PlexinD1 glycoprotein controls migration of positively selected thymocytes into the medulla

Abstract

Precise intrathymic cell migration is important for thymocyte maturation and organ architecture. The orchestration of thymocyte trafficking, however, is not well understood at a molecular level. Here, we described highly regulated plexinD1 expression on CD4+CD8+ double positive (DP) thymocytes. PlexinD1 expression was further affected by the engagement of T cell receptor complex. Activation of plexinD1 via the ligand, semaphorin 3E, repressed CCL25 chemokine signaling via its receptor CCR9 in CD69+ thymocytes. In the absence of plexinD1, CD69+ thymocytes remained in the cortex, maturing to form ectopic single positive (SP) thymocyte clusters in Plxnd1-deficient fetal liver cell-transplanted mice. As a consequence, the boundary between DP and SP thymocytes at corticomedullary junctions was disrupted and medullary structures formed under the thymic capsule. These results demonstrate the importance of plexinD1 in directing migration of maturing thymocytes via modulation of biological responses to chemokine gradients.

Figure 1. Global expression analysis of DP to CD8 SP development in N15 TCR tg^+/+ RAG-2^−/− H-2^b mice

A. Thymocytes of N15TCRtg^+/+ RAG2^−/− mice were stained with anti-CD8α-FITC and anti-CD4-PE. Three different populations were sorted based on the CD4 cell surface expression level: DP is CD8^highCD4^high, Int is CD8^high/CD4^int (intermediate population) and CD8 SP is mature CD8 single positive. Total RNA was isolated from sorted thymocytes and used for microarray analysis as described in the supplementary methods. B. Hierarchical clustering was performed with genes that show significant changes (LCB >2.0). C. Gene expression Patterns I and II as noted in B. The 391 genes showing an increase during maturation were clustered in pattern I. The 328 genes that decreased (LCB <−2.0) with maturation cluster to make up pattern II. R: Receptor. Pattern III and IV are altered in the Int population only and documented in Fig. S1 and http://cvc.dfci.harvard.edu/share_folder/N15RAG-2ko_microarray_data/.

Figure 2. Expression and regulation of plexin D1 on thymocytes

A. A lysate of surface biotinylated thymocytes from 3–4 week old B6 mice was prepared. Following preclearing with goat IgG agarose, samples representing 2 × 10⁷ cells were incubated with preimmune Ab, affinity-purified anti-plexin D1 IC or with the same Ab plus a 150-fold molar excess of the immunizing peptide or an irrelevant control peptide. After SDS/PAGE and transfer, the membrane was probed with steptavidin-HRP. B. Western blotting using affinity-purified anti-plexin D1 IC on lysates of 293T cells transfected with human plexin D1 cDNA, 293-T cells transfected with an irrelevant control cDNA, primary thymocytes, and NIH3T3 mouse fibroblasts separated by SDS-PAGE (8%). C. Paraffin sections were stained with indicated anti-plexin D1 EC, IC or preimmune serum as a control. Sections were counterstained with Mayer’s hematoxylin. D. Single cell suspensions were prepared from the indicated organs of C57BL/6 mice. Total cell lysates were resolved on 6% SDS-PAGE and blotted with anti-plexin D1 IC and anti-actin antiserum (B: bone marrow; T: thymus; L: lymph node; S: spleen). E. Total RNAs were isolated from sorted thymocyte populations and used for cDNA synthesis. Synthesized cDNA was used for Realtime PCR analysis according to the manufacturer’s instructions (ABI). Normalized values using GAPDH real time PCR values with the same cDNA as a control were converted to % of maximum value (value of DP thymocyte population). Error bars are the standard error of 3 independent experiments. F. Single cell suspensions of B6 thymocytes were prepared and then stimulated with the indicated combinations of mAbs for 6, 12 and 24 h. Total cell extracts were prepared from stimulated cells after removal of dead cells using Histopaq (Sigma, MO) and analyzed by western blot with anti-plexin D1 IC. This figure is representative of 3 independent experiments. The graph shows quantitated-plexin D1 band intensities normalized as % of untreated sample.

Figure 3. Sema3E-Fc binds the plexinD1 ectodomain

A. Purified plexinD1 ectodomain (plexin D1-Fc), sema3E (sema3E-Fc) and sema3F (sema3F-Fc) proteins were analyzed by Coomassie brilliant blue staining after being resolved by SDS-PAGE. The interaction of plexinD1-Fc and sema-Fc proteins was analyzed using BIAcore as described in supplementary methods. B. PlexinD1-Fc was conjugated to UltraLink Biosupport (Pierce, NJ). Sema3E-Fc and sema3F-Fc were incubated with crosslinked plexinD1-Fc. After washing, proteins were eluted by boiling in protein sample buffer and analyzed by Coomassie brilliant blue staining after resolution on SDS-PAGE. Control is unconjugated UltraLink Biosupport resin. 3E is sema3E-Fc and 3F is sema3F-Fc as a negative control. C. Total thymocytes were stained for CD4, CD8β and CD69. The upper left panel shows the CD4/CD8β pattern and the numbers on gated regions are % of total thymocytes. The upper right histograms show the expression of CD69 in the indicated subpopulations. The darker gray area of the histogram is CD69 staining and the light gray indicates control antibody staining. Lower panels show CD69⁻ and CD69⁺ thymocytes stained with purified sema3E-Fc (10μg/ml)(blue) or Fc-only protein (red) developed with anti-mouse IgG_2c-PE. The fluorescence of sema3E-Fc is plotted in the indicated thymocyte populations. The numbers on gated regions are % of total thymocytes. The histogram for CD69⁺ DN cells are omitted because of the low cell number. D. Indicated regions were collected as described in Experimental Procedures. Total RNAs were isolated and used for cDNA synthesis. Real time PCR was performed as described in Fig 2E. Error bar is standard error of 4 independent experiments. S/C: subcapsullary zone and thymic capsule; C: Cortex; J: corticomedullary junction; M: medulla.

Figure 4. Inhibition of chemokine-mediated migration by sema3E-Fc

A. Migration assays were performed in 24-well transwell plates with a 5μm pore size as depicted. Cells were added to the insert and the indicated chemokines were added to the bottom well with purified sema3E-Fc or Fc-only protein. B. After 2 hrs, cells that had migrated to the bottom well were harvested and stained with anti-CD4, anti-CD8α, anti-CD69, and anti-TCRβ mAbs. Mock in the top panel is a sample without chemokine treatment. CCL25 was added to the bottom well in the left set of panels and CCL21 to the right set. The concentration of S3E (sema3E-Fc) is indicated on the right of each panel. In the middle panels, the CD4/CD8 patterns of the CD69⁺ cells that migrated to the bottom of the well are shown for the CCL25- and CCL21-treated cultures. More than two-thirds of CD69⁺ thymocytes were DP in CCL25-treated and less than one-quarter were DP in CCL21-treated cultures. C. An aliquot of cells harvested from B was counted and the number of CD69⁺ thymocytes migrating in response to CCL25 or CCL21 in the presence of various concentrations of sema3E was determined. Error bars are the standard error of one triplicate experiment. D. Single cell suspensions were prepared from B6 thymus, stained with the indicated mAbs and analyzed on a flowcytometer. Data were analyzed using FlowJo (TreeStar, Or). The upper left panel is a CD4/CD8 profile. CD69 expression in each gated population is depicted with dark gray above control staining in the upper right panels. In the lower four sets of panels, CCR7 and CCR9 expression in each population (gated on CD69, CD4 and CD8 expression) are shown. The upper row in each set of four shows samples stained with isotype control mAbs while the lower row shows those stained with anti-CCR7 and anti-CCR9 mAbs. The numbers in the panels are % of gated populations.

Figure 5. Aberrations in localization of CD69⁺ thymocytes and medullary structure in Plxnd1^−/− fetal liver cell transplanted thymus

A. Cryosections of reconstituted mice were stained with anti-CD4/CD8α/CD69 mAbs. Stained sections were analyzed using a Carl Zeiss LSM510 confocal microscope. The blue dashed line follows the corticomedullary junction and solid white line shows the thymic capsule. This figure is representative of four independent transplantation experiments. CD69 is depicted as white spots in the upper left section of each quartet and as blue spots in the merged image in the lower left section of the same quartet. CD4 is green and CD8 is red. (Note that CD69⁺ cells co-expressing CD4 and/or CD8 appear in different colors.) The sections are from B6 mice, mice reconstituted with Plxnd1⁺ or Plxnd1^−/− fetal liver cells (left, middle and right panels, respectively). Cryosections were stained with anti CD4 (Green)/CD8α (Red)/ER-TR5 (Blue; a marker of mTEC) mAbs. B. Stained sections were analyzed as described in A and represent magnified images of inserts (a) and (b) shown in Fig. S6A. The solid line is the corticomedullary junction and the dashed line is the boundary of SP thymocytes. The graph depicts the distribution of SP thymocytes in the cortex (ER-TR5⁻ area), analyzed by plotting SP thymocyte numbers based on the size of SP thymocyte clusters in the cortex. SP thymocyte clusters and number of cells per cluster were counted from 5 different slides. The error bar is the standard error. C. Medullary structures were visualized by staining paraffin sections of transplanted mice with hematoxylin and eosin. The indicated inserts in the chimeric thymi from Plxnd1^−/− fetal liver cell donors are magnified and the arrows highlight contact between the medulla and the thymic capsule.

Figure 6. Aberrations in localization of CD69⁺ thymocytes and corticomedullary structure in Sema3e^−/− thymus

A. Cryosections were stained as described in Fig. 5A. The blue dashed line follows the corticomedullary junction and the solid white line outlines the thymic capsule. This figure is representative of three independent analyses. CD69 is depicted as the white signal in the lower left section of each quartet. CD4 is green and CD8 is red. The sections are from Sema3e^−/− mice and Sema3e^+/+ littermate controls. White scale bar = 100μm. B. Medullary structures from Sema3e^−/− mice and Sema3e^+/+ littermate controls were visualized by staining paraffin sections with hematoxylin and eosin. Scale bar = 1mm.

Figure 7. Model of plexin D1 function in intrathymic migration

A. Plexin D1 inhibits CCR9 signaling in activated CD69⁺ DP thymocytes. As described in the text, this inhibition is selective for CCR9 and probably linked with other adhesion molecules. B. Because plexinD1 inhibits migration signals mediated through CCR9 via CCL25, activated CD69⁺ DP thymocytes rapidly migrate toward the medulla in response to CCL19/21 rather than responding to the influence of CCL25 in the cortex. Yellow cells are plexinD1⁺ thymocytes; red cells are plexinD1⁻ thymocytes; DC are dendritic cells; mTEC are medullary thymic epithelial cells; cTEC are cortical thymic epithelial cells. SCZ: subcapsular zone. For simplicity, death by neglect of CD69⁻ thymocytes or by negative selection of CD69⁺ thymocytes has been omitted.