Visualizing dendritic cell networks in vivo


In the steady state, dendritic cells (DCs) in the lymph node induce T cell tolerance to self antigens. Innate signals trigger the maturation of tissue DCs, which migrate into lymph nodes and activate T cells. To examine DCs in vivo, we produced transgenic mice whose DCs expressed enhanced yellow fluorescent protein. Two-photon microscopy of lymph nodes in live mice showed that most of the steady-state DCs were enmeshed in an extensive network and remained in place while actively probing adjacent T cells with their processes. Mature DCs were more motile than steady-state DCs and were rapidly dispersed and integrated into the sessile network, facilitating their interaction with migrating T cells.


Dendritic cells (DCs) are specialized antigen-presenting cells that are found in nonlymphoid and lymphoid organs and are essential in initiating immune responses and maintaining tolerance1,2,3,4,5,6. In nonlymphoid organs, DCs serve as sentinels that capture and then carry antigens to lymphoid organs, where they interact with specific T cells. In contrast, DCs that reside in lymphoid organs such as lymph nodes receive antigens from lymph, blood or immigrant DCs, relaying antigen from the periphery7,8,9,10,11. The outcome of antigen presentation by DCs depends on the state of DC terminal differentiation or maturation1,2,3,4,5,6. In the steady state, immature DCs capture, process and present a variety of self antigens to T cells; sources of antigen include serum proteins, extracellular matrix components and dead cells11,12,13,14,15,16,17,18. Antigen presentation by steady-state DCs establishes T cell tolerance to prevent adverse reactions to self when the same antigens are presented in immunizing conditions6. In contrast, DCs that have been stimulated to mature by inflammatory mediators, such as Toll-like receptor ligands, induce strong effector T cell responses1,2,3,4,5,6.

Two-photon microscopy has been used to study mature DCs actively participating in immune responses in explanted intact lymph nodes or surgically exposed lymph nodes in anaesthetized mice19,20,21,22. DCs were labeled with exogenous fluorescent dyes, and DC maturation was induced during isolation or by deliberate treatment with adjuvants19,20,21,22. In those studies mature DCs 'crawled' in random directions, albeit at a slower speed than T cells, and they rapidly extended and retracted long dendritic processes. However, steady-state DCs were not examined.

To examine DC activity in the steady state, we produced transgenic mice expressing enhanced yellow fluorescent protein (EYFP) in CD11c+ DCs. We selected founders that had sufficient expression for imaging deep in the lymph node by two-photon laser-scanning microscopy and verified that the expression was restricted to the appropriate cells. Here we report on the activity of steady-state and mature DCs in various regions in the lymph nodes of live mice.


CD11c-EYFP–transgenic mice

We produced transgenic mice expressing an EYFP23 reporter under the control of the CD11c promoter24. We examined intact explanted lymph nodes from the progeny of two of the founders by two-photon fluorescence microscopy to determine their suitability for deep tissue imaging. CD11c-EYFPhi mice had many bright fluorescent cells at depths of 175–300 μm; these cells were not visible in CD11c-EYFPlo mice (Fig. 1a and Supplementary Videos 1 and 2 online). In agreement with results of the two-photon experiments, brightly stained cells (EYFP+ cells) were also present in frozen sections from lymph nodes only in the CD11c-EYFPhi mice, and these cells expressed endogenous CD11c, as determined by immunofluorescence (Fig. 1b,c). By flow cytometry, we found that the EYFP+ cells bright enough to be visualized intravitally were CD11c+CD19CD3CD86+ major histocompatibility complex (MHC) class II–positive DCs (Fig. 1d and Supplementary Fig. 1 online). A small percentage of B cells and T cells also expressed EYFP, but the amount of EYFP expressed by these cells was at least one log lower than that expressed by DCs and was consequently below the threshold of detection in the microscope (determined by comparison with CD11c-EYFPlo; Fig. 1 and Supplementary Fig. 1 online). Thus, the fluorescent cells visualized by two-photon microscopy in CD11c-EYFPhi mice corresponded to DCs with high expression of CD11c, which could be visualized in lymph nodes at depths of up to 300 μm (Supplementary Videos 1 and 2 online).

Figure 1: Brightness and surface markers of CD11c-EYFP cells.

(a) Two-photon fluorescence images descending along the z-axis through explanted lymph nodes from CD11c-EYFPhi and CD11c-EYFPlo mice, with EYFP in green and second harmonics from collagen fibers in blue; depth from the apex of the lymph node (in μm) is above the images. Only CD11c-EYFPhi cells can be visualized deep in the lymph node. Scale bars, 50μm. (b) Fluorescence micrographs of lymph node sections from wild-type (WT), CD11c-EYFPlo and CD11c-EYFPhi mice stained with anti-CD3 (red) and anti-B220 (blue); EYFP fluorescence was visualized directly. Scale bars, 200 μm. (c) Confocal micrographs of lymph node sections immunostained for CD11c in red, with EYFP fluorescence in green. Scale bars, 10 μm. (d) Flow cytometry histograms showing surface marker expression on EYFP+ cells from CD11c-EYFPhi mice, gated on cells bright enough to detect intravitally (Supplementary Fig. 1 online). Results are representative of at least two experiments. MHC II, MHC class II. Control, isotype control antibody.

EYFP+ DC distribution in lymph nodes

In lymph node sections, many EYFP+ DCs were present in the perifollicular zone and T cell area, and occasional cells were present in the B cell follicle and subcapsular space (Figs. 1 and 2); these cells were not macrophages or granulocytes because they did not have high expression of the cell surface markers MOMA-1 (Fig. 2a), CD11b (Fig. 2b), F4/80 (Fig. 2c) or Gr-1 (Fig. 2d). Unlike mature DCs that recently immigrated to the lymph node20,25, steady-state DCs were not preferentially associated with high endothelial venules (HEVs; Fig. 2e,f and Supplementary Video 3 online). By immunofluorescence, EYFP+ DCs were heterogeneous for expression of MHC class II, the DC marker DEC-205 and CD86 (Fig. 2g–i). There was a partial overlap between populations of cells expressing EYFP and MHC class II (Fig. 2g) as well as between populations that expressed EYFP and DEC-205 (Fig. 2h). As noted above, by flow cytometry, which is a more sensitive technique, all EYFP+ cells expressed surface MHC class II, but they showed a broad range of expression of this antigen (Fig. 1d and Supplementary Fig. 1 online). We conclude that transgenic CD11c-EYFP DCs resemble their wild-type counterparts in expressing heterogeneous amounts of traditional markers of DC maturation26,27.

Figure 2: CD11c-EYFP mice examined by immunofluorescence.

Confocal micrographs of lymph node sections stained for various cell surface markers in red (above images) and counterstained in blue with CD3 or B220 to provide anatomical definition in the lymph node. Squares linked to arrows indicate areas magnified at right. Scale bars, 200 μm (low-resolution images) or 100 μm (high-magnification insets). Results are representative of at least three experiments.

In vivo imaging of steady-state DCs

We used two-photon intravital microscopy to examine the inguinal lymph nodes of anaesthetized mice. We maintained the inguinal lymph node at 37 °C during imaging using a stage warmer and an objective heater and confirmed the temperature with a temperature probe close to the lymph node. This system enabled the acquisition of three-dimensional volumes in 15–30 s, including collagen fibers visualized through second-harmonic imaging (Supplementary Video 4 online) and three fluorescence channels including enhanced cyan fluorescent protein (ECFP), enhanced green fluorescent protein (EGFP) and EYFP (Fig. 3a and Supplementary Video 3 online). Blood flow was verified by intravenous injection of rhodamine dextran (Fig. 3b and Supplementary Video 3 online). We found abundant rolling and attached cells in blood vessels that we verified as HEVs by in vivo direct staining with Alexa 546–conjugated MECA-79 antibody, an HEV marker (Fig. 3c), or by morphological criteria (Supplementary Video 3 online). Lymph flow was also maintained, based on drainage of subcutaneously injected rhodamine dextran through the efferent lymphatics (data not shown). Our blood and lymph flow results were similar to those reported before28 and verified that the inguinal lymph node preparation could be used to assess the in vivo dynamics of DCs in physiological conditions. Finally, we also examined popliteal lymph nodes20 and found no differences in DC activity in the two preparations (data not shown).

Figure 3: Two-photon intravital microscopy of inguinal lymph nodes.

(a) Three-dimensional reconstruction from two-photon microscopy data showing the anatomic distribution of perifollicular and deeper EYFP+ DCs (green) relative to the lymph node capsule (second-harmonics signals from collagen are in blue). (b) Three-dimensional reconstruction from two-photon microscopy data with EYFP+ DCs (green) and blood vessels (red) visualized by intravenous injection of rhodamine-labeled dextran. (c) Maximum intensity projection of a 20-μm-thick volume showing EYFP+ DCs and HEVs (red) visualized by injection of labeled MECA-79 antibody. Scale bar, 50 μm.

Two-photon intravital microscopy of the inguinal lymph node showed distinct populations of DCs (Fig. 4 and Supplementary Video 1 online). The most superficial EYFP cells were in the subcapsular sinus (Fig. 4a). These DC had few dendrites and multiple large ruffles, which they extended and retracted rapidly (Supplementary Videos 5 and 6 online). Subcapsular DCs were intercalated among collagen fibers (visualized by second harmonics) and relatively sessile phagocytic macrophages, visualized by their uptake of 66-kilodalton rhodamine-dextran29. The subcapsular DCs resembled EYFP cells on serosal surfaces outside the lymph node (Supplementary Video 7 online), which may suggest that they were recent immigrants from the periphery30.

Figure 4: Anatomic distribution and activity of DCs in the lymph node.

(af) Inguinal lymph nodes from CD11c-EYFP mice examined by intravital microscopy. EYFP fluorescence is in green. Scale bars, 50 μm. (a) Subcapsular DCs surrounded by macrophages that have taken up rhodamine-labeled dextran. (b) DCs in the T cell area (T-zone) of the lymph nodes, with transferred EGFP+ T cells 'false-colored' red. (c) DCs form a network surrounding the B cell follicles. (d) DC cluster at the boundary between T cell and B cell areas (T-B interface). (e) Inguinal lymph node diagram indicates the anatomical locations of the DCs in the surrounding images. (f) DCs in a B cell follicle (B-zone), with transferred EGFP+ B cells 'false-colored' blue. (g) 'Crawling' speeds of DCs vary with anatomical location. Each point represents the average speed of one cell as calculated by the movement of the centroid; cells with velocities less than 2 μm/min are considered sessile. Horizontal lines indicate median speeds. Cells tend to arrest when incorporated into cell networks, although individual fast-moving cells do occur in all locations. Results summarize at least three imaging sessions in each location.

In contrast, most of the DCs in the lymph node parenchyma were in extensive networks that are difficult to appreciate by conventional histology or intravital studies of immigrant mature DCs19,20,21. T cell zones and B cell follicles were demarcated by injection of CD11c-EYFP mice with EGFP-expressing T or B cells, respectively. DCs in deep paracortical networks interspersed with T cells showed extensive 'probing', and more than 95% seemed to be firmly tethered to the network and were thus sessile (Fig. 4b and Supplementary Videos 8 and 9 online).

Three different populations of DCs were associated with the B cell follicles: sessile perifollicular DCs between the subcapsular space and B cell follicle; DC clusters at the junction of T and B cell zones; and scattered DCs in the B cell follicle. The perifollicular DCs (Fig. 4c and Supplementary Videos 4 and 10 online) were relatively 'dim' for EYFP, showed 'probing' motions and seemed well positioned to acquire antigens from lymph. In the junction between the T and B cell zones, DCs were present in sessile clusters that varied in size up to 80 μm across and contained sessile EYFP lymphocytes (Fig. 4d and Supplementary Video 11 online; discussed below); DCs in clusters had very dynamic membrane extensions and seemed to actively capture lymphoid cells from their surroundings (Supplementary Video 12 online). EYFP-bright DCs were sparsest in the B cell follicles but they were on average the most motile DC population in the lymph node (Fig. 4f,g). We conclude that in the steady state, DCs in lymph nodes form previously unappreciated groups of cells that can be distinguished based on their activity. These include a minority of highly motile discrete cells and extensive assemblies of interconnected sessile cells in networks and clusters.

Quantitative analysis of steady-state DC movement

Measurement of DC motility is challenging because active 'probing' of the DC means that the three-dimensional center of fluorescence (centroid) for a DC is continually and rapidly shifting independent of displacement (actual movement through the tissue). This can be shown with different colors to demonstrate the outline of a single DC at 30-second intervals (Supplementary Fig. 2 online). Because these movements are rapid, the rate at which three-dimensional image volumes are acquired in the tissue has a large effect on the measured instantaneous and average speeds. This makes it difficult to compare speed measurements between different studies and also means that studies with high sampling rates may overestimate DC speed. To derive speeds that are related to displacement in the tissue, we applied a filter that rejected small movements in each step and only recorded movements that were greater than a threshold of 2.5 μm (Supplementary Fig. 2 online). Although we obtained three-dimensional data sets every 30 s, this filter made the three-dimensional analysis of the tracks relatively independent of the sampling rate and yielded a speed better related to displacement of the DC over time, a parameter that should be more comparable between studies (Supplementary Fig. 2 online). Other studies have made use of plots of displacement versus the square root of time to derive roughly linear plots consistent with a random 'walk'20,21,31. The slope of these plots can be used to determine a motility coefficient31, which is independent of the data acquisition rate. We used displacement as a filter at the single-cell level by defining a 'sessile' cell as a cell that did not leave a sphere 20 μm in diameter within 30 min.

Most steady-state DCs in T cell zones were sessile, with a median speed slower than 1 μm/min, whereas their motility coefficient was 1.2 μm2/min (Fig. 4g, Supplementary Fig. 2 online and data not shown). Perifollicular DCs were mostly sessile, with a median speed of 1.6 μm/min (Fig. 4g). DCs in the B cell follicle were the fastest moving, with a median speed of 4 μm/min and relatively few sessile cells (Fig. 4g). There were sparse individual DCs that were 'crawling' in the follicle and may have been traversing it (Supplementary Video 10 online). In almost all cases, cells that were sessile by displacement thresholds had a nonzero speed, mostly because of 'probing' movements too large to be filtered out by the 2.5-μm threshold.

Steady-state lymphocyte-DC interaction

The DC clusters in the T cell–B cell interface areas seemed to actively capture round cells (Supplementary Video 12 online), which often remained in the clusters for the duration of imaging (>30 min). To determine the cellular composition of DC clusters, we labeled thick sections of fixed inguinal lymph nodes with fluorescent antibodies and examined three-dimensional reconstructions of confocal images (Supplementary Video 13 online). Clusters contained EYFP DCs, B cells and T cells (Supplementary Fig. 3 online). Because of their tight packing, it was impossible to count the DCs in a cluster with the EYFP signal alone. We therefore visualized nuclei with the nuclear stain TO-PRO-3 in combination with CD3 and CD19 immunostaining (Fig. 5) and generated reconstructions as described above; clusters typically contained up to eight DCs plus three to five B and T cells per DC on the interior (n = 54 clusters measured). Lymphocytes in these clusters were atypically round and sessile, rather than being polarized and mobile like B and T cells in the B cell follicles and T cell zones. Because DC clusters were not always present, and because they incorporate both B and T cells, we speculate clusters might be involved in later stages (such as B cell–T helper cell interactions) of ongoing immune responses.

Figure 5: DC clusters include more than one DC.

Series of confocal images optically sectioning an inguinal lymph node, with EYFP in green, DNA in red (TO-PRO 3), and CD3 and CD19 in blue. Arrows indicate nuclei from DCs in a single cluster. DCs (A–H) show no CD3 or CD19 staining. Numbers above images indicate the depth of the optical plane in the tissue section.

Mature DC interactions with endogenous networks

As steady-state DCs acted differently than mature migratory DCs observed before19,20,21, we directly compared these two cell populations by intravital imaging of the inguinal lymph node. We injected purified, lipopolysaccharide (LPS)–activated ECFP+ DCs32 intradermally into the flanks of CD11c-EYFP mice as described20. We examined inguinal lymph nodes histologically and by intravital microscopy from 12 to 72 h after DC injection. Transferred DCs were sparse in the inguinal nodes for up to 18 h; by 24 h there was visible swelling of the lymph node, and ECFP DCs had migrated into the lymph node and settled most densely at the interface between the B and T cell zones (Fig. 6a). Although this location brought them near HEVs, ECFP DCs were not directly associated with HEVs, based on immunostaining with MECA-79 antibody (Fig. 6a). DCs were present throughout the T cell area at 24 h and later time points (Fig. 6a).

Figure 6: Transferred ECFP DCs gradually incorporate into the endogenous DC network.

Splenic DCs were purified by immunomagnetic selection from the spleens of ECFP+ mice injected subcutaneously with a tumor secreting Flt3 ligand, and 4 × 106 to 6 × 106 cells were injected intradermally along with LPS into the flanks of CD11c-EYFP mice. Inguinal lymph nodes were then examined by immunofluorescence and intravital microscopy. (a) Confocal micrographs of inguinal lymph nodes immunostained for HEVs with MECA-79 antibody. Scale bars, 200 μm (low-resolution panels) or 100 μm (high-magnification insets); optical thickness is 10 μm in the ECFP channel and 1.2 μm in all other channels. (b) Plots of displacement versus the square root of time, used to calculate the motility coefficient of transferred ECFP+ cells and endogenous EYFP+ cells. Inguinal lymph nodes were imaged by multiphoton microscopy and cells were tracked as described above; the motility of transferred DCs gradually decreases as the cells incorporate into the endogenous DC network. (c) 'Scattergram' of the average velocities of individual transferred and endogenous DCs (time, horizontal axis). Horizontal lines indicate median speeds. Data in b and c summarize at least four imaging fields examined in each time point.

We tracked ECFP+ and EYFP+ cells in the T cell zones as described above (Supplementary Fig. 2 online). In agreement with the histology results, ECFP+ DCs were not associated with HEVs and by 48–72 h were evenly dispersed throughout the T cell area (Supplementary Video 14 online). At 24 and 48 h, the transferred mature ECFP DCs moved faster than the EYFP+ DCs (P < 0.001), whereas at 72 h there was no significant difference (Fig. 6b,c and Supplementary Video 14 online). The motility coefficient for the mature DC decreased with time after transfer, as reported before20 (from 1.88 μm2/min to 1.36 μm2/min), and both the immigrant and transferred populations contained sessile cells (<2 μm/min) at all time points (Fig. 6c). Furthermore, the median speed of the endogenous EYFP+ DCs increased with time after injection from 1 μm/min to 2.1 μm/min between 24 and 72 h, whereas the motility coefficient decreased from 1.2 μm2/min to 0.9 μm2/min. This suggests that the endogenous cells had more 'probing' and local dynamics without a corresponding increase in movement through the tissue. This increase in local 'probing' may have been due to the LPS injected along with the activated DCs20. At 96 h, the numbers of transferred DC in the lymph node dropped off sharply, probably indicating cell death. We conclude that mature DCs migrating to the lymph node are initially more motile than steady-state DCs, but they eventually join the sessile DC networks after distributing themselves throughout the node.


We found that in the steady state, only a minor fraction of the DCs in the lymph nodes 'crawled'. Most DCs formed dense networks of cells that touched each other with the tips of their processes, which showed rapid and extensive 'probing'. The DC networks surrounded the B cell follicles on all sides and extended into the T cell zones, but were particularly dense in the border zone between the T and B cell follicle, where T cell–dependent immune responses are initiated33,34. In this area, DCs often adhered closely to one another, creating tightly packed clusters that incorporated T and B lymphocytes, which we speculate may be involved in later stages of the immune response. Both networks and clusters were dynamic, in that 'crawling' DCs could be seen joining groups of sessile cells. When we compared endogenous DCs with mature DC immigrants, we found that the LPS-activated mature DCs moved faster than endogenous DCs at 24 and 48 h as they dispersed into the T cell zone. Nevertheless, many of the immigrant DCs were sessile at all time points, indicating that these immigrants had joined the endogenous DC network. Our finding are consistent with the observation that the speed of mature DC movement slows with time after arrival in the lymph node20,21. We speculate that mature DCs stop migrating as they join DC networks in the lymph node and that this change in activity may be required for efficient antigen presentation in vivo.

Whereas DCs in the T cell zone are referred to anatomically as 'interdigitating reticulum cells'35, based on their reticular appearance in fixed tissue sections, and are described as 'sessile'36 because they do not recirculate, the microscopic dynamics of steady-state lymph node DCs were unknown. Our experiments differ from other published reports on DC movement in vivo that showed no evidence of a sessile network and instead reported random DC 'crawling' with average speeds of 2.7–6.6 μm/min in lymph nodes19,20,21,28,31. However, all of those studies involved labeling of a small number of mature activated DCs, and therefore the steady-state DC network could not have been appreciated.

In the steady state, DC networks present T cells with an enormous surface area rich in the self MHC class II–peptide complexes required to sustain T cell longevity37,38 and to maintain self-tolerance29. Self and foreign antigens that arrive in lymph nodes from the lymph are processed simultaneously by many DCs in the network7,8,9,10,39, much like antigen delivery by antibody to DEC-205 (anti-DEC-205; refs. 40,41); this maximizes the surface area for presentation of such antigens.

We found that mature DCs that arrived in lymph nodes from tissues were more motile on average than steady-state DCs. The higher average speed of immigrant DCs may be promoted by stimulation of cytoskeletal dynamics by Toll-like receptors or chemokine signaling30. We propose that faster-moving mature DC traverse the network before joining it, whereupon they become sessile. In the network, DCs are well positioned to form stable interactions with antigen-specific T cells or to relay antigen to other DCs in the network by releasing exosomes or apoptotic bodies9,42. When such cell fragments are ingested by several DCs in the network, the antigen-presenting surface area is expanded, increasing the probability that a T cell migrating through the lymph node would meet a DC carrying its cognate antigen.

The backbone of the DC network is likely to be the reticular fiber network, a system of collagen bundles that is enveloped by reticular fibroblasts and undergoes considerable remodeling in response to inflammatory signals from tissues43,44. Immigrant DCs may migrate when in contact with T cells and other DCs and join the network only when they locate stromal cells or attachment points to collagen bundles that would provide greater anchorage.

Earlier studies suggested that DCs typically contact 500 T cells per hour (ref. 19); higher-resolution imaging showed that the 'probing' motion of DC processes enables DCs to contact up to 5,000 T cells per hour (ref. 21). This 'probing' activity is dependent on the small GTPases Rac1 and Rac2, and genetic experiments have confirmed the importance of these pathways for presentation to T cells45. Our data suggest that the DCs maintain a fixed position in the network and scan the passing T cells by 'probing' with their dendrites. The arrangement of DCs in a network throughout the T cell area forces T cells to pass among many 'probing' DCs on their course through the lymph node. This is different from models in which both T cells and DCs are engaged in extensive random searches, which are based on prior data with immigrant DCs21. Given the high motility coefficient of T cells, there is little disadvantage to fixing DCs in the network. However, it is important that mature DCs carrying antigen into the lymph node from the periphery disperse throughout the T cell zones to increase the probability that a randomly migrating T cell will contact a DC presenting its cognate antigen. We propose that the higher motility of mature DCs functions to distribute the DCs and the antigen they carry throughout the T cell zones and thereby maximize the likelihood of antigen-specific DC–T cell interactions.


Transgenic mice.

EYFP-Venus23 was cloned into a CD11c promoter24. The linearized construct was injected into C57BL/6-CBA F1 fertilized female pronuclei, and progeny positive by PCR for the transgene were backcrossed to C57BL/6 for at least six generations. C57BL/6 mice expressing ECFP on the β-actin promoter were purchased from Jackson Laboratories46. Mice were used for experiments at 4–6 weeks of age; all mice were housed in specific pathogen–free conditions and were treated in accordance with protocols approved by the Rockefeller University and New York University School of Medicine Institutional Animal Care and Use Committees.


Lymph nodes were fixed in PBS with 4% paraformaldehyde plus 10% sucrose and were cryoprotected in PBS plus 30% sucrose before being embedded in optimum cutting temperature compound and being frozen. Frozen tissue was sectioned (20 or 50 μm in thickness) on a microtome and was fixed in acetone. All incubations were done in a humidified chamber. Sections were blocked in 5% BSA and 10% serum in PBS in the presence of anti-CD16/CD32 (BD Biosciences) and were sequentially blocked with excess streptavidin and biotin (Vector Laboratories). For peroxidase amplification with biotinyl tyramides (NEN), we quenched endogenous peroxidase activity with 3% H2O2 in PBS before the initial blocking step, according to the NEN protocol. The antibodies used were directly labeled anti-CD3ε–Alexa 647, anti-B220–Alexa 647 (CalTag), biotinylated anti-CD3ε, anti-CD1b, anti-CD11c, anti-CD19, anti-CD80, anti-CD86, anti-B220, anti-F4/80, anti-Gr-1, anti-I-Ab, anti-PECAM, anti-rat IgM (BD Biosciences), MOMA-1 (BMA Biomedicals), anti-CD11c, anti-CD45.2 (EBioscience), unconjugated MECA-79 (BD Biosciences) and biotin and Alexa 647 conjugates of anti-DEC-205 produced and conjugated by the Rockefeller University Monoclonal Antibody Core Facility (New York, New York). Streptavidin-indocarbocyanine was from Jackson ImmunoResearch; TO-PRO-3 was from Molecular Probes. Sections were mounted in Fluoromount-G (Southern Biotech) and were stored at 4 °C.

Confocal microscopy.

Confocal images were acquired on a Zeiss LSM 510 system with 458-, 488-, 514-, 543- and 633-nm excitation lines at the Rockefeller University Bio-Imaging Facility; EYFP and ECFP fluorescence were visualized directly in all images. Tiled images were obtained with the motorized stage using a 40 × or 63 × objective to capture the entire cut surface of the lymph node, and images were exported into Adobe Photoshop for final processing. Three-dimensional reconstructions were generated from Z-stacks with Imaris (Bitplane).

Flow cytometry.

Single-cell suspensions of lymph nodes were pretreated with anti-CD16/CD32 and were stained with CD11c-allophycocyanin and biotinylated antibodies specific for CD3, CD4, CD8a, CD11b, CD19, B220, CD86, DEC-205 and MHC class II (I-Ab; Becton Dickinson Pharmingen). Living cells were gated based on forward- and side-scatter on a FACSVantage SE (Becton-Dickinson) and data were analyzed with FlowJo (Tree Star).

Cell transfer.

To demarcate B cell follicles, we transferred 1 × 108 B cells purified from EGFP+ C57BL/6 mice by magnetic depletion with anti-CD43 (Miltenyi Biotec). CD11c+ DCs were purified by immunomagnetic selection from the spleens of ECFP+ mice injected with cells from a tumor expressing Flt3 ligand and 4 × 106 to 6 × 106 cells injected intradermally along with 1 ng/μl LPS (Sigma) at four sites surrounding the inguinal lymph node. To image the macrophages in the subcapsular sinus and blood vessels throughout the lymph node, we injected 1 mg of rhodamine-dextran (66 kilodalton; Molecular Probes) in 100 μl saline subcutaneously in the base of the tail. To examine high endothelial veins in vivo, we injected 10 μg of Alexa 594–labeled MECA-79 antibody intravenously.


Mice were anaesthetized with 100 mg ketamine, 15 mg xylazine and 2.5 mg acepromazine per kg body weight and were kept anaesthetized with hourly injections of half this dose. Mice were restrained on a stage warmer at 37 °C (BioTherm Micro S37; Biogenics) and the abdominal skin was incised from the edge of the rib cage through the midline to the thigh. A skin flap was separated from the abdominal muscle to expose the inguinal lymph node.

To stabilize, moisturize and maintain the lymph node at 37 °C, we placed the skin flap on a thermoconductive base made of silicone elastomer (Sylgard 184; Dow Corning) surrounding a core of thermoconductive putty (T-putty 502; Thermagon). The connective and adipose tissue covering the lymph node was removed by microsurgery that kept blood and lymph vessels intact, and a saline-filled chamber consisting of a cover-slip glued to a nylon washer was mounted on the skin flap over the lymph node. The chamber was laterally secured by three 'insect pins' that pierced the skin and lodged in the silicone base. The temperature was monitored with a probe (Bioptechs) placed near the lymph node to ensure that it was at 37 ± 1 °C. To compensate for anesthesia-induced respiratory depression, we provided mice with 100% oxygen by mask throughout the imaging session.

Intravital microscopy.

The lymph node was imaged with a Bio-Rad Radiance multiphoton microscope fitted with a Tsunami pulsed laser (Spectraphysics) and controlled by Lasersharp2000 software (Bio-Rad) and a Nikon 40 × objective (water immersed; numerical aperture, 0.8). For imaging of EYFP, EGFP and ECFP, the excitation wavelength was set to 910 nm; for rhodamine, 870 nm was used. Band-pass filters optimized for the detection of EYFP (550/30) and ECFP (510/30) and a 530 long-pass dichroic mirror were used to detect fluorescent proteins; EGFP appeared in both channels. To create time-lapse sequences, we typically scanned volumes of tissue 50 × 400 × 400 μm at 3-μm Z-steps and 30-second intervals.

Data analysis.

Cell movement was analyzed with Volocity software (Improvision). For cell speeds, the coordinates of each cell were calculated and tracked over time; to eliminate motion artifacts caused by dendrite 'probing', pulse and breathing, displacements smaller than 2.5 μm were filtered out of the cell tracks. Despite efforts to filter high-frequency movement related to 'probing', DCs that were distinctly sessile still seemed to have velocities up to 2 μm/min; this velocity was therefore used as the threshold below which cells were considered sessile. As cell speed showed an asymmetrical distribution, data are presented in 'scattergrams' and were compared with a nonparametric 'bootstrapping' procedure based on 10,000 resampling iterations47. P values less than 0.05 were considered significant.

Note: Supplementary information is available on the Nature Immunology website.


  1. 1

    Banchereau, J., Pascual, V. & Palucka, K.A. Autoimunity through cytokine-induced dendritic cell activation. Immunity 20, 539–550 (2004).

    CAS  Article  Google Scholar 

  2. 2

    Guermonprez, P., Valladeau, J., Zitvogel, L., Thery, C. & Amigorena, S. Antigen presentation and T cell stimulation by dendritic cells. Annu. Rev. Immunol. 20, 621–667 (2002).

    CAS  Article  Google Scholar 

  3. 3

    Lanzavecchia, A. & Sallusto, F. Regulation of T cell immunity by dendritic cells. Cell 106, 263–266 (2001).

    CAS  Article  Google Scholar 

  4. 4

    Liu, Y.J. Dendritic cell subsets and lineages, and their functions in innate and adaptive immunity. Cell 106, 259–262 (2001).

    CAS  Article  Google Scholar 

  5. 5

    Mellman, I. & Steinman, R.M. Dendritic cells: specialized and regulated antigen processing machines. Cell 106, 255–258 (2001).

    CAS  Article  Google Scholar 

  6. 6

    Steinman, R.M., Hawiger, D. & Nussenzweig, M.C. Tolerogenic dendritic cells. Annu. Rev. Immunol. 21, 685–711 (2003).

    CAS  Article  Google Scholar 

  7. 7

    Itano, A.A. et al. Distinct dendritic cell populations sequentially present antigen to CD4 T cells and stimulate different aspects of cell-mediated immunity. Immunity 19, 47–57 (2003).

    CAS  Article  Google Scholar 

  8. 8

    Manickasingham, S. & Reis e Sousa, C. Microbial and T cell-derived stimuli regulate antigen presentation by dendritic cells in vivo. J. Immunol. 165, 5027–5034 (2000).

    CAS  Article  Google Scholar 

  9. 9

    Inaba, K. et al. Efficient presentation of phagocytosed cellular fragments on the major histocompatibility complex class II products of dendritic cells. J. Exp. Med. 188, 2163–2173 (1998).

    CAS  Article  Google Scholar 

  10. 10

    Belz, G.T. et al. Distinct migrating and nonmigrating dendritic cell populations are involved in MHC class I-restricted antigen presentation after lung infection with virus. Proc. Natl. Acad. Sci. USA 101, 8670–8675 (2004).

    CAS  Article  Google Scholar 

  11. 11

    Cyster, J.G. Chemokines and the homing of dendritic cells to the T cell areas of lymphoid organs. J. Exp. Med. 189, 447–450 (1999).

    CAS  Article  Google Scholar 

  12. 12

    Crowley, M., Inaba, K. & Steinman, R.M. Dendritic cells are the principal cells in mouse spleen bearing immunogenic fragments of foreign proteins. J. Exp. Med. 172, 383–386 (1990).

    CAS  Article  Google Scholar 

  13. 13

    Bujdoso, R., Hopkins, J., Dutia, B.M., Young, P. & McConnell, I. Characterization of sheep afferent lymph dendritic cells and their role in antigen carriage. J. Exp. Med. 170, 1285–1301 (1989).

    CAS  Article  Google Scholar 

  14. 14

    Liu, L.M. & MacPherson, G.G. Antigen acquisition by dendritic cells: intestinal dendritic cells acquire antigen administered orally and can prime naive T cells in vivo. J. Exp. Med. 177, 1299–1307 (1993).

    CAS  Article  Google Scholar 

  15. 15

    Holt, P.G., Schon-Hegrad, M.A. & Oliver, J. MHC class II antigen-bearing dendritic cells in pulmonary tissues of the rat. Regulation of antigen presentation activity by endogenous macrophage populations. J. Exp. Med. 167, 262–274 (1988).

    CAS  Article  Google Scholar 

  16. 16

    Vermaelen, K.Y., Carro-Muino, I., Lambrecht, B.N. & Pauwels, R.A. Specific migratory dendritic cells rapidly transport antigen from the airways to the thoracic lymph nodes. J. Exp. Med. 193, 51–60 (2001).

    CAS  Article  Google Scholar 

  17. 17

    Huang, F.P. et al. A discrete subpopulation of dendritic cells transports apoptotic intestinal epithelial cells to T cell areas of mesenteric lymph nodes. J. Exp. Med. 191, 435–444 (2000).

    CAS  Article  Google Scholar 

  18. 18

    Belz, G.T. et al. The CD8α+ dendritic cell is responsible for inducing peripheral self-tolerance to tissue-associated antigens. J. Exp. Med. 196, 1099–1104 (2002).

    CAS  Article  Google Scholar 

  19. 19

    Bousso, P. & Robey, E. Dynamics of CD8+ T cell priming by dendritic cells in intact lymph nodes. Nat. Immunol. 4, 579–585 (2003).

    CAS  Article  Google Scholar 

  20. 20

    Mempel, T.R., Henrickson, S.E. & Von Andrian, U.H. T-cell priming by dendritic cells in lymph nodes occurs in three distinct phases. Nature 427, 154–159 (2004).

    CAS  Article  Google Scholar 

  21. 21

    Miller, M.J., Hejazi, A.S., Wei, S.H., Cahalan, M.D. & Parker, I. T cell repertoire scanning is promoted by dynamic dendritic cell behavior and random T cell motility in the lymph node. Proc. Natl. Acad. Sci. USA 101, 998–1003 (2004).

    CAS  Article  Google Scholar 

  22. 22

    Miller, M.J., Safrina, O., Parker, I. & Cahalan, M.D. Imaging the single cell dynamics of CD4+ T cell activation by dendritic cells in lymph nodes. J. Exp. Med. 200, 847–856 (2004).

    CAS  Article  Google Scholar 

  23. 23

    Nagai, T. et al. A variant of yellow fluorescent protein with fast and efficient maturation for cell-biological applications. Nat. Biotechnol. 20, 87–90 (2002).

    CAS  Article  Google Scholar 

  24. 24

    Brocker, T., Riedinger, M. & Karjalainen, K. Driving gene expression specifically in dendritic cells. Adv. Exp. Med. Biol. 417, 55–57 (1997).

    CAS  Article  Google Scholar 

  25. 25

    Bajenoff, M., Granjeaud, S. & Guerder, S. The strategy of T cell antigen-presenting cell encounter in antigen-draining lymph nodes revealed by imaging of initial T cell activation. J. Exp. Med. 198, 715–724 (2003).

    CAS  Article  Google Scholar 

  26. 26

    Wilson, N.S. et al. Most lymphoid organ dendritic cell types are phenotypically and functionally immature. Blood 102, 2187–2194 (2003).

    CAS  Article  Google Scholar 

  27. 27

    Shortman, K. & Liu, Y.J. Mouse and human dendritic cell subtypes. Nature Rev. Immunol. 2, 151–161 (2002).

    CAS  Article  Google Scholar 

  28. 28

    Miller, M.J., Wei, S.H., Cahalan, M.D. & Parker, I. Autonomous T cell trafficking examined in vivo with intravital two-photon microscopy. Proc. Natl. Acad. Sci. USA 100, 2604–2609 (2003).

    CAS  Article  Google Scholar 

  29. 29

    Kang, Y.S. et al. SIGN-R1, a novel C-type lectin expressed by marginal zone macrophages in spleen, mediates uptake of the polysaccharide dextran. Int. Immunol. 15, 177–186 (2003).

    CAS  Article  Google Scholar 

  30. 30

    Ohl, L. et al. CCR7 governs skin dendritic cell migration under inflammatory and steady-state conditions. Immunity 21, 279–288 (2004).

    CAS  Article  Google Scholar 

  31. 31

    Miller, M.J., Wei, S.H., Parker, I. & Cahalan, M.D. Two-photon imaging of lymphocyte motility and antigen response in intact lymph node. Science 296, 1869–1873 (2002).

    CAS  Article  Google Scholar 

  32. 32

    Okabe, M., Ikawa, M., Kominami, K., Nakanishi, T. & Nishimune, Y. 'Green mice' as a source of ubiquitous green cells. FEBS Lett. 407, 313–319 (1997).

    CAS  Article  Google Scholar 

  33. 33

    Jacob, J., Kassir, R. & Kelsoe, G. In situ studies of the primary immune response to (4-hydroxy-3-nitrophenyl)acetyl. I. The architecture and dynamics of responding cell populations. J. Exp. Med. 173, 1165–1175 (1991).

    CAS  Article  Google Scholar 

  34. 34

    Liu, Y.J., Zhang, J., Lane, P.J., Chan, E.Y. & MacLennan, I.C. Sites of specific B cell activation in primary and secondary responses to T cell-dependent and T cell-independent antigens. Euro. J. Immunol. 21, 2951–2962 (1991).

    CAS  Article  Google Scholar 

  35. 35

    Friess, A. Interdigitating reticulum cells in the popliteal lymph node of the rat. An ultrastructural and cytochemical study. Cell Tissue Res. 170, 43–60 (1976).

    CAS  Article  Google Scholar 

  36. 36

    Fossum, S. Lymph-borne dendritic leucocytes do not recirculate, but enter the lymph node paracortex to become interdigitating cells. Scand. J. Immunol. 27, 97–105 (1988).

    CAS  Article  Google Scholar 

  37. 37

    Takeda, S., Rodewald, H.R., Arakawa, H., Bluethmann, H. & Shimizu, T. MHC class II molecules are not required for survival of newly generated CD4+ T cells, but affect their long-term life span. Immunity 5, 217–228 (1996).

    CAS  Article  Google Scholar 

  38. 38

    Kirberg, J., Berns, A. & von Boehmer, H. Peripheral T cell survival requires continual ligation of the T cell receptor to major histocompatibility complex-encoded molecules. J. Exp. Med. 186, 1269–1275 (1997).

    CAS  Article  Google Scholar 

  39. 39

    Inaba, K. et al. High levels of a major histocompatibility complex II-self peptide complex on dendritic cells from the T cell areas of lymph nodes. J. Exp. Med. 186, 665–672 (1997).

    CAS  Article  Google Scholar 

  40. 40

    Bonifaz, L. et al. Efficient targeting of protein antigen to the dendritic cell receptor DEC-205 in the steady state leads to antigen presentation on major histocompatibility complex class I products and peripheral CD8+ T cell tolerance. J. Exp. Med. 196, 1627–1638 (2002).

    CAS  Article  Google Scholar 

  41. 41

    Hawiger, D. et al. Dendritic cells induce peripheral T cell unresponsiveness under steady state conditions in vivo. J. Exp. Med. 194, 769–779 (2001).

    CAS  Article  Google Scholar 

  42. 42

    Thery, C. et al. Molecular characterization of dendritic cell-derived exosomes. Selective accumulation of the heat shock protein hsc73. J. Cell Biol. 147, 599–610 (1999).

    CAS  Article  Google Scholar 

  43. 43

    Kaldjian, E.P., Gretz, J.E., Anderson, A.O., Shi, Y. & Shaw, S. Spatial and molecular organization of lymph node T cell cortex: a labyrinthine cavity bounded by an epithelium-like monolayer of fibroblastic reticular cells anchored to basement membrane-like extracellular matrix. Int. Immunol. 13, 1243–1253 (2001).

    CAS  Article  Google Scholar 

  44. 44

    Katakai, T., Hara, T., Sugai, M., Gonda, H. & Shimizu, A. Lymph node fibroblastic reticular cells construct the stromal reticulum via contact with lymphocytes. J. Exp. Med. 200, 783–795 (2004).

    CAS  Article  Google Scholar 

  45. 45

    Benvenuti, F. et al. Requirement of Rac1 and Rac2 expression by mature dendritic cells for T cell priming. Science 305, 1150–1153 (2004).

    CAS  Article  Google Scholar 

  46. 46

    Hadjantonakis, A.K., Macmaster, S. & Nagy, A. Embryonic stem cells and mice expressing different GFP variants for multiple non-invasive reporter usage within a single animal. BMC Biotechnol. 2, 11 (2002).

    Article  Google Scholar 

  47. 47

    Efron, B. Bootstrap methods: another look at the jackknife. Ann. Statist. 7, 1–26 (1979).

    Article  Google Scholar 

Download references


We thank R. Steinman for discussions and W. Gan for guidance on two-photon microscopy. Supported by the National Institutes of Health (AI55037 to M.L.D. and AI051573 to M.C.N.), Irene Diamond Foundation (M.L.D.), Rothschild Foundation (G.S.), Medical Scientist Training Program (GM07739 to R.L.L.), German Research Foundation (DU 548/1-1 to D.P.) and Howard Hughes Medical Institute (M.C.N.).

Author information



Corresponding authors

Correspondence to Michael L Dustin or Michel C Nussenzweig.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Supplementary information

Supplementary Fig. 1

Flow cytometric analysis of EYFP cells visible by two-photon microscopy. (PDF 366 kb)

Supplementary Fig. 2

Quantitative analysis of DC movement. (PDF 500 kb)

Supplementary Fig. 3

CD11c+ DCs form clusters with T and B cells in the steady state. (PDF 2360 kb)

Supplementary Video 1

Z-stack of two-photon images taken at depths of 0-300 μm through a B cell follicle in the LN of a living mouse. EYFP DCs are green, adoptively transferred EGFP B cells are false-colored cyan. Four populations of DCs are visible. Bright extended subcapsular DCs (arrows); dimmer DCs in a network surrounding the follicle; scattered compact DCs in the follicles (circles) and an extensive network of DCs in the T cell zone of the paracortex. (MOV 3861 kb)

Supplementary Video 2

Three-dimensional reconstruction of the different DC populations around a B cell follicle based on the data presented in Video 1. Colors are as in Video 1. (MOV 3453 kb)

Supplementary Video 3

No close association of resident CD11c-YFP cells with HEVs. Blood vessels were visualized with 66 kDa TRITC-dextran. The HEVs (crossing from top left to bottom right), can be distinguished from arterioles (crossing in the perpendicular direction) based on their irregular epithelium and the shadows of lymphocytes slowly rolling along it. YFP cellular debris is seen on the middle left. A mixture of slowly moving and sessile CD11c-YFP cells is seen scattered in the general area of the HEVs, but not immediately juxtaposed to them. Also seen are CFP lymphocytes (cyan) which exhibit a tighter morphology than DCs and can reach higher maximum speeds. (MOV 2956 kb)

Supplementary Video 4

Perifollicular DCs in relation to the LN capsule. A 3-dimensional reconstruction of a lymph node, showing the spatial relation of the perifollicular network of EYFP DCs (green) to the fibrous capsule (blue, second harmonics signal from collagen). (MOV 2480 kb)

Supplementary Video 5

Subcapsular sinus DCs and macrophages. Subcapsular sinus EYFP DCs (green) navigating among the less mobile macrophages, visualized by their phagocytosis of subcutaneously injected 66 kDa rhodamine-dextran (red). As in most videos below, a two-dimensional projection of a 50 μm thick volume is shown. (MOV 2793 kb)

Supplementary Video 6

Subcapsular DCs probing movement. At higher resolution, the probing movement of subcapsular DCs can be better appreciated. Note the characteristic ruffle shaped extensions of these cells. Dimmer, less motile perifollicular DCs are seen in the background. (MOV 3227 kb)

Supplementary Video 7

Tissue DC crawling. EYFP DCs (green) show fast crawling on the serosal surface lining the fat pads that neighbor the inguinal LN. Intravenously injected rhodamine dextran (red) flows in thin capillaries, bright red spots represent endocytosed dextran within endothelial cells. The morphology of these cells resembles that of subcapsular DCs in the LN. (MOV 1642 kb)

Supplementary Video 8

T cell zone DCs. EYFP DCs (green) form an extensive network in the T cell zone of the LN. Note that the DCs exhibit extensive probing movements but show little crawling. (MOV 3825 kb)

Supplementary Video 9

T cell zone DC network dynamics. A time sequence depicting a two-dimensional projection of a 50 μm volume in the interface of the T cell and B cell zones. The T cell zone is located below and to the left of the B cell follicle. The behavior of CD11c-EYFP DCs (green) in the network was followed. The great majority of the cells are laterally stable, exhibiting only probing movement, but occasionally a cell could be seen repositioning within the network (red circles). Adoptively transferred EGFP B cells are false-colored cyan. (MOV 2577 kb)

Supplementary Video 10

Different movement patterns of follicular and perifollicular DCs. Different behaviors of EYFP DCs (green) and ECFP B cells (cyan) in a B cell follicle (top right) and the perifollicular network (bottom left). Whereas some crawling movement is observed in the follicle, the DCs in the perifollicular network appear more dendritic and mainly probe with their processes. (MOV 1039 kb)

Supplementary Video 11

DC cluster Z stack. A Z-stack (45 μm deep, at 1 μm intervals) through several DC clusters in the T-B interface area. Each cluster is made up of numerous tightly apposed DCs enveloping lymphocytes. (MOV 2407 kb)

Supplementary Video 12

The behavior of DC clusters in the T-B interface zone (50 μm thick volume). Cluster position was stable for the duration and remained the same when the area was imaged again 4 hours later. DCs can be seen joining (blue circles) or leaving (red circles) the cluster. The shadows of lymphocytes can be seen drawn into the clusters enveloped in DC processes (yellow circles). (MOV 3422 kb)

Supplementary Video 13

Three-dimensional reconstruction of a DC cluster from the T-B interface area. The source data are confocal images that optically sectioned an immunofluorescently stained frozen section. EYFP is green, CD3 is red, and B220 is blue. Several DCs form a cluster that encloses numerous B and T cells. (MOV 2920 kb)

Supplementary Video 14

Resident LN DCs vs. transferred mature DCs of splenic origin. Transferred ECFP+ DCs (cyan), immunomagnetically purified from the spleen of a mouse treated with a FLT3-L-secreting tumor are seen here moving among resident CD11c-EYFP cells (yellow). The sequence was recorded at a depth of 180-230 μm 48 h after cell transfer. In this particular field, transferred cells crawled slightly faster than residents. Most of the movement consists of process probing, with cells occasionally flowing their soma and nucleus into one of the processes. To emphasize cell morphology, the CFP channel is shown alone on the right. (MOV 2619 kb)

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Lindquist, R., Shakhar, G., Dudziak, D. et al. Visualizing dendritic cell networks in vivo. Nat Immunol 5, 1243–1250 (2004).

Download citation

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing