Medullary stromal cells synergize their production and capture of CCL21 for T-cell emigration from neonatal mouse thymus

Abstract

The release of newly selected αβT cells from the thymus is key in establishing a functional adaptive immune system. Emigration of the first cohorts of αβT cells produced during the neonatal period is of particular importance, because it initiates formation of the peripheral αβT-cell pool and provides immune protection early in life. Despite this, the cellular and molecular mechanisms of thymus emigration are poorly understood. We examined the involvement of diverse stromal subsets and individual chemokine ligands in this process. First, we demonstrated functional dichotomy in the requirement for CCR7 ligands and identified CCL21, but not CCL19, as an important regulator of neonatal thymus emigration. To explain this ligand-specific requirement, we examined sites of CCL21 production and action and found Ccl21 gene expression and CCL21 protein distribution occurred within anatomically distinct thymic areas. Although Ccl21 transcription was limited to subsets of medullary epithelium, CCL21 protein was captured by mesenchymal stroma consisting of integrin α7+ pericytes and CD34+ adventitial cells at sites of thymic exit. This chemokine compartmentalization involved the heparan sulfate-dependent presentation of CCL21 via its C-terminal extension, explaining the absence of a requirement for CCL19, which lacks this domain and failed to be captured by thymic stroma. Collectively, we identified an important role for CCL21 in neonatal thymus emigration, revealing the importance of this chemokine in initial formation of the peripheral immune system. Moreover, we identified an intrathymic mechanism involving cell-specific production and presentation of CCL21, which demonstrated a functional synergy between thymic epithelial and mesenchymal cells for αβT-cell emigration.

Graphical abstract

Figure 1.

CCR7L deficiency prolongs the intrathymic dwell time Of neonatal SP thymocytes. (A-C) Flow cytometric analysis of thymocytes in plt/plt (n = 16) and +/plt (n = 16) littermate postnatal day–10 neonatal mice. cSP4 were gated as CD4⁺TCRβ^hiCD25⁻Foxp3⁻, SP8 as CD8⁺TCRβ^hi, immature SP as CD69⁺CD62L⁻, and mature SP as CD69⁻CD62L⁺. Percentages of cSP4 and SP8 expressing Ki67 are also shown in panels B and C. (E-D) Rag2GFP levels in cSP4 and SP8 thymocytes from Rag2GFPplt/plt mice (n = 5; red) and Rag2GFP+/plt controls (n = 7; blue). Gray histograms indicate nonfluorescent control cells. (F) Numbers and frequencies of M2a, M2b, and M2c subsets of mature CD69⁻CD62L⁺ cSP4 and SP8 in Rag2GFP+/plt (n = 16) controls and Rag2GFP plt/plt mice (n = 16). Bar chart indicates percentages of each subset in Rag2GFP+/plt (blue bars) and Rag2GFP plt/plt (red bars). (G) Rag2GFP levels in the M2c subset of cSP4 and SP8 thymocytes from +/plt controls (n = 7; blue) and plt/plt (n = 5; red) mice. Bar charts show percentages of M2a, M2b, and M2c subsets of cSP4 and SP8 in +/plt (blue bars) and plt/plt (red bars). For analysis of data in panels F and G, multiple comparison analysis was achieved by a 2-way analysis of variance followed by Sidak’s posttest in GraphPad Prism to determine statistical differences. In all cases, error bars represent mean ± SEM. Flow cytometric data representative of at least 3 independent experiments. **P < .01, ***P < .001, ****P < .0001. DP, double positive; MFI, mean fluorescence intensity.

Figure 2.

CCL21 is important for neonatal SP thymocyte egress. (A) Flow cytometric analysis and quantitation of CD4/CD8 thymocyte subsets from postnatal day–10 (P10) Ccl21a^−/− (n = 15; red bars) and Ccl21a^+/− (n = 16; blue bars) littermate controls. Quantitation of immature CD69⁺CD62L⁻ and mature CD69⁻CD62L⁺ subsets of cSP4 (B) and TCRβ^hi SP8 (C) thymocytes in Ccl21a^−/− (n = 15; red bars) and Ccl21^+/− (n = 16; blue bars) littermate P10 neonatal mice. (D) Numbers and frequencies of M2a, M2b, and M2c subsets of cSP4 and SP8 thymocytes in Ccl21a^+/− (blue bars) and Ccl21a^−/− (red bars) mice. For analysis of data in panel D, multiple comparison analysis was achieved by a 2-way analysis of variance followed by Sidak’s posttest in GraphPad Prism to determine statistical differences. In all cases, error bars represent mean ± SEM. Flow cytometric data representative of 3 independent experiments. *P < .05, ***P < .001, ****P < .0001.

Figure 3.

Thymus-specific CCL21 deficiency decreases RTE frequency in WT peripheral tissues. (A) Schematic of the experimental approach used to measure thymic output from Ccl21a-deficient thymus. Freshly isolated E17 CD45.2⁺ Ccl21a^+/− or Ccl21a^−/− thymic lobes were grafted under the kidney capsule of CD45.1⁺ WT hosts. Spleens were harvested from host mice 7 days after surgery. (B) Quantitation of total thymocyte numbers, and the number and proportion of CD4⁺CD8⁺ double-positive (DP) thymocytes in E17 Ccl21a^+/− (n = 10) or Ccl21a^−/− (n = 11) thymic lobes before transplantation. Flow cytometric detection and quantitation of donor thymus–derived CD45.2⁺TCRβ^hi T cells (C) and CD45.2⁺TCRβ^hi cSP4 and CD45.2⁺TCRβ^hi SP8 T cells (D) in the spleens of WT mice that received either Ccl21a^+/− (n = 8; blue bars) or Ccl21a^−/− (n = 8; red bars) grafts. Error bars represent mean ± SEM. Flow cytometric data representative of at least 3 independent experiments. *P < .05, **P < .01.

Figure 4.

Ccl21a gene expression in the neonatal thymus is restricted to mTECs. (A) Flow cytometric analysis and quantitation of tdTomato expression in total CD45⁺ thymocytes and the indicated thymic stromal populations in postnatal day–10 (P10) heterozygous Ccl21a^tdTom mice (n = 8). Thymic mesenchyme was identified as CD45⁻CD31⁻EpCAM1⁻ cells. Red lines indicate tdTomato staining levels in heterozygous Ccl21a^tdTom mice; gray histograms indicate nonfluorescent control cells. Flow cytometric plots and analysis representative of 3 independent experiments. (B) Analysis of Ccl21a^tdTom expression in P10 EpCAM1⁺Ly51⁺UEA1⁻ cTECs and EpCAM1⁺Ly51⁻UEA1⁺ mTECs. (C) Ccl21a^tdTom expression after subdivision of total mTECs into MHCII^lowCD80^low (mTEC^lo) and MHCII^hiCD80^hi (mTEC^hi) subsets (n = 10). In all cases, error bars represent mean ± SEM. For analysis of data in panel A, multiple comparison analysis was achieved by a 1-way analysis of variance followed by Tukey’s posttest in GraphPad Prism to determine statistical differences. In panel D, confocal microscopy was used to show tdTomato expression in heterozygous Ccl21a^tdTom P10 mice alongside identification of cTECs with either anti-CD205 (green) or mTECs with UEA1 (green); dotted line denotes the CMJ. Scale bars represent 50 μm. Images shown are representative of 3 mice. ****P < .0001. C, cortex; M, medulla; MHC, major histocompatibility complex.

Figure 5.

CCL21 protein is presented by thymic mesenchyme at sites of thymic exit. (A) Confocal images of thymus sections from postnatal day–10 (P10) WT mice stained with antibodies to the endothelial marker CD31 (white), mTEC marker ERTR5 (green), and CCL21 protein (red). Blue dotted line indicates the CMJ, and blue arrows indicate vessels investigated at higher magnification in panel C. (B) Image of a thymus section from a Ccl21a knockout P10 mouse stained with anti-CCL21 (red), anti-CD31 (white), and ERTR5 (green). Note the absence of CCL21 staining. Scale bars in panels A and B denote 50 μm. (C) High-power images of CCL21⁻ (upper panels) and CCL21⁺ (lower panels) vessels identified by blue arrowheads in panel A. Images show individual channels for ERTR5 (green), CD31 (white), and CCL21 protein (red), as well as a combined image showing all markers simultaneously. Scale bars denote 25 μm. Data are representative of 4 mice from 2 separate experiments. (D) Schematic diagram and flow cytometric analysis of thymic mesenchymal populations associated with thymic blood vessels. Schematic is based on findings of Sitnik et al and demonstrates CD34⁻integrin α7⁺ pericytes and CD34⁺integrin α7⁻ adventitial mesenchymal cells that surround thymic blood vessels. Flow cytometric analysis shows identification of these populations in P10 WT thymus. (E) Flow cytometric analysis of presentation of CCL21-mRFP by indicated thymic stromal populations in plt/plt P10 thymus suspensions. Gray histograms represent control staining seen in the absence of CCL21-mRFP. Bar chart indicates percentages of CCL21-mRFP⁺ cells within each stromal subset. (F) Flow cytometric analysis of stromal cell presentation of full-length CCL21-mRFP (red), full-length CCL19-mRFP (blue), or tCCL21-mRFP (green) by CD34⁻integrin α7⁺ pericytes and CD34⁺integrin α7⁻ adventitial mesenchymal cells. Gray filled histograms represent staining levels observed when no chemokines were added. Bar chart shows mRFP mean fluorescence intensity (MFI) for each fluorescent chemokine and indicated stromal cell type. For analysis of data in panels E and F, multiple comparison analysis was achieved by a 1-way analysis of variance (ANOVA) followed by Tukey’s post-test (E) or 2-way ANOVA followed by Sidak’s posttest (F) in GraphPad Prism to determine statistical differences. All data shown representative of 3 independent experiments, with a total of 7 to 11 mice for each analysis. Error bars represent mean ± SEM. *P < .05, **P < .01, ***P < .001, ****P < .0001.

Figure 6.

Thymic heparan sulfate is restricted to adventitial mesenchyme and pericytes. Flow cytometric analysis (A) and quantitation (B) of heparan sulfate expression by CD45⁺ thymocytes, CD31⁺ endothelium, EpCAM-1⁺ TECs, CD34⁻integrin α7⁺ pericytes, and CD34⁺integrin α7⁻ adventitial mesenchyme in postnatal day–10 (P10) WT thymus. Gray filled histograms indicate staining levels where no anti–heparan sulfate primary antibody was added. (B) Analysis of mean fluorescence intensity (MFI) of heparan sulfate (HS) on CD34⁻integrin α7⁺ pericytes (orange bar) and CD34⁺integrin α7⁻ adventitial mesenchyme (purple bar) after gating on HS-expressing cells. Data shown from 3 separate experiments with a total of mice. Error bars in panel B represent mean ± SEM. For analysis of data in panel B, multiple comparison analysis was achieved by a 1-way analysis of variance followed by Tukey’s posttest in GraphPad Prism to determine statistical differences. (C) Confocal images of thymus sections from a P10 WT mouse stained with anti-CD31 (blue), anti-HS (green), and anti-CCL21 protein (red). White dotted line indicates the CMJ. Scale bar denotes 20 μm. Images shown representative of 5 mice. *P < .05, ** P < .01, ****P < .0001. C, cortex; M, medulla.

Figure 7.

Heparan sulfate mediates CCL21 presentation at sites of thymic exit. (A) Confocal image of a blood vessel in a WT postnatal day–10 (P10) thymus section stained with anti-CD31 (red) and either anti–heparan sulfate (green; upper images) or an antibody to detect Δ-heparan sulfate (green; lower images). (B) Confocal images as in panel A, but sections were treated with heparinase III (H’ase III) enzyme before antibody staining. Scale bars denote 20 μm. Images typical of 2 separate experiments involving at least 3 mice. (C) Flow cytometric analysis of heparan sulfate expression by pericytes and adventitial mesenchyme before (blue histogram and blue bar) and after (red histogram and red bar) heparinase III treatment. (D) Flow cytometric analysis of Δ-heparan sulfate expression by pericytes and adventitial mesenchyme before (blue histogram and blue bar) or after (red histogram and red bar) heparinase III treatment. Bar charts in panels C and D show mean fluorescence intensity (MFI) expression levels of heparan sulfate and Δ-heparan sulfate, respectively. Data from 3 experiments with a minimum of 8 mice. (E) Flow cytometric analysis of CCL21m-RFP chemokine presentation by adventitial mesenchyme (upper panels) and pericytes (lower panels) in digested P10 WT thymus samples before (blue line) and after (red line) heparinase III treatment. Gray histograms represent control staining where no chemokine was added. Data from 3 separate experiments and 11 mice. Error bars represent mean ± SEM. Paired Student t tests were performed for statistical analysis of data in panel E. ****P < .0001.