|Home | About | Journals | Submit | Contact Us | Français|
New preparations, fluorescent probes and imaging techniques are providing the means to observe the behavior of cells in the tissue environment of lymphoid organs. In particular, when combined with two-photon laser microscopy, intravital imaging of surgically exposed lymph nodes provides a unique view of lymphocyte migration and antigen presentation as it occurs within the living animal. The view is emerging that lymphocytes migrate randomly within lymphoid organs, and that lymphocyte contact with antigen-presenting cells may be a stochastic process rather than one guided by chemokine gradients.
The immune system consists of a distributed network of trillions of cells that must operate independently to provide antigen specificity and yet function in a coordinated manner to defend us from a wide variety of pathogens. Communication between cells can be initiated by direct cell contact, or can take place at some distance within the tissue environment via chemokines.
Over the past 20 years, remarkable progress in molecular immunology has defined the mechanism of antigen recognition and identified a growing cast of molecules and signaling pathways that link the T-cell receptor to the nucleus. However, we still understand very little about the basics of motility, compartmentalization and antigen recognition in vivo, because these events occur within densely packed lymphoid organs [1–3]. How do T cells, B cells and dendritic cells (DCs) move within the native tissue environment? How are T and B cell compartmental boundaries established and maintained? How does a T cell locate antigen within the lymph node — by following chemokine gradients or by random collisions with antigen-presenting cells (APCs)?
There is increasing recognition that events defined in vitro may not correspond to the physiological situation in vivo . For example, contact between a T cell and an APC leads to a redistribution of surface molecules and formation of the ‘immunological synapse’ [5–7]. This type of molecular redistribution has also been studied in vitro in planar lipid bilayers with defined molecular constituents. Yet it is still unclear whether stable synapses occur in the environment of the lymph node, or whether antigen recognition naturally involves short-lived serial encounters.
Lymphoid organs have remained a black box into which defined cell populations can be induced to home, but from which we have been able to obtain only ‘snapshot’ views, by extracting cells or analyzing fixed tissue. Clearly, there exists a strong need for imaging approaches to visualize living cells within intact lymphoid tissue.
To illuminate the black box of the tissue environment, new in vitro preparations have been developed that more faithfully represent the native tissue environment of intact lymphoid organs. These include 3D collagen gel matrices , monolayers of endothelial cells (ECs) bathed with flowing solution to mimic the forces experienced by cells , cultures of clusters or reaggregated tissue fragments [10•,11], engineered tissue surrogates , and whole lymph node explants (Figure 1; [13,14••,15••]). All of these preparations lack intact vascular and lymphatic vessels, and therefore cannot be applied to investigate processes such as lymphocyte trafficking. Furthermore, the lack of blood and lymphatic vessels may disrupt the distribution of important soluble factors or alter the physiological levels of tissue oxygenation.
To overcome this limitation, several promising methodologies have been developed to visualize cells within the in vivo tissue environment. Non-invasive methods include bioluminescence imaging of cells engineered to express luciferase , magnetic resonance imaging microscopy to track cells labeled with superparamagnetic particles , and positron-emission tomography (PET; ). Although these methods can be applied to intact animals, all three require cell engineering to derive populations that can be detected and presently lack single-cell resolution. Instead, optical techniques offer cellular, and even sub-cellular, levels of resolution. Intravital preparations of exposed lymphoid organs permit light microscopic imaging in the native tissue environment with intact circulatory elements, but require anesthesia and surgery to bring objective lenses close enough to the tissue.
To image cells at depths of more than about 50 μm, two-photon microscopy is the technique of choice. When combined with fluorescent probes, confocal microscopy or two-photon microscopy can reveal single cells at the plane of focus, either by imaging through a pinhole in the case of confocal microscopy, or by selectively exciting the fluorophore only at the plane of focus in the case of two-photon microscopy. Recently, we provided a detailed comparison of confocal and two-photon microscopy as applied to the imaging of T cells in situ [19•]. Advantages of two-photon microscopy include less photodamage, greater sensitivity and deeper imaging within tissue, which is possible because the near-infrared illumination penetrates tissue more effectively.
Advances in imaging techniques and fluorescent markers to label or genetically tag cells or specific proteins are now making it possible to ‘see’ events in real time that could previously only be inferred. Confocal and two-photon imaging approaches have recently provided the first glimpse of lymphocyte dynamics within the tissue environment. We used two-photon microscopy to visualize the behavior of individual T and B lymphocytes in explanted lymph node [14••] and spleen  maintained in culture. Using an in vivo adoptive transfer approach (Figure 2), T and B cells labeled with green or red CellTracker™ dyes (Molecular Probes Inc., Eugene, OR, USA) homed to appropriate locations and exhibited vigorous motility within the intact lymph node, with velocities that averaged 12 μm/minute and 6 μm/minute, respectively. T cells were observed to migrate in a ‘stop-and-go’ fashion, similar to the behavior of T cells in a collagen gel matrix culture system, with alternating episodes of rapid motion, when cells were elongated, followed by momentary pauses when cells rounded up. The period of these cycles averaged 1–2 minutes. In the absence of antigen, very few T cells were truly stationary, although pressure on the lymph node or accumulated photodamage caused the cells to stop moving. In healthy preparations, a 3D random walk in all three axes emerges over time because of T-cell turning. In the absence of antigen, naïve T cells from ovalbumin-specific DO11.10 transgenic TCR mice moved randomly without evidence of chemokine gradients.
Very different results were obtained in a parallel study that used a similar adoptive transfer approach, but different imaging and tissue culture conditions. Naïve T cells were immobile when lymph nodes were maintained in culture with atmospheric oxygen tension and imaged using confocal microscopy. Possible reasons for the differences observed have been discussed previously [19•,21,22], and include differences in imaging depth, tissue handling, photodamage and oxygen tension in the tissue. Our study [14••] used lymph nodes immersed in 95% O2, 5% CO2, as previous work on brain slice preparations clearly demonstrated enhanced survival and function when tissue preparations are maintained in vitro under these conditions. Stoll et al. [15••], however, chose to use 20% O2, reasoning that lymph nodes may possess low oxygen tension under physiological conditions .
In an effort to resolve the discrepancy regarding the motility of naïve T cells in vivo, we adopted two-photon microscopy to perform intravital imaging of the inguinal lymph node in an anesthetized mouse [24••]. In this preparation, a simple surgical procedure exposed the inguinal node and allowed lymphocytes to be tracked in vivo as they move within the microcirculation, home into the lymph node and migrate within the T-cell zone of the lymph node. Care was taken to maintain intact circulation of blood and lymph, and to avoid microdissection by imaging through naturally occurring windows in the fat pads that lie on top of the node. Under these conditions, T cells exhibited vigorous motility and migrated randomly without evidence for collective drift or motion along putative chemokine gradients. As observed in explanted lymph nodes, naïve T cells moved in a stop-and-go manner, elongating while moving rapidly ahead and then pausing every 1–2 min on average. Overall, the average velocity of movement was 11 μm/min, very similar to that of T cells in explanted nodes at the same temperature. Figure 3 illustrates a field of T cells with four cell tracks highlighted in a depth-encoded representation of T-cell positions. In this instance, the mouse was breathing room air during the entire experiment; in other experiments we used a mask to deliver a stream of 95% O2, 5% CO2 to maintain respiratory drive during long-term measurements. The key point is that vigorous motility in a random walk characterizes the behavior of naïve T cells in vivo. We postulate that T cells distribute autonomously through the T-cell zone, and that the search for antigen is a stochastic process.
Antigen challenge dramatically alters the behavior of T cells, leading to cell enlargement, expression of new genes, secretion of cytokines and cell proliferation. Two different approaches have been used to visualize changes in T-cell dynamics evoked by antigen. When antigen-specific T cells were transferred into animals that had been injected subcutaneously with specific antigen, clusters and swarms of enlarged T cells were observed one day following adoptive transfer [14••]. At later times, cells divided and resumed a vigorous pattern of motility. Using an alternative method of antigen priming, in which APCs were differentiated in vitro from bone marrow cells, pulsed with antigen and then injected subcutaneously, Stoll et al. [15••] observed contact between T cells and APCs that lasted >15 hours in a one-to-one pattern of association.
Recently, we have pursued an in vivo labeling method to visualize antigen-primed DCs interacting with CD4+ T cells (MJ Miller, SH Wei, I Parker, MD Cahalan, unpublished data). If the T cells can be likened to swimming fish, DCs behave effectively as nets; they make contact with T cells by throwing out long membrane tethers and rapidly reeling them back in, constantly changing their shape and greatly expanding their capture radius. It appears from these early studies that the initial encounter between a T cell and a DC relies upon dynamic cell behaviors that are finely tuned to optimize the chance of random collisions.
Two-photon microscopy represents an optimal technique for tracking the behavior of living cells deep within the tissue environment. It is already feasible to image T cells and other cells of the immune system within the circulation, or in the tissue environment of lymph node, spleen, Peyer's patch, thymus and peripheral tissues. Video presentations of the data demonstrate the dynamic behavior of T cells and B cells as they migrate within the lymph node, and of DCs as they interact with T cells during antigen presentation. Two-photon imaging will be adaptable to a wide variety of new probes for second messengers and gene expression, and to a broad range of processes both physiological and pathological. Combined with intravital imaging of surgically exposed lymphoid organs, two-photon imaging is providing a unique view of lymphocyte dynamics in vivo.
A recent study used two-photon microscopy to examine the interaction of dendritic cells labeled in vitro with motile CD8+ T cells in an explanted lymph node preparation [25••]. T cells made stable, long-lasting contacts with antigen-pulsed DCs, rather than a series of short contacts.
Papers of particular interest, published within the annual period of review, have been highlighted as:
• of special interest
•• of outstanding interest