The migration of developing T cells (thymocytes) between distinct thymic microenvironments is crucial for their development. Ex vivo studies of thymus tissue explants suggest two distinct migratory behaviors of thymocytes in the thymus. In the cortex, thymocytes exhibit a stochastic migration, whereas medullary thymocytes show confined migratory behavior. Thus far, it has been difficult to follow all thymocytes in an entire thymus and relate their differentiation steps to their migratory dynamics. To understand the spatial organization of the migratory behavior and development of thymocytes in a fully functional thymus, we developed transgenic reporter lines for the chemokine receptors ccr9a and ccr9b, as well as for rag2, and used them for noninvasive live imaging of the entire thymus in medaka (Oryzias latipes). We found that the expression of these two chemokine receptors in the medaka juvenile thymus defined two spatially distinct subpopulations of thymocytes. Landmark events of T cell development including proliferation, somatic recombination, and thymic selection can be mapped to subregions of the thymus. The migratory behavior of thymocytes within each of the subpopulations is equally heterogeneous, and specific migratory behaviors are not associated with particular domains in the thymus. During the period when thymocytes express rag2 their migratory behavior was more homogeneous. Therefore, the migratory behavior of thymocytes is partly correlated with their developmental stage rather than being defined by their spatial localization.

The thymus is the primary lymphoid organ for T lymphocyte development. Lymphoid precursors that originate in the hematopoietic organ and express the appropriate chemokine receptors migrate toward and then enter the thymus (reviewed in Refs. 1, 2). Upon entry, developing T cells (thymocytes) migrate between distinct thymic regions, each of which constitutes a specific signaling environment. The migration in and out of these defined microenvironments is necessary for the orderly sequence of events associated with T cell development (reviewed in Ref. 3).

Chemokines and their receptors play an important role in guiding thymocytes toward different microenvironments (1, 4). In situ expression analyses in mice have shown that stromal cells in the thymus express distinct chemokines and thereby define different thymic microenvironments (5). The sequential expression of chemokine receptors by thymocytes apparently directs their migratory path through these microenvironments (6). However, information on dynamic aspects of thymocyte movements in each thymic region cannot be obtained by looking at distributions of thymocytes in fixed tissue. The size and position of the mouse thymus hinder the study of dynamics of thymocyte trafficking in an entire thymus in vivo; therefore, to visualize dynamic aspects of thymocyte development, two-photon imaging techniques have been used in thymic explant preparations (715). These studies have shown that thymocytes behave differently in distinct thymic microenvironments. Cortical thymocytes migrate in a random walk and, after receiving positive selection signals, migrate toward the medulla with increased speed (8, 9, 11, 12, 15). Once in the medulla, thymocytes interact with medullary thymic epithelial cells and dendritic cells (DCs), and they move faster while displaying a confined migratory behavior (12). It is poorly understood to what extent the native environment of the organ, including chemokine production and distribution, is disrupted in this experimental set up and how this influences thymocyte motility (16). Hence, the type of migratory pattern that thymocytes adopt in a thymus with an intact architecture and chemokine milieu is unclear.

T cell development is conserved throughout jawed vertebrates (1721); therefore, we chose to use medaka juvenile thymus as a model to study this process. The optical transparency and the anatomic location of the thymus in medaka larvae allow the visualization and tracking of thymocytes within a full thymus in a noninvasive and quantitative manner. To study the migratory movements displayed by all thymocytes in a fully functional thymus, that nonetheless has a simple enough architecture to allow the study of the full organ, we developed multiple medaka transgenic reporter lines. In this study, we characterized the spatial organization of developmental events, such as proliferation, rearrangement of TCRs, and selection in the medaka thymus. We found that thymocytes are spatially organized into two thymic subpopulations that differ in the expression of the chemokine receptor ccr9b. Thymic immigrants expressing ccr9a home into the cortical thymus in the outer zone and proliferate. Once rag2 expression is switched off, they turn on expression of ccr9b and move toward the inner zone of the thymus, where they undergo thymic selection. At this stage, ccr9b-expressing thymocytes leave the inner zone of the thymus briefly to interact with dendritic cells located in the vicinity. Next, we studied the dynamics of migratory behavior of thymocytes by tracking cells in an entire thymus for the first time. Analysis of the migratory paths of these two thymic subpopulations and a population representing a defined developmental stage (i.e., rag2-expressing thymocytes) suggest that the migratory behavior of thymocytes is affected by their developmental stages and cannot be segregated based on distinct thymic localizations.

Larvae (10–14 d postfertilization [dpf]) of the medaka Cab inbred strain were used for all experiments. Four different transgenic reporter lines were generated for this study. Detailed descriptions of the reporter constructs for ccr9a, ccr9b, lck, and rag2 are schematically depicted in Supplemental Figs. 1–3. The cxcr3a:gfp transgene was used to monitor macrophages and DCs as described previously (22).

The autoimmune regulator aire single guide RNA (sgRNA) target site (5′-TTTCCCTCTGGTCTACGGTTTGG-3′) was predicted and evaluated for off-target sites by using the CCTop web tool (23) with the following parameters: PAM = ‘NGG’; total mismatches ≤ 4; core-length = 12; core mismatches ≤ 2. Cas9 mRNA and sgRNA generation and in vitro cleavage assay were performed as described previously (23). Detailed description of the donor plasmid is schematically depicted in Supplemental Fig. 3; 150 ng/μl Cas9 mRNA, 15 ng/μl of each sgRNA (aire- and donor-specific) and 10 ng/μl donor plasmid were coinjected at the one-cell stage. Approximately 20 injected larvae were selected for rising. The offspring was genotyped by PCR with primers p1: 5′-GGTTTCATAGACGTAGCTTTGCTT-3′; p2: 5′-TTTGGATAAGTTGCTCCAGAAGGA-3′; and p3: 5′-GCTCGACCAGGATGGGCA-3′. Two aire:gfp knock-in fish were analyzed in this work; both F1 fish showed similar expression patterns.

RNA in situ hybridization was performed as described previously (24, 25) using digoxigenin-labeled antisense probes. Combined fluorescent whole-mount in situ hybridization and immunostaining was performed as described previously (26). Probes are listed in Table I.

Table I.
Gene-specific in situ hybridization probes used in this study
GeneEnsembl ID/Accession NumberNucleotides X to X
aire ENSORLG00000016279 842–1508 
ccr9a ENSORLG00000007672 1–1159 
ccr9b ENSORLG00000012439 83–729 
ccl25a ENSORLG00000018031 1–327 
foxn1 ENSORLG00000008113 463–1239 
gata3 NM_001104718 169–842 
nur77a ENSORLG00000015557 1–636 
nur77b ENSORLG00000015279 3–663 
rag2 ENSORLG00000011979 64–1230 
GeneEnsembl ID/Accession NumberNucleotides X to X
aire ENSORLG00000016279 842–1508 
ccr9a ENSORLG00000007672 1–1159 
ccr9b ENSORLG00000012439 83–729 
ccl25a ENSORLG00000018031 1–327 
foxn1 ENSORLG00000008113 463–1239 
gata3 NM_001104718 169–842 
nur77a ENSORLG00000015557 1–636 
nur77b ENSORLG00000015279 3–663 
rag2 ENSORLG00000011979 64–1230 

Immunostaining was performed as described previously (26). Briefly, medaka hatchlings at 10–12 dpf were fixed with 4% paraformaldehyde in 1× PBS, 0.1% Tween-20 (1xPTw) for at least two days at 4°C. For primary Ab incubation, anti–proliferating cell nuclear Ag (PCNA; PC10, sc-56, 1:200 dilution; Santa Cruz Biotechnology), anti-phosphohistone H3 (Ser10, 06-570, 1:500 dilution; Millipore) Abs were incubated with the samples in the incubation buffer (1% sheep serum, 0.8% Triton X-100, 1% BSA in 1xPTw) at 4°C in the dark for 3 d with gentle agitation on a turning wheel. Secondary Abs, Alexa Fluor 647 goat anti-rabbit and anti-mouse IgGs (A 21245 and A21236, respectively), Alexa Fluor 488 goat anti-rabbit IgG (A11034), and Alexa Fluor 546 goat anti-mouse IgG (A11030, all at 1:200 dilution; Invitrogen) were used for the incubation with samples in the incubation buffer at 4°C in the dark for 2 d. Nuclei were counterstained with DAPI (D9564, 1:200 dilution from 5 mg/ml; Sigma-Aldrich). Isotype controls of primary Abs of mouse (1040, 1:200 dilution; Life Technologies) and rabbit (086199, 1:200 dilution; Life Technologies) were used to confirm the specificity of phosphohistone H3 (pH3) and PCNA labeling. The fluorescence of the transgenic GFP or TagRFP was sufficient to detect under the confocal microscope without immunostaining. Whole-larvae images were taken using a Leica TCS SPE or Zeiss LSM 780 NLO microscope.

Transgenic ccr9b:tagRFP medaka larvae (13–14 dpf) were stained for 1 h in a 50-μg/ml Acridine Orange (Sigma) solution prepared with embryo rearing medium (ERM) media. After incubation, the larvae were washed three times for 10 min with ERM and imaged directly afterward.

Medaka larvae (10–14 dpf) were anesthetized with tricaine methanesulfonate (Fluka, A5040), mounted in 1.5% low melting agarose (Peqlab) on glass-bottom culture dishes (MatTek Corporation), and covered with ERM containing 0.001% tricaine methanesulfonate. Time-lapse experiments for the analysis of thymocyte trafficking were completed with a PerkinElmer Ultraview VoX or Ultraview ERS Spinning disk confocal microscope using a 40× water-immersion objective (LD C-Apochromat, 1.1 NA, Corr; Zeiss). In general, the full thymus was imaged with a time interval <15 s (intervals >15 s would not allow for accurate tracking of double-labeled thymocytes) for consecutive periods of 30 min with z-stacks of 65–85 μm spanning the whole thymus area (z-space 1 μm). To avoid bleed-through between the two channels imaged an emission discrimination filter was used for all spinning disk imaging. Two photon images of the thymus of live and fixed samples, time lapse imaging of the cxcr3a:gfp;ccr9b:tagRFP double transgenic line and whole-larvae images were taken using a Zeiss LSM 780 NLO microscope. Thymus z-stacks were acquired using a 40× water-immersion objective (LD C-Apochromat, 1.1 NA, Corr, Zeiss). For time-lapse imaging, z-stacks of 65-85μm spanning the whole thymus area (z-space 1 μm) were acquired every 2 min for a period of 60 min or more. To image a whole larvae 5-6 separate z-scans of ∼ 300 μm (z-space, 10 μm) with 15% x-y overlap were taken and then stitched using the microscope operating software. Green (e.g., 488 nm) and red (e.g., 561 nm) channels were acquired sequentially by line using Gallium-Arsenide-Phosphide detectors. Imaging of the head region was performed with a Zeiss Lightsheet Z.1 microscope.

Time-lapse images of the ccr9a:h2b-gfp;ccr9b:tagRFP double-transgenic larvae at 11–12 dpf, taken with Spinning Disk confocal microscopes, were analyzed with Imaris software. The four-dimensional tracking for GFP channel (spot diameter, 3 μm; MaxDistance, 4 μm; and MaxGapSize, 3 μm) was used to identify the x, y, and z coordinates for each cells at each given time point. Single-cell tracks generated using Imaris software were examined manually. Additional filters were used to discriminate between single-positive (GFP+TagRFP) and double-positive (GFP+TagRFP+) cells in the thymus area. Tracked cells outside the thymus were excluded manually using the Edit Spots option. The Imaris software was used to calculate the motility parameters straightness (calculated as the ratio of a cell’s net distance traveled to its total path length), mean square displacement (MSD) over time, and average cell speed. Only cells with minimum track duration of 600 s were included in the calculation of the MSD over time. The squared displacement values were calculated using Imaris Fit MSD curves option. The obtained data were exported into an Excel file for analysis. GraphPad Prism software (version 6) was used for graphing and statistical analysis. Unpaired, two-tailed Student t tests were used to compare the means of different data sets.

Cells in the developing mouse thymus express the chemokine genes Ccl25, Ccl21, and Cxcl12 and their receptors (Ccr9, Ccr7, and Cxcr4, respectively) (5, 27). In the genomes of the teleost fish zebrafish and medaka, orthologs for both ccl25 and cxcl12 can be found; in each case, they are duplicated (25, 28). Thymic stromal cells express ccl25a, but not ccl25b. The expression of cxcl12a is restricted to the region surrounding the thymus, whereas cxcl12b is not expressed in the area surrounding the thymus nor inside the organ (25, 29). In teleost fish, Cxcl12a and Ccl25a cooperate in guiding lymphoid precursors to settle in the thymus (25, 29). A whole-mount in situ hybridization survey of chemokine receptors in medaka larvae (10–12 dpf) revealed that only two chemokine receptors—ccr9a and ccr9b—are expressed in the thymus (25). To better understand the expression pattern of these chemokine receptors, we performed RNA in situ hybridization on sections and found distinct expression patterns for each of these genes. ccr9a is uniformly expressed in the organ, whereas expression of ccr9b is restricted to the cells that are located in the central area of the thymus (Supplemental Fig. 1A, 1B). To track the movements of thymocytes, we constructed expression reporters for ccr9a (ccr9a:h2b-gfp) and ccr9b (ccr9b:tagRFP) and used them for in vivo imaging (Supplemental Fig. 1C, 1D; Supplemental Video 1). The transgenes mirrored the expression patterns of the ccr9a and ccr9b genes as determined by in situ hybridization. We combined these transgenes in one line to compare the expression patterns in the same fish (Fig. 1A, 1B). In ccr9a:h2b-gfp,ccr9b:tagRFP double-transgenic larvae (10–12 dpf), up to 800 hematopoietic cells throughout the thymus expressed GFP (which we will, for simplicity, refer to as expressing ccr9a; Fig. 1C). A subset of the cells expressing GFP also expressed TagRFP (i.e., expressed ccr9b; Fig. 1B, 1C). These expression patterns identified two spatially organized thymocyte subpopulations (Fig. 1D): cells expressing ccr9a and not ccr9b in the outer zone of the thymus, and cells expressing both chemokine receptors in the inner zone of the thymus (which we will call ccr9a and ccr9a/b cells, respectively). The proportion of ccr9a/b cells among the total number of thymocytes labeled by ccr9a was 36.8 ± 4% (n = 7, mean ± SD). The ccr9a/b cells showed a further graded subdivision by level of ccr9b expression (Fig. 1E, middle panel). Cells with higher TagRFP levels (ccr9a/bhigh) were preferentially located in the center of the thymus, whereas cells expressing lower levels of TagRFP (ccr9a/blow) were less concentrated at the center (Supplemental Video 2). Occasionally, ccr9a/b cells also appeared in the outer zone for short periods (Supplemental Video 3). These results are consistent with a role for the chemokine receptor ccr9b in the positioning of thymocytes in the medaka juvenile thymus, which is in agreement with the general function of the chemokine system in controlling the migratory patterns and positioning of immune cells in the lymphoid organs (30).

FIGURE 1.

Distribution and movements of cells expressing ccr9a and ccr9b in the double-transgenic (ccr9a:h2b-gfp;ccr9b:tagRFP) fish at 11 dpf. (A) Lateral (top panels) and dorsal (bottom panels) views of the head regions of the same double transgenic fish. The ccr9a-expressing cells (green) are located in the thymus, skin, and other tissues, whereas ccr9b-expressing cells (red) are preferentially located in the thymus. Major anatomic landmarks are highlighted. The white circle indicates the position of the thymus. Scale bar, 100 μm. (B) Three-dimensional rendering of a thymus illustrating the trafficking of hematopoietic cells in the extrathymic region (derived from Supplemental Video 1). White arrows indicate migration paths of thymus colonization. Red arrows indicate emigration paths of cells from the thymus into the periphery. The yellow dashed line demarcates the thymus. Anterior is to the left. Scale bar, 30 μm. (C) Number of ccr9a-positive and ccr9a/b double-positive hematopoietic cells in the transgenic thymus at 11–12 dpf (n = 7; mean ± SD). All hematopoietic cells express GFP. We found no cells expressing only TagRFP. (D) Scheme depicting two zones in the juvenile thymus. Cells in the outer zone express only ccr9a, whereas cells in the inner zone express both chemokine receptors ccr9a and ccr9b. (E) Trajectories of individual hematopoietic cells in the double transgenic thymus at 11 dpf over a period of 30 min (derived from Supplemental Video 2). Trajectories of all ccr9a-positive but ccr9b-negative cells (blue) in the thymus are shown in the top and bottom panels in blue. Color-coded trajectories of ccr9a/b double-positive (red) cells are shown in the middle and bottom panels. Colors of trajectories indicate the red fluorescence (ccr9b) intensity of each tracked cell as shown in the color scale bar. Positive cells with higher TagRFP signal were enriched in the center of the thymus. A merge of all tracks is shown in the bottom panel. Lateral (left panels) and side views (right panels) of whole thymus. Scale bars, 30μm. Average cell speed (F) and straightness (G) of thymic immigrants during the migration from the extrathymic region into the thymus. (H) Still photographs from a time-lapse recording illustrating the migration of two ccr9a-expressing cells into the ventral side of a thymus (Supplemental Video 4). Numbers indicate time in minutes. The green and blue lines indicate the trajectories of tracked cells. Scale bar, 20μm. (I) Expression pattern of chemokine ccl25a in the medaka larva at 10 dpf. The position of the thymus is demarcated by a yellow dashed line. Red arrows indicate the ccl25a extrathymic expression domain. Scale bar, 60μm.

FIGURE 1.

Distribution and movements of cells expressing ccr9a and ccr9b in the double-transgenic (ccr9a:h2b-gfp;ccr9b:tagRFP) fish at 11 dpf. (A) Lateral (top panels) and dorsal (bottom panels) views of the head regions of the same double transgenic fish. The ccr9a-expressing cells (green) are located in the thymus, skin, and other tissues, whereas ccr9b-expressing cells (red) are preferentially located in the thymus. Major anatomic landmarks are highlighted. The white circle indicates the position of the thymus. Scale bar, 100 μm. (B) Three-dimensional rendering of a thymus illustrating the trafficking of hematopoietic cells in the extrathymic region (derived from Supplemental Video 1). White arrows indicate migration paths of thymus colonization. Red arrows indicate emigration paths of cells from the thymus into the periphery. The yellow dashed line demarcates the thymus. Anterior is to the left. Scale bar, 30 μm. (C) Number of ccr9a-positive and ccr9a/b double-positive hematopoietic cells in the transgenic thymus at 11–12 dpf (n = 7; mean ± SD). All hematopoietic cells express GFP. We found no cells expressing only TagRFP. (D) Scheme depicting two zones in the juvenile thymus. Cells in the outer zone express only ccr9a, whereas cells in the inner zone express both chemokine receptors ccr9a and ccr9b. (E) Trajectories of individual hematopoietic cells in the double transgenic thymus at 11 dpf over a period of 30 min (derived from Supplemental Video 2). Trajectories of all ccr9a-positive but ccr9b-negative cells (blue) in the thymus are shown in the top and bottom panels in blue. Color-coded trajectories of ccr9a/b double-positive (red) cells are shown in the middle and bottom panels. Colors of trajectories indicate the red fluorescence (ccr9b) intensity of each tracked cell as shown in the color scale bar. Positive cells with higher TagRFP signal were enriched in the center of the thymus. A merge of all tracks is shown in the bottom panel. Lateral (left panels) and side views (right panels) of whole thymus. Scale bars, 30μm. Average cell speed (F) and straightness (G) of thymic immigrants during the migration from the extrathymic region into the thymus. (H) Still photographs from a time-lapse recording illustrating the migration of two ccr9a-expressing cells into the ventral side of a thymus (Supplemental Video 4). Numbers indicate time in minutes. The green and blue lines indicate the trajectories of tracked cells. Scale bar, 20μm. (I) Expression pattern of chemokine ccl25a in the medaka larva at 10 dpf. The position of the thymus is demarcated by a yellow dashed line. Red arrows indicate the ccl25a extrathymic expression domain. Scale bar, 60μm.

Close modal

The ccr9a and ccr9a/b subpopulations in the thymus may represent thymocytes at distinct stages of T cell development. The first step in T cell development is the migration of lymphoid progenitors into the thymus (31, 32). To understand which subpopulation enters the thymus, we used time-lapse imaging of the ccr9a:h2b-gfp,ccr9b:tagRFP double-transgenic larvae (Supplemental Videos 1, 4). At 10–12 dpf, the medaka thymus is partially encapsulated (Supplemental Fig. 1E), but not yet vascularized (data not shown). Consistent with previous studies (29, 33), time-lapse imaging showed that cells from the extrathymic region migrated into the thymus to colonize the thymus. Cells entered the thymus with an average speed of 12.3 ± 4.9 μm/min (Fig. 1F) in a straight path (Fig. 1G, 1H; Supplemental Video 4) through routes on the ventrolateral side of the thymus (Fig. 1B, white arrows; Supplemental Fig. 1E). These routes were defined by the expression of the chemokine ccl25a (Fig. 1I, red arrows). Approximately 85% of thymic immigrants were ccr9a cells (n = 62), the rest were ccr9a/b cells that expressed very low levels of TagRFP. These cells had previously left and now reentered the thymus. A similar mode of thymocyte migratory behavior has been described as in-out-in in zebrafish (29).

ccr9a/b cells were seen egressing the thymus using the same migratory paths that had been used to colonize the thymus (Fig. 1B, red arrows; Supplemental Video 4). To test whether ccr9a/b cells that leave the thymus and do not reenter are naive T cells that have completed their intrathymic development, we used a lck:gfp transgenic reporter (Supplemental Fig. 2A). The lck gene is expressed by thymocytes and naive T cells outside the thymus (34, 35). In lck:gfp;ccr9b:tagRFP double-transgenic larvae, a subset of cells outside the thymus that express ccr9b also express lck, suggesting that they are indeed naive T cells (Supplemental Fig. 2B). Other ccr9b-expressing cells, which had macrophage/DC morphology, lacked expression of lck, confirming the notion that they are not T cells (Supplemental Fig. 2B).

In summary, thymic immigrants express ccr9a and colonize the juvenile thymus using different routes. Thymocytes at more advanced developmental stages additionally express ccr9b and leave the thymus via several routes, including some that were also used for entry.

The thymic development of T cells includes several steps such as proliferation, specification, rearrangement of TCRs and selection—processes that require the dynamic relocation of thymocytes in and out of different thymic microenvironments (1, 36, 37). To determine the stage relative to the onset of ccr9b expression and the location in the thymus where these processes occur, we used various markers to visualize them. We will discuss them in the order in which they have been demonstrated to occur in mammalian thymocytes.

To identify proliferating cells, we used two approaches. First, we identified mitotic cells in ccr9a:h2b-gfp;ccr9b:tagRFP transgenic fish by nuclear morphology and found that cell divisions are seen exclusively in thymocytes expressing only ccr9a, and not in ccr9a/b cells (Fig. 2A; Supplemental Video 5). Cell divisions are also seen in ccr9a-expressing cells directly outside the thymus (Supplemental Video 1). As a second approach, we stained fixed transgenic larvae for the M-phase markers pH3 and PCNA. Again, only ccr9a+ cells expressed mitotic markers (Fig. 2B, 2C, left panel), whereas ccr9b-expressing cells were neither pH3- nor PCNA-positive (Fig. 2D). Most of the mitotically active ccr9a+ cells were located in the ventrolateral region of the thymus (Fig. 2C, left panel)—that is, in the region where many cells first enter the thymus. Some cells that expressed mitotic markers, presumably thymic stromal cells, did not express ccr9a. These cells were evenly distributed between thymic compartments (Fig. 2C, right panel). In summary, thymocytes appear to divide soon after entering the thymus, while expressing only ccr9a and residing in the outer region, and enter a mitotically quiescent state once they move to the center of the thymus and become ccr9a/b cells.

FIGURE 2.

Spatial distribution of dividing cells in the thymus. (A) Time-lapse images at higher magnification (numbers indicate time in minutes) derived from Supplemental Video 5 shows a cell division event (white arrows) in the thymus of a double-transgenic (ccr9a:h2b-gfp;ccr9b:tagRFP) larva at 11 dpf. Scale bar, 5 μm. (B) Thymus of a ccr9a:gfp (green) transgenic larva at 11 dpf stained for pH3 (magenta) and PCNA (blue); one plane (z = 1μm) of a Z-stack spanning the whole thymus. The center of the thymus shows little or no PCNA or pH3 staining. The inset shows 6-fold magnification of the area outlined in the top panel. The position of the thymus is demarcated by a yellow dashed line. Scale bars at top and bottom panels, 20 and 5 μm, respectively. (C) Distribution of pH3-positive cells in a ccr9a:gfp transgenic thymus. Two-photon confocal images of 15 thymuses were divided into eight segments (see schematic above right panel). The numbers of dividing ccr9a-positive (pH3+GFP+; left panel) and dividing ccr9a-negative (pH3+GFP) cells (right panel) in each segment were counted. Each dot represents an individual pH3+ cell. The horizontal lines indicate the averages of the number of pH3+ cell in each segment. The overall number of pH3+GFP cells per segment showed no statistically significant differences. (D) Thymus of a ccr9b:tagRFP (red) transgenic larva at 11 dpf stained for pH3 (magenta) and PCNA (blue); one plane (z = 1 μm) of a Z-scan spanning the whole thymus area. The inset shows 6-fold magnification of the area outlined in the top panel. There is no colocalization between TagRFP and pH3 or PCNA (n = 7). The yellow dashed line demarcates the position of the thymus. Scale bars at top and bottom panels, 20 and 5 μm, respectively.

FIGURE 2.

Spatial distribution of dividing cells in the thymus. (A) Time-lapse images at higher magnification (numbers indicate time in minutes) derived from Supplemental Video 5 shows a cell division event (white arrows) in the thymus of a double-transgenic (ccr9a:h2b-gfp;ccr9b:tagRFP) larva at 11 dpf. Scale bar, 5 μm. (B) Thymus of a ccr9a:gfp (green) transgenic larva at 11 dpf stained for pH3 (magenta) and PCNA (blue); one plane (z = 1μm) of a Z-stack spanning the whole thymus. The center of the thymus shows little or no PCNA or pH3 staining. The inset shows 6-fold magnification of the area outlined in the top panel. The position of the thymus is demarcated by a yellow dashed line. Scale bars at top and bottom panels, 20 and 5 μm, respectively. (C) Distribution of pH3-positive cells in a ccr9a:gfp transgenic thymus. Two-photon confocal images of 15 thymuses were divided into eight segments (see schematic above right panel). The numbers of dividing ccr9a-positive (pH3+GFP+; left panel) and dividing ccr9a-negative (pH3+GFP) cells (right panel) in each segment were counted. Each dot represents an individual pH3+ cell. The horizontal lines indicate the averages of the number of pH3+ cell in each segment. The overall number of pH3+GFP cells per segment showed no statistically significant differences. (D) Thymus of a ccr9b:tagRFP (red) transgenic larva at 11 dpf stained for pH3 (magenta) and PCNA (blue); one plane (z = 1 μm) of a Z-scan spanning the whole thymus area. The inset shows 6-fold magnification of the area outlined in the top panel. There is no colocalization between TagRFP and pH3 or PCNA (n = 7). The yellow dashed line demarcates the position of the thymus. Scale bars at top and bottom panels, 20 and 5 μm, respectively.

Close modal

To determine the relative onset of V(D)J recombination, we used expression of rag2 as a proxy because expression levels of the Rag1 and Rag2 genes correlate directly with V(D)J recombination activity in mice (38). We created a rag2 reporter line (rag2:gfp-pest), which labeled thymocytes with different GFP intensities, in a pattern resembling the expression of the endogenous rag2 gene (Supplemental Fig. 3A, 3B). The transgenic thymus contained up to 250 rag2-expressing cells with varying levels of fluorescence (Fig. 3A, 3B). The rag2+ cells with higher fluorescence levels were preferentially located laterally and in the outer zone of thymus, whereas thymocytes in the inner zone were either negative or had very low expression levels (Fig. 3A). Analysis of fish carrying both the rag2:gfp-pest and ccr9b:tagRFP transgenes revealed that rag2 and ccr9b were coexpressed only by a subset of cells located near the interface of the inner and outer zone that expressed low levels of both (Fig. 3C). Therefore, it is likely that cells lose rag2 expression as they gain ccr9b expression.

FIGURE 3.

Spatial distribution of rag2-expressing cells in the thymus. (A) Three-dimensional rendering of a thymus in a double transgenic rag2:gfp-pest;ccr9b:tagRFP larva at 11 dpf. Most rag2+ cells (yellow) are located in the outer zone of the thymus, whereas ccr9b+ cells (red) are found in the inner zone. Scale bars, 30 μm. (B) Number of cells expressing high or low levels of GFP in rag2:gfp-pest transgenic larvae (n = 10; mean ± SD). (C) Coexpression of rag2 (GFP-PEST; yellow) and ccr9b (TagRFP; red) in a double-transgenic (rag2:gfp-pest;ccr9b:TagRFP) larva. The fluorescence intensities of GFP and TagRFP were quantified along the white lines (bottom panel) in five selected cells as shown in the top right panel. Scale bars, 5μm.

FIGURE 3.

Spatial distribution of rag2-expressing cells in the thymus. (A) Three-dimensional rendering of a thymus in a double transgenic rag2:gfp-pest;ccr9b:tagRFP larva at 11 dpf. Most rag2+ cells (yellow) are located in the outer zone of the thymus, whereas ccr9b+ cells (red) are found in the inner zone. Scale bars, 30 μm. (B) Number of cells expressing high or low levels of GFP in rag2:gfp-pest transgenic larvae (n = 10; mean ± SD). (C) Coexpression of rag2 (GFP-PEST; yellow) and ccr9b (TagRFP; red) in a double-transgenic (rag2:gfp-pest;ccr9b:TagRFP) larva. The fluorescence intensities of GFP and TagRFP were quantified along the white lines (bottom panel) in five selected cells as shown in the top right panel. Scale bars, 5μm.

Close modal

The conclusion that ccr9b is activated late during somatic recombination, and is therefore a marker for the final stages of thymocyte maturation in the thymus, is also confirmed by the spatial expression patterns of other markers for T cell differentiation, such as gata3, tcrb, and nur77b (Table I; Supplemental Fig. 4A, 4B). In mice, the transcription factor Gata3 is involved in multiple steps of intrathymic T cell development, including T cell commitment and β-selection (39). As β-selection begins, TCR-β is highly expressed (40), whereas Nur77 is upregulated after TCR stimulation (41, 42). In medaka, gata3 expression can be detected throughout the whole juvenile thymus, but most prominently in the outer zone (Supplemental Fig. 4A). The highest levels of tcrb expression can be found in the inner zone (Supplemental Fig. 4A), colocalizing with ccr9b (Supplemental Fig. 4B). nur77b was expressed more in a heterogeneous pattern, with some patches of very high expression within a larger area of low expression; however, the entire expression domain was also confined to the inner zone of the thymus (Supplemental Fig. 4A), indicating that TCR stimulation occurs in that area.

After somatic recombination and before leaving the thymus, each thymocyte must pass thymic selections; those that fail the selections die and are cleared in the thymus (43). To define the position of cells undergoing thymic selection, we analyzed the expression pattern of the ortholog of aire in the medaka thymus. In mice, the transcription factor Aire promotes clonal deletion of self-reactive thymocytes (44) and is expressed in thymic medullary epithelial cells and in DCs (45). To determine the expression domain of aire, we used RNA in situ hybridization and aire:gfp knock-in fish (generated by CRISPR/Cas9; see 2Materials and Methods and Supplemental Fig. 3C, 3D). By in situ hybridization, we detected patches of aire-expressing cells in the dorsocentral region of the juvenile thymus (Supplemental Fig. 3F). This pattern was also seen in the aire:gfp knock-in larvae (Supplemental Fig. 3E). At 12 dpf, aire+ cells were located in the dorsal side and in the inner zone of the thymus (Supplemental Fig. 3E); this indicates that mature medullary regions have formed in the juvenile thymus.

As we have shown above, thymocytes activate ccr9b expression during or after the late phase of somatic recombination and continue to express it until after they have left the thymus. We would therefore expect some ccr9b-expressing thymocytes to be undergoing apoptosis because of selection. We stained transgenic ccr9b:tagRFP larvae with the fluorescent dye acridine orange to determine the correlation between expression of ccr9b and cell death. For this assay, we used slightly older larvae (13–14 dpf) because the higher frequency of cell death allowed a more reliable statistical analysis. We found that 65 ± 13% (mean ± SD; n = 10) of apoptotic cells in the thymus were ccr9b+ (Fig. 4A), of which 62% were located in the inner zone of the thymus. This result indicates that cell death is not exclusively seen in ccr9b+ cells. Because fewer than 40% of thymocytes express ccr9b at this stage, cell death is overrepresented 1.5-fold in this population. Although the majority of thymocytes are eliminated during thymic selection (43), we found an average of only eight apoptotic cells per thymus (Fig. 4B). This is not necessarily a paradox, because dying thymocytes are normally cleared through immediate engulfment and removal by macrophages (46). We visualized this process in vivo by combining the transgenic cxcr3a:gfp reporter line that labels mononuclear phagocytic cells, including macrophages and DCs (22), with the ccr9b:tagRFP reporter (Fig. 4C). Phagocytes expressing cxcr3a:gfp within the thymus were mainly located near the interface of the inner and outer zone of the thymus (Fig. 4D). Phagocytes in the thymus and outside the thymus actively establish contacts with and engulf surrounding ccr9b+ cells. Therefore, the observation that the number of dying thymocytes detected in the thymus is low can be explained by the efficient engulfment and removal of apoptotic cells by thymic phagocytes.

FIGURE 4.

Thymocyte interactions with phagocytic cells. (A) The left panel shows the proportion of acridine orange stained cells in the thymus that were ccr9b positive. The right panel shows representative images of acridine orange staining (orange) in the transgenic ccr9a:tagRFP (red) larvae at 12–13 dpf. Two frames from a Z-stack spanning the whole thymus are shown. Images are representative of 10 stained transgenic larvae. Scale bar, 20 μm. (B) Number of acridine orange stained cells in the thymus (n = 10). (CG) Overview and high magnification of thymus of a double transgenic (cxcr3a:gfp; ccr9b:tagRFP) fish at 13–14 dpf. (C) Three-dimensional rendering of the thymus to show the positioning of cxcr3a-expressing cells (macrophages and DCs; white) and ccr9b-expressing cells (thymocyte; red) in the thymus. Scale bar, 20 μm. (D) One frame from a Z-stack spanning the thymus. Macrophages and DCs are located near the interface of outer and inner zone of the thymus. Yellow arrows indicate phagocytes that have engulfed ccr9b cells. White arrows indicate ccr9b cells outside the thymus. The white dashed line demarcates the position of the inner zone of the thymus. Scale bar, 30 μm. (E) An example of the in-out-in migratory behavior displayed by a ccr9b+ cell. During its migration, the cell interacts intermittently with DCs that are located in the outer zone of the thymus (from Supplemental Video 6). Colors of trajectories indicate the time line, as indicated in the color scale bar. The white dashed line demarcates the position of the inner zone of the thymus. Scale bar, 10 μm. (F) An example of DC (white) interaction with ccr9b-expressing thymocyte (asterisks) derived from Supplemental Video 7. Scale bar, 5 μm. (G) Time course of a macrophage (asterisk) engulfing a ccr9b-positive cell (yellow arrow); derived from Supplemental Video 8. Numbers indicate time in minutes. Scale bar, 5 μm.

FIGURE 4.

Thymocyte interactions with phagocytic cells. (A) The left panel shows the proportion of acridine orange stained cells in the thymus that were ccr9b positive. The right panel shows representative images of acridine orange staining (orange) in the transgenic ccr9a:tagRFP (red) larvae at 12–13 dpf. Two frames from a Z-stack spanning the whole thymus are shown. Images are representative of 10 stained transgenic larvae. Scale bar, 20 μm. (B) Number of acridine orange stained cells in the thymus (n = 10). (CG) Overview and high magnification of thymus of a double transgenic (cxcr3a:gfp; ccr9b:tagRFP) fish at 13–14 dpf. (C) Three-dimensional rendering of the thymus to show the positioning of cxcr3a-expressing cells (macrophages and DCs; white) and ccr9b-expressing cells (thymocyte; red) in the thymus. Scale bar, 20 μm. (D) One frame from a Z-stack spanning the thymus. Macrophages and DCs are located near the interface of outer and inner zone of the thymus. Yellow arrows indicate phagocytes that have engulfed ccr9b cells. White arrows indicate ccr9b cells outside the thymus. The white dashed line demarcates the position of the inner zone of the thymus. Scale bar, 30 μm. (E) An example of the in-out-in migratory behavior displayed by a ccr9b+ cell. During its migration, the cell interacts intermittently with DCs that are located in the outer zone of the thymus (from Supplemental Video 6). Colors of trajectories indicate the time line, as indicated in the color scale bar. The white dashed line demarcates the position of the inner zone of the thymus. Scale bar, 10 μm. (F) An example of DC (white) interaction with ccr9b-expressing thymocyte (asterisks) derived from Supplemental Video 7. Scale bar, 5 μm. (G) Time course of a macrophage (asterisk) engulfing a ccr9b-positive cell (yellow arrow); derived from Supplemental Video 8. Numbers indicate time in minutes. Scale bar, 5 μm.

Close modal

Time-lapse imaging of the ccr9b:tagRFP;cxcr3a:gfp double transgenic larvae revealed that the in-out-in migratory behavior of ccr9b+ cells allows them to intermittently interact with cxcr3a-expressing cells (that from morphology appear to be DCs), both in the outer zone (Fig. 4E, Supplemental Video 6) and outside the thymus (Fig. 4F, Supplemental Video 7). Phagocytes can also engulf cells distant from their cell bodies by extending long protrusions (Fig. 4G, Supplemental Video 8), which allow them to clear cells from regions devoid of macrophages, such as the inner zone of the thymus (e.g., in Supplemental Video 9). Thus, there are two distinct groups of cxcr3a-expressing phagocytes in the thymus. A first group patrols the thymus and recognizes and engulfs dying thymocytes by migrating toward them. Another group of thymic cxcr3a+ phagocytes remains stationary but is highly interactive, and ccr9a/b thymocytes move in their direction to establish contact.

Ex vivo analyses in mice have shown that migratory patterns of thymocytes vary between distinct thymic microenvironments (9, 11, 12, 15). To describe the migratory behavior of thymocytes in medaka in a quantitative manner, we used time-lapse imaging of ccr9a:h2b-gfp,ccr9b:tagRFP double-transgenic larvae (10–12 dpf) and measured the average speed and straightness of migration for each of the tracked cells. The average speed for both ccr9a and ccr9a/b cells was 2 μm/min, ranging from <1 to >6 μm/min in each population (Fig. 5A). If all ccr9a and ccr9a/b cells are divided into a faster and a slower moving population using the average speed (>3 and <3 μm/min, respectively), the faster-moving cells would be seen to be confined to certain areas of the entire ccr9a and ccr9a/b territories. The fast-moving ccr9a cells were found mainly in the dorsal and ventral side of the subcapsular area (Fig. 5B, top panel), whereas the fast ccr9a/b cells were mainly near the interface of the inner and outer zone of the thymus (Fig. 5B, bottom panel). In both subpopulations, faster-moving cells showed longer displacement length than slower moving cells did (Fig. 5C). Thus, the fastest ccr9a and ccr9a/b cells can be distinguished from the remaining cells, both by their location and by their significantly higher displacement.

FIGURE 5.

Heterogeneous migratory behavior of thymocytes expressing ccr9a and ccr9a/b in the double-transgenic (ccr9a:h2b-gfp;ccr9b:tagRFP) fish at 11 dpf. (A) Average cell speed of individual thymocytes over a 30-min period. Number of tracked cells for ccr9a and ccr9a/b are 653 and 334, respectively. (B) Trajectories of ccr9a (top panel) and ccr9a/b cells (bottom panel) with average cell speed less than (blue) or greater than (yellow) 3 μm/min. (C) Displacement length of ccr9a and ccr9a/b cells with average cell speed less than (blue) or greater than (yellow) than 3 μm/min. Error bars show SE in mean. (D) Straightness of individual thymocytes with average cell increased to 3 μm/min. (E) Trajectories of ccr9a (left panel) and ccr9a/b cells (right panel). Lateral view of thymus. Color-coded trajectories indicate the straightness values according to (D). MSD graphs (F) and track plots of ccr9a/b (G) and ccr9a (H) cells with different straightness values. In track plots, the cell trajectories were aligned for the starting position. The red line in (F) represents the linear regression (R) analysis for each plot. ***p < 0.001.

FIGURE 5.

Heterogeneous migratory behavior of thymocytes expressing ccr9a and ccr9a/b in the double-transgenic (ccr9a:h2b-gfp;ccr9b:tagRFP) fish at 11 dpf. (A) Average cell speed of individual thymocytes over a 30-min period. Number of tracked cells for ccr9a and ccr9a/b are 653 and 334, respectively. (B) Trajectories of ccr9a (top panel) and ccr9a/b cells (bottom panel) with average cell speed less than (blue) or greater than (yellow) 3 μm/min. (C) Displacement length of ccr9a and ccr9a/b cells with average cell speed less than (blue) or greater than (yellow) than 3 μm/min. Error bars show SE in mean. (D) Straightness of individual thymocytes with average cell increased to 3 μm/min. (E) Trajectories of ccr9a (left panel) and ccr9a/b cells (right panel). Lateral view of thymus. Color-coded trajectories indicate the straightness values according to (D). MSD graphs (F) and track plots of ccr9a/b (G) and ccr9a (H) cells with different straightness values. In track plots, the cell trajectories were aligned for the starting position. The red line in (F) represents the linear regression (R) analysis for each plot. ***p < 0.001.

Close modal

Next, we analyzed the behavior of the remaining cells in more detail to determine whether they could be further subdivided into groups with different behaviors and, if so, to test whether this was associated with different locations. We split them into subgroups based on the straightness values of their tracks and plotted the tracks for each subgroup separately (Fig. 5D, 5E). The straightness distribution did not differ significantly between the ccr9a and ccr9a/b populations (Fig. 5D). Among the ccr9a cells, those with the straightest paths (red in Fig. 5D, 5E) were more likely to be moving near the surface of the thymus (i.e., in the same region as the fastest moving cells described above; yellow tracks in Fig. 5B). No significant differences between the other three ccr9a subpopulations, or between the four subpopulations of ccr9a/b thymocytes were seen. We also calculated the mean displacement of cells belonging to each subgroup over a period of 10 min (Fig. 5F; note the difference in scale on the y-axes of the plots). The four groups displayed slightly different behaviors, from cells with low straightness and confined (straightness <0.1) or stochastic (straightness values of 0.1–0.2) migratory behavior, to cells with higher straightness values (0.4–1.0) that exhibited more directed migration (Fig. 5F). The track plot of ccr9a (Fig. 5G) and ccr9a/b (Fig. 5H) cells in each subgroup also qualitatively confirms the range of migratory behaviors. Apart from the ccr9a cells with the straightest paths being near the surface, we found no other relationship between straightness and location of the slower cells in the thymus.

It seemed possible that different migratory behaviors could correspond to different developmental stages of the thymocytes. To assay the behavior of a population representing a more defined stage, we followed the movements of rag2+ cells (Fig. 6A). The average speed of rag2+ cells (1.3 ± 0.5 μm/min, mean ± SD; n = 258 tracks) was at the lower end of the range of the overall population of ccr9a cells (1.85 ± 0.72 μm/min, mean ± SD; n = 653 tracks) and that of ccr9a/b cells (2.12 ± 0.68 μm/min, mean ± SD; n = 334 tracks). Thus, while thymocytes express rag2, they slow down. Two distinct migratory behaviors were observed in the rag2+ cells (Fig. 6B): ∼69% of the cells migrated with straightness up to 0.2 with a random-walk movement (e.g., in Fig. 6C, 6D, top panel; Supplemental Video 10); the rest of the rag2-expressing cells migrated with a straightness of >0.2 and thus moved in a more directed path (Fig. 6D, bottom panel). Splitting the ccr9a population in terms of straightness also at 0.2 shows that ∼65% of tracks have a straightness <0.2. The behavior of rag2+ cells is therefore shifted slightly toward reduced speed and reduced directed migration, suggesting that the rag2 population does correspond to a more (yet not completely) homogeneous subpopulation of ccr9a. We found that thymocytes in different thymic zones migrate in a heterogeneous fashion, whereas cells in a narrower developmental window will migrate more homogeneously.

FIGURE 6.

The migratory behavior of rag2-expressing cells. (A and B) Color-coded tracks of rag2+ cells in the transgenic rag2:gfp-pest larva over a period of 30 min. (A) The migration paths of rag2+-expressing cells in Fig. 3A. Colors of trajectories show the cell speed of each track at each time point, as indicated in the color scale bar. Scale bars, 30 μm. (B) The migration track of a rag2+-expressing cell in (A, inset). Colors of trajectories show the time line, as indicated in the color scale bar (Supplemental Video 10). (C) The frequency distribution of track straightness of individual rag2+ cells (n = 258 cells from three thymuses). (D) MSD graphs of rag2+ cells with straightness lower or higher than 0.2, respectively. The red line represents the linear regression (R) analysis for each plot.

FIGURE 6.

The migratory behavior of rag2-expressing cells. (A and B) Color-coded tracks of rag2+ cells in the transgenic rag2:gfp-pest larva over a period of 30 min. (A) The migration paths of rag2+-expressing cells in Fig. 3A. Colors of trajectories show the cell speed of each track at each time point, as indicated in the color scale bar. Scale bars, 30 μm. (B) The migration track of a rag2+-expressing cell in (A, inset). Colors of trajectories show the time line, as indicated in the color scale bar (Supplemental Video 10). (C) The frequency distribution of track straightness of individual rag2+ cells (n = 258 cells from three thymuses). (D) MSD graphs of rag2+ cells with straightness lower or higher than 0.2, respectively. The red line represents the linear regression (R) analysis for each plot.

Close modal

In vivo imaging helps to decipher the spatiotemporal dynamics of cell migration within the crowded cellular environment of the thymus. As the mouse thymus cannot be accessed for intravital imaging with currently available techniques, it has not been possible to study the dynamics of thymocyte migration at during T cell development in a thymus without disturbing its architecture or changing the chemokine milieu (16). Alternative animal models such as zebrafish and medaka allow direct in vivo analysis to study cellular dynamics of T cell development in a noninvasive manner. For instance, a transgenic medaka fish for rag1 (rag1:egfp) has been used to examine the migration of lymphoid progenitors into the thymus (33). This study suggested that rag1:egfp-expressing cells within the thymus display random-walk motility. However, whether all thymocytes exhibit similar migratory behavior or stochastic migration is only characteristic for rag1-expressing cells has not been addressed. Recently, zebrafish was used as a model to describe aspects of early thymopoiesis, including thymus homing and emigration of cells from the thymus rudiment (29). In this work, transgenic zebrafish lines expressing fluorescent proteins under the regulatory elements of ikaros and foxn1 genes were used to demonstrate a dynamic lymphoepithelial interaction during early thymopoiesis (2.5–5 dpf), particularly when the thymus anlage is forming (i.e., 2.5–3 dpf). However, the thymic rudiment at 2.5-5 dpf in both zebrafish and medaka only contains a small number of thymocytes, and the migratory behavior of thymocytes inside the thymic rudiment cannot be studied in a quantitative manner. At the stage imaged in the study presented here (10–14 dpf), the thymus contains a large number of hematopoietic cells and is still transparent, unlike at later stages of maturation during which a noninvasive in toto imaging of the thymus is not possible. This allowed us to examine the migratory behavior of all thymocytes simultaneously in a densely packed thymus. We show that landmark T cell developmental stages like homing, proliferation, somatic recombination, and cell death of thymocytes preferentially take place within specific areas of the medaka juvenile thymus. In addition, we analyzed the dynamic of thymocyte-myeloid cell interactions during thymic selections.

The expression of chemokine and chemokine receptors define distinct thymic microenvironments in the mouse thymus (5). In contrast to the mouse developing thymus, which expresses multiple chemokines (5), only one chemokine (ccl25a) is expressed in the medaka juvenile thymus (25, 29). Interference with expression of ccl25a alters thymus homing (25, 29), and ectopic expression of ccl25a leads to disorientation of ccr9a-expressing cells (25). In mice, the chemokine CCL25 has a single receptor, CCR9, which is expressed in thymocytes at different developmental stages and is dramatically upregulated by pre-TCR signaling (4749). The migration of mouse thymocytes from the outer cortex into the subcapsular zone is CCR9 dependent (50). The CCL25/CCR9 pair is also required for attraction of lymphoid progenitors into the mouse thymus (27). In medaka and zebrafish, the ccr9 gene is duplicated (21). Both Ccr9a and Ccr9b are highly similar to mammalian CCR9 (25). However, whether they activate the same or distinct downstream signaling pathways is unknown.

Based on the expression patterns of ccr9a and ccr9b, two zones can be defined in the medaka juvenile thymus in which events of T cell development, such as proliferation, recombination, and selection, are spatially organized. Fig. 7 summarizes a model for the route of developing thymocytes, as derived from the results presented here. Thymic immigrants expressing ccr9a enter through the ventrolateral side of the thymus in the outer zone, which is probably equivalent to the mouse cortex. In this area, lymphoid precursors undergo mitosis. After the proliferative phase, they activate the expression of ccr9b. Because rag2 cells are mostly found in the outer zone of the thymus and express ccr9a, and not ccr9b, somatic recombination likely occurs in this region. Because some ccr9a/blow cells in the inner zone of the thymus expressed rag2, albeit very weakly, we conclude that the expression of ccr9b is activated at late stages of somatic recombination, when cells move from the outer into the inner zone. The presence of dying ccr9a/b cells, together with the engulfment and removal of ccr9a/b cells by cxcr3a-expressing phagocytes, suggests that a subset of ccr9a/b thymocytes undergo thymic selection in the inner zone of the thymus, where nur77b and tcrb are highly expressed. In a previous study (33), the lack of UEA-1 staining in the medaka juvenile thymus was used to argue that thymic medullary epithelial cells might not develop before 2 wk postfertilization. However, the expression of aire by some cells located in the dorsal and inner zone of the thymus indicates that mature medullary regions have already developed in the medaka juvenile thymus. At some point after the thymic selection, cells leave the thymus through various routes, including the same ventrolateral routes used for entry. The presence of lck-positive cells outside the thymus that also express ccr9b suggests that these cells have just left the thymus, having completed their intrathymic development.

FIGURE 7.

Cell migration during T cell development in the medaka juvenile thymus. The thymus is not vascularized at 10–12 dpf. (i) Immigrant precursors expressing chemokine receptor ccr9a migrate from the extrathymic region to colonize the outer zone of the thymus. In the outer zone, the early T cell precursors subsequently proliferate (ii) and then start to express rag2 (iii). Thymocytes expressing both ccr9a and ccr9b (ccr9a/b) are located in the inner zone of the thymus (iv). A subset of ccr9a/b cells interact with DCs and macrophages, which are mainly located near the interface of the inner and outer zone of the thymus (v), as well as outside the thymus (vi). Phagocytes can also engulf ccr9a/b cells distant from their cell bodies by extending long protrusions (vii).

FIGURE 7.

Cell migration during T cell development in the medaka juvenile thymus. The thymus is not vascularized at 10–12 dpf. (i) Immigrant precursors expressing chemokine receptor ccr9a migrate from the extrathymic region to colonize the outer zone of the thymus. In the outer zone, the early T cell precursors subsequently proliferate (ii) and then start to express rag2 (iii). Thymocytes expressing both ccr9a and ccr9b (ccr9a/b) are located in the inner zone of the thymus (iv). A subset of ccr9a/b cells interact with DCs and macrophages, which are mainly located near the interface of the inner and outer zone of the thymus (v), as well as outside the thymus (vi). Phagocytes can also engulf ccr9a/b cells distant from their cell bodies by extending long protrusions (vii).

Close modal

An unexpected feature of the medaka juvenile thymus is the spatial organization of the intrathymic macrophages and DCs. In mice, DCs are located preferentially in the medulla (43, 51), whereas in medaka, cxcr3a+ DCs are enriched at the interface between the outer and inner zone of the thymus. They are not found in the inner zone of the thymus, possibly because they lack expression of ccr9b. The dying ccr9a/b thymocytes that are removed from this zone may be those that have failed thymic selection. Phagocytes detect and engulf dying cells in this region in two different ways. They can engulf ccr9a/b thymocytes via directed migration or by extending long protrusions with phagocytic cups at their tips. Conversely, ccr9a/b thymocytes, by using an in-out-in type of migratory behavior, are able to migrate toward phagocytic cells located in the outer zone and outside the thymus. After reaching a DC, a thymocyte remains in close proximity to the phagocytic cell. Because thymic DCs are able to induce negative but not positive selection (52) and thymocyte migration toward the DCs is a characteristic behavior during negative selection (53), the migration of ccr9b cells toward cxcr3a+ DCs outside the inner zone suggests that negative selection can also occur in the dorsal side of the thymus. This observation is consistent with the expression pattern of the aire gene in this region.

Ex vivo studies in mice have shown that differences in speed and directionality between cortical and medullary thymocytes are more evident when the movement of thymocytes in a specific developmental stage is analyzed (12, 15). In medaka, thymic immigrants migrate with high velocity toward the thymus in a straight path (Ref. 33 and this work). The extrathymic expression domain of ccl25a suggests that this chemokine guides thymic immigrants to the ventral side of the thymus. Inside the juvenile thymus, thymocytes migrate with lower directionality and velocity. Thymocytes show an overall heterogeneous migratory behavior, lacking any discernible global pattern; a small subset of cells that move faster is found in each zone of the thymus. Most thymocytes move randomly or show a confined migratory pattern, although some also display directed migration. Nevertheless, rag2+ cells, which represent a narrower differentiation stage, displayed a more homogeneous migratory behavior. The majority of rag2+ cells displayed random-walk movement. The same migratory behavior was also noticed in rag1-expressing thymocytes (33), suggesting that stochastic cell migration is a characteristic behavior for thymocytes during somatic recombination. The observed direct migration of ccr9a/b cells toward the DCs and the fact that they have contact with multiple DCs seems to be a characteristic behavior of thymocytes during negative selection (12). Overall, these findings suggest that, during somatic recombination and thymic selection, thymocytes exhibit characteristic behaviors. Therefore, the heterogeneous behavior of ccr9a+ and ccr9a/b+ cells within the whole thymus can be explained by the fact that these two chemokine receptors are expressed in cells that are at different developmental stages. It is worth noting that maturation of the thymus is associated with changes in the size and shape of the organ; therefore, some aspects of thymocytes’ migration in the adult thymus might differ from the juvenile thymus. This finding is consistent with the observation that thymocytes expressing rag1:egfp moved at a higher velocity in adults (33).

It has been proposed that the migratory behavior of T cells in a lymphoid organ is mostly dependent on the intricate microarchitecture of the organ, the chemokine milieu and the inherent migratory activity of cells (54). The medaka juvenile thymus, with a relatively simple architecture and chemokine milieu (i.e., it expresses uniformly one chemokine), allowed us to visualize migratory behavior of thymocytes at a point when it may be more strongly determined by cell-intrinsic mechanisms than the thymic environment.

We thank the Advanced Light Microscopy Facility (AMLF) at the European Molecular Biology Laboratory Heidelberg for continuous support; PerkinElmer and Zeiss for support of the AMLF; Joanna Natalia Buffoni and Sinja Kraus for care of fish; Thomas Boehm, Darren Gilmour, and Francesca Peri for critical comments on the manuscript; and all members of the Leptin and Wittbrodt Laboratories for suggestions and encouragement.

This work was supported by the EMBO and the European Molecular Biology Laboratory (EMBL) (to the laboratory of M.L.) and grants and fellowships from the EMBL-European Union Marie Curie Action Cofund (to B.B.), Marie-Curie Initial Training Network FishForPharma Grants FP7-PEOPLE-2011-ITN and PITN-GA-2011-289209 (to P.K.), and cofinanced EMBO-Marie Curie and Human Frontier Science Program (HFSP) long-term fellowships (to M.R.). This work was also supported by the German Research Foundation (to the laboratory of J.W.), a Japan Society for the Promotion of Science postdoctoral fellowship for research abroad and a HFSP long-term fellowship (both to D.I.), and a CellNetworks postdoctoral fellowship (to T.T.).

The online version of this article contains supplemental material.

Abbreviations used in this article:

aire

autoimmune regulator

DC

dendritic cell

dpf

days postfertilization

ERM

embryo rearing medium

MSD

mean square displacement

PCNA

proliferating cell nuclear Ag

pH3

phosphohistone H3

sgRNA

single guide RNA.

1
Takahama
Y.
2006
.
Journey through the thymus: stromal guides for T-cell development and selection.
Nat. Rev. Immunol.
6
:
127
135
.
2
Anderson
G.
,
Lane
P. J.
,
Jenkinson
E. J.
.
2007
.
Generating intrathymic microenvironments to establish T-cell tolerance.
Nat. Rev. Immunol.
7
:
954
963
.
3
Petrie
H. T.
2003
.
Cell migration and the control of post-natal T-cell lymphopoiesis in the thymus.
Nat. Rev. Immunol.
3
:
859
866
.
4
Dzhagalov
I.
,
Phee
H.
.
2012
.
How to find your way through the thymus: a practical guide for aspiring T cells.
Cell. Mol. Life Sci.
69
:
663
682
.
5
Bleul
C. C.
,
Boehm
T.
.
2000
.
Chemokines define distinct microenvironments in the developing thymus.
Eur. J. Immunol.
30
:
3371
3379
.
6
Bunting
M. D.
,
Comerford
I.
,
McColl
S. R.
.
2011
.
Finding their niche: chemokines directing cell migration in the thymus.
Immunol. Cell Biol.
89
:
185
196
.
7
Bousso
P.
,
Bhakta
N. R.
,
Lewis
R. S.
,
Robey
E.
.
2002
.
Dynamics of thymocyte-stromal cell interactions visualized by two-photon microscopy.
Science
296
:
1876
1880
.
8
Ladi
E.
,
Herzmark
P.
,
Robey
E.
.
2008
.
In situ imaging of the mouse thymus using 2-photon microscopy.
J. Vis. Exp.
(
11
).
9
Ueda
Y.
,
Katagiri
K.
,
Tomiyama
T.
,
Yasuda
K.
,
Habiro
K.
,
Katakai
T.
,
Ikehara
S.
,
Matsumoto
M.
,
Kinashi
T.
.
2012
.
Mst1 regulates integrin-dependent thymocyte trafficking and antigen recognition in the thymus.
Nat. Commun.
3
:
1098
.
10
Bhakta
N. R.
,
Oh
D. Y.
,
Lewis
R. S.
.
2005
.
Calcium oscillations regulate thymocyte motility during positive selection in the three-dimensional thymic environment.
Nat. Immunol.
6
:
143
151
.
11
Ehrlich
L. I.
,
Oh
D. Y.
,
Weissman
I. L.
,
Lewis
R. S.
.
2009
.
Differential contribution of chemotaxis and substrate restriction to segregation of immature and mature thymocytes.
Immunity
31
:
986
998
.
12
Le Borgne
M.
,
Ladi
E.
,
Dzhagalov
I.
,
Herzmark
P.
,
Liao
Y. F.
,
Chakraborty
A. K.
,
Robey
E. A.
.
2009
.
The impact of negative selection on thymocyte migration in the medulla.
Nat. Immunol.
10
:
823
830
.
13
Phee
H.
,
Dzhagalov
I.
,
Mollenauer
M.
,
Wang
Y.
,
Irvine
D. J.
,
Robey
E.
,
Weiss
A.
.
2010
.
Regulation of thymocyte positive selection and motility by GIT2.
Nat. Immunol.
11
:
503
511
.
14
Sanos
S. L.
,
Nowak
J.
,
Fallet
M.
,
Bajenoff
M.
.
2011
.
Stromal cell networks regulate thymocyte migration and dendritic cell behavior in the thymus.
J. Immunol.
186
:
2835
2841
.
15
Witt
C. M.
,
Raychaudhuri
S.
,
Schaefer
B.
,
Chakraborty
A. K.
,
Robey
E. A.
.
2005
.
Directed migration of positively selected thymocytes visualized in real time. [Published erratum appears in 2005 PLoS Biol. 3: e373.]
PLoS Biol.
3
:
e160
.
16
Benechet
A. P.
,
Menon
M.
,
Khanna
K. M.
.
2014
.
Visualizing T Cell Migration in situ.
Front. Immunol.
5
:
363
.
17
Rodewald
H. R.
2008
.
Thymus organogenesis.
Annu. Rev. Immunol.
26
:
355
388
.
18
Langenau
D. M.
,
Zon
L. I.
.
2005
.
The zebrafish: a new model of T-cell and thymic development.
Nat. Rev. Immunol.
5
:
307
317
.
19
Hirano
M.
,
Das
S.
,
Guo
P.
,
Cooper
M. D.
.
2011
.
The evolution of adaptive immunity in vertebrates.
Adv. Immunol.
109
:
125
157
.
20
Boehm
T.
,
Iwanami
N.
,
Hess
I.
.
2012
.
Evolution of the immune system in the lower vertebrates.
Annu. Rev. Genomics Hum. Genet.
13
:
127
149
.
21
Bajoghli
B.
2013
.
Evolution and function of chemokine receptors in the immune system of lower vertebrates.
Eur. J. Immunol.
43
:
1686
1692
.
22
Aghaallaei
N.
,
Bajoghli
B.
,
Schwarz
H.
,
Schorpp
M.
,
Boehm
T.
.
2010
.
Characterization of mononuclear phagocytic cells in medaka fish transgenic for a cxcr3a:gfp reporter.
Proc. Natl. Acad. Sci. USA
107
:
18079
18084
.
23
Stemmer
M.
,
Thumberger
T.
,
Del Sol Keyer
M.
,
Wittbrodt
J.
,
Mateo
J. L.
.
2015
.
CCTop: An Intuitive, Flexible and Reliable CRISPR/Cas9 Target Prediction Tool.
PLoS One
10
:
e0124633
.
24
Bajoghli
B.
,
Aghaallaei
N.
,
Czerny
T.
.
2005
.
Groucho corepressor proteins regulate otic vesicle outgrowth.
Dev. Dyn.
233
:
760
771
.
25
Bajoghli
B.
,
Aghaallaei
N.
,
Hess
I.
,
Rode
I.
,
Netuschil
N.
,
Tay
B. H.
,
Venkatesh
B.
,
Yu
J. K.
,
Kaltenbach
S. L.
,
Holland
N. D.
, et al
.
2009
.
Evolution of genetic networks underlying the emergence of thymopoiesis in vertebrates.
Cell
138
:
186
197
.
26
Inoue
D.
,
Wittbrodt
J.
.
2011
.
One for all--a highly efficient and versatile method for fluorescent immunostaining in fish embryos.
PLoS One
6
:
e19713
.
27
Calderón
L.
,
Boehm
T.
.
2011
.
Three chemokine receptors cooperatively regulate homing of hematopoietic progenitors to the embryonic mouse thymus.
Proc. Natl. Acad. Sci. USA
108
:
7517
7522
.
28
Nomiyama
H.
,
Osada
N.
,
Yoshie
O.
.
2013
.
Systematic classification of vertebrate chemokines based on conserved synteny and evolutionary history.
Genes Cells
18
:
1
16
.
29
Hess
I.
,
Boehm
T.
.
2012
.
Intravital imaging of thymopoiesis reveals dynamic lympho-epithelial interactions.
Immunity
36
:
298
309
.
30
Griffith
J. W.
,
Sokol
C. L.
,
Luster
A. D.
.
2014
.
Chemokines and chemokine receptors: positioning cells for host defense and immunity.
Annu. Rev. Immunol.
32
:
659
702
.
31
Moore
M. A.
,
Owen
J. J.
.
1967
.
Experimental studies on the development of the thymus.
J. Exp. Med.
126
:
715
726
.
32
Ritter
M. A.
1978
.
Embryonic mouse thymus development: stem cell entry and differentiation.
Immunology
34
:
69
75
.
33
Li
J.
,
Iwanami
N.
,
Hoa
V. Q.
,
Furutani-Seiki
M.
,
Takahama
Y.
.
2007
.
Noninvasive intravital imaging of thymocyte dynamics in medaka.
J. Immunol.
179
:
1605
1615
.
34
Yamada
A.
,
Takaki
S.
,
Hayashi
F.
,
Georgopoulos
K.
,
Perlmutter
R. M.
,
Takatsu
K.
.
2001
.
Identification and characterization of a transcriptional regulator for the lck proximal promoter.
J. Biol. Chem.
276
:
18082
18089
.
35
Langenau
D. M.
,
Ferrando
A. A.
,
Traver
D.
,
Kutok
J. L.
,
Hezel
J. P.
,
Kanki
J. P.
,
Zon
L. I.
,
Look
A. T.
,
Trede
N. S.
.
2004
.
In vivo tracking of T cell development, ablation, and engraftment in transgenic zebrafish.
Proc. Natl. Acad. Sci. USA
101
:
7369
7374
.
36
Yang
Q.
,
Jeremiah Bell
J.
,
Bhandoola
A.
.
2010
.
T-cell lineage determination.
Immunol. Rev.
238
:
12
22
.
37
Yui
M. A.
,
Rothenberg
E. V.
.
2014
.
Developmental gene networks: a triathlon on the course to T cell identity.
Nat. Rev. Immunol.
14
:
529
545
.
38
Oettinger
M. A.
,
Schatz
D. G.
,
Gorka
C.
,
Baltimore
D.
.
1990
.
RAG-1 and RAG-2, adjacent genes that synergistically activate V(D)J recombination.
Science
248
:
1517
1523
.
39
Ho
I. C.
,
Tai
T. S.
,
Pai
S. Y.
.
2009
.
GATA3 and the T-cell lineage: essential functions before and after T-helper-2-cell differentiation.
Nat. Rev. Immunol.
9
:
125
135
.
40
Rothenberg
E. V.
,
Moore
J. E.
,
Yui
M. A.
.
2008
.
Launching the T-cell-lineage developmental programme.
Nat. Rev. Immunol.
8
:
9
21
.
41
Osborne
B. A.
,
Smith
S. W.
,
Liu
Z. G.
,
McLaughlin
K. A.
,
Grimm
L.
,
Schwartz
L. M.
.
1994
.
Identification of genes induced during apoptosis in T lymphocytes.
Immunol. Rev.
142
:
301
320
.
42
Moran
A. E.
,
Holzapfel
K. L.
,
Xing
Y.
,
Cunningham
N. R.
,
Maltzman
J. S.
,
Punt
J.
,
Hogquist
K. A.
.
2011
.
T cell receptor signal strength in Treg and iNKT cell development demonstrated by a novel fluorescent reporter mouse.
J. Exp. Med.
208
:
1279
1289
.
43
Klein
L.
,
Kyewski
B.
,
Allen
P. M.
,
Hogquist
K. A.
.
2014
.
Positive and negative selection of the T cell repertoire: what thymocytes see (and don’t see).
Nat. Rev. Immunol.
14
:
377
391
.
44
Mathis
D.
,
Benoist
C.
.
2009
.
Aire.
Annu. Rev. Immunol.
27
:
287
312
.
45
Mittaz
L.
,
Rossier
C.
,
Heino
M.
,
Peterson
P.
,
Krohn
K. J.
,
Gos
A.
,
Morris
M. A.
,
Kudoh
J.
,
Shimizu
N.
,
Antonarakis
S. E.
,
Scott
H. S.
.
1999
.
Isolation and characterization of the mouse Aire gene.
Biochem. Biophys. Res. Commun.
255
:
483
490
.
46
Surh
C. D.
,
Sprent
J.
.
1994
.
T-cell apoptosis detected in situ during positive and negative selection in the thymus.
Nature
372
:
100
103
.
47
Love
P. E.
,
Bhandoola
A.
.
2011
.
Signal integration and crosstalk during thymocyte migration and emigration.
Nat. Rev. Immunol.
11
:
469
477
.
48
Norment
A. M.
,
Bogatzki
L. Y.
,
Gantner
B. N.
,
Bevan
M. J.
.
2000
.
Murine CCR9, a chemokine receptor for thymus-expressed chemokine that is up-regulated following pre-TCR signaling.
J. Immunol.
164
:
639
648
.
49
Wurbel
M. A.
,
Malissen
B.
,
Campbell
J. J.
.
2006
.
Complex regulation of CCR9 at multiple discrete stages of T cell development.
Eur. J. Immunol.
36
:
73
81
.
50
Benz
C.
,
Heinzel
K.
,
Bleul
C. C.
.
2004
.
Homing of immature thymocytes to the subcapsular microenvironment within the thymus is not an absolute requirement for T cell development.
Eur. J. Immunol.
34
:
3652
3663
.
51
Steinman
R. M.
,
Hawiger
D.
,
Nussenzweig
M. C.
.
2003
.
Tolerogenic dendritic cells.
Annu. Rev. Immunol.
21
:
685
711
.
52
Brocker
T.
1999
.
The role of dendritic cells in T cell selection and survival.
J. Leukoc. Biol.
66
:
331
335
.
53
Dzhagalov
I. L.
,
Chen
K. G.
,
Herzmark
P.
,
Robey
E. A.
.
2013
.
Elimination of self-reactive T cells in the thymus: a timeline for negative selection.
PLoS Biol.
11
:
e1001566
.
54
Munoz
M. A.
,
Biro
M.
,
Weninger
W.
.
2014
.
T cell migration in intact lymph nodes in vivo.
Curr. Opin. Cell Biol.
30
:
17
24
.

The authors have no financial conflicts of interest.