Key Points
Lymph carries the “omic,” vesicular, and immune cell signature of the draining organs.
Lymph analysis provides precise and “undiluted” biochemical and cellular information.
A protocol for lymph collection from mouse and rat lymphatics is reported.
Abstract
Unlike the blood, the interstitial fluid and the deriving lymph are directly bathing the cellular layer of each organ. As such, composition analysis of the lymphatic fluid can provide more precise biochemical and cellular information on an organ's health and be a valuable resource for biomarker discovery. In this study, we describe a protocol for cannulation of mouse and rat lymphatic collectors that is suitable for the following: the “omic” sampling of pre- and postnodal lymph, collected from different anatomical districts; the phenotyping of immune cells circulating between parenchymal organs and draining lymph nodes; injection of known amounts of molecules for quantitative immunological studies of nodal trafficking and/or clearance; and monitoring an organ’s biochemical omic changes in pathological conditions. Our data indicate that probing the lymphatic fluid can provide an accurate snapshot of an organ’s physiology/pathology, making it an ideal target for liquid biopsy.
Introduction
Lymph is derived from the ultrafiltrate of plasma proteins and molecules, as well as from the extracellular fluid that bathes each parenchymal organ (1–7). As much as 70% of the plasma in the capillaries’ arterial end will extravasate through a filtration process driven by the net balance between the hydrostatic and oncotic pressures, and only a small fraction of the filtrated fluid will return to the blood circulation through the capillaries’ venule end. The majority of the ultrafiltrate will give rise to the interstitial fluid, which will be further enriched with products of tissue metabolism and catabolism, altogether generating the lymphatic fluid (8–14). The lymph cellular fraction is predominantly composed of immune cells, which after patrolling parenchymal organs, enter the lymphatic capillaries and collectors to travel to the draining lymph node (14). Apoptotic cells have also been reported in lymphatic fluid (15). Lymph vesicular and exosome fraction have also been reported in both human and mice (16, 17).
The lymphatic vascular network begins as blind-ended initial lymphatics/lymphatic capillaries in the parenchymal space of almost all tissues. These lymphatic capillaries can also form a complex honeycomb-shaped plexus in some tissues. The lymphatic capillaries usually contain large gaps in the lymphatic endothelium of their walls that serve as the primary valves and portals for immune cell entry. The lymphatic capillaries will connect to the next downstream component of the lymphatic vascular network termed precollectors that have a typically continuous endothelium layer, contain secondary valves in their lumen to minimize lymph backflow, and normally are not muscularized. The precollectors lead to the network of collecting lymphatics that are larger, contain secondary valves, and muscularized with unique lymphatic muscle that may have phasic pumping activity. These collectors anastomose in a complex arboreal fashion to eventually transport lymph to the afferent prenodal lymphatics (18, 19). The initial lymph and immune cells are thus transported down the lymphatic network into the progressively larger lymphatic vessels, which directly enter into the draining lymph node. Each node receives lymph from an anatomically defined region of the body, and all lymph passes through at least one node, but often several (18, 19).
For several reasons, there is a keen interest in cannulating lymphatic collectors and analyzing the lymphatic fluid:
1) Lymphatic cannulation and lymph collection will allow the qualitative and quantitative phenotypical analysis of lymph-carried immune cells (20). This immunological area still requires more in-depth investigation, because the majority of information on the immune cell type in the lymph still relies on ovine and bovine studies that were performed decades ago (20). Although immune cell trafficking in lymphatics has been visualized by two-photon microscopy, a more quantitative analysis of lymph-carried immune cells in peripheral lymphatics is still missing.
2) Analysis of the lymphatic fluid collected during physiological and pathological conditions will determine the molecular signature of the lymph transported to the draining node. This is an important area of investigation, which is also very much understudied and will prove very useful in the discovery of new biomarkers. Indeed, in contrast to plasma, which is not in direct contact with the cellular layers of parenchymal organs, the lymph is derived directly from the organ interstitial fluid. As such, the molecular signature of different pathological conditions is quantitatively more represented in the lymph than the plasma (3–6, 13, 21).
3) Lymph collection from different parenchymal organs will allow qualitative and quantitative analysis of the proteome, lipidome, and metabolome transported to the draining node from different anatomical districts. This is also an important analysis toward deepening our understanding of the differential “omic” composition and tissue Ag specificity of different anatomical areas.
4) Finally, for immunological studies and immunization protocols, the ability to directly inject known amounts of Ag into the lymphatic vessel can be useful for quantitative studies of the nodal immune cell populations, as compared with s.c. or i.p. injections, which result in an unknown amount of immunogen transported to the draining node.
Materials and Methods
Reagents
Equipment
The following equipment was used: stereomicroscope (SMZ1270; Nikon), glass pipette (part number GCP-40-60; Living Systems Instrumentation), 1-ml syringes (catalog no. 305620; Becton Dickinson), nylon thread (part number THR-G; Living Systems Instrumentation), scissors (ROBOZ-RS5882), forceps (Dumont no. 5 tip size, 0.25 × 0.01 mm tip size), forceps (Dumont no. 5 forceps, 0.05 × 0.01 mm tip size), Vannas Scissors (SCS-VAN-ST-2.5MM; Living Systems Instrumentation), petri dishes (catalog no. 350044; Becton Dickinson), standard 0.65-ml plastic tubes (catalog no. 87003-290; VWR International), Hamilton syringe (5 μl, 33G; Hamilton), and Leica CM3050 S Cryostat Olympus FV1200 Confocal Laser Scanning Microscope custom-made cannulation chambers (Supplemental Fig. 1). The chamber can be used for in vivo fixation for immunohistochemical analysis or for in vivo luminal injection into cannulated vessels (e.g., transfection, long-term molecule injection, etc.). Although this chamber can be used for the injection of larger vessels, nonetheless, the chamber was designed and optimized for vessels <1 mm in diameter, which are typically very difficult to cannulate. This chamber allows intraluminal cannulation while the vessel is securely connected to the glass cannula supported by the chamber. The chamber is built in aluminum and consists of a body that is 1 inch wide, 3 inches long, and 1/4 inch thick, connected to a vertical arm that is 1 inch wide, 1 inch high, and 1/4 inch thick (Supplemental Fig. 1). The aluminum should be anodized to prevent corrosion. A 5/8-inch hole is drilled in the center, 1/2 inch from the end of the horizontal arm opposite the vertical arm. A 1-inch-diameter hole is drilled in the vertical arm, where it connects with the horizontal arm to host the Luer (Supplemental Fig. 1A). Beginning 1 inch from the vertical arm of the chamber, a 1/8-inch-wide v-groove channel is cut into the centerline of the top of the horizontal arm (Supplemental Fig. 1A). The channel depth progresses from 0 inches to a depth of 3/16 inch upon connection to the 5/8-inch hole. A commercially available polypropylene female Luer-1/4-28 bulkhead-1/16-inch hose barb (catalog no. NC1346312; Fisher Scientific) is inserted into the hole drilled in the vertical arm. Polyethylene tubing (1/16 inch cat. no. 2815892301; Agilent Technologies) is connected to the hose barb on the higher end and to a glass pipette (Part number-GCP-40-60; Living System Instrumentation) at the distal end. The pipette is then secured within the channel using standard aquarium-grade silicone adhesive (cat. no. 19313650; Fisher Scientific) and extends into the 5/8-inch hole.
A round glass coverslip (25-mm diameter) is secured at the bottom of the horizontal arm using silicone adhesive to cover the bottom part of the 5/8-inch hole. This forms a reservoir that can be filled with liquid (PBS or DMEM) during the cannulation procedure to keep the vessel moist. Once completed, the inflow end of the lymphatic vessel of interest is cannulated with the glass pipette (Part number-GCP-40-60; Living Systems Instrumentation) within the chamber bath. All injections can be performed once the vessel is cannulated with its inflow end secured to the glass pipette.
Animal preparation
Female and male C57BL/6J mice (8–12 wk old) were obtained from The Jackson Laboratory and were bred in the animal facilities at the Albert Einstein College of Medicine. In some experiments, the mice were injected with LPS (2 mg/kg body weight) i.p. 16–18 h prior to lymphatic cannulation. Male Sprague Dawley rats weighing 233–558 g were housed at Texas A&M University in an environmentally controlled vivarium approved by the American Association for Accreditation of Laboratory Animal Care. All rodents were allowed full access to food and water with a 12-h light/dark cycle and constant temperature and humidity.
All animal protocols in this study were approved by the Albert Einstein College of Medicine and the Texas A&M University Laboratory Animal Care Committee and adhered to both institutional and federal guidelines. Mice were anesthetized with a combination of ketamine (80 mg kg−1) and xylazine (10 mg kg−1) injected i.p. Surgery was initiated 10–15 min after injection when the animal no longer reacted to a toe pinch. Rats were anesthetized by an i.m. injection of Innovar-Vet (0.3 ml/kg) and diazepam (2.5 mg/kg i.m. catalog no. NDC0409-3213-12; Hospira). Half-supplemental dosages of Innovar-Vet were given as needed (Innovar-Vet is a combo solution of droperidol, 20 mg/ml, catalog no. D1414; Sigma-Aldrich, and fentanyl, 0.4 mg/ml, catalog no. F3886; Sigma-Aldrich).
Proteomic analysis
Equal protein amounts from mouse lymph (10 μg), prepared in technical triplicates, were reduced, alkylated, and subjected to an “in-solution” trypsin/Lys-C/Glu-C digestion protocol (21). The digested peptides were extracted for nano–liquid chromatography (LC)/mass spectrometry (MS)/MS analysis, as previously described for human lymph samples (213–6, 22). Raw files from each technical and biological replicate were filtered, de novo sequenced, and assigned a protein identifier using PEAKS 8.0 software (Bioinformatics Solutions, Waterloo, ON, Canada) by searching against the mouse (Mus musculus) Swiss-Prot database (March 2017; 91,343 entries). The parent mass tolerance was set to 15 parts per million (ppm) using a monoisotopic mass, and the fragment ion mass tolerance was set to 0.05 Da, as described previously for human proteomic profiling (21). Protein identifications were accepted if they could be assigned with a confidence score (−10 logP [lgP]) >15 for peptides and −10 lgP >15 for proteins, which corresponds to a minimum of one peptide per protein after data were filtered for <1.0% false discovery rate (FDR) for peptides and <1% FDR for proteins (corresponding to p < 0.05). Networks, functional analyses, and biochemical and cellular pathways were generated with the ingenuity pathway analysis (IPA; Ingenuity Systems, Redwood City, CA) and applied to the 335 proteins extracted from mouse lymph (indexed in Supplemental Table I). A right-tailed Fisher exact test was used to calculate p values.
Metabolites and lipid extraction for LC/MS analysis
Metabolites were extracted from 10 μl of lymph using methanol/acetonitrile/dH2O (5:3:2). Samples were vortexed for 10 min, incubated for 20 min at 4°C, and centrifuged at 16,000 × g for 15 min at 4°C. Supernatants were collected and dried in a rotational vacuum concentrator, and the pellets were resuspended in 5% formic acid for subsequent analyses.
For lipid extraction, 10 μl of lymph samples were combined with 30 μl of chloroform/methanol (2:1 v/v). In each sample, 1 μl of glucoraphanin was injected as an internal standard. The suspension was vortexed and incubated on ice for 20 min. Samples were centrifuged for 15 min at 15,000 × g, and the organic layer was re-extracted with 2:1 chloroform/methanol. The chloroform layers were combined and dried in a rotational vacuum concentrator and reconstituted in 50 μl of isopropanol (IPA).
UHPLC-MS/MS
Lipidomics analysis
Lipid extracts were separated by Reprosil C18 column (2.0 mm × 150 mm, 2.5 μm). A binary solvent system was used, in which mobile phase A consisted of ACN/H2O (60:40), 10 mM ammonium acetate, and mobile phase B of IPA/ACN (90:10), 10 mM ammonium acetate. A gradient with flow rate of 200 μl/min was used. Samples were eluted with a linear gradient from 32% B at 0 min, 40% B at 1 min, a hold at 40% B at 1.5 min, 45% B at 4 min, 50% B at 5 min, 60% B at 8 min, 70% B at 11 min, and 80% B at 14 min. The total run time was 23 min. The source parameters were as follows: mass range: 250 3000 amu for full MS; ion source settings: spray voltage = 3 kV (both electrospray ionization positive and negative), vaporizer = 370°C, ion transfer tube = 285°C, S-lens = 45%, sheath gas = 60, auxiliary gas = 20, and sweep gas = 1; data acquisition settings: automatic gain control (MS) = 1 × 105, automatic gain control (MS2) = 1 × 105, mass range = 250 1200 Da, fixed first mass = 75 Da, and apex trigger = N/A; and initial (generic) acquisition settings (that were further optimized): dynamic exclusion = 6 s, isolation window = 1.0 Da, top-four experiments, injection time = 100 ms, resolution MS = 70,000 (full width at half maximum at m/z 200), resolution MS2 = 17,500 (full width at half maximum atm/z 200), and normalized collision energy = 20.
A pooled plasma sample was prepared as quality control to assess the stability of the instrument and ensure the reliability of the data. Quality control sample was run before and after the sequence and in every eight sample runs in the sequence to ensure the reproducibility of the data.
Data software
Lipid identification was performed with LipidSearch Software. Search parameters were the following: precursor mass tolerance = 3 ppm, product mass tolerance = 7 ppm, and M-score threshold = 3. For metabolomics analysis, raw files were exported, converted into mzXML format through MassMatrix, and then processed by Maven software (http://maven.princeton.edu/and [http://metlin.scripps.edu/download/]) (23). MS chromatograms were elaborated for peak alignment, matching and comparison of parent and fragment ions, and tentative metabolite identification (within a 2-ppm–mass deviation range between observed and expected results against the imported Kyoto Encyclopedia of Genes and Genomes database). Results were graphed with GraphPad Prism 5.01 (GraphPad Software). Statistical analyses were performed with the same software. Data are presented as the means ± SD of three biological replicates.
CSF tracer injection
CSF tracer injections were performed as previously described (13
Tissue collection and processing for imaging
Mice were euthanized with 10% Euthasol by i.p. injection. The mice were then transcardially perfused with cooled heparinized PBS (10 U/ml). Deep cervical lymph nodes and/or skullcaps were removed and fixed in 4% paraformaldehyde overnight.
Confocal imaging and analysis
Stained lymph node sections were imaged on an Olympus FV1200 confocal laser scanning microscope. Six sections were selected per animal and imaged with a 10× objective at resolution of 800 × 800 pixels. Multiple fields of view were tiled if needed. A z-stack was acquired at an interval of 5 μm through the entire section, typically eight optical sections. Within Olympus FluoView, a maximum projection was calculated and tiles were stitched if needed. Image processing and quantitative analysis were performed in the FIJI distribution of ImageJ (24, 25). For percent area coverage of lymph node sections, a binary threshold was determined visually and applied identically to all images, then as region of interest was drawn based on DAPI and Lyve1 to measure total node area, and the percent coverage of tracer in the thresholded image was measured within the region of interest. Data were collected and analyzed in GraphPad Prism 7 (GraphPad Software)
Meningeal whole mounts were also imaged on an Olympus FV1200 confocal laser scanning microscope. A 6 × 10 tile array of 640 × 640 pixels was acquired at 10× magnification in a z-stack at 5 μm intervals for 22 optical sections total to capture the entire sample throughout the tile scan. Maximum projections and tile stitching were performed in Olympus FluoView, and image processing and analysis were performed in FIJI.
Flow cytometry analysis
Mesenteric lymph nodes were harvested from untreated C57/Bl6 mice and LPS-treated C57/Bl6 mice (16 h after i.p. injection of 50 μg ultrapure LPS, Escherichia coli O111:B4; InvivoGen). Lymph nodes were pulled apart into small pieces with forceps and incubated in 400 U/ml of collagenase D (Roche) in complete media (RPMI 1640/10% FBS) for 30 min at 37°C. After incubation, a single-cell suspension was obtained by homogenizing tissue between frosted glass slides (12-550-343; Fisherbrand) and passing the cells through a 70-μm filter (08-771-2; Falcon). Cells were washed with 1 ml FACS buffer (PBS, 1% FCS, and 0.02% sodium azide) and pelleted by centrifugation in 1.5-ml safe lock tubes (022363204; Eppendorf) at 300 rpm for 5 min at 4°C before proceeding with staining. Cells collected from two to four afferent mesenteric lymphatic vessels from untreated or LPS-treated C57/Bl6 mice were combined in complete media, counted on the hemocytometer, and pelleted by centrifugation at 350 rpm for 10 min at 4°C. For staining, cells were first incubated with LIVE/DEAD Blue (L23105; Invitrogen) in PBS for 30 min on ice while protecting the cells from light, as per the manufacturer’s instructions. Cells were then washed with cold FACS buffer, pelleted by centrifugation, and incubated with 1 μg of CD16/CD32 Fc block (BD Biosciences) in FACS buffer for 5 min on ice. Combinations of fluorophore-conjugated Abs (summarized in Supplemental Table II) were added to cells in a final volume of 100 μl of FACS buffer and incubated on ice for 25 min, protected from light. Cells were then washed with FACS buffer, pelleted by centrifugation, and resuspended in cold FACS buffer for acquisition on the BD LSR II.
Solute injection into the lymphatic collectors
Lymph from anesthetized rodents was prepared for cannulation as outlined above (up to the paragraph on fat removal and lymphatic exposure). After fat removal, the vessel was maintained at its initial natural length and position to avoid any potential mechanical stress. A segment of each 0.5-cm-long prenodal lymphatic vessel (80–120 μm in diameter) was exposed by isolating the vessel from the surrounding fat tissue under a dissecting microscope (Fig. 5A, 5B). The exposed lymphatic vessel was then cut, producing two free ends, one connected with the mesenteric lymph node (nodal-end) and the other connected with the intestinal loop (intestinal-end). The intestinal-end was cannulated and tied onto the glass pipette, previously connected with one end of perfusion tubing system with a 10-1 suture (catalog no. 7707G; Ethilon, Somerville, NJ) (Fig. 5A–C). The perfusion tubing system was then mounted on the vessel preparation board (Fig. 1). The cannulating pipet, previously heparinized (heparin sodium 1:100 in Dulbecco PBS; catalog no. NOC63323-540-11; APP Pharmaceuticals), was filled with the perfusion solution. Perfusion was performed at 3 cm/water pressure. During the perfusion, the perfusion tubing system, a 1-ml syringe tube (BD309602; Franklin Lakes, NJ), was secured at 3 cm above the cannulating pipette to ensure a 3 cm/water perfusion pressure (Supplemental Videos 1, 2).
Results
Developing a method for prenodal lymph collection from different anatomical districts in mice
After shaving the fur from the designated anatomical area, an incision is made through the skin, exposing the underlying fascia and muscle layers. Collecting lymphatic vessels are identified as the pale vessels running in parallel to small arterial and vein vessels (Fig. 1A–C) under the dissecting microscope. Fat tissue, particularly abundant around the mesenteric vessels, but less so in other locations, is carefully removed (Fig. 1D), and ∼1-cm length of the lymphatic vessel is securely tied/ligated at both the proximal and distal ends using nylon thread (Fig. 1E–G). Once the lymphatic is securely ligated, the vessel is cut at the proximal and distal ends above the ligatures and then transferred to the custom-made vessel chamber (Fig. 1H, Supplemental Fig. 1), which holds a glass pipette (tip diameter of 40–60 μm) (Fig. 1I–M) connected to the polyethylene tubing attached to the male Luer stop. The chamber as well as the glass pipette are prefilled with warm 1× PBS supplemented with 1× protease inhibitor mixture solution for proteomic, lipidomic, and metabolomic studies or 1× MEM with 10% FBS for FACS analysis of immune cells. Two strategies were used to collect lymphatic fluid. In the first strategy, the knot/suture in the proximal end of the vessel is removed, and the glass pipette inserted into the vessel to flush out the lymph after removing the distal knot (Fig. 1H–M). Alternatively, the lymphatic vessel is transferred into a petri dish and, upon removal of the knots, at the proximal and distal ends, the vessel is held toward the plate, and the lymphatic fluid pressed out (Fig. 1N–P). Using the cannulation method, we were able to successfully collect ∼1 μl (1.5 ± 0.7 μl) of lymphatic fluid from four to five mouse collectors harvested from one mouse, ∼30–40% less amount (0.4 ± 0.1 μl) when using the pressing method (Fig. 1O, 1P). In terms of cellularity, 29,500 ± 10,476 cells were collected following cannulation (of four to five lymphatic collectors) and 5916 ± 4626 upon lymphatic pressing (of four to five lymphatic collectors). In conclusion, for each mouse or rat, we normally cannulate four to five lymphatics in one sitting, and this yields an amount of ∼1 μl of lymph. All metabolomic and proteomic experiments described below have been performed using 10 μl of lymph (collected from 10 to 12 mice/rats), whereas FACS analysis was performed using 2–3 μl of lymph (collected from two to three mice).
Lymphatic cannulation and lymph collection. (A) Microvascular network artery, vein, and lymphatic vessel (original magnification ×6.3). (B) A suitable mesenteric lymphatic collector identified for cannulation (original magnification ×200). (C) Magnified (original magnification ×400) portion of the same lymphatic collector. (D) By using surgical forceps, the fat tissue around the lymphatic collector has been removed. (E–G) By using fine forceps, the nylon thread has been passed underneath the lymphatic vessel to make knots on the lymphatic segment to be cannulated. (H) Proximal and distal ends of the lymphatic collector are securely ligated and the collector is placed in the cannulation chamber. (I) The knot is removed at the distal end of the lymphatic collector, and by holding the edge of the vessel using forceps, the vessel is cannulated with the pipette that was previously secured to the cannulation chamber. (J) Enlargement of (I). (K) Following cannulation, the knot at the proximal end of the vessel is cut, and pressure is applied to the cannula to flush out the lymphatic fluid. (L and M) Enlargement of (J) to show detail of cannulated lymphatic. (N–P) Lymphatic fluid being pressed out of the lymphatic collector.
Analysis of rat mesenteric lymph for fatty acid, cholesterol, and metabolites
The lacteal is the lymphatic capillary present inside the villi of the small intestine, where triglycerides, formed by phospholipids, cholesterol ester and apolipoprotein B48, are uptaken. The lacteals will merge into progressively larger lymphatic that will drain to the mesenteric node. As such, the prenodal mesenteric lymph offers an important snapshot of the lipids, which are transported from the intestine/enterocytes into the lymph. To cannulate a mesenteric lymphatic, a small loop of intestine, measuring ∼4–5 cm in length, is exteriorized through a mesenteric incision, and a section of the mesenteric loop containing the collecting lymphatic vessels is positioned in a custom-made dissection board (Fig. 2A, 2B, Supplemental Fig. 1C–E) within the field of view of a stereomicroscope and continuously suffused with 1× PBS to hydrate the exposed area (Fig. 2C). The afferent lymphatic is cannulated as previously shown in Fig. 1 and the collected lymph analyzed for lipid contents following removal of the cellular and vesicular fractions. The efferent lymphatic exiting the mesenteric node can be easily identified and harvested for postnodal lymph analysis (Fig. 2D).
Untargeted metabolomic analyses of lipids and metabolites in mesenteric lymph (A) A mesenteric loop is exteriorized and positioned in the (B) U-shaped part of the dissection board (Supplemental Fig. 1). (C) Enlargement of (B) to resolve veins and lymphatic vessels. (D) Imaging showing both afferent (arrows) end efferent (asterisk) mesenteric lymphatics. (E and F) Relative quantification of lipids, such as fatty acid, steroids, and cholesterol intermediates present in mesenteric prenodal lymph (46–48). Values represent the peak area intensity measurements. Data are reported as mean ± SD of three independent experiments. (G–I) Relative quantification of different metabolites present in the mesenteric prenodal lymph. Each metabolite was detected using untargeted metabolomic analyses (46–48). Values represent the peak area intensity measurements. Data are reported as mean ± SD of three independent experiments.
Lipids were extracted from the lymph, collected from 10 to 12 animals (10 μl) using methanol/chloroform extraction and, following chromatographic separation, analyzed by MS/MS. Unambiguous species identification was carried out by analyzing the retention time and the associated fragmentation MS2 profiles by employing the direct comparison against the same parameters, acquired from chemically defined standards.
Both fatty acids and products from cholesterol biosynthesis could be detected in the mesenteric lymph, validating the cannulation method as suitable to study efficiency of lipid absorption under different conditions (Fig. 2E, 2F). Importantly, because chylomicrons are extensively remodeled throughout their transport from the lymphatic system into the blood stream into the liver, this method allow profiling of the early chylomicrons as they are released from the gut epithelial cells into the mesenteric lymph.
As a proof of principle of the suitability of mesenteric cannulation to analyze the biochemistry of gut-absorbed products, we also analyzed hydrophilic metabolites (Fig. 2G–I). Metabolites from the glycolysis pathway, urea cycle, and citric acid cycle could all be detected in the collect lymph (Fig. 2G–I). Unambiguous species identification was carried out by analyzing the retention time and the corresponding MS2 fragmentation profiles using the direct comparison against the same parameters acquired from chemically defined standards.
Analysis of mouse lymph draining to the deep cervical nodes
The glymphatic system is a CNS clearance conduit, which consists of a para-arterial space functioning as a waste clearance pathway to remove fluid and extracellular proteins/molecules from the interstitial compartments of the brain and spinal cord (26–28) Astrocytes extend their processes to ensheathe the brain’s vasculature and clear the interstitial fluid and waste product from the perivascular glymphatic space (29, 30). Additionally, bona fide lymphatic vessels have been shown to line the dural sinuses and meningeal membranes (Fig. 3A) (31, 32). As previously determined, these and other brain-draining lymphatics drain transported lymph directly to the cervical nodes (33–39). Indeed, we confirmed that fluorescent OVA, directly injected into the cisterna magna, could be readily visualized in the deep cervical nodes after 1 h (Fig. 3B, 3C). Because the ability to probe the lymph draining from the brain could be used as a liquid biopsy for brain cellular/metabolic activities in both physiological and pathological conditions, we set up a method to cannulate afferent lymphatic draining to the deep cervical nodes (Fig. 3D, 3E). Proteomic analysis of afferent lymph confirms that soluble proteins, directly injected in the cerebral fluid, can drain to the afferent lymphatic entering the deep cervical node (Fig. 3B). It also indicates that the lymphatic fluid entering the cervical nodes can convey a proteomic sampling of the brain interstitial fluid to the nodal immune cells (Fig. 3F).
Structure of meningeal lymphatics and drainage of cerebrospinal fluid to deep cervical nodes. (A) Representative confocal image showing immunostained lymphatic vessels along the transverse sinus (vertical) and superior sagittal sinus (horizontal) in the mouse meninges. Lyve1+ lymphatic vessels run along the sinuses and drain some of the tracer delivered into the cerebrospinal fluid via injection to the cisterna magna. (B) Representative confocal image showing accumulation of CSF tracer (OVA-conjugated Alexa Fluor 647) in the deep cervical lymph nodes 1 h after cisterna magna injection. (C) Quantification of area coverage by tracer in deep cervical lymph nodes 1 h after intracisternal injection (27.00 ± 8.34, n = 8 mice). (D) Imaging of an afferent lymphatic entering the deep cervical node (E) enlargement of (D) and (F) TAU peptides derived from tryptic digest of the collected mouse cervical lymph proteome. (G) Leg preparation before harvesting the afferent popliteal lymphatic. (H) imaging of the afferent popliteal lymphatic. Harvesting and cannulation is performed as described in Fig. 1.
Ag injection s.c. is a common immunization protocol; however, the amount of Ag reaching the draining node is not easily quantifiable. As such, lymph collection from the vessel draining the site of injection into the regional node can prove very useful for quantitative immunological studies. Lymphatic preparation from a mouse prenodal collector draining the interstitial fluid from the skin of the foot and leg skin is also reported (Fig. 3G, 3H). Lymphatic harvesting, cannulation, and lymph collection from the skin are identical to what is reported in Fig. 1.
Sampling of mouse mesenteric lymph in an inflammatory colitis model
In the next series of experiments, we tested whether the sampling of the lymphatic fluid would provide a snapshot on any changes in the interstitial fluid/lymph following an inflammatory condition. To this goal, we set up the dextran sulfate sodium (DSS) colitis model. C57Bl6 mice are either given access to sterilize water or water supplemented with 40–50 kDa DSS for 7 d. Following treatment, mesenteric lymphatics are cannulated (as in Figs. 1, 2) and lymphatic fluid analyzed by proteomic analysis after removal of the cellular and vesicular fractions (Fig. 4A). A volcano plot combining data from five independent experiments indicates several proteins upregulated above the statistical significance (Fig. 4B). Pathway analysis of the control and DSS proteome (Fig. 4C, 4D) indicates upregulation of several proinflammatory pathways, including immune cell trafficking (p value 1.04 × 10−26 and z-score 4.94), angiogenesis-related pathways (p value 2.58 × 10−21 and z-score 3.75), APC activation (p value 3.15 × 10−20 and z-score 3.20 Fig. 4E), inflammation (p value 1.72 × 10−19 and z-score 2.78), chemotaxis (p value 1.47 × 10−17 and z-score 2.70), carbohydrate metabolism (p value 5.47 × 10−17 and z-score 2.42), amino acid metabolism (p value 5.32 × 10−16 and z-score 2.21), proteolysis (p value 1.06 × 10−13 and z-score 2.48), and oxidative stress (p value 1.06 × 10−13 and z-score 2.48) (Fig. 4C, 4D). Additionally, molecules associated with MHC class I and MHC class II pathways, NFκΒ, STAT, and ERK activation are also upregulated in the DSS samples as compared with controls (Fig. 4C, 4D, Supplemental Table I). The data indicate that lymph sampling can provide a molecular signature of organ’s pathological conditions.
Proteomic profiling of mouse mesenteric lymph. (A) Heat map of representative biological quadruplicates of mesenteric lymph proteins as identified and quantified by label free proteomic analysis (all proteins are reported in Supplemental Table I). Only proteins that passed a selected significance statistical threshold (ANOVA, p < 0.05 and FDR < 1% for protein and peptide expression) are represented in the heat maps. (B) Volcano plot indicating statistical significance in the fold changes observed between the healthy and DSS lymph proteome. (C and D) IPA predicted the canonical biochemical pathways of the mouse lymph proteome (p < 0.05, FDR < 1% for the proteins reported in Supplemental Table I). The major molecular and cellular networks predicted by IPA to be associated with the DSS mouse lymph proteome depicts a wide variety of pathways associated with an inflammatory state. (E) View of IPA-predicted dendritic cell activation pathway. ROS, reactive oxygen species.
Sampling of rat pre- and postnodal lymph
The ability to sample the pre- and postnodal lymphatic fluid draining to each node is fundamental to the quantitative analysis of the nodal clearance capacity as well as measurement of fluid homeostasis throughout the body (22). To analyze the efficiency of nodal clearance, we established a controlled system in which known amounts of fluorophore-labeled proteins, fluorescent dextran beads, or fluorescent bacteria are perfused into a cannulated prenodal second-order mesenteric afferent lymphatic collector (Fig. 5A–C). The mesenteric vascular hierarchy, used for both blood and lymphatic vessels, indicates that a first-order vessel is the afferent prenodal vessel that enters the mesenteric lymph node and a second-order vessel is one branch upstream from the prenodal afferent lymphatic. The cannulation was also performed using methylene blue to confirm no lymphatic fluid/dye leakage for the full duration of the perfusion (Fig. 5C). All tracers are injected at an intraluminal pressure of 3 cm H2O to mimic the average basal intralymphatic pressure (40–45) (Fig. 5D, Supplemental Videos 1, 2). Fifteen to sixty minutes after infusion, a sample of postnodal lymph is collected from the common mesenteric efferent lymph trunk. Quantitative fluorescence measurement details the efficiency of the lymph node in clearing the fluorophore-labeled HEL, fluorescent dextran beads, or fluorescent bacteria from the afferent lymph (Fig. 5E–J). These data demonstrate that the cannulation protocol described in this study can be employed to quantitatively analyze pre- and postnodal lymph.
Delivery of fluorochrome-labeled proteins in the efferent lymphatics. (A) A mesenteric afferent lymphatic is exposed and prepared as in Figs. 1 and 2. The nodal end of the isolated mesenteric lymphatic is cannulated and tied with a suture onto the glass pipette, previously connected with a perfusion tubing system. (B) A solution of fluorochrome-labeled BSA is perfused into the prenodal lymphatic (at 3 cm water pressure) through the glass pipette using lymphatic physiological pumping activity. (C) A solution of methylene blue is perfused into the prenodal lymphatic (at 3 cm water pressure). Brightfield microscopy shows both the efferent lymphatic and the draining lymph node. (D) Schematic of the lymphatic perfusion. (E) Fluorescence trace and bar graph of HEL/FITC protein as injected in the prenodal lymph and as detected in the postnodal lymph after 30 min. (F) Bar graphs depict the percentages of fluorochrome-labeled protein detected in the prenodal lymph and collected in the postnodal lymph. Measurements were collected from four separate rats and statistical analysis performed using a two-tailed t test; average and SD are reported. (G) Fluorescent dextran beads (2 μm), as injected in the prenodal lymph and retrieved in the postnodal lymph. (H) Bar graphs depict the percentages of fluorochrome-labeled beads detected in the prenodal lymph and collected in the postnodal lymph. Measurements were collected from four separate rats and statistical analysis performed using a two-tailed t test; average and SD are reported. (I) Fluorescent Staphylococcus aureus, as injected in the prenodal lymph and retrieved in the postnodal lymph. (J) Bar graphs depict the percentages of fluorochrome-labeled bacteria detected in the prenodal lymph and collected in the postnodal lymph. Measurements were collected from four separate rats and statistical analysis performed using a two-tailed t test; average and SD are reported.
Analysis of mouse lymph circulating immune cells
A pivotal function of the afferent lymphatic is to transport immune cells, patrolling peripheral organs to the draining nodes. To this day, very scarce information exists on the immunophenotyping of the immune cells present in the efferent lymph (20). To determine whether the amount of lymph collected from afferent lymphatic cannulation is sufficient to detect immune cells, the cell pellet retrieved from 10 μl of collected lymph was stained with a panel of T cell, B cell, and dendritic cell Abs (Supplemental Table II). Phenotyping was performed both on lymph collected under nonstimulatory conditions or following LPS injection (50 μg) 16 h prior collecting the lymph. Between 50 and 80% of collected cells were of bone marrow origins (CD45+) in lymph collected from unstimulated or LPS-stimulated conditions, respectively (Fig. 6A). The presence of tissue-specific cells and apoptotic parenchymal cells circulating in the afferent lymph has been reported before (20), which could account for the CD45-negative− population. For each lymphatic collector, between 15,000 and 25,000 cells could be harvested (Fig. 6B). Of the CD45+ cells, around 30–40% were B cells, 15–20% T cells, and the rest non–T/B cells (Fig. 6C). Of the T cells subpopulations, 70–80% were CD4+ T cells and the rest CD8+ T cells both under unstimulated and LPS-stimulated mice (Fig. 6C). Dendritic cells were <10% of the non–T/B cells in the lymph of unstimulated mice; however, they represented almost 80% of the non–T/B cells in the lymph of LPS-stimulated mice (Fig. 6C).
Flow cytometry analysis of cells isolated from murine mesenteric lymphatics. (A) Representative flow cytometry plots and gating strategy to detect live CD45+ hematopoietic cells in mesenteric afferent lymphatic vessels (red dot plot) in untreated (top panel) and LPS-treated (16 h; bottom panel) mice. Dot plots (red) are overlaid with contour plots of mesenteric lymph node cells (MLN; gray). Cell frequencies (in percentages) of indicated parent populations in the mesenteric lymphatic vessels are denoted on the plots in red. (B) Number of cells per volume (1 μl) of lymph fluid collected (untreated, n = 4 mice; LPS-treated, n = 4 mice). (C) Summary of cell frequencies (right) of indicated immune cell populations [gated as in (A)] detected in mesenteric lymphatic vessels isolated from untreated (gray, n = 4) and LPS-treated mice (red, n = 4). Live cells (Live), doublet discriminated cells negative for LIVE/DEAD stain; T cells, CD45+CD3+CD19−; B Cells, CD45+CD19+CD3−; non–T/B cells, CD45+CD3−CD19−; CD8 T cells, CD45+CD3+CD19−CD8+; CD4 T cells, CD45+CD3+CD19− CD8−; dendritic cells (DCs), CD45+CD3−CD19−CD11c+; nondendritic cells (Non-DCs), CD45+CD3−CD19−CD11c−CD11b+. Bars, mean ± SEM. Statistics, Student t test (Holm–Sidak method). *p < 0.05.
Altogether, our results indicate that the lymph harvested even from one collector is sufficient to perform quantitative FACS immune profiling in both physiological and inflammatory conditions.
Discussion
There are several advantages in being able to cannulate the lymphatic system:
The lymph is the fluid that is directly derived from the interstitial fluid, which, in contrast to the blood, bathes each organ’s cellular layers. Thus, the lymph is the best medium to analyze the metabolic/catabolic product secreted from the parenchymal cells. Unlike the blood, in which metabolic and catabolic products from all organs are combined, the lymph can actually provide the organ’s specific metabolic signature.
Cell-released exosomes are collected in the interstitial fluid that forms the lymph. Because the plasma contains a mixture of exosomes derived from different organs, the lymph is the best media to analyze tissue-specific exosomes.
The lymph carries the proteomic signature of the organs from which it originates. The self- and nonself-proteomes draining into the collecting node transport critical immunological information to the nodal APCs.
Tissue-bound dendritic cells, which collect immunological information from the organ of origin, circulate to the draining node through the afferent lymph.
With this protocol, we were able to successfully collect ∼1 μl of lymphatic fluid from each mouse (harvesting from a total of four to five collectors) and 2–3 μl in rats using the cannulation procedure (Figs. 1–5). The collected amount is ∼30–40% less when using the pressing method. In each experimental setting, we would normally collect lymph from four to five lymphatic vessels from each mouse/rat. Importantly, we have demonstrated that this protocol is very useful for lymph analysis (Figs. 2–6). Indeed, 10 μl of lymph was sufficient to perform either a comprehensive proteomic (Fig. 4, Supplemental Table I), lipidomic (Fig. 2), or metabolomic analysis (Fig. 2). The performed proteomic analyses (Fig. 4, Supplemental Table I) significantly overlap with what was previously published from our laboratory and others (7, 9, 11, 21). Additionally, the metabolomics studies (Fig. 2) indicated that even few microliters of lymphatic fluid can identify lipids and metabolites. Thus, overall, our analysis demonstrated that the amount of lymph collected using this cannulation protocol would be a very useful tool for dissecting the molecular signature of the lymph transported to the draining node and for biomarker discovery in different pathological conditions.
In addition, the collected microliters were also sufficient for FACS analysis (Fig. 6). The phenotypic analysis of lymph circulating immune cells remains elusive. Indeed, only a few detailed studies of immune cells in bovine and ovine peripheral lymph have been reported (20). As such, this protocol will allow for the qualitative and quantitative analysis of lymph-carried immune cells collected during physiological and pathological conditions. Our data indicate the following: 1) the majority of the cells in the lymphatic vessels of naive and LPS-treated mice are hematopoietic-derived cells, particularly T and B cells, and 2) inflammatory stimuli (LPS) increase the number of circulating immune cells in general and dendritic cells in particular (Fig. 6). Thus, lymph collection will prove very useful for the phenotypic and functional analysis of patrolling immune cells and cells circulation to the draining nodes.
Considering the significance of protein biomarker discovery in different pathological conditions, there is a great interest in liquid biopsies. As such, peripheral blood is the most commonly analyzed biological fluid for protein, lipid, carbohydrate, and electrolyte evaluation in physiological and pathological conditions. However, in peripheral blood, the tissue biomarkers are greatly diluted, which can hinder early diagnosis. Indeed, tissue-specific Ags have a 100–500-fold lower concentration in the blood than in the lymph. This is due to the fact that the lymphatic capillaries have open ends and directly collect the interstitial fluid, which is present among the cellular layers of each parenchymal organ. The merging of tissue capillaries into a lymphatic collector preserves the tissue’s omic signature, which can be sampled following cannulation of the collector. As such, we anticipate that the protocol will be very valuable for investigators interested in analyzing lymphatic contents in physiological and pathological conditions including cancer, metabolic disorders, and acute or chronic inflammatory diseases.
Disclosures
The authors have no financial conflicts of interest.
Footnotes
This work was supported by National Institutes of Health Grants AG045223 and AI137198 (to L.S.) and 1U01HL123420 (to D.C.Z. and A.A.G.).
The online version of this article contains supplemental material.
Abbreviations used in this article:
- DSS
- dextran sulfate sodium
- FDR
- false discovery rate
- IPA
- ingenuity pathway analysis
- LC
- liquid chromatography
- MS
- mass spectrometry
- ppm
- part per million
- UHPLC
- ultra-high performance LC.
- Received April 1, 2019.
- Accepted August 8, 2019.
- Copyright © 2019 by The American Association of Immunologists, Inc.