Lymph nodes are highly organized structures specialized for efficient regulation of adaptive immunity. The blood and lymphatic systems within a lymph node play essential roles by providing functionally distinct environments for lymphocyte entry and egress, respectively. Direct imaging and measurement of vascular microenvironments by intravital multiphoton microscopy provide anatomical and mechanistic insights into the essential events of lymphocyte trafficking. Lymphocytes, blood endothelial cells, and lymphatic endothelial cells express sphingosine 1-phosphate receptor 1, a key G protein-coupled receptor regulating cellular egress and a modulator of endothelial permeability. Here we report the development of a differential vascular labeling (DVL) technique in which a single intravenous injection of a fluorescent dextran, in combination with fluorescent semiconductor quantum dot particles, differentially labels multiple blood and lymphatic compartments in a manner dependent on the size of the fluorescent particle used. Thus DVL allows measurement of endothelial integrity in multiple vascular compartments and the affects or pharmacological manipulation in vascular integrity. In addition, this technique allows for real-time observation of lymphocyte trafficking across physiological barriers differentiated by DVL. Last, single-field fluid movement dynamics can be derived, allowing for the simultaneous determination of fluid flow rates in diverse blood and lymphatic compartments.
- intravital two-photon excitation microscopy
- lymph node
- fluorescence labeling
- microfluidic dynamics
lymph nodes provide a complex, orchestrated environment to regulate the adaptive immune response (18). Lymphocytes must be able to enter the lymph node, move to their proper location to receive antigenic stimulation, then properly leave the lymph node into the circulation once again. Initially, lymphocytes arrive in the lymph node cortex predominantly by migrating out of the blood into the lymph node parenchyma though specialized blood endothelium of high endothelial venules (6). Cells then follow specific chemokine gradients to their respective B-cell follicles or T-cell paracortex, where antigen presentation can occur (10, 13, 24, 26).
Lymph nodes additionally influence the extent and efficiency of the immune system by regulating the recirculation of naive and effector cells into the periphery (11, 19, 36). One molecule that controls the egress of lymphocytes from the lymph node is the lysophospholipid sphingosine 1-phosphate (S1P), which binds to a family of G protein-coupled receptors, S1P1–5 (7, 11, 27). In particular, S1P1 is required on lymphocytes for their proper egress from the lymph node recirculation (8, 9, 20, 39–41, 43). While S1P1 is expressed on lymphocytes, it is also highly expressed on both blood and lymphatic vasculature within the lymph node, where it plays a role in the maintenance of proper vascular integrity. The control of lymphocyte egress reflecting the inherent property of S1P1 receptor to respond to its ligands and to modulate signaling efficacy is intrinsic to both lymphocytes and endothelial cells (32, 33, 35–38, 42).
Previous methods describing the microenvironments within lymph nodes and related cellular activities have had some limitations. Histological section analysis is a snapshot of cellular dynamics, while lymph node explants disrupt both lymphatic and blood flow. Additionally, ex vivo approaches may disrupt chemokine gradients that regulate lymphocyte entry, egress, and motility (40). Intravital studies of lymphatic sinuses by retrograde labeling with fluorophore-conjugated antibodies against lymphatic endothelium-specific hyaluronan receptor (LYVE-1) require prolonged duration for adequate staining (8). The lack of specific, simple, and rapid method to fluorescently discriminate between blood and lymphatic compartments within an intact vascular system suitable for intravital two-photon microscopy led us to develop the differential vasculature labeling (DVL) technique, allowing simultaneous and differential labeling of both blood and the lymphatic systems, ensuring normal intravital flow conditions. This approach takes advantage of the differences in particle distribution across blood and lymphatic endothelia (17, 21–23, 29–31). Therefore, differential partitioning of fluorescent particles of distinct sizes, following a single intravenous injection, rapidly label multiple vascular compartments, allowing for differential blood and lymphatic system labeling and direct measurement of fluid flow rates by a single-frame vascular flow analysis.
MATERIALS AND METHODS
Mice and fluorescent dyes.
Experiments were conducted on either C57Bl/6J mice or mice expressing S1P1-enhanced green fluorescent protein (eGFP) fusion protein from the S1P1 locus, as previously described (2). Mice were housed in specific pathogen-free conditions at The Scripps Research Institute. All experiments were performed in accordance with protocols approved by the Institutional Animal Care and Usage Committee.
Fluorescently labeled wheat germ agglutinin (WGA), quantum dots (qdots), and dextrans were purchased from Life Technologies. All fluorescent markers within a single experiment were injected as a mixture intravenously. Particle sizes, fluorescence emission maxima, and amount of fluorophore administered to mice are summarized in Table 1. WGA labeling was performed on explanted lymph nodes, as previously described (34, 43). Briefly, nodes were incubated in 0.1 mg/ml WGA-Alexa Fluor 594 in PBS on ice for 60 min and then washed several times with fresh PBS. For in vivo visualization of lymphatic sinuses, mice were injected subcutaneously at the base of tail with 20 μg of Alexa Fluor 546-conjugated anti-LYVE-1 Ig (R&D Systems, no. 223322) 16 h before the experiment. W146 was administered intraperitoneally to mice at a dose of 10 mg/kg 20 min before imaging.
Intravital preparation, two-photon excitation microscopy, imaging, and data analysis.
The fluorescent particles chosen for DVL were as follows: 1) qdots: semiconductor crystals with size in three dimensions limited to atomic sizes in the order of several nanometers (28); 2) WGA: lectin that selectively binds to N-acetylglucosamine and N-acetylneuraminic acid (sialic acid) residues (44); 3) dextrans: complex, branched glucon (polysaccharide made of many glucose molecules) composed of chains of varying lengths (from 3 to 2,000 kDa) (14, 15).
All two-photon excitation microscopy (TPEM) intravital imaging was conducted on surgically exposed inguinal nodes (25). Mice were anesthetized by intraperitoneal injection of 80 mg/kg ketamine and 10 mg/kg xylazine, followed by subsequent administration of 20–40 mg/kg ketamine to maintain anesthesia during imaging. A cutaneous incision was made in the abdominal area, and the skin flap was retracted, exposing the inguinal node for intravital imaging, and adipose tissue surrounding the node was removed. Only mice with intact vasculature after the surgery were used for imaging, as assessed by 1) the absence of blood extravasation from injured vessels; and 2) no extravasation of qdots and other larger fluorescent particles (>3.6 nm in diameter). Animals lacking fluid flow within either blood or lymphatic vessels were not used. To ensure stable and consistent vascular flow imaging, a custom-made imaging assembly, consisting of imaging chamber, base, and holding frame, was constructed. The mouse was placed on the imaging assembly base, and the imaging chamber was placed over the node and attached to the holding frame, securing the mouse in the imaging assembly. The assembly with the mouse secured was situated over a 37°C heating pad and placed under the microscope objective. The objective lens was mounted onto a PI piezoelectric actuator for real-time, computer-controlled precision adjustments to compensate for tissue deformation and related drift, keeping the imaging window relatively constant. The imaging chamber was supplied with constant superfusion of oxygenated (95% O2 plus 5% CO2) RPMI media at temperature maintained between 36 and 37°C. Explanted node imaging was conducted on the same imaging chamber superfused with oxygenated and warmed RPMI media.
An average volume imaged for time-lapsed microscopy was 400 μm × 400 μm × 50 μm (X × Y × Z), at 512 × 512 pixels (8 bits per pixel) with a Z resolution of 2.5 μm and temporal resolution of 15–20 s. A typical single time point three-dimensional imaging volume was 800 μm × 800 μm × 350 μm (X × Y × Z). Two-photon imaging was conducted on a Leica SP5 confocal microscope. The femtosecond pulsed radiation generated by a Ti:Sapphire femtosecond oscillator (Newport, Irvine, CA) was directed into the SP5 scan head through an auxiliary optical input. An Olympus water immersion objective lens with 2-mm working distance and 0.95 numerical aperture was used for both sample excitation and collection of backscattered fluorescence signal. The signal spectrum was subdivided into four regions with three long-pass dielectric mirrors at 560-, 593-, and 665-nm wavelengths and further filtered and detected via non-descanned mode by four photomultiplier tubes.
Postimaging processing, three-dimensional-four dimensional rendering, and data analysis were performed using Volocity (PerkinElmer), Imaris (Bitplane), and MatLab (Mathworks).
Single-field velocity analyses.
In a conventional TPEM time-lapsed cell tracking, the objects under investigation are moving significantly slower than the scanning rate of the confocal microscopy. Thus velocity can be determined as a function of movement in the time between acquisitions of the individual Z-Stacks. However, if the velocity of individual particles is comparable to the scanning rate, such that the particle observed travels a significant distance during a single XY scanning, resulting in an oval appearance of round cells (the elongation artifact), then the scanning rate of a single XY frame itself becomes of determining factor with both spatial (resolution) and temporal (scan rate) parameters.
Single-field velocity analyses is as follows. If two particles (A and B) travel on the same vector, with particle A traveling at a known velocity VA, while particle B of known diameter d travels at an unknown velocity VB, such that VA > VB then the duration when particle A first reaches particle B and when it eventually evens up with it is time t. During the time t, particle A travels distance S, while particle B consequently travels distance S-d, hence the unknown velocity of particle B is defined as follows: In formula 1, VA is the known velocity of particle A, S is the distance particle A traveled while passing particle B, and d is the diameter of particle B.
This is a specific case in which both parties involved travel in the same direction, but in the single-field velocity analysis, such colinearity is disturbed by the local flow direction variations, calling for the necessity of including the two constituents, which comprise the elongation artifact in the above given formula using Pythagorean theorem. In the following formula, the vascular flow velocity Vc (velocity of cells in the vasculature) is given as: In formula 2, VS is the scanning velocity, S is the distance encompassing elongation artifact in the Y scan direction, d is the thickness of the elongation artifact (diameter of the particle), and h is the spatial deviation between the moving particle and the direction of the Y scan.
There are several prerequisites that need to be satisfied for this model to work, the Y scanning velocity should be faster than the fluid (blood) flow, and yet be close enough to produce substantial overlap between the XY scanning and the movement within the vasculature. This can be achieved simply by changing the scan frequency. The Y scan direction must also be along the vessel, close to the direction of flow.
The known parameters in this model are the imaging scanning rate set by the user during imaging, the resolution necessary to the calculated distance of the overlap between the Y scan and the moving particle in vasculature and the diameter of the particle.
In a typical vascular system, a number of factors influence fluid velocity. The complexity of the vascular labyrinth is one prominent factor that explains the heterogeneity observed in the flow velocities in different vessels. In a unit time t fluid in a vessel moving at velocity V travels distance V/t; thus within the cylinder with diameter d, the mass of the liquid with density ρ encompassing V/t distance is (ρdV)/t = (ρπr2 × V)/t or ρdV = ρπr2 × V. After branching, the fluid mass in each of the vessels is ρd1V1 = ρπr12 × V1 and ρd2V2 = ρπr22 × V2, respectively. As mass is conserved, the mass of fluid that flows during the unit time though the vessel before branching must be equal to the sum of masses of liquid in both of the branched vessels: ρd0V0 = ρd1V1 + ρd2V2 (ρπr02 × V0 = ρπr12 × V1 + ρπr22 × V2, thus r02 × V0 = r12 × V1 + r22 × V2). In the case that the diameters of all vessels are the same d0 = d1 = d2, then V0 = V1 + V2 [V1 = V2 = (V0/2)]. Thus the velocity of fluid after branching drops by a factor of two, although the diameters of the vessels remain constant. If the diameters d1 and d2 are not the same as d0, then d0V0 = d1V1 + d2V2, hence V1 and V2 are not equal (V1 <> V2), and each must be <V0. This model is consistent with the empiric measurements made in our experimental system.
Differential vascular labeling.
We sought to positively identify all vascular spaces in lymph node by examining the lymph nodes from S1P1-eGFP knock-in mice stained with WGA, previously shown to colocalize with LYVE-1 in lymphatic endothelium (43). WGA binds to the extracellular matrix surrounding both blood and lymphatic vessels, in contrast to the endothelial S1P1-eGFP expression. To visualize blood spaces within the lymph node, mice were injected intravenously with qdot 705 before node excision and WGA staining (Fig. 1A). Veins and arteries showed distinct cobblestone or linear S1P1-eGFP expression patterns, respectively, as well as obvious differences in gross vessel shape and WGA staining (Fig. 1A, Supplemental Video S1; the online version of this paper contains supplemental data). Intravital imaging of an inguinal node also demonstrated cobblestone-like expression of S1P1-eGFP on thin-walled structures containing slower moving blood, consistent with the flow rate of veins, while thick-walled vessels exhibiting linear S1P1-eGFP expression patterns contained fast moving blood, consistent with blood flow characteristics of arteries (Fig. 1B). Real-time, time-lapsed imaging of qdots in arterial blood also enabled us to visualize the shear flow and flow disturbance of erythrocyte movement at the vessel junction, as well as the contraction and the expansion of the S1P1-eGFP-rich vessel walls (Supplemental Video S2).
Selective permeability of fluorescent particles.
In the process of testing various fluorescent markers in intravital imaging (Table 1), we noticed that fluorescent particles differentially partitioned into distinct vascular compartments within the lymph node. When fluorescent particles were injected immediately before the surgical preparation for intravital imaging, capillaries (3.7 ± 0.8 μm in diameter), but not larger vessels (control 19.4 ± 6.2 μm in diameter), allowed certain particles to penetrate efficiently, while others could not. Specifically, qdot 705 and qdot 625 did not efficiently enter the capillaries, whereas qdot 585 and Alexa Fluor 594-conjugated BSA both readily distributed into these small vessels (Fig. 2A). Smaller dextran particles exhibited fluorescence profile within capillaries similar to that of larger qdots (Fig. 2A). In addition to possessing differences in vascular partitioning within the lymph node, different fluorescent particles also cleared from the blood and different rates. While large particles undergo moderate reduction in fluorescence signal in blood, the small dextrans have a faster rate of signal depletion (Fig. 2B). To determine whether increasing vascular permeability would change the rate of fluorophore depletion from the blood, we treated mice with 10 mg/kg of the S1P1-selective antagonist W146 and examined the rate of fluorophore clearance 20 min later. W146 treatment enhanced the rate of qdot extravasation, but had little effect on the diffusion of dextrans (Figs. 2B and 3, Supplemental Video S3). The lack of effect of W146 on the disappearance of dextrans from blood suggests that their diffusion is already maximal in mice that have normal blood vascular integrity.
Selectivity of particle penetration into the lymphatic system from blood.
The rapid loss of fluorescent dextrans from the blood circulation was associated with a concomitant increase in fluorescence in large S1P1-eGFP+ ducts that appeared to be lymphatic vessels, without significant accumulation in the lymph node interstitium (Fig. 4A and Supplemental Video S4). To determine whether these areas were indeed lymphatic vessels, mice were injected with Alexa Fluor 546-labeled anti LYVE-1 antibody at the base of tail, and the draining node was imaged 16 h later. A distinct special relationship of dextran signal with lymphatic sinus endothelial LYVE-1 signal was observed, with dextran fluorescence being contained within LYVE-1-positive vessels (Fig. 4B). Both dextran 10K and 3K exhibited time-dependent redistribution from inside the capillaries into the lymphatic system, with dextran 3K diffusing more rapidly than dextran 10K (Fig. 4C and Supplemental Video S5). While S1P1 antagonism increases blood vascular permeability, resulting in significant extravasation of qdots from blood into the interstitium, it did not allow for larger qdots to enter into the lymphatics (Fig. 5). Thus, intravenous injection of fluorescent, small molecular weight dextrans, together with large fluorescent qdots, differentially and simultaneously labels both the blood and the lymphatic vasculature (Supplemental Videos S6.1 and S6.2). The retention of fluorescent dextran signal within the lymph typically allows for 90 min of imaging, permitting facile visualization of lymphocyte interactions with lymph node structures evolved in cell entry and/or egress.
Determination of flow velocities within blood vessels.
Single-field analysis of fluid flow dynamics was measured as described above (Supplemental Videos S7.1 and S7.2). The analysis of blood flow within vessels before and after branching (Fig. 6) shows changes in flow velocity. In Fig. 5, the blood flows downstream into the branching vessels. The blood flow velocity before branching of the vessel (55.35 μm in diameter) is 207.4 ± 17.8 μm/s, whereas the blood flow after branching, in the vessel of 66.32 μm in diameter, is 114.04 ± 4.31 μm/s. (Fig. 5B, a single XY scan though the vessel presented in Fig. 5A). These data are consistent with theoretical predictions based on the law of conservation of mass. Further analysis of several lymph node vascular structures revealed some heterogeneity in blood and lymphatic flow rates. For accurate analysis, the XY scan rates used were adjusted in accordance with the fluid velocity to achieve substantial overlap between moving negative contrast red blood cells and the Y scan. The velocity of axial blood flow within an artery with diameter of 79.1 ± 1.6 μm in the inguinal lymph node was calculated to be 405.8 ± 29.1 μm/s, with variability reflecting the flow disturbance observed at the arterial junctions. Within the lymphatic vessels with diameter of 151.93 ± 2.63 μm, the overall lymphocyte velocity was calculated by simple tracking to be 19.5 ± 0.3 μm/min, whereas, at the areas of lymphatic valves, where increased fluid flow must occur due to narrowing of the vessel diameter down to 55.13 ± 1.73 μm, the cellular velocity increased to 96.8 ± 13.8 μm/min.
Previous methods of discriminating blood and lymphatic vasculature within the lymph node have utilized qdots or other fluorescent dyes to define blood vessels while using LYVE-1-specific antibodies to label the lymphatic vessels. While this allows for easy visualization of fluid within the blood vessels, defining lymphatic fluid has generally been more difficult, requiring surgical exposure of spinotrapezius muscle and direct injection of fluorescent particles within 100 μm of a large arteriole/venule pair, followed by electrical stimulation of the muscle to facilitate lymph flow (16). Here we demonstrate that intravenous injection of various fluorescent particles showed discriminative localization and/or selective retention within blood and lymphatic compartments without direct and uncontrolled arteriovenous-lymphatic connections.
Generally, larger particles, such as qdots, remain predominantly within the blood vasculature, with the smaller qdots efficiently entering into the smallest diameter capillaries within the lymph node. Small molecular weight fluorescent dextran particles leak out of blood vasculature in a time-dependent manner and rapidly accumulate in the lymphatic system, allowing for functional labeling of the lymphatic system. The maintenance of qdots within the blood was shown to be dependent on vascular integrity; disrupting normal vascular integrity by S1P1 chemical antagonism led to rapid depletion of qdot 705 from blood. Regardless of such increased extravasation, qdot 705 failed to penetrate into the lymphatic system, as did fluorescently-labeled dextrans.
The DVL is based on selective labeling of the fluid within the vasculature, which permits the single-field vascular fluid analysis by intravital TPEM. Single-field vascular fluid flow analysis within intact blood and lymphatic vessels provides a convenient and quick method to determine rapid fluid flow rate in vascular subcompartments and changes therein in response to pathological or pharmacological perturbations in live animals utilizing deep tissue scanning microscopy method.
The enhanced anatomical discrimination achieved with the use of DVL enables the simultaneous intravital study of several vascular compartments, including arteries, veins, capillaries, and lymphatic vessels of varying sizes. The ability to readily distinguish these vascular compartments should allow for the investigation of several other pathologies, including inflammatory disorders and cancer (1, 3–5, 12). Discrimination of fine vascular structures by DVL would enable study of the contributing factors of angiogenesis and lymphangiogenesis in vasculature-dependent processes, such as tumor growth, cellular mimicry, and intravasation, leading to metastasis. Furthermore, DVL may allow a technically difficult TPEM analysis of fluid flow rates in small peripheral vessels, as well as neovasculature and neolymphatics of a developing tumor. The simplicity and resolution of this technique make it a valuable tool in studying cellular migration, endothelial integrity, and vascular fluid flow rate.
No conflicts of interest, financial or otherwise, are declared by the author(s).
G.S. and H.R. conception and design of research; G.S. and N.L. performed experiments; G.S. analyzed data; G.S. and H.R. interpreted results of experiments; G.S. prepared figures; G.S. and H.R. drafted manuscript; G.S., S.C., P.G.-C., N.L., and H.R. edited and revised manuscript; G.S., S.C., P.G.-C., N.L., and H.R. approved final version of manuscript.
- Copyright © 2012 the American Physiological Society