonary vessels. Shear stress can promote the production of NO in endothelial cells. In 2000, in situ fluorescence procedures using IPL showed that shear stress was able to induce endothelial cell NO using a NO probe (diaminofluorescein diacetate), which was preceded by intracellular changes in Ca²⁺ detected using a fluorescent probe (Fluo-3) [22].

To observe the lung in a more physiological manner, it should be imaged in vivo. As discussed earlier, challenges faced with lung imaging compared to other organs are its intrinsic motion and the maintenance of breathing after opening the thorax. Different approaches have been developed to address these issues including (a) maintaining mechanical ventilation or reestablishment of spontaneous breathing, (b) the structure and composition of the window in contact with the lung, and (c) the management of cardiorespiratory motion.

Access to the lung requires opening of the thorax, which will cause lung collapse in the absence of positive pressure ventilation. Therefore, most of the preparations will use mechanical ventilation. Animals are anesthetized and tracheally intubated to be ventilated with room air or enriched oxygen. The animal is placed on a warming pad set to 37 °C to help maintain body temperature. To prevent dehydration, physiologic crystalloid solutions should be administered every hour or continuously. In some experiments, the thorax is closed after placing the window. To recover spontaneous breathing, the removal of the air introduced in the pleural cavity is required by vacuum or syringe suction. Mechanical ventilation is maintained until the lung is able to re-expand. Fingar et al. used a thoracic window implanted in rats to follow for 2 weeks the progression of pulmonary edema and alveolar flooding after lung injury by monitoring the leakage of vascular dyes [23]. Thoracic windows have been maintained in dog and rabbits for several months [6, 7]. Implanting a window has the advantage of not interfering with physiological respiration. However, because the motion is not controlled, it is not suited to high-resolution acquisition.

Once the thorax is open, it is important for high-resolution images to have stable windows with good optical properties and minimal interference with the normal structure and function of the lung. Most of the thoracic windows have used a metallic structure, first described by Terry et al. [4], to be introduced in between two adjacent ribs. The air from the pleural cavity is removed by suction or a syringe to bring the lung close to the observation window, which can be a cover glass [4, 24] or a transparent membrane made of Cronar [6] or Teflon [25]. To prevent the lung from dehydrating or cooling, the Teflon membrane used by Kuhnle et al. is covered by a warmed and bubbled Tyrode’s solution. Metal and glass windows are efficient but their rigidity can induce trauma upon the delicate lung surface. Other approaches use less invasive methods, particularly in the mouse. Tabuchi et al. [26] used a polyvinylidene membrane sealed with glue over the ribs, and Kreisel et al. [27] attached the lung tissue to the bottom of a coverglass with tissue adhesive (VetBond). However, irritation may be produced from the moving lung touching the membrane potentially invoking an inflammatory response.

One of the first approaches used on cats [3] and rats [23] was to use paralyzing agents to halt respiration. Another way to increase the time of stabilization is to temporarily suspend the respiration for 30 s [28] or in one of the two lungs by clamping the bronchus while ventilating the other lobe [29]. However, in these techniques where the respiration is blocked, the observed tissue will suffer from impaired oxygenation, which will undoubtedly affect the observed physiology. Without interrupting respiration, another technique has been to image the lung once every respiratory cycle, when the lung stops moving for a moment and comes back to the previous cycle position, at the end of the expiratory phase. This approach has been applied by Kuhnle et al. [25] and Tabuchi et al. [26]. This timing can be achieved by matching the ventilation rate and the acquisition. Indeed, imaging every 0.5 s with a ventilator rate of 120 breaths/min by Kreisel et al. [27] enables image acquisition once every breath cycle. More recently, in an effort to obtain more physiologic imaging, Fiole et al. did not use any stabilization procedure in the mouse lung during imaging, but instead corrected the images post hoc [30]. Every minute, a series of images was acquired and just one image was retained without any deformation. Lung structure was used as a frame of reference to select the correlated images by computer analysis.

Wagner et al. [8] addressed in an efficient way the obstacle of cardiorespiratory movement in live animal imaging using a thoracic window with built-in suction, providing enough stabilization for real-time microscopy. A similar technique, adapted from Wagner’s thoracic window coupling suction to gently immobilize the lung on a glass coverslip has been recently adopted and miniaturized for mice by Looney et al. [24]. Different mouse windows were developed by groups in Germany [26], Japan [29] and USA [24, 27, 31]. The application of these lung intravital microscopy methods to mice has allowed access to the tools available with transgenic animals.

One advantage of fluorescence intravital microscopy is the high-resolution enabled by all fluorescent techniques. Compared to other techniques described for fluorescent imaging, lung intravital imaging makes it possible to observe lung injury under physiological conditions and with maintenance of the lung microenvironment. The preservation of blood and lymphatic circulation is one the main attractive features of IVM, which is important to study vascular permeability and leukocyte recruitment. It is consequently one of the most powerful approaches to study processes in lung injury animal models at a cellular and molecular scale.

One of the major limitations of intravital microscopy is the restricted ability to image deep in tissues. Imaging with two-photon excitation is confined mainly to 30–100 μm below the pleural surface, accessing only the most superficial layer of the lung. It may be a concern if the injury and inflammation in this superficial layer differ from the rest of the tissue. Moreover, the surgical preparation needed to access the lung could induce trauma that could have deleterious effects upon the microcirculation. Studies must be evaluated carefully considering the influences that may alter the normal physiology of the lung. Lastly, even though these IVM methods enable the acquisition of lung images up to several hours, it is currently unsuited for repeated observations in rodents.

With these advanced fluorescence microscopy methods, it is possible to generate high-resolution images of several z-stack positions to generate 3D reconstitutions. The methods described also allow for imaging over several hours to generate time-lapsed data. It is then possible to analyze, localize, and quantify different parameters in four dimensions (3D plus time) such as cell velocity, colocalization, shape, volume, number, intensity, and color. Fluorescence is defined as the emission of light from a fluorescent probe after its excitation by an external light source of defined wavelength. The visualization of tissues, cells or proteins can be achieved by tagging molecules with specific fluorophores which can be selectively excited and specifically visualized. Here, we describe some useful applications for lung injury models.

Visualizing the live lung enables the real-time observation of structural changes occurring during injury. For example, changes in lung vessel diameter by intravital microscopy was monitored during sepsis in rats [32] or after hypoxia in mice [26]. To visualize the lung structure several tools are available. Good resolution of the lung matrix can be obtained without using any extrinsic dyes due to the second harmonic generation effect. In a tissue, the specific molecular structure of collagen fibers generates an ultraviolet second harmonic light when excited with an infrared laser of a two-photon microscope. Reporter mice ubiquitously expressing fluorescent proteins also enable the imaging of stromal cells like the ubiquitously expressed actin-CFP reporter mice [29] or the mTmG reporter mice. A tdTomato fluorophore is expressed ubiquitously in the mTmG mice, and the localization of the fluorescent proteins to membrane structures outlines cell morphology and allows resolution of fine cellular processes (mT). This mouse can be crossed with a Cre-recombinase reporter mouse to target GFP expression in specific cell types (mG). Since the alveoli are surrounded by a dense meshwork of capillaries, labeling of the blood circulation by intravascular injection of tagged albumin, polysaccharides (dextran) or untargeted quantum-dots [33] will also produce an excellent outline of the alveolar structure.

Lung edema and vascular leakage is a characteristic feature of lung injury. Labeled albumin or labeled polysaccharides (dextran) injected intravenously can also be used to monitor vascular leakage in vivo. Indeed, under homeostatic conditions blood vessels limit the passage of dextran larger than 70 kDa, but during inflammatory conditions, dextran up to 2000 kDa can leak from the intravascular compartment [34]. Dextran efflux and vascular permeability can be quantified by measuring the changes in fluorescent intensity (Fig. 8.2a). The sensitivity of the fluorophore leakage can be modulated by using molecules of different sizes. Fingar et al. used a rat model of lung injury induced by oleic acid or compound 48/80 to directly measure in vivo the kinetics and magnitude of pulmonary vessel leakage and the development of edema. Leakage of intravascular FITC-albumin or rhodamine dye can be observed and quantified [23]. This same method was used to measure FITC dextran leakage after PMA or cigarette smoke-induced lung inflammation [31]. Looney et al. measured in mice the dynamic leakage of Texas Red dextran into the extravascular compartment during lung injury after intratracheal administration of LPS. Interestingly, in vivo imaging revealed a differential rate of vascular leakage across the imaged alveoli [24], an observation that could not have been made using measurement of global lung vascular permeability.

Another typical feature of lung injury is the rapid and massive recruitment of neutrophils into the lungs. Intravital microscopy is one of the most powerful techniques to study the anatomical location and dynamic influx of immune cells, and it is especially valuable for observing trafficking of cells from the circulation to peripheral tissues. Neutrophil recruitment into the lung is different from other vascular beds, and intravital microscopic approaches allow dissecting of these mechanisms in detail. Indeed, the size of a neutrophil (6–8 µm in diameter) can be bigger than the diameter of 50 % of the lung capillaries (2–15 µm). This may explain why neutrophils are sequestered in the lung. Sequestration of neutrophils has been observed in live rabbits by Kuebler et al. [35], and labeled leukocytes confirmed that lung capillaries are the predominant site of leukocyte sequestration. Neutrophil sequestration is accompanied by a morphological change into elongated shapes that have been observed in lung slices by actin labeling and confocal microscopy [36]. The same group made important discoveries about the requirement of selectins in sequestration and emigration of neutrophils in the lung [37]. Different tools are available to study neutrophils in vivo. Cells can be isolated and fluorescently labeled with vital dyes before infusion into recipient animals for fluorescence microscopy. Such a method was used by Presson et al. to observe the migration of Rhodamine-6G in vivo-labeled leukocytes into the rat lungs after PMA or cigarette smoke exposure [31]. The same group also demonstrated in a mouse IVM model the role of nitric oxide in neutrophil lung infiltration during sepsis using iNOS knockout mice [38]. Alternatively, the injection of antibodies tagged with a range of fluorescent dyes allows for labeling of neutrophils. However, caution must be taken when using this technique as antibodies can induce the activation or the depletion of the targeted cells. Anti-Gr-1, for example, can deplete neutrophils if high doses are used. In addition, the use of transgenic mice expressing fluorescent proteins in a specific cell lineage is a powerful method for specific labeling of leukocytes (Fig. 8.2c). A variety of strains have been created to track neutrophils. One example is the lysozyme-M (LysM)-green fluorescent protein (GFP) mouse that is characterized by bright green neutrophils and monocytes that are dim green. Using LysM-GFP mice, Kreisel et al. [27] observed by two-photon intravital imaging the mechanisms of neutrophil extravasation in bacterial pneumonia and ischemia-reperfusion after murine lung transplantation. A large pool of resident lung neutrophils was observed that rapidly increased in number after inflammatory challenge. Neutrophils clustered around monocytes, and the depletion of monocytes reduced this clustering phenomenon and reduced neutrophil extravasation. In the same mouse model of ischemia-reperfusion injury. They established that alveolar macrophages and their cell membrane associated protein DAP12 were important for the production of the chemokine CXCL2 and subsequent neutrophil extravasation [33]. However, LysM is expressed in the lung by both neutrophils and macrophages. To obtain fluorescent expression that is more restricted to neutrophils, the MRP8 promoter has been used [39]. Table 8.1 describes mouse strains commonly used to visualize cells in lung injury models.