Effects of fluorescent and nonfluorescent tracing methods on lymphocyte migration in vivo




The use of fluorescent dyes to monitor in vivo cellular migration and proliferation has greatly expanded, but little is known about their potential influence on cell migration.


Adoptive transfer studies of lymphocytes labeled with various dyes were performed, and their in vivo homing was compared with that of coinjected unlabeled control cells. In addition, in vitro migration and binding studies were performed to analyze the various steps of transmigration separately.


These data showed that the intracellular fluorescent dyes calcein acetoxymethyl ester, 2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein acetoxymethyl ester, 5-chloromethylfluorescein diacetate, 5-(and-6)-carboxyfluorescein diacetate, succinimidyl ester, and fluorescein isothiocyanate affect in vivo homing of especially B lymphocytes to lymphoid organs, without any direct effect on in vitro chemotactic or adhesive activity. The only label that did not affect migration was the extracellular and nonfluorescent molecule biotin, provided that the labeling was performed at room temperature. Interestingly, by using the highly versatile congenic Ly5.1-Ly5.2 system, we also demonstrated intrinsic differences in lymphocyte migration based on allelic differences.


Our data showed that fluorescent labeling of lymphocytes has a severe effect on their homing capacity in vivo. Labeling of cells with biotin appeared to be a good alternative for this purpose; however, if direct fluorescence is required, the negative effects on cell migration should be considered. © 2004 Wiley-Liss, Inc.

To understand the mechanisms that underlie the function of the immune system, immunologists have searched for methods to trace individual lymphocytes in their migration to lymphoid organs and in their subsequent interactions with other cells. In previous studies, radioactive markers were the method of choice to label lymphocytes, using autoradiographic techniques to trace them in tissues (1–3). These elaborate procedures have become obsolete with the development of fluorescent labels: powerful new techniques such as fluorescence microscopy and flow cytometry have made it relatively easy to follow, quantify, and phenotype adoptively transferred lymphocytes ex vivo (4–6). Currently, many fluorescent dyes exist, each with its specific features and (dis)advantages (7): some dyes have the ability to bind DNA and thus stain the nucleus (e.g., Hoechst 33342), some are highly lipophilic and localize in the cell membrane (e.g., 1,1′dioctadecyl- 3,3,3′-tetramethylindocarbocyanine perchlorate and PKH26), and others are lipophilic enough to cross the cell membrane and localize in the cytoplasm (e.g., fluorescein isothiocyanate [FITC] and 5-[and-6]-carboxyfluorescein diacetate, succinimidyl ester [CFSE]). Most of the labels that have been considered suitable for lymphocyte proliferation and migration studies are characterized by their nontoxicity, high fluorescence, and persistent staining: labels such as PKH26, calcein acetoxymethyl ester (Calcein), 2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein acetoxymethyl ester (BCECF), and CFSE have been shown to adequately label lymphocytes, and labeled cells have been shown to localize in peripheral lymphoid organs after in vivo transfer (6, 8, 9). Nonetheless, it remains obscure whether the applied labels leave the migratory behavior of cells unaltered.

Other methods to trace cells after in vivo transfer without the necessity of cellular labeling depend on the use of karyotypes or allelic differences between congenic mouse strains that can be traced with monoclonal antibodies. A widely used example of the latter are the murine allotypes Ly5.1 and Ly5.2 of the CD45 molecule (10, 11), which have been applied in adoptive transfer studies (12) and bone marrow transplantation (13). Using this system, donor cells require minimal handling before transfer, thus decreasing potential negative effects on the behavior of the cells.

With the aim of investigating the extent to which cell handling and labeling can affect in vivo lymphocyte migration, we combined Ly5 congenic mice with cell labeling in short-term homing assays. Therefore, labeled and unlabeled lymphocytes were transferred simultaneously into the same recipient. Because both populations of cells originated from the same donor, they were intrinsically identical and differed only in the absence or presence of a cellular label. Any interfering effect of a label of interest could thus be directly monitored because there was an internal control for the effect of each label. This approach offers new insights into the effect of (non-)fluorescent dyes on cell migration in vivo.



C57BL/6-Ly5.1 (B6-Ly5.1) and C57BL/6-Ly5.2 (B6-Ly5.2) mice were bred and maintained under conventional conditions in the laboratory animal facilities at the VU Medical Center, Amsterdam, The Netherlands. Mice carrying the Ly5.1 and Ly5.2 alleles were generated by using the F1 progeny of an intercross of B6-Ly5.1 and B6-Ly5.2 mice. These mice were designated B6-Ly5.1-Ly5.2. All mice were used at the age of 8 to 15 weeks. All experiments were approved by the VU ethical committee for experimental procedures in animals.

Preparation of Lymphocytes

Single-cell suspensions from freshly isolated spleen, peripheral lymph nodes (axillary, brachial, and inguinal), and mesenteric lymph nodes were obtained by mincing and gently pressing the tissue through a fine nylon mesh, while frequently rinsing with 2% new-born calf serum (NBCS)/phosphate buffered saline (PBS). Peripheral blood was collected by heart puncture and collected in PBS with 20 U/ml heparin. Erythrocytes from spleen and blood were lysed by incubation for 5 min at room temperature in ACK lysis buffer (150 mM NH4Cl, 1.0 mM KHCO3, and 0.1 mM Na2 ethylene-diaminetetra-acetic acid, pH 7.4) and washed and resuspended in 2% NBCS/PBS. More than 90% of these cells were viable as determined by trypan blue exclusion.

Lymphocyte Labeling


Lymphocytes were washed and resuspended in 0.1% bovine serum albumin (BSA)/PBS to a concentration of 1 × 107 cells/ml. Labeling with CFSE was performed according to instructions of the manufacturer (Molecular Probes, Eugene, OR). Briefly, CFSE (50-mM stock in dimethyl sulfoxide) was added to a final concentration of 0.5 μM, and the cells were incubated for 15 min at 37°C while shaking. Then cells were washed with an excess of warm 0.1% BSA/PBS and reincubated in this buffer for another 30 min at 37°C to ensure complete modification of the label, as recommended by the manufacturer. Cells were washed extensively and resuspended in the appropriate buffer for further use.

5-Chloromethylfluorescein diacetate (CMFDA).

Treatment was similar to CFSE labeling, except that the cells were labeled at 5 × 107 cells/ml in 1% BSA/PBS with a final concentration of 1.0 μM CMFDA (Molecular Probes) for 30 min, as recommended by the manufacturer's protocol for labeling.

Calcein and BCECF.

Lymphocytes were washed two times with PBS and resuspended to a concentration of 2 to 5 × 107 cells/ml PBS. After 15 min of preincubation at 37°C, Calcein or BCECF (Molecular Probes; 1-mM stock in dimethyl sulfoxide) was added to a final concentration of 0.2 μM, and cells were incubated for 20 min at 37°C while shaking. Free label was removed by washing the cells first through a cushion of fetal calf serum and subsequently with PBS and resuspended in the appropriate buffer for further use.

Biotin and FITC.

Labeling was comparable with Calcein and BCECF, except that biotinylation of cells was performed by incubating cells in PBS with 80 μg/ml D-biotinoyl-ϵ-aminocaproic acid N-hydroxy-succinimide ester (Biotin-7-NHS; Boehringer Mannheim, Mannheim, Germany) for 15 min at room temperature or 37°C while shaking; FITC labeling was performed by 30 min of incubation at 37°C in PBS with a 1:30 dilution of a filtered solution of 1 mg/ml FITC (Sigma, St. Louis, MO) in PBS.

In all cases, the mean fluorescence intensity of the labeled population exceeded that of nonlabeled cells by 100-fold or more. None of these labels significantly affected the viability of the cells, as determined by staining with trypan blue after the labeling procedure.

In Vivo Migration

Labeled lymphocytes were mixed in a 1:1 ratio with control lymphocytes, which originated from the same cell suspension and had been treated similarly as the labeled cells, except that no label was added during the procedure. This mixture of labeled and unlabeled cells was washed once with PBS, and 2 to 5 × 107 of viable cells was injected intravenously into recipient mice that were matched by sex and age but had the opposite Ly5 phenotype (i.e., B6-Ly5.1 mice received cells from B6-Ly5.2 mice, or vice versa). Experiments performed with CMFDA and CFSE were performed two times, and those with BCECF, Calcein, Biotin, and FITC were performed three times. In other experiments, a 1:1 mixture was made of (nonlabeled) lymphocytes from B6-Ly5.1 and B6-Ly5.2 mice and subsequently injected into sex- and age-matched B6-Ly5.1-Ly5.2 recipients. The in vivo migration of Ly5.1 versus Ly5.2 lymphocytes was analyzed in three independent experiments. After 2 h, recipient mice were killed, and lymphoid organs and blood were collected and analyzed for the presence of donor cells by means of flow cytometry. The ratio of labeled to unlabeled or of B6-Ly5.1 to B6-Ly5.2 donor cells before injection was determined by flow cytometry and used to correct for injected ratios.

In Vitro Migration

Lymphocytes from lymph nodes were labeled with the various labels and mixed in a 1:1 ratio with unlabeled control cells as described above; the mixture was resuspended in RPMI-1640 medium containing 2% newborn calf serum and 10 mM Hepes, pH 7.4, and incubated for 10 min at 37°C, and 1 × 106 cells in 100 μl of medium was added to the top chamber of a 6.5-mm diameter, 5-μm pore polycarbonate Transwell insert (Costar, Cambridge, MA) and incubated in triplicate with 250 ng/ml of recombinant mouse stromal-derived factor 1α (SDF-1α; R&D Systems, Minneapolis, MN) in the bottom chamber for 3 h at 37°C and 5% CO2. The number and phenotype of cells that had migrated specifically in response to SDF-1α to the bottom chamber were determined by flow cytometry. This experiment was performed once for each of the indicated labels, with all conditions in triplicate.

Binding Assay for High Endothelial Venule (HEV)

Freshly isolated lymph node cells were labeled with CFSE as described above and mixed with an equal amount of unlabeled control cells. In some experiments, this mixture was enriched for B cells by T-cell depletion. For enrichment, cells were incubated with rat anti-mouse antibodies against CD3, CD4, CD8, and CD90, washed, and subsequently incubated with magnetic goat anti-rat immunoglobulin G microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany). The uncoated fraction of B cells was subsequently separated from bead-coated T cells by passing the suspension over an LS+ separation column on a MidiMACS magnet (Miltenyi Biotec), according to the manufacturer's instructions. The unbound fraction consisted of more than 92% B cells, as determined by CD19 expression.

Lymphocyte adherence to HEV in vitro was assayed by the method of Butcher et al. (14), a modification of an assay originally described by Stamper and Woodruff (15). Briefly, lymphocyte suspensions of total lymphocytes or purified B cells, containing 3 × 105 cells/100 μl in culture medium (RPMI, 5% fetal calf serum, 20 mM Hepes, pH 7.4), were incubated at 8°C for 30 min on unfixed frozen sections (8 μm) of control lymph nodes. During the assay, lymph node sections were rotated at 100 rpm. After completion of the assay, unbound cells were decanted, and sections were fixed in PBS containing 2% formaldehyde, 0.25% glutaraldehyde, 2.5 mM CaCl2, and 2.5 mM MgCl2. After 1 h, the sections were transferred to PBS without fixatives, and cell binding to HEV was assessed microscopically under darkfield or fluorescent illumination. The numbers of unlabeled and fluorescently labeled cells per HEV were determined, and the ratio of labeled to control cells was calculated and corrected for the ratio applied to the sections. A minimum of 25 HEVs was analyzed per condition. This experiment was performed once for total lymphocytes and once for purified B cells.

Flow Cytometry

The number and phenotype of lymphocytes that had been subjected to in vitro or in vivo migration were analyzed on a FACScan cytometer (Becton Dickinson, Mountain View, CA) using CellQuest software (Becton Dickinson). Depending on the experiment and the organ, 1 × 104 to 2 × 105 of total cells was collected per sample to analyze sufficient donor cells from the recipient's organs. Lymphocytes were analyzed after gating on forward and side scatter. For the detection of biotin-labeled cells, cells were incubated with streptavidin-CyChrome (Pharmingen, San Diego, CA).

The following monoclonal antibodies were used to identify lymphocyte subsets: AL1-4A2 (mouse anti-mouse Ly5.2/CD45.2), A20 (mouse anti-mouse Ly5.1/CD45.1), KT3.1 (rat anti-mouse CD3ϵ), RA3-6B2 (rat anti-mouse B220/CD45R), Mel-14 (rat anti-mouse CD62L), GK1.5 (rat anti-mouse CD4), and 53-6.7 (rat anti-mouse CD8). Culture supernatants of hybridoma cells were purified with protein G–Sepharose (Pharmacia, Uppsala, Sweden) and subsequently labeled with Biotin-7-NHS (Boehringer) or Alexa-488 reagent (Molecular Probes), according to standard protocols. Primary antibodies were detected with phycoerythrin-conjugated rabbit F(ab′)2 fragment anti-rat immunoglobulin G (Jackson, West Grove, PA) or streptavidin-CyChrome.


In Vivo Effect of Calcein, BCECF, CFSE, and CMFDA

Despite their widespread use in various applications, little is known about the potential effects that fluorescent dyes can have on the physiology of cells. To address this question more specifically, we investigated the direct effect of four widely used fluorescent dyes on lymphocyte migration in vivo. Labels tested were the intracellular fluorescent dyes Calcein, BCECF, CMFDA, and CFSE. All four labels are characterized by the fact that they are nonfluorescent in their native form and can freely pass through cell membranes; once inside the cytosol, these labels are modified by nonspecific esterases and thereby become highly fluorescent molecules that are retained inside the cell. Figure 1A shows the chemical structure of these dyes, where they are depicted in their membrane-permeant forms.

Figure 1.

Use of fluorescent dyes in lymphocyte migration in vivo. A: Chemical structure of the four fluorescein-based dyes Calcein, BCECF, CMFDA, and CFSE depicted in their membrane-permeant forms. Excitation and emission wavelengths are given for their fluorescent forms. B: Schematic based on fluorescence-activated cell sorting of the experimental setup followed in this study: in this example, CFSE-labeled Ly5.1+ lymphocytes are mixed with nonlabeled Ly5.1+ control cells (left panel). Two hours after injection into a Ly5.2+ recipient, these donor cells can be clearly distinguished in the host's lymphoid organs by staining for Ly5.1 (middle panel). When gated on this population and stained for the appropriate lymphocyte subset, the ratio of CFSE-labeled and control T or B cells can be easily determined (right panel) and compared with the ratio after injection.

When cells are labeled and transferred in vivo, they can be followed for several days to weeks, depending on the dye (6). We tested to the extent to which these labels affect homing of lymphocytes into secondary lymphoid organs by direct comparison of their migratory capacity with that of unlabeled control cells. Hereto, B6-Ly5.2 mice were injected with a 1:1 mixture of labeled and unlabeled Ly5.1+ lymphocytes. After 2 h, the distribution of the transferred Ly5.1+ donor cells was monitored in the host's lymphoid organs, and the numbers and phenotypes of the labeled cells were compared with the coinjected control cells. Figure 1B illustrates this experimental setup, in which the migration of CFSE-labeled B cells to the recipient's spleen is depicted and compared with control B cells.

Using this approach, we were able to calculate the ratio of labeled to control cells in all isolated organs for T and B cells; a ratio of 1 indicates that the label had no effect on the localization of the cells, whereas a ratio lower than 1 designates that the labeled cells migrated less efficiently than control cells. No effect of labeling could be found on the distribution of T cells because labeled and unlabeled T cells migrated equally well, with the exception of a slightly increased migration when labeling with Calcein (Fig. 2). However, all four labels had a dramatic inhibitory effect on homing of B cells to peripheral lymph nodes and mesenteric lymph nodes: depending on the dye, only 25% to 60% of labeled cells, compared with control cells, could be retrieved from the host lymph nodes (Fig. 2). In case of CMFDA and CFSE, this inhibition of B cells was not restricted to lymph nodes because the number of B cells labeled with these dyes in the spleen and blood was significantly smaller than that of control cells.

Figure 2.

Effects of four fluorescent dyes on B- and T-lymphocyte migration. Fluorescently labeled and nonlabeled Ly5.1+ lymphocytes were coinjected intravenously into Ly5.2+ recipients. After 2 h, the ratio of labeled to nonlabeled Ly5.1+ donor cells was calculated in the lymphoid organs of the host. Data are shown for T cells (gray bars) and B cells (black bars) and have been corrected for the ratio of labeled to nonlabeled cells after injection. The horizontal line represents the ratio for equal migration of labeled and nonlabeled cells. Means and standard deviations are shown for three mice per group and are representative of at least two independent experiments. MLN, mesenteric lymph node; PLN, peripheral lymph node.

In Vivo Effect of FITC and Biotin

These data clearly show that lymphocyte labeling with Calcein, BCECF, CMFDA, and CFSE is not without effect on the homing capacity of (B) lymphocytes to lymphoid organs. What all four dyes have in common is that they need intracellular modification to become fluorescent and cell impermeable. It could well be that this intracellular modification affects the cell's behavior in general and its homing potential in particular. To rule out any negative effects of intracellular conversion, we tested the usefulness of FITC as a cell marker. FITC in structure is closely related to CMFDA, but it lacks the two acetate groups of CMFDA and has a thiocyanate group instead of a chloromethyl group (Figs. 1A and 3A). Because FITC can freely diffuse into the cell, it can bind intra- and extracellularly to free amine groups, but it does not have to be modified to become fluorescent. However, despite these beneficial characteristics, transfer of FITC-labeled cells in vivo showed that FITC labeling had a strong negative effect on the homing of B cells to lymph nodes (Fig. 3B), similar to the other dyes (Fig. 2). The effect on T cells was negligible, except that in some experiments a reduced migration to mesenteric lymph nodes was seen, but this was not a consistent finding. Hence, despite the fact that FITC labeling does not cause the intracellular formation of possibly toxic byproducts, it cannot be considered an inert fluorescent label and should be regarded as inappropriate for in vivo lymphocyte homing studies when migration of B cells is involved.

Figure 3.

Effect of FITC and biotin labeling on lymphocyte migration in vivo. A: Chemical structures of FITC (with excitation and emission wavelengths) and biotin as examples of dyes that do not require intracellular modification. B: When compared with nonlabeled control cells, FITC labeling negatively affects B-cell migration to lymphoid organs. Biotin does not show this effect, provided that the labeling is performed at 20°C. Data are shown for T cells (gray bars) and B cells (black bars) and have been corrected for the ratio after injection. The horizontal line represents the ratio for equal migration of labeled and nonlabeled cells. Means and standard deviations are shown for three mice per group and are representative of three independent experiments. MLN, mesenteric lymph node; PLN, peripheral lymph node.

To avoid the use of labels that can affect the intracellular condition of the cell, we tested the effect of the extracellular label biotin. This molecule is often used to label proteins or cells for biochemical purposes, but it can be applied for cell labeling and subsequent in vivo tracking (16). Biotin is a small polar molecule that cannot permeate the cell, but it can very efficiently bind to cell surface molecules due to its ability to react with free amine groups (Fig. 3A). Although biotin itself is not fluorescent, biotin-labeled cells can be detected with fluorescently conjugated streptavidin. After transfer in vivo, we found that biotin, in contrast to all previously described labels, did not affect the migratory capacity of lymphocytes in vivo: biotin-labeled T and B cells migrated to lymphoid organs as efficiently as nonlabeled control cells (Fig. 3B). However, when labeling was performed at 37°C instead of room temperature, we also found a negative effect of biotin because biotinylated B and T cells clearly migrated differently from similarly treated control cells (Fig. 3B). Altogether, these findings indicate that labeling lymphocytes with biotin at room temperature is a suitable method for tracking naive lymphocytes in vivo because binding of biotin to cell surface molecules does not influence the entry of cells into lymph nodes or spleen, and it does not remove cells from the circulation.

Effect of Calcein, BCECF, CFSE, and CMFDA on Transmigration

Although all fluorescent labels negatively influenced lymphocyte migration in vivo, it is not clear which underlying molecular events caused this homing defect. Lymphocyte migration to secondary lymphoid organs is a multistep process (17), and the effect of these dyes might be explained by their interaction with molecules that are important for one of these steps. To test whether these labels negatively affect the chemotactic and transmigratory capacity of lymphocytes, a mixture of labeled and nonlabeled lymphocytes was applied to a transwell with a chemotactic gradient of SDF-1, a chemokine that efficiently attracts T and B cells (18). In contrast to the in vivo data, fluorescently labeled cells transmigrated as efficiently as control cells to the lower compartment of the transwell. No specific inhibitory effect of the label could be found on the migratory capacity of T or B cells (Fig. 4), and similar results were found for FITC (data not shown). These data indicate that the fluorescent labeling does not affect the ability of lymphocytes to respond to a chemotactic signal.

Figure 4.

Fluorescent labeling does not inhibit the chemotactic capacity of lymphocytes. A mixture of fluorescently labeled and nonlabeled lymphocytes was transferred to a transwell insert and subjected to 250 ng/ml SDF-1α in the lower compartment. After 3 h of incubation at 37°C, the number and phenotype of transmigrated cells were determined by fluorescence-activated cell sorting analysis. The ratio of labeled to nonlabeled cells was calculated for T cells (gray bars) and B cells (black bars) and corrected for the initial ratio. Data are the mean and standard deviation of three transwells.

Effect of CFSE on HEV Binding

Another aspect of the multistep lymphocyte migration process is the tethering and rolling on the endothelium of the HEV (17, 19). This selectin-dependent initial adhesion step can be studied in vitro with the HEV-binding assay. Therefore a mixture of CFSE-labeled and control cells was incubated on fresh lymph node sections to allow binding to the natural ligands expressed on the HEV. However, no negative effect of the dye could be observed because CFSE-labeled cells bound as effectively as nonlabeled control cells (Fig. 5). Even when purified B cells were used, no effect could be observed, indicating that CFSE labeling does not disturb the binding capacity of lymphocytes to HEVs.

Figure 5.

CFSE labeling does not inhibit the capacity of lymphocytes to bind to HEVs. Mixtures of CFSE-labeled and nonlabeled cells were incubated on unfixed lymph node sections to allow binding to the natural ligands expressed on the endothelium of an HEV. Each dot represents the ratio of labeled to control cells bound to a single HEV by using total lymphocytes or purified B cells. The horizontal bar indicates the mean of these ratios. HEV-bound cells were counted on at least three separate sections from six different peripheral lymph nodes.

Homing of Ly5.1+ Versus Ly5.2+ Lymphocytes In Vivo

Although we cannot establish the underlying molecular mechanism, it is clear that in vitro labeling of lymphocytes before transfer in vivo can have strong effects on the migratory capacity of cells. To maintain the advantage of tracing two different populations of donor cells in one acceptor mouse without the need for fluorescent labeling, we further exploited the Ly5 allogenic system. The F1 progeny of an intercross between B6-Ly5.1 and B6-Ly5.2 mice, which express Ly5.1 and Ly5.2 on all hematopoietic cells, was used to trace injected Ly5.1 and Ly5.2 cells. After transfer of Ly5.1+ and Ly5.2+ lymphocytes into these mice, all three populations could simply be recognized by fluorescence-activated cell sorting analysis after staining simultaneously for the Ly5.1 and Ly5.2 allotypes (Fig. 6A). With this approach, donor cells did not require extensive handling before use, thereby decreasing the potential negative effects on their migration and function. Surprisingly, extensive transfer experiments indicated that even untreated Ly5.1+ and Ly5.2+ lymphocytes did not have identical migratory capacities (Fig. 6B): more Ly5.1+ T cells and Ly5.2+ B cells ended up in the lymph nodes than did their Ly5 counterparts. These results were consistently found in repetitive experiments and could not be attributed to the health status of one of the donor mice; in addition, mice were matched for sex and age in all experiments. This signifies that, despite the small genetic difference between B6-Ly5.1 and B6-Ly5.2 mice, their lymphocytes show small intrinsic differences that affect their migration.

Figure 6.

Ly5.1+ and Ly5.2+ lymphocytes do not show identical localization in lymphoid organs after transfer into a B6-Ly5.1-Ly5.2 recipient. A: Schematic based on fluorescence-activated cell sorting of the experimental setup followed in this study. Left: The injected mixture of Ly5.1+ and Ly5.2+ cells is shown. Right: When these cells are injected into a B6-Ly5.1-Ly5.2 recipient, they can be easily distinguished from host cells by double staining for Ly5.1 and Ly5.2. B: A mixture of unlabeled lymphocytes isolated from B6-Ly5.1 and B6-Ly5.2 mice was coinjected intravenously into B6-Ly5.1-Ly5.2 recipients. After 2 h, the ratio of Ly5.1+ to Ly5.2+ cells was determined in the lymphoid organs of the B6-Ly5.1-Ly5.2 host. Data are shown for T cells (gray bars) and B cells (black bars) and have been corrected for the ratio after injection. The horizontal line represents the ratio for equal migration of Ly5.1+ to Ly5.2+ cells. Means and standard deviations are shown for three mice per group and are representative of three independent experiments.


Although fluorescent dyes are widely used to track lymphocyte migration in vivo, this study shows that cell labeling can severely affect the migratory behavior of lymphocytes. When used according to standard protocols and applied in short-term homing assays, the fluorescent labels Calcein, BCECF, CMFDA, CFSE, and FITC displayed a strong negative effect on B-cell migration to lymph nodes. These data clearly indicate that fluorescent dyes are not inert and can actually interfere with cellular functions, which is supported by in vitro data (20–22). It should be noted that the concentrations of the fluorescent dyes used in this study were not excessively high and were in most cases even below the generally recommended concentrations. For instance, CFSE was used at a concentration of 0.5 μM in the present study, whereas others have used this label up to 5 to 10 μM (6, 20, 23, 24). It could be anticipated that such high concentrations might result in even stronger side effects than those described here and could even affect other cell types than B cells. In addition, it is conceivable that fewer side effects of these labels can be achieved by lowering the temperature during the labeling procedure, because we also found negative effects of biotinylating cells when labeling at 37°C but not at room temperature (Fig. 3B). However, such alterations to the conventional protocols will most likely have unfavorable effects on the labeling efficiency because the labeling process depends on the activity of intracellular enzymes. Therefore, we suggest that labeling cells at room temperature with the nonfluorescent, extracellular label biotin is the most appropriate method to label lymphocytes for homing assays because it did not show any negative effects on in vivo migration (Fig. 3B), as has been reported previously for additional in vitro assays (25, 26).

To investigate the level at which the fluorescent dyes inhibit the migration of lymphocytes to lymph nodes, in vitro experiments were performed that allowed examination of the individual steps of the migration process. The HEV-binding assay clearly showed that CFSE labeling does not interfere with the L-selectin–dependent binding of (B) lymphocytes to their natural ligands on HEVs (Fig. 5). Other labels were not tested, but because it has been conclusively shown that FITC has no negative effect on this process (14), we assumed that the effect of the different labels on in vivo lymphocyte migration has its origin on a different level than on the L-selectin–dependent interaction with the endothelium.

The consecutive steps in lymphocyte migration involve the chemokine-dependent arrest of the rolling lymphocytes and their subsequent transmigration. To mimic this process, transwell assays were performed, and these studies clearly showed that none of the applied labels interfered with the chemotactic ability of T or B cells (Fig. 4). Nonetheless, it should be recognized that this approach did not provide an accurate representation of the in vivo situation, particularly due to the lack of adhesion molecules. Addition of natural ligands to the transwell, such as intracellular adhesion molecule 1 (27), could make this assay more sensitive and might reveal yet unobserved differences in the process of transmigration.

Despite the fact that these in vitro studies did not identify a mechanism responsible for the decreased lymphocyte migration, B cells appeared to be mostly affected in their homing capacity. However, Calcein did show some stimulatory effect on T cells in vivo (Fig. 2A) and in vitro (Fig. 4). Because it has been shown that B cells are more brightly labeled than T cells when using Calcein (8) or FITC (4), this does signify that B cells can respond differently from T cells to fluorescent dyes. It is conceivable that these labels interfere with B-cell–specific molecules that are directly or indirectly related to HEV-dependent migration. A potential candidate could be the B-cell–specific chemokine receptor CXCR5, because changes to its sensitivity for its ligand BLC can have direct effects on entry into lymphoid organs (16). Alternatively, a Calcein-specific effect on the T-cell chemokine receptor CCR7 or its signaling pathways could be the reason for the slight increase in T-cell migration.

Further, it should be noted that decreased numbers of CFSE- and CMFDA-labeled B cells were also found in spleen and blood, although this decrease was less apparent than observed in lymph nodes (Fig. 2). This would argue that CFSE- and CMFDA-labeled B cells are not only less capable of entering lymph nodes but may be even more prone to cell death upon transfer. For the dyes needing intracellular modification, this may be the result of the formation of toxic byproducts such as formaldehyde and acetic acid during hydrolysis of the acetoxymethyl and acetate esters of these dyes (28).

To avoid the use of intrusive cellular labels, we also compared the migratory behavior of the congenic Ly5.1-Ly5.2 system, but to our surprise these genetically almost identical mice did show differences in lymphocyte distribution upon in vivo transfer (Fig. 6B). This could be due to a different activation state of the transferred lymphocytes because this approach required that the two populations of injected cells originate from distinct donors. However, we corrected for this by using donor mice that were matched for sex, age, and health status, and the same differences could be observed in all experiments. Therefore, it well may be that in these experiments depletion from the circulation rather than disturbed migration was the cause of these effects because the unequal distribution was found not only found in lymph nodes but also in blood and spleen. Although functional differences between congenic mouse strains would not be expected due to their genetic resemblance, it has been shown that the small differences between these Ly5 allotypes can cause differences in in vitro immune responses (29) and can even decrease long-term donor engraftment upon congenic bone marrow transplantation (30). These findings indicate that even B6-Ly5.1 and B6-Ly5.2 mice are functionally more distinct than would be expected from the small genetic difference in the Ly5 gene only.

In conclusion, our data show that adoptive transfer experiments can be negatively influenced by the method chosen to trace the transferred cells. It will be important to find out whether these and other labels can also affect other cellular functions or different cell types. Taken together, it is clear that only labeling of cells with biotin does not interfere with the lymphocyte's physiology or migratory capacities. Biotin labeling is fast and easy but does require additional stainings for detection. When direct fluorescent labeling is required to follow cells in vivo, e.g., in proliferation studies or with intravital microscopy and biotin can not be used, one should be aware of negative side effects of labeling on cell function.


We thank Dr. Elga de Vries and Dr. Sarah Floris for generously providing us with the fluorescent dyes Calcein, BCECF, and CMFDA and Dr. Wendy Unger for CFSE. The B6-Ly5.1 and B6-Ly5.2 mice were obtained from Dr. Joop Brandenburg from the EUR, Rotterdam, The Netherlands.