Investigating the Role of Toll-Like Receptors in Models of Arthritis

Rheumatoid arthritis (RA) is a chronic autoimmune disease characterized by persistent synovial inflammation leading to tissue destruction and progressive loss of joint function. Here we describe two methods that can be used to assess the contribution of toll-like receptors (TLRs), and their potential ligands, to RA pathogenesis. We focus on the antigen-induced model of murine arthritis and human synovial tissue explant models. Both enable detection of TLR, and TLR ligand, expression, as well as investigation of the effect of inhibition of these molecules. Each offers a unique insight into disease; with murine models allowing kinetic analysis in live animals and explant models allowing examination of inflamed human tissue, which together can help us to dissect the role of TLRs in the onset and progression of RA.


Introduction
The hallmarks of rheumatoid arthritis (RA) include synovial inflammation and destruction of joint cartilage and bone; mediated by persistent production of pro-inflammatory cytokines and matrix metalloproteinases (MMPs). Compelling evidence supports a role for TLRs in contributing to the aberrant inflammatory response observed in RA. On one hand, ex vivo and in vitro studies using human tissue and cells have shown expression and functionality of specific TLRs in RA joints. On the other hand, many in vivo experimental models of arthritis have demonstrated TLR ligand requirement for disease induction as 1 1 1 1,* 1 % disease incidence; (4) onset of disease occurs at a defined time, facilitating kinetic studies; (5) disease severity can be controlled by the dose of intra-articularly injected antigen; (6) episodes of exacerbation and remission occurring in RA patients can be mimicked by controlled rechallenge with antigen. However, both animal models of RA progress significantly more rapidly than the human disease and are characterized mostly by acute inflammatory responses, necessitating complementary approaches to examine some aspects of disease.
Ex vivo models of the human disease that consists of the culture of cells from RA synovial membranes from patients undergoing joint replacement surgery can be helpful in obtaining a picture of late stage human disease. Originally described by Brennan et al. [ 35 ], this system led to the discovery that arthritic joints have elevated levels of pro-inflammatory cytokines [ 36 ], and provided the rationale for testing TNF-α blockade in RA [ 37 ]. Here, we provide a detailed description of the protocol to isolate, phenotype and culture RA membrane cells, which represent a mixed population of all synovial cell types that spontaneously produce high amounts of pro-inflammatory mediators. Furthermore, we describe the protocol for TLR activation, inhibition and expression as well as cytokine level quantification. A major advantage of this model is that since the cells continue to release cytokines in short term culture (~up to 3 days), presenting an opportunity to study pathological processes that drive inflammation and allowing the study of disease intervention and efficacy of novel therapeutics. Two prominent disadvantages are that the nature of this model requires disruption of intact tissue and therefore certain important cell contact dependant processes. Secondly, not all cell populations from the intact synovium are fully represented following dissociation (e.g., neutrophils and endothelial cells). The current protocol may be modified to favor retention of additional subsets.   Table 1 ) with levels of LPS < 10 pg/ml. Store at −80 °C. 1. Preparation of antigen. Transfer 10 ml of FCA in a small, sterile plastic vial. In a separate vial, make an mBSA stock solution by dissolving 20 mg mBSA in 9 ml of sterile water and then add 1 ml of 10× PBS to obtain a 10 ml of a 2 mg/ml mBSA solution. Aspirate this solution with a 10 ml syringe, attach a 23 G × 1 in. needle on the syringe, press the syringe plunger to inject with force the solution in the 10 ml of CFA previously prepared. Emulsify this white oil-water emulsion by repeatedly aspirating and flashing with a 1 ml syringe to homogeneity or until the emulsion is thick enough to remain in the vessel when inverted (see Note 2).
3. Immunization of mice (see Note 5). Gently shave the rumps of the mice with electric clippers to completely remove the fur at the base of the tail. Inject 100 μl in total of the emulsion intradermally at two sites at the base of the tail using a 1 ml syringe with a 23 G × 1 in. needle (see Notes 6 and 7).
4. Induction of arthritis. Seven days later, mice are sedated (see step 2) and unilateral arthritis is induced by intra-articular injection of mBSA in PBS into the right knee joint. Prepare a 40 mg/ml mBSA stock solution using sterile water and dilute it 1:2 (vol:vol) in PBS. Before injection, http://eproofing.springer.com/books/printpage.php?token=hcKiChH62c8ED0xcrdXaZX-PfMgJgXJPl8cIy3jyrSo sterilize the solution using a 0.20 μm syringe filter. Gently shave the right leg of the mice with electric clippers to completely remove the fur around the knee joint.
Inject 200 μg of mBSA in PBS (10 μl total volume) into the intra-articular space of the knee joint (see Fig. 2 and Note 8). Control mice are injected intra-articularly with 10 μl of PBS, while the contralateral, left joint functions as untreated control.

Fig. 2
Intra-articular injection into the cavity of the knee joint. Syringe positioning to locate correct injection site (A) and insert the needle into the cavity of the knee joint (B) 5. Monitor mice every day after the intra-articular injection. Knee swelling should be evident 24 h after the injection (see Note 9).

TLR-Induced Arthritis
1. Preparation of DAMP. In order to avoid TLR activation by endotoxin contamination of endogenous TLR ligands, a LAL test should be carried out to quantify endotoxin levels in recombinant DAMPs according to the manufacturer's instructions (LAL assay QCL-1000™).
Only preparations with endotoxin levels <1 EU/ml should be injected in the mouse. Dilute DAMP of choice (see Table 1 ) in sterile PBS to the desired concentration (e.g., 0.1 mg/ml FBG).
2. Induction of arthritis. Mice are sedated (see Subheading 3.1.1 , step 2) and unilateral arthritis is induced by intra-articular injection of the DAMP of choice into the right knee joint. Gently shave the right leg of the mice with electric clippers to completely remove the fur around the knee joint. Inject 1-10 μg of DAMP in PBS (10 μl total volume) into the intra-articular space of the knee joint (see  Table 1 ) is shown in Fig. 3 as http://eproofing.springer.com/books/printpage.php?token=hcKiChH62c8ED0xcrdXaZX-PfMgJgXJPl8cIy3jyrSo an example.

Fig. 3
Sections of the knee joints of wild type (A-B) and TLR4−/− (C) mice 3 days after intra-articular injection of PBS (A) or 1 μg FBG (B-C) stained with H and E. Sections show inflammatory cell infiltration, mild synovitis, and pannus formation exclusively in wild type mice injected with FBG [ 24 ] 3. Monitor mice every day after the intra-articular injection. Knee swelling should be evident 24 h after the injection (see Note 9).

Immunohistochemistry
1. Mouse knee joints are excised 1, 3 or 7 days after intra-articular injection by removing the skin and subcutaneous tissue, cutting longitudinally the muscles that cover the front and side of the femur and the side of the tibia and cutting the femur 1-2 mm above the knee joint and the tibia 1-2 mm below the patella. Carefully remove muscle tissue in excess without damaging the knee joint.
2. Fix the freshly isolated knee joints in 10 % (vol/vol) neutral buffered formalin for 48 h at room temperature.
3. Decalcify the knee joints in 10 % EDTA/PBS for 4 weeks, changing the solution three times per week (see Note 10).
4. Embed the tissue in paraffin wax using a cycle on an automatic tissue processing machine (e.g., a representative cycle is ethanol 70 % for 90 s at 40 °C, 5× ethanol 100 % for 90 s at 40 °C, 3× xylene 90 s at 40 °C and paraffin for 90 s at 63 °C).
5. Cut coronal tissue sections at a thickness of 4 μm at seven depths throughout the joint, 80 μm apart. Mount sections onto glass microscope slides made to ensure firm electrostatic attraction of paraffin sections (e.g., Superfrost™ Plus Slides). Let the tissue sections air-dry for 1 h or until dry and place them in an oven at 60 °C overnight. This will help with adherence of the sections to the http://eproofing.springer.com/books/printpage.php?token=hcKiChH62c8ED0xcrdXaZX-PfMgJgXJPl8cIy3jyrSo slides.
7. H and E staining: place slides in hematoxylin for 6 min and 30 s and rinse under tap water for 2 min; dip slides 2-3 times in 0.3 % acid alcohol for 40 s and rinse under tap water for 2 min; dip slides 8-10 times in ammonia water for 1 min and rinse under tap water for 1 min; finally, place slides in eosin for 1 min and 45 s and rinse under tap water for 3 min. See Fig. 4a as an example of the results that can be produced. 10. Remove slides from xylene and apply DPX mountant and coverslips using an automated coverslipping machine (see Note 11).
11. Histological analysis of H and E and safranin-O stained sections is performed using a light microscope, a camera and image acquisition software (e.g., BX51 microscope, Olympus; 18.2 Color Mosaic camera, Diagnostic Instruments; Spot Advanced or DP Manager acquisition software) (see Note 12). Score histopathological changes using the following parameters as http://eproofing.springer.com/books/printpage.php?token=hcKiChH62c8ED0xcrdXaZX-PfMgJgXJPl8cIy3jyrSo previously described [ 39 ]. Grade the influx of inflammatory cells into synovium (infiltrate) and the joint cavity (exudate) with an arbitrary scale from 0 (no inflammation) to 3 (severe inflammation). Determine chondrocyte death as the percentage of cartilage area containing empty lacunae in relation to the total area and cartilage surface erosion as the amount of cartilage lost in relation to the total cartilage area. Assess bone destruction in ten different areas of the total knee joint section and grade it on an arbitrary scale of 0 (no damage) to 3 (complete loss of bone structure). Calculate the mean score for each mouse in an experimental group by averaging the histopathological scores in at least five section depths per joint.

Real-Time PCR
3.3.1. RNA Preparation 1. Excise mouse knee joints at day 1, 3, or 7 after intra-articular injection (see Note 13), carefully removing muscle tissue, and immediately freeze them in liquid nitrogen. Maintain tissues at −80°C until pulverization is carried out using a BioPulverizer (see Fig. 1 ) following manufacturer's instructions.
2. Lyse pulverized tissue by adding 700 μl of RLT buffer (included in the RNeasy mini kit) per sample.
3. Extract and purify total RNA according to manufacturer's instructions (RNeasy Mini kit) including homogenization of tissue lysate with a shredder (QIAshredder Homogenizer) (see Note 14).
4. Assess total RNA concentration and purity by measuring the sample absorbance at 260 nm and the ratio of absorbance at 260 and 280 nm respectively, using a spectrophotometer.  Fig. 5 . 3. Lyse pulverized tissue by resuspending it in ice-cold T-PER tissue protein extraction reagent containing protease inhibitor in a 1.5 ml Eppendorf tube at a final concentration of 100 mg/ml. 10. Remove the isobutanol, rinse the top of the gel with deionized water and carefully remove any residual water with Whatman paper.
11. Prepare the stacking gel by mixing 0.625 ml 4× stacking buffer with 0.325 ml acrylamide-bis solution, 1.5 ml water, 12.5 μl APS, and 2.5 μl TEMED. Pour the stacking gel on top of the separating gel, insert a comb and let polymerize at room temperature for 30 min.
12. Remove the comb, rinse the wells with deionized water and fill them with 1× running buffer.
13. Place the gels in the tank filled with 1× running buffer and load samples for analysis of tenascin-C (or other endogenous TLR ligands). Include one well with pre-stained molecular weight marker and one with mouse embryonic fibroblast cell lysates as positive control.
14. Finalize the assembly of the gel unit and connect to the power supply. Run the gel at 100-200 V for 1-1.5 h.

Immunoblotting
1. Transfer the protein samples that have been separated by SDS-PAGE to nitrocellulose membranes using a semi-dry blotting system (Bio-Rad).
2. Soak two sponges, two sheets of Whatman paper and one nitrocellulose membrane per gel in icecold 1× transfer buffer.
3. Disconnect the gel unit from the power supply and disassemble it. Remove and discard the stacking gel. Lay sponges, paper, membrane and gel on the surface of the blotting device in the following order to form a sandwich: one sponge, one sheet of Whatman paper, one membrane, one gel, one sheet of Whatman paper, and one sponge. PBS-T at room temperature on a shaker set at 70 rpm.
8. Incubate the membrane with a 1:50,000 dilution of the secondary antibody in antibody diluent for 1 h at room temperature on a shaker set at 70 rpm.
9. Discard the secondary antibody and wash the membrane three times for 10 min each with 10 ml PBS-T at room temperature on a shaker set at 70 rpm. 10. Mix 1 ml of each solution of the ECL substrate and add it immediately to the membrane and incubate for 2 min at room temperature on a shaker set at 70 rpm.
11. Drain excess ECL substrate and place the membrane between two transparent plastic leaves in an X-ray film cassette and expose to an X-ray film for 1 min at a start. Increase or decrease exposure time if signal is too weak or too strong, respectively.
12. Strip the membrane by incubating it with stripping buffer for 15 min at room temperature on a shaker set at 70 rpm.
13. Remove the stripping buffer and store it at 4 °C for subsequent use. Block the membrane twice for 5 min with blocking buffer at room temperature on a shaker set at 70 rpm.
14. Discard the blocking buffer and replace it with 1:200 dilution of the anti-actin antibody in antibody diluent. Incubate overnight at 4 °C on a shaker set at 70 rpm.
15. Discard the primary antibody and wash the membrane three times for 10 min each with 10 ml PBS-T at room temperature on a shaker set at 70 rpm. 16. Incubate the membrane with a 1:5000 dilution of the secondary antibody in antibody diluent for 1 h at room temperature on a shaker set at 70 rpm.
17. To detect actin, follow steps 9-11 as in Subheading 3.4.2 . An example of the results obtained is showed in Fig. 6 .

Fig. 6
Tenascin-C protein levels in the joint were assessed 1, 3 and 7 day after intra-injection by western blot analysis of equal amounts of total joint lysate with a rat monoclonal antibody to mouse tenascin-C (TN-C). Mouse embryonic fibroblast cell lysate was used as a positive control (+) and non-injected mice as a negative control (c  Fig. 7 ).

Fig. 7
To test the potential for cleavage of cell surface markers by enzymes commonly used for synovial tissue digestion, PBMCs were treated with a range of collagenases for 90 min and staining intensity for CD4 (top row CD4-APC-H7) and CD56 (lower row CD56-PE) was compared to that obtained following no enzyme treatment. To further increase the surface area for optimal digestion, the tissue pieces are forcibly pressed using the flat end of a 5 ml syringe plunger for 2 min.
2. Using the forceps, scrape the finely chopped synovial tissue from the petri dish into a small conical flask. Wash the petri dish with 2 ml RPMI/enzyme mix using a Gilson pipette, there should be no visible tissue remaining on the plate. Mix tissue/enzyme thoroughly. Add the remaining RPMI/enzyme mix to the flask and transfer the flask to a shaking water bath at 175 strokes per minute for 1-2 h at 37 °C. At the end of first hour, check the tissue. It should appear "gloopy" or "stringy." The pinkish color of RPMI in media should discolor. If the tissue is still visibly clumpy, continue incubating for up to 60 min. It is critical not to go over 2 h, as this will begin to affect viability. Swirl every 20-30 min by hand.
3. Transfer 20 ml of ice-cold medium (RPMI with 10 % FBS) to the flask containing the digested tissue to terminate the digestion.

Sieve cells through sterilized beaker covered with Microsieves 200
μm material allowing the digested synovium to pass through. It is important to aggressively force through the tissue clumps using a Corning Cell lifter, again pressing with the rubber end of a plunger from a 2 ml syringe.
5. Wash the flask with a further 10 ml of ice-cold medium and rinse membrane covered beaker. Remove any remaining tissue from the Microsieve membrane and place in 10 ml of medium (RPMI containing 10 % FBS) in a 10 cm tissue culture dish and leave overnight in an incubator; cells still remaining within the tissue will egress overnight from the tissue and adhere to plastic and can be further passaged to yield synovial fibroblasts.
6. Transfer filtered cells from beaker into a 50 ml falcon tube, add medium (RPMI containing 10 % FBS) to make up to 50 ml and spin at 360 × g for 10 min. Resuspend the pellet with 10 ml Red Blood Cell Lysis buffer, incubate for 5 min at room temperature, terminate the reaction by the addition of 40 ml RPMI, and spin for 5 min at 360 × g. Resuspend cells in 50 ml RPMI and spin for 5 min again. If the pellet is not clean (has fibers) sieve cells through a 70 μm cell strainer to get single cell suspension and spin at 360 × g for 5 min. Cells are now ready for phenotype analysis or cell culture.   5. Add antibody cocktails (see Table 2 ) in a total volume of 20 μl directly on top of FcR block without washing. Incubate at 4 °C for a further 20 min.

FACS Analysis
Using antibody panels A and B (see Table 2 ), it is possible to identify the main immune cell types within http://eproofing.springer.com/books/printpage.php?token=hcKiChH62c8ED0xcrdXaZX-PfMgJgXJPl8cIy3jyrSo the typical RA synovial tissue. After exclusion of dead cells and debris, first gate on CD45+ and CD45populations to facilitate identification of further subsets (see Fig. 8 and Notes 22-25). CD45+ cells usually make up the bulk (approx. 60 %) but there is a significant degree of heterogeneity in the cellular composition of RA synovial membrane tissue from each RA donor, as demonstrated in Fig. 9 , with CD45+ cells ranging between 20 and 80 % of all viable cells recovered. The remaining CD45-population is largely CD90+, suggestive of a fibroblast-like phenotype (see Fig. 9a ), and comprises approximately 20 % of all viable cells. Other major cell subsets within the CD45+ gate (hematopoietic in origin) are depicted in Fig. 9b . These are mostly macrophages (CD14+, HLADR+) with a lesser proportion of CD3+ T cells (1-20 %). CD11c+, HLADR+, CD14− cells, likely conventional dendritic cells, are few in number (<5 %) and neutrophils (CD14−, CD15+) even less (<2 %). This may reflect the "dampened" biological activity of the tissue collected at joint replacement surgery (i.e., end stage RA), where the tissue may be more quiescent after long-term chronic inflammation. To confirm this, it will be of interest to compare cellular composition of synovial biopsy tissue collected in early onset RA to late stage surgery.  2. To assess TLR ligand activity in this system, incubate cells with 50 μl of TLR ligand agonists or antagonist (as previously described [ 5 ]). For example stimulation with FSL-1 10 ng/ml, poly(I:C) HMW 20 ng/ml, LPS 10 ng/ml, flagellin 10 ng/ml, R847 1 μg/ml, R848 1 μg/ml, or ssDNA/LyoVec 500 ng/ml can be used in order to modify spontaneous cytokine production.
3. Harvest cell supernatants 24-72 h later. We routinely measure cytokines after 48 h and store supernatants at −20 °C until cytokine assays are performed.

Adenoviral Infection
Using AdEasy adenoviral constructs of the TLR adapter molecules MyD88 and Mal, we have previously shown that these molecules control spontaneous cytokine and MMP production in synovial membrane cultures [ 7 ]. A detailed protocol for the generation of adenoviral vectors can be found elsewhere [ 41 ].
Here we describe how to transfect RA synovial membrane cells. around the knee joint. The patella (or knee cap) should be visible under the skin as white.
Person A scruffs the mouse and turns it onto its back with one hand and gently pulls the leg nearly straight with the other hand. The hind foot is held down with the thumb while the knee is supported from underneath with the index finger.
Person B aligns the needle perpendicular to the leg and over the top of the patella and, by pressing down the needle gently, identifies the groove between the femur and the tibia (see Fig. 2a ). The needle is maintained in line with this groove while slid slightly back before raising it to approximately a 45° angle and inserting it into the cavity of the joint; all in one smooth motion (see Fig. 2b ). Ten microliter of solution are dispensed.
The two most frequent incorrect targets of the injection are the bone and the site immediately below the patella. The person performing the injection will encounter resistance in the former case while notice the needle moving loosely around in the latter case. If the solution comes straight out in the process of injection or if in an incorrect location, remove the needle and begin the process again. If the skin has been damaged, manipulate surrounding shaven skin to have intact layer over the knee. No more than three attempts should be made on the knee, after which intra-articular administration should be stopped if unsuccessful and the mouse allowed to recover. 9. Clinical evaluation of arthritis may include the use of a caliper to measure knee joint swelling. This consists of measuring the distance between the medial and lateral aspects of each knee joint at the level of the patellar ligament and is expressed as knee diameter. However, in our experience we have found this technique not very reliable. Alternatively, water displacement [ 42 ] is a more accurate and reliable method that can be used to measure knee swelling. However, this requires the use of a plethysmometer, which may not be readily available.
10. Tissues such as joints containing calcified areas need to be decalcified before processing in order to become soft enough for sectioning. It is possible to test biochemically for calcium in solution to assess endpoint of decalcification.
11. DPX is a synthetic resin mounting media composed of Distyrene, a plasticizer, and xylene. It preserves the stain and dries quickly, enabling slides to be screened immediately. It has a low viscosity, allowing the medium to flow easily and prevents air bubbles from becoming trapped. Application of coverslips to microscope slides can also be done manually by adding a drop of mounting media at base of slide, placing the coverslip at an angle as to begin to spread the media and carefully lower the coverslip onto the tissue while the media spreads underneath. Allow slides to dry under a hood.
12. Histological analysis should be performed by an investigator who is blinded to the experimental groups.
13. During the first 7 days after intra-articular injection of mBSA, the acute phase of the arthritic response can be investigated. In order to evaluate the chronic phase of the arthritic response, disease can be allowed to progress for 2-3 weeks after intra-articular injection of antigen.
14. Isolation of total RNA from tissue lysates requires homogenization to reduce viscosity caused by high-molecular-weight cellular components and cell debris. Unlike traditional methods that use syringes and needles, QIAshredder spin columns require a single and fast centrifugation step and http://eproofing.springer.com/books/printpage.php?token=hcKiChH62c8ED0xcrdXaZX-PfMgJgXJPl8cIy3jyrSo reduce loss of sample material. 15. Oligo(dT) primers anneal to any mRNA with a poly(A) tail, generating full length copies of the mRNA. Alternatively, random primers can be used. These are random combinations of nucleotides, 6 or 9 bp long, that enable transcription of 5! ends of long genes, generating cDNAs that may not be full length copies of the entire gene. 16. If performing the real-time PCR in a Rotor-Gene 6000 instrument using 0.1 ml strip tubes, there is no need to centrifuge the tubes before starting the reaction as the machine spins samples for the entire duration of the PCR.
17. Choose the acrylamide-bis solution percentage most appropriate to the endogenous TLR ligand of interest.
18. There are numerous commercial sources of collagenase used by various laboratories to digest synovial tissue. However, great care has to be taken to ensure each batch of enzyme is tested for endotoxin contamination and not simply rely on manufacturer's product sheets. We batch test using Endpoint Chromogenic LAL Assays (Lonza) and reject if contamination exceeds 1 EU/ml. In addition to exclude any possibility that low levels of endotoxin could stimulate the highly sensitive macrophage population within RA synovial membranes, we perform mock enzymatic digestions (i.e., incubation of human monocytes with collagenase NB1 8 mg/ml for 2 h followed by three washes, culture of monocytes for 18 h and measurement of monocyte activation by assaying for TNF-α production by ELISA). Many commercial sources of collagenase induce up to 20 ng/ml of TNF-α under these conditions. We observe no such effect using NB1 or Liberase TL in any batches assayed.
19. The choice of enzyme to digest synovial tissue is also critical depending on the end point assay each researcher uses. We have found Collagenase A from Roche cleaves CD56, a NK cell marker which then may give a false under representation of NK cells within the tissue. Another popular source of enzyme is Roche Liberase TL, which we found to cleave CD4 (as demonstrated in Fig. 7 ), leading to an underestimation of this cell population. As new technologies such as CyTOF and multicolor FACS are developed which allow multiple parameter analysis of mixed cells populations, it is critical that the integrity of cell surface receptors remain intact. After extensive testing, we have found NB1 with its high collagenase and low neutral protease activity was the optimal collagenase for tissue digestion while maintaining cell surface receptor expression. However the high specificity of this particular enzyme results in a lower cell yield and careful thought is needed as RA synovial membranes can yield very little cells and if cytokine analysis is the endpoint, a collagenase preparation such as Liberase TL may be a more appropriate digestive enzyme of choice.
20. For clear definition of positive staining and to assist set up of the flow cytometer, it is suggested to leave one sample unstained (in addition to those for staining panels A and B).
21. The inclusion of FcR block is necessary to avoid nonspecific staining of all FcR bearing cells, in particular macrophages, which are abundant in RA synovial membranes. 22. It is essential to include a viability stain as up to 20 % of total cells do not remain intact after tissue dissociation. Therefore only collect data from viable cells. 23. For low abundance cell types, it is suggested to gate on the appropriate subset and collect a minimum of 5000 events therein.
24. While it is preferable to analyze cells directly after staining, if a cytometer is not available, samples may be kept at 4 °C in the dark for up to a few hours, or, alternatively, they can be fixed post staining and analyzed the following day. 25. NK cells are also detectable with the staining panels detailed here (CD45+, CD3− and CD56+) when using collagenase NB-1 for tissue digestion. Other collagenases (e.g., Liberase TL) cleave CD56, rendering it undetectable on the cell surface (see Fig. 7 and Note 18).