We previously described a fibronectin/VLA-5-dependent impairment of NOD thymocyte migration, correlated with partial thymocyte arrest within thymic perivascular spaces. Yet, NOD thymocytes still emigrate, suggesting the involvement of other cell migration-related alterations. In this context, the aim of this work was to study the role of extracellular matrix ligands, alone or in combination with the chemokine CXCL12, in NOD thymocyte migration. Intrathymic contents of CXCL12, fibronectin, and laminin were evaluated by immunohistochemistry while the expression of corresponding receptors was ascertained by flow cytometry. Thymocyte migration was measured using Transwell chambers and transendothelial migration was evaluated in the same system, but using an endothelial cell monolayer within the chambers. NOD thymocytes express much lower VLA-5 than C57BL/6 thymocytes. This defect was particularly severe in CD4+ thymocytes expressing Foxp3, thus in keeping with the arrest of Foxp3+ cells within the NOD giant perivascular spaces. We observed an enhancement in CXCL12, laminin, and fibronectin deposition and colocalization in the NOD thymus. Furthermore, we detected altered expression of the CXCL12 receptor CXCR4 and the laminin receptor VLA-6, as well as enhanced migratory capacity of NOD thymocytes toward these molecules, combined or alone. Moreover, transendothelial migration of NOD thymocytes was diminished in the presence of exogenous fibronectin. Our data unravel the existence of multiple cell migration-related abnormalities in NOD thymocytes, comprising both down- and up-regulation of specific responses. Although remaining to be experimentally demonstrated, these events may have consequences on the appearance of autoimmunity in NOD mice.

The role of the thymus in autoimmune type 1 diabetes has been extensively studied in the NOD mouse. Destruction of pancreatic β cells is a T cell-dependent process (1, 2). Mature single-positive thymocytes generated by positive selection from immature cells in NOD fetal thymic organ cultures could destroy pancreatic β cells (3). However, thymectomy at weaning leads to a dramatic increase in the appearance of diabetes, suggesting the existence of thymic-dependent suppressor mechanisms (4).

NOD mouse thymus also displays abnormalities in the microenvironmental compartment, including altered distribution of epithelial cell subsets, enhanced deposition of extracellular matrix (ECM)3 and the presence of giant perivascular spaces (PVS) (5, 6, 7). Since these giant PVS contain large amounts of mature T cells, we previously raised the hypothesis of an abnormal NOD thymocyte migration (8). Accordingly, NOD thymocytes present a defect in the membrane expression of VLA-5 (the α5β1 integrin or CD49e/CD29, a fibronectin receptor), being first observed in late CD4CD8 double-negative cells and more pronounced in mature CD4 and CD8 single-positive subsets. Importantly, NOD thymocyte adhesion to thymic epithelial cells or to fibronectin is diminished; the fibronectin-driven migration of NOD thymocytes is significantly impaired and the giant PVS are filled with VLA-5-negative thymocytes that accumulate in these areas (9, 10). Rather surprisingly, however, intrathymic FITC injection, which allows the identification of FITC-stained recent thymic emigrants in peripheral lymphoid organs, showed similar rates of thymocyte emigration in NOD recent thymic emigrants when compared with controls (9), indicating that other cell migration-related ligands and receptors might be also involved in driving NOD thymocyte migration.

Chemokines are attractive cytokines constitutively expressed in the thymus and preferentially attract distinct CD4/CD8-defined thymocyte subpopulations expressing their corresponding receptors. The most common example, CXCL12, is mainly found in the subcapsular area and in the medulla of the thymic lobules. This chemokine attracts all thymocyte subpopulations that express its receptor CXCR4 (11) and also plays a role in thymocyte emigration (12, 13). Interestingly, CXCL12 binds and is presented by fibronectin and laminin, suggesting the involvement of chemokine-ECM combined interactions in thymocyte migration and development in normal and pathological conditions (14, 15, 16, 17).

Herein, we studied the role of ECM ligands as fibronectin and laminin, alone or in combination with CXCL12, in NOD thymocyte migration. In addition to fibronectin (8), the intrathymic contents of laminin and CXCL12 were augmented in the NOD thymus, when compared with controls. Although fibronectin-driven migration was largely decreased, migratory responses to laminin and CXCL12 (alone or in combination) were enhanced in NOD thymocytes. Importantly, the contact with fibronectin was able to impair transendothelial migration of NOD thymocytes. Conjointly, these data reinforce the importance of fibronectin/VLA-5 interactions in NOD thymocyte migration and export (including regulatory T cells) and unravel how complex the control is of thymocyte migration events in this mouse model.

Female C57BL/6 and NOD mice ages 12–15 wk (nondiabetic) were obtained and maintained under specific-pathogen free conditions at Necker Hospital (Paris, France). Animals were handled according to the rules precluded by the European Community. Three to seven mice were used in each type of experiment described below.

Appropriate dilutions of the following Abs were used: allophycocyanin/anti-CD4, Pacific blue/anti-CD4, PerCP/anti-CD8α, PE/anti-CD49d, PE/anti-CD49e, PE/anti-CD49f, FITC/anti-CXCR4, PE/anti-CXCR4, FITC/ anti-CD25, FITC/rat IgG2a, and PE/rat IgG2a (BD Pharmingen/BD Biosciences) and allophycocyanin/anti-Foxp3, purified/anti-Foxp3, and allophycocyanin/rat IgG2a (eBioscience). Purified/anti-CXCL12 and FITC/donkey anti-goat were obtained from Santa Cruz Biotechnology. Purified rabbit antisera specific for fibronectin and laminin were purchased from Novotec, rabbit anti-cytokeratin immunoserum was a DakoCytomation product, and rhodamine/goat anti-rabbit Ig was obtained from Sigma-Aldrich. Cellular fibronectin and laminin were purchased from Sigma-Aldrich and mouse recombinant CXCL12 was obtained from R&D Systems. In some blocking experiments, purified anti-CD49e was applied (50 μg/ml) to inhibit fibronectin-VLA-5-mediated interactions.

Thymi were removed from NOD and C57BL/6 mice (at least five animals per group), embedded in Tissue-Tek (Miles), and frozen at −70°C. Five-micrometer-thick cryostat sections were settled on poly-l-lysine (Sigma-Aldrich)-covered glass slides, and acetone fixed for 10 min. Slides were washed with PBS, followed by treatment with 1% BSA. Samples were incubated with primary Abs overnight at 8°C or for 1 h at room temperature and subsequently submitted to corresponding secondary Abs for 30 min at room temperature. Immunostained samples were analyzed by confocal microscopy using the Zeiss LSM 510 or LSM 5 Pascal devices. Negative controls, in which primary Abs were replaced by unrelated Igs or in which the secondary Ab was used alone, did not generate any significant labeling. Images obtained after double labeling for a given ECM ligand or cytokeratin and CXCL12 were quantified using the Metamorph software (Molecular Devices). Quantitative fluorescence analyses were performed by transforming specific staining in pixels and dividing the total pixel numbers by the area analyzed, obtaining the numbers of pixels/μm2. Three to five low-magnification micrographs from independent cryosections of each thymus were analyzed for the expression of the fibronectin, laminin, or cytokeratin alone or in colocalization with CXCL12. Microscopic fields thus selected comprised both cortical and medullary regions of the thymic lobules. In the case of cytokeratin labeling (alone or colocalized with CXCL12), giant PVS in NOD thymi were not included for morphometric studies.

Thymi were vigorously lacerated in 24-well plates (Costar; Corning) using a syringe plunger. Cells were washed and maintained in PBS-5% FCS for cell counting and subsequently submitted to four- or six-color immunofluorescence staining as previously described (16). Cells were then evaluated by flow cytometry in a FACSCalibur or FACSCanto II device (BD Biosciences).

Thymocyte migratory activity was assessed in the Transwell system. Briefly, 5-μm pore size Transwell plates (Costar; Corning) were coated with 10 μg/ml fibronectin or laminin, during 1 h at 37°C and later blocked with 10 μg/ml BSA. Cells (2.5 × 106 in 100 μl of RPMI 1640/1% BSA) were added in the upper chambers. After 3 h of incubation at 37°C in a 5% CO2 humidified atmosphere, migration was defined by counting the cells that migrated to the lower chambers containing migration medium alone (1% BSA-RPMI 1640) or the medium containing the chemokine CXCL12 (200 ng/ml). Cells were then labeled with appropriate Abs and analyzed by flow cytometry.

Transendothelial migration assays were done as described elsewhere (18). The mouse thymic endothelioma cell line tEnd.1 was provided by Dr. T. C. Barja-Fidalgo (University of Rio de Janeiro, Brazil). The cells were grown in RPMI 1640 medium supplemented with 10% FCS, 2 mM glutamine, 100 U/ml penicillin, and 100 U/ml streptomycin (all from Invitrogen Life Technologies). To mimic thymocyte export from the organ and to evaluate the role of the ECM molecules studied herein in the emigration process, we performed a modified ECM-transendothelial migration assay (19). The superior compartments of 8-mm pore-sized Transwell chambers were coated with BSA, fibronectin, or laminin. Endothelial cells were plated in the inferior compartment of the membranes, which were placed upside down, and cells were allowed to adhere during 2 h. Forty-eight hours later, thymocytes were added to the superior compartment of the Transwell, so that they could migrate overnight. In this case, thymocytes were first in contact with ECM molecules (or BSA, used herein as an unrelated control protein) and then with the endothelial cell layer, mimicking their tissue-to-blood emigration. Migrating cells were then counted and analyzed by flow cytometry.

Results were analyzed using Student’s t test and differences between C57BL/6 and NOD mice, unless otherwise indicated, were considered statistically significant when p < 0.05.

As we previously reported (9, 10), NOD thymocytes expressed much lower amounts of VLA-5 than C57BL/6 thymocytes. The decreased expression of VLA-5 was first observed in CD4CD8 cells, being most pronounced in mature CD4+ and CD8+ single-positive subsets of NOD thymocytes, which coincides with the accumulation of mature thymocytes within the NOD PVS, as previously reported by our group (8). Such differences are specific to NOD, because VLA-5 density is lower in NOD thymocytes, even when compared with several other normal mouse strains (9, 10). We further studied discrete CD4+ mature thymocyte subpopulations expressing Foxp3, which are considered as central (20). Multicolor flow cytometry analysis showed that there was an increase in the relative numbers of CD4+CD8CD25+ cells in the NOD thymus, as compared with controls (Fig. 1,A). The majority of these cells expressed Foxp3, but both subsets CD4+CD8CD25+Foxp3+ and CD4+CD8CD25Foxp3+ exhibited a clear-cut decrease in VLA-5 expression in NOD thymocytes (as compared with C57BL/6 animals), as seen by the flow cytometric profiles, as well as the corresponding relative numbers (Fig. 1 B). Of note, within each experimental group the relative numbers of VLA-5+ cells in the two Foxp3+ subsets were similar: ∼40–50% in normal mice and 4–5% in NOD animals.

FIGURE 1.

Accumulation of CD4+Foxp3+ cells in PVS of NOD mice. A, Enhanced numbers of CD4+CD8CD25+ cells in NOD thymus when compared with C57BL/6, analyzed by multicolor flow cytometry. As shown in B, CD4+CD8CD25+Foxp3+ thymocytes from NOD mice (filled histograms) express lower amounts of VLA-5 when compared with controls (open histograms). Similar results are also observed in CD4+CD8CD25Foxp3+ thymocytes. The graphics shown in the bottom part of this panel further reveal that the relative numbers of these subpopulations are significantly decreased. Results are expressed as mean ± SE. ∗, p < 0.05. C, Confocal analyses show the presence of Foxp3+ thymocytes (green) in the medullary region of the C57BL/6 thymus and the accumulation of these cells in a giant PVS of a NOD thymus (D), limited here by a fibronectin staining (red). Original magnification, ×400.

FIGURE 1.

Accumulation of CD4+Foxp3+ cells in PVS of NOD mice. A, Enhanced numbers of CD4+CD8CD25+ cells in NOD thymus when compared with C57BL/6, analyzed by multicolor flow cytometry. As shown in B, CD4+CD8CD25+Foxp3+ thymocytes from NOD mice (filled histograms) express lower amounts of VLA-5 when compared with controls (open histograms). Similar results are also observed in CD4+CD8CD25Foxp3+ thymocytes. The graphics shown in the bottom part of this panel further reveal that the relative numbers of these subpopulations are significantly decreased. Results are expressed as mean ± SE. ∗, p < 0.05. C, Confocal analyses show the presence of Foxp3+ thymocytes (green) in the medullary region of the C57BL/6 thymus and the accumulation of these cells in a giant PVS of a NOD thymus (D), limited here by a fibronectin staining (red). Original magnification, ×400.

Close modal

It should be pointed out that, in respect to their intrathymic distribution, Foxp3+ thymocytes were largely encountered within the limits of the NOD giant PVS, as revealed by double-labeling confocal microscopic analysis (Fig. 1, C and D). Thus, the intra-PVS concentration of these particular cell subsets in the NOD thymus followed the general rule of sequestration of mature thymocytes within NOD giant PVS, as we had previously defined for both CD4 and CD8 single-positive subpopulations (5, 8).

We previously showed that the defect in VLA-5 expression on NOD thymus was not seen in the thymic epithelium (9). Herein, we further investigated VLA-5 expression in hemopoietic-derived components of the thymic microenvironment and we observed that the percentages of VLA-5 in CD19+ B lymphocytes and CD11c+ dendritic cells (but not CD11b+ macrophages) were significantly lower in NOD mice as compared with age/sex-matched controls (data not shown).

In contrast to the pattern seen for VLA-5 (Fig. 2,A), the expression of the other integrin-type fibronectin receptor VLA-4 (α4β1 or CD49d/CD29) was enhanced in mature NOD CD4+ single-positive cells, as compared with controls. Such an increase, although rather slight, was statistically significant. By contrast, the expression of this receptor in CD8+ single-positive NOD thymocytes was diminished, as compared with controls (Fig. 2 A).

FIGURE 2.

Membrane expression of receptors for fibronectin, laminin, and CXCL12 in NOD mouse thymocytes. The histograms in A show the expression of the integrin α-chains for the fibronectin receptors VLA-4 and VLA-5 (CD49d and CD49e, respectively) and the laminin receptor VLA-6 (CD49f) in total thymocytes. Gray and black dashed lines represent unrelated Ig-staining controls of C57BL/6 and NOD thymocytes, respectively. As compared with controls, NOD thymocytes exhibit a clear decrease in VLA-5, a slight decrease in VLA-4, and a clear-cut enhancement of VLA-6. Regarding VLA-5, such a decrease can also be seen when the percentage numbers of VLA-5+ thymocytes were calculated, as revealed in the graphics seen below the corresponding histograms. When percentages of NOD VLA-4+ thymocytes were compared with their C57BL/6 counterparts, statistically significant differences were seen, with a decrease in VLA-4+ cells in the CD4+CD8+ double-positive (DP), as well as in CD8 single-positive (SP), and an increase in the CD4 subpopulation. With respect to the expression of the laminin receptor VLA-6, except for the CD4CD8 compartment, virtually all thymocytes expressed this receptor in their membranes. Nevertheless, when the MFI was evaluated, we found in NOD thymocytes a significant increase in the membrane density of this receptor, as compared with the values seen in C57BL/6 thymocytes. The histograms of cytofluorometry for CXCR4 detection in C57BL/6 (open) and NOD (filled) are shown on the left side of B. Here, a decrease in CXCR4 membrane expression is seen in NOD thymocytes, as compared with controls. Gray and black dashed lines in the histogram represent unrelated Ig-staining controls of C57BL/6 and NOD thymocytes, respectively. On the right side of this panel, evaluation of the percentages of NOD thymocytes expressing CXCR4 revealed a significant decrease in both CD4 and CD8 single-positive mature subsets, as compared with controls. C, The simultaneous expression of CXCR4 with VLA-5, VLA-4, or VLA-6 in CD4/CD8-defined thymocyte subsets. The upper part of this panel shows the corresponding cytofluorometric dot plots, which clearly reveal a decrease in VLA-5 expression among CXCR4+ NOD thymocytes and an increase in the density of VLA-6 in this particular subset. The bottom graphics show the relative cells numbers of VLA-5, VLA-4, or VLA-6 within the CXCR4+ subsets. Note that in most cases, the relative numbers of cells expressing a given receptor pair was diminished or unchanged in NOD thymocytes, as compared with controls. However, the membrane density of VLA-6 was enhanced in all CXCR4+ CD4/CD8-defined NOD thymocyte subsets. In all panels, percentages or MFI of positive cells were expressed as mean ± SE. ∗, p < 0.05.

FIGURE 2.

Membrane expression of receptors for fibronectin, laminin, and CXCL12 in NOD mouse thymocytes. The histograms in A show the expression of the integrin α-chains for the fibronectin receptors VLA-4 and VLA-5 (CD49d and CD49e, respectively) and the laminin receptor VLA-6 (CD49f) in total thymocytes. Gray and black dashed lines represent unrelated Ig-staining controls of C57BL/6 and NOD thymocytes, respectively. As compared with controls, NOD thymocytes exhibit a clear decrease in VLA-5, a slight decrease in VLA-4, and a clear-cut enhancement of VLA-6. Regarding VLA-5, such a decrease can also be seen when the percentage numbers of VLA-5+ thymocytes were calculated, as revealed in the graphics seen below the corresponding histograms. When percentages of NOD VLA-4+ thymocytes were compared with their C57BL/6 counterparts, statistically significant differences were seen, with a decrease in VLA-4+ cells in the CD4+CD8+ double-positive (DP), as well as in CD8 single-positive (SP), and an increase in the CD4 subpopulation. With respect to the expression of the laminin receptor VLA-6, except for the CD4CD8 compartment, virtually all thymocytes expressed this receptor in their membranes. Nevertheless, when the MFI was evaluated, we found in NOD thymocytes a significant increase in the membrane density of this receptor, as compared with the values seen in C57BL/6 thymocytes. The histograms of cytofluorometry for CXCR4 detection in C57BL/6 (open) and NOD (filled) are shown on the left side of B. Here, a decrease in CXCR4 membrane expression is seen in NOD thymocytes, as compared with controls. Gray and black dashed lines in the histogram represent unrelated Ig-staining controls of C57BL/6 and NOD thymocytes, respectively. On the right side of this panel, evaluation of the percentages of NOD thymocytes expressing CXCR4 revealed a significant decrease in both CD4 and CD8 single-positive mature subsets, as compared with controls. C, The simultaneous expression of CXCR4 with VLA-5, VLA-4, or VLA-6 in CD4/CD8-defined thymocyte subsets. The upper part of this panel shows the corresponding cytofluorometric dot plots, which clearly reveal a decrease in VLA-5 expression among CXCR4+ NOD thymocytes and an increase in the density of VLA-6 in this particular subset. The bottom graphics show the relative cells numbers of VLA-5, VLA-4, or VLA-6 within the CXCR4+ subsets. Note that in most cases, the relative numbers of cells expressing a given receptor pair was diminished or unchanged in NOD thymocytes, as compared with controls. However, the membrane density of VLA-6 was enhanced in all CXCR4+ CD4/CD8-defined NOD thymocyte subsets. In all panels, percentages or MFI of positive cells were expressed as mean ± SE. ∗, p < 0.05.

Close modal

A distinct expression profile was seen in respect to the VLA-6 laminin receptor (α6β1 integrin or CD49f/CD29). Although the percentages of VLA-6+ cells were slightly diminished within the double-negative subset, we found that the membrane density of VLA-6 was largely enhanced in all CD4/CD8-defined NOD thymocyte subpopulations, when compared with controls, as ascertained by the mean fluorescence intensity (MFI; Fig. 2 A).

Given the combined role of ECM ligands and the chemokine CXCL12 upon thymocyte migration (21), we further examined the expression of CXCR4 in NOD thymocytes. CXCR4 is expressed by all thymocyte subpopulations. Nevertheless, we observed a discrete decrease in the expression of this receptor in total NOD thymocytes when compared with controls, particularly in the mature subsets being statistically significant in CD8 single- positive cells (Fig. 2 B).

We next evaluated the density of ECM ligands and CXCL12 in the NOD mouse thymus. As shown in Fig. 3, A–D, NOD thymi presented enhanced deposition of fibronectin, laminin, and CXCL12 when compared with controls. This enhancement was significantly different as ascertained by computer-based quantitation. Additionally, colocalization of CXCL12 with any of the ECM molecules was enhanced (Fig. 3, G and H).

FIGURE 3.

Enhanced deposition of fibronectin, laminin, and CXCL12 in the NOD thymus. A and B, Typical immunohistochemical profiles showing the deposition of fibronectin (red) and CXCL12 (green) in C57BL/6 and NOD thymi, respectively. C and D, Correspond to the deposition of laminin (red) and CXCL12 (green) in C57BL/6 and NOD thymi, respectively. E and F, The colocalization of epithelial (cytokeratin-positive) cells with CXCL12 in thymi from C57BL/6 and NOD mice, respectively. F, The dotted line defines the limits of a giant PVS located on the right side of the micrograph. CXCL12 was also detected within the giant PVS of the NOD thymus as well as in the endothelial cells (white arrow). G, Bars correspond to quantitative analysis of selected microscopic fields of thymi from C57BL/6 and NOD mice (n = 5 thymi/group) in terms of fibronectin, laminin, and CXCL12, with the results expressed in pixels/μm2. H, The percentages of CXCL12 colocalization with fibronectin or laminin are higher in NOD thymi when compared with C57BL/6 (∗, p < 0.05). I, Although the general amount of thymic epithelial cell profiles (defined by anticytokeratin labeling, seen in red fluorescence) did not differ in C57BL/6 vs NOD thymi, there is a significant increase in the colocalization of CXCL12 (labeled with green fluorescence) in cytokeratin+ thymic epithelial cells, as revealed by the yellow color recorded. As detailed in Materials and Methods, quantitative fluorescence analyses were performed by transforming specific staining in pixels and dividing the total pixel numbers by the area analyzed, obtaining the numbers of pixels/μm2. Three to five low-magnification micrographs from independent cryosections of each thymus were analyzed for the expression of each marker, alone or in colocalization. In the case of cytokeratin labeling, giant PVS in NOD thymi were not included for morphometric studies. Original magnification, ×400.

FIGURE 3.

Enhanced deposition of fibronectin, laminin, and CXCL12 in the NOD thymus. A and B, Typical immunohistochemical profiles showing the deposition of fibronectin (red) and CXCL12 (green) in C57BL/6 and NOD thymi, respectively. C and D, Correspond to the deposition of laminin (red) and CXCL12 (green) in C57BL/6 and NOD thymi, respectively. E and F, The colocalization of epithelial (cytokeratin-positive) cells with CXCL12 in thymi from C57BL/6 and NOD mice, respectively. F, The dotted line defines the limits of a giant PVS located on the right side of the micrograph. CXCL12 was also detected within the giant PVS of the NOD thymus as well as in the endothelial cells (white arrow). G, Bars correspond to quantitative analysis of selected microscopic fields of thymi from C57BL/6 and NOD mice (n = 5 thymi/group) in terms of fibronectin, laminin, and CXCL12, with the results expressed in pixels/μm2. H, The percentages of CXCL12 colocalization with fibronectin or laminin are higher in NOD thymi when compared with C57BL/6 (∗, p < 0.05). I, Although the general amount of thymic epithelial cell profiles (defined by anticytokeratin labeling, seen in red fluorescence) did not differ in C57BL/6 vs NOD thymi, there is a significant increase in the colocalization of CXCL12 (labeled with green fluorescence) in cytokeratin+ thymic epithelial cells, as revealed by the yellow color recorded. As detailed in Materials and Methods, quantitative fluorescence analyses were performed by transforming specific staining in pixels and dividing the total pixel numbers by the area analyzed, obtaining the numbers of pixels/μm2. Three to five low-magnification micrographs from independent cryosections of each thymus were analyzed for the expression of each marker, alone or in colocalization. In the case of cytokeratin labeling, giant PVS in NOD thymi were not included for morphometric studies. Original magnification, ×400.

Close modal

Interestingly, in addition to CXCL12 localization in the thymic epithelium (revealed with an anti-cytokeratin Ab), we found some cells expressing CXCL12 in giant PVS in NOD thymi, mainly at the endothelial cell layer (Fig. 3,F). We also performed morphometric analyses to demonstrate that, even in the absence of an increase of the thymic epithelial cell (TEC) network, colocalization of CXCL12 in TEC was enhanced on NOD mouse thymus (Fig. 3 I).

Since thymocyte migration is a guided response to the combined stimuli of at least ECM molecules and chemokines (22), we also checked the expression of both receptors in the same cells. We observed a decrease in the relative numbers of NOD thymocytes simultaneously expressing VLA-4/CXCR4, VLA-5/CXCR4, and VLA-6/CXCR4 receptor pairs, when compared with controls. This was particularly significant for VLA-5+CXCR4+, being clear-cut when the corresponding cytofluorometric dot pots were analyzed (Fig. 2,C, upper and lower panels). Regarding the simultaneous expression of VLA-6 and CXCR4, despite the reduced percentage numbers of CXCR4+VLA-6+ cells in the NOD thymus, we found a significant enhancement in terms of VLA-6 fluorescence mean intensity in all NOD CXCR4+CD4/CD8-defined thymocyte subsets, as compared with controls (Fig. 2 C).

We also phenotyped regulatory CD4+CD8CD25+Foxp3+ cells for the expression of fibronectin (VLA-4 and VLA-5), laminin (VLA-6), and CXCL12 (CXCR4) receptors. The corresponding relative cell numbers, summarized in Table I, show that in NOD mice, there is a huge decrease in the percentages of VLA-5+ thymocytes expressing the CD4+CD8CD25+Foxp3+ phenotype. By contrast, there is a significant increase in the percentages of cells expressing CXCR4. Nevertheless, when evaluating the simultaneous expression of VLA-5 and CXCR4, we found a 4-fold decrease in the percentages of NOD thymocytes expressing the VLA-5+CXCR4+CD4+CD8CD25+Foxp3+ phenotype.

Table I.

VLA and CXCR4 expression in CD4+CD8CD25+Foxp3+ thymocytes from control and NOD micea

CD4+CD8CD25+Foxp3+C57BL/6NOD
VLA-4+ 81.1 ± 2.9 73.6 ± 0.6b 
VLA-5+ 38.4 ± 1.6 5.3 ± 0.2b 
VLA-6+ 98.5 ± 0.3 99.1 ± 0.3 
CXCR4+ 49.9 ± 1.3 59.0 ± 1.3b 
VLA-4+/CXCR4+ 42.0 ± 1.9 44.6 ± 0.6 
VLA-5+/CXCR4+ 19.9 ± 1.0 4.2 ± 0.3b 
VLA-6+/CXCR4+ 45.4 ± 1.7 54.9 ± 0.6b 
CD4+CD8CD25+Foxp3+C57BL/6NOD
VLA-4+ 81.1 ± 2.9 73.6 ± 0.6b 
VLA-5+ 38.4 ± 1.6 5.3 ± 0.2b 
VLA-6+ 98.5 ± 0.3 99.1 ± 0.3 
CXCR4+ 49.9 ± 1.3 59.0 ± 1.3b 
VLA-4+/CXCR4+ 42.0 ± 1.9 44.6 ± 0.6 
VLA-5+/CXCR4+ 19.9 ± 1.0 4.2 ± 0.3b 
VLA-6+/CXCR4+ 45.4 ± 1.7 54.9 ± 0.6b 
a

Results are expressed as percentage mean ± SE.

b

Differences in NOD cells were statistically significant when compared to controls. Data were derived from three control and three NOD mice.

Overall, the data presented above reveal a differential pattern of the expression of ECM/CXCL12 receptor pairs in NOD thymocytes, which in turn could be related to their migratory response to CXCL12 and ECM molecules. We thus evaluated NOD thymocyte migration toward these stimuli. We previously showed that thymocytes from 8- to 10-wk-old NOD mice migrate less than thymocytes from BALB/c and C57BL/6 mice toward fibronectin, an alteration related to the VLA-5 expression defect (9, 10). This was confirmed herein in older 12- to 15-wk-old animals (Fig. 4,A). Interestingly, when Transwell chambers were coated with laminin, NOD thymocytes migrated more than controls. Similarly, CXCL12-driven migration was also enhanced as compared with controls (Fig. 4 A).

FIGURE 4.

Altered thymocyte migration in NOD mice. A, Bars show the total numbers of migrating thymocytes through fibronectin or laminin, or toward CXCL12, as well as CXCL12 combined with one given ECM protein. Although fibronectin-driven migration is clearly diminished in NOD thymocytes, it is enhanced when laminin is applied as the haptotactic stimulus. Similarly, CXCL12-driven migration is higher in NOD thymocytes, as compared with controls. When the chemokine was used in combination with a given ECM protein, the resulting migration was diminished or increased when fibronectin or laminin was, respectively, applied. Note that in these experiments, the total numbers of migrating cells, in both ECM/chemokine combined stimulations, were clearly higher then the sum of the two stimuli applied alone. This was seen with regard to both C57BL/6 and NOD thymocytes. B, Bars show the relative numbers (percentages of input) of migrating CD4/CD8-defined thymocyte subsets. In general, the alteration patterns of NOD vs C57BL/6 seen in A were also found with regard to thymocyte subsets: decrease when fibronectin was applied, increase when laminin or CXCL12 was applied. Combined stimulation followed the same trends seen when the given ECM was used alone, although synergy was seen with respect to most thymocyte subsets from both normal and NOD mice. Experiments using anti-VLA-5 mAb treatment (seen in C) revealed that fibronectin-driven migration was largely abrogated by anti-VLA-5 treatment in control thymocytes (>50% reduction), whereas in NOD mice, reduction was not statistically significant. When both fibronectin plus CXCL12 were applied, there was a large decrease of migrating thymocytes from control mice in the presence of anti-VLA-5 Ab (>60% reduction). In NOD thymocytes, reduction was also significant, but to a much lesser extent (<35%). In all panels, values correspond to specific migration after subtracting the numbers of migrating cells obtained for each individual in wells coated with BSA, and results are expressed as mean ± SE. ∗, p < 0.05.

FIGURE 4.

Altered thymocyte migration in NOD mice. A, Bars show the total numbers of migrating thymocytes through fibronectin or laminin, or toward CXCL12, as well as CXCL12 combined with one given ECM protein. Although fibronectin-driven migration is clearly diminished in NOD thymocytes, it is enhanced when laminin is applied as the haptotactic stimulus. Similarly, CXCL12-driven migration is higher in NOD thymocytes, as compared with controls. When the chemokine was used in combination with a given ECM protein, the resulting migration was diminished or increased when fibronectin or laminin was, respectively, applied. Note that in these experiments, the total numbers of migrating cells, in both ECM/chemokine combined stimulations, were clearly higher then the sum of the two stimuli applied alone. This was seen with regard to both C57BL/6 and NOD thymocytes. B, Bars show the relative numbers (percentages of input) of migrating CD4/CD8-defined thymocyte subsets. In general, the alteration patterns of NOD vs C57BL/6 seen in A were also found with regard to thymocyte subsets: decrease when fibronectin was applied, increase when laminin or CXCL12 was applied. Combined stimulation followed the same trends seen when the given ECM was used alone, although synergy was seen with respect to most thymocyte subsets from both normal and NOD mice. Experiments using anti-VLA-5 mAb treatment (seen in C) revealed that fibronectin-driven migration was largely abrogated by anti-VLA-5 treatment in control thymocytes (>50% reduction), whereas in NOD mice, reduction was not statistically significant. When both fibronectin plus CXCL12 were applied, there was a large decrease of migrating thymocytes from control mice in the presence of anti-VLA-5 Ab (>60% reduction). In NOD thymocytes, reduction was also significant, but to a much lesser extent (<35%). In all panels, values correspond to specific migration after subtracting the numbers of migrating cells obtained for each individual in wells coated with BSA, and results are expressed as mean ± SE. ∗, p < 0.05.

Close modal

When CXCL12 was applied combined with a given ECM ligand, the resulting migration varied, depending on which ECM molecule was applied. When fibronectin and CXCL12 were used together, NOD thymocytes migrated less than controls. Nevertheless, the opposite was found when CXCL12 was used in combination with laminin (Fig. 4 A).

Interestingly enough, Fig. 4 A also shows that the combination of CXCL12 with fibronectin or laminin resulted in migration levels that were far higher than the sum of migrating cells obtained with CXCL12 or a given ECM ligand alone. Such synergy was seen in both control and NOD thymocytes.

It should be noted that these results do represent the migration patterns specifically driven by each chemokine and/or ECM ligand(s). To avoid the possibility of measuring a stimulus-independent increase in the motility of NOD thymocytes, we discounted, in each animal, the values recorded after migration toward BSA from those values obtained as a result of each specific stimulus (alone or combined).

When examining the migratory capacity of CD4/CD8-defined thymocyte subsets, overall we observed similar differences to those seen in the total population. Except for CD4CD8 cells, fibronectin-driven migration was diminished in all other stages of thymocyte differentiation, with differences being more striking in mature subsets. An opposing profile was seen when NOD thymocytes migrated through laminin: except for double-negative cells, migration was enhanced in NOD thymocytes, being significantly different from controls in the double-positive and CD4 single-positive subsets (Fig. 4 B). With respect to migration toward CXCL12 alone, NOD thymocyte migration was higher than controls in all CD4/CD8-defined subpopulations.

When CXCL12 was applied in combination with fibronectin or laminin, significant differences between NOD and control thymocytes were observed: when fibronectin plus CXCL12 was used, resulting migration was decreased in NOD thymocytes, except for the double-negative subset. By contrast, in the laminin plus CXCL12 combination CD4+CD8+ and mature single-positive cells from NOD mice migrated significantly more than corresponding controls (Fig. 4 B).

Of note, when CD4/CD8-defined subsets were evaluated, the concept of synergy initially seen in respect to total migrating cells remained valid, being observed in virtually all subsets from normal and NOD thymocytes.

It should also be pointed out that treatment of thymocytes with anti-VLA-5 mAb largely impaired fibronectin-driven migration of normal thymocytes, but no statistically significant effect was seen in NOD thymocytes (Fig. 4,C). When fibronectin plus CXCL12 were applied in combination, NOD thymocyte migration was diminished in the presence of anti-VLA-5 Ab. Yet, in this experimental condition, the anti-VLA-5-induced impairment was much more important in normal thymocytes (Fig. 4 C).

We previously showed that VLA-5-negative thymocytes are accumulated in the giant PVS in NOD thymi, although similar rates of thymocyte emigration were recorded in short-term experiments when compared with control animals (9). Since the migration capacity of NOD thymocytes driven by laminin and/or CXCL12 was enhanced, we decided to investigate the role of these molecules in NOD thymocyte emigration, by using a transendothelial migration assay.

As shown in Fig. 5, we first observed that the thymic endothelial cell line tEnd.1 expressed CXCL12, as well as laminin and fibronectin (Fig. 5, A–C). When thymocytes were allowed to migrate through the endothelial cell monolayer, no differences were observed when we compared NOD thymocyte transmigration to controls (data not shown). However, when the superior surface of the Transwell membrane was coated with fibronectin while the inferior surface was covered by the endothelial monolayer, NOD thymocyte migration was significantly diminished as compared with controls (Fig. 5,D). By contrast, no differences were seen when laminin was used as coating molecule, nor when CXCL12 was applied in the bottom chamber of the Transwells (Fig. 5,D), suggesting that the encounter of migrating cells with fibronectin before reaching the endothelial layer may negatively interferes with NOD thymocyte migration. Furthermore, although this blockage was observed in all thymocyte subpopulations, it was particularly striking in the mature CD4+ and CD8+ single-positive cells (Fig. 5 E), which are those physiologically able to be exported from the thymus.

FIGURE 5.

Transendothelial migration of NOD thymocytes is diminished in the presence of exogenous fibronectin. A–C, Correspond to the immunofluorescence detection of fibronectin, laminin, and CXCL12, respectively, in the thymic endothelial cell line tEnd-1. Labeling after using appropriate Ig-matched unrelated Igs did not yield any significant fluorescence signal (data not shown). D, The total numbers of thymocytes that migrated through an endothelial cell layer previously coated with fibronectin or laminin. NOD thymocytes migrated significantly less than controls when fibronectin (but not laminin) was applied. Also similar were the migration values recorded when CXCL12 was applied alone in the bottom well of the migration chambers. As seen in E, the fibronectin-related decrease of transendothelial migration seen with NOD thymocytes was statistically significant with regard to the CD4-CD8 double-positive (DN), as well as CD4 and CD8 single-positive (SP) cells. Although to a lesser extent, a decrease in transendothelial NOD thymocyte migration was also seen when fibronectin was applied onto the endothelial cell layer and CXCL12 in the bottom wells. Such a decrease was seen in term of total thymocytes as well as in CD4/CD8-defined subsets (F). Results are expressed as mean ± SE. ∗, p < 0.05.

FIGURE 5.

Transendothelial migration of NOD thymocytes is diminished in the presence of exogenous fibronectin. A–C, Correspond to the immunofluorescence detection of fibronectin, laminin, and CXCL12, respectively, in the thymic endothelial cell line tEnd-1. Labeling after using appropriate Ig-matched unrelated Igs did not yield any significant fluorescence signal (data not shown). D, The total numbers of thymocytes that migrated through an endothelial cell layer previously coated with fibronectin or laminin. NOD thymocytes migrated significantly less than controls when fibronectin (but not laminin) was applied. Also similar were the migration values recorded when CXCL12 was applied alone in the bottom well of the migration chambers. As seen in E, the fibronectin-related decrease of transendothelial migration seen with NOD thymocytes was statistically significant with regard to the CD4-CD8 double-positive (DN), as well as CD4 and CD8 single-positive (SP) cells. Although to a lesser extent, a decrease in transendothelial NOD thymocyte migration was also seen when fibronectin was applied onto the endothelial cell layer and CXCL12 in the bottom wells. Such a decrease was seen in term of total thymocytes as well as in CD4/CD8-defined subsets (F). Results are expressed as mean ± SE. ∗, p < 0.05.

Close modal

We then evaluated the membrane expression of VLAs in those cells migrating through the inserts containing fibronectin and endothelial cells. We found that NOD transendothelial-migrating cells were VLA-5-enriched, when compared with nonmigrating cells or with all cells examined before migration (Table II). This was strikingly different from what was seen when C57BL/6 thymocytes were allowed to migrate through inserts containing fibronectin and endothelial cells: no differences in VLA-5 expression were found before vs after migration.

Table II.

VLA-5 expression on thymocytes from control and NOD mice before and after transendothelial migrationa

CD4CD8CD4+CD8+CD4+CD8CD4CD8+
C57BL/6NODC57BL/6NODC57BL/6NODC57BL/6NOD
Before migration (%) 86.4 ± 2.0 53.8 ± 4.7 92.7 ± 0.8 48.2 ± 0.8 93.4 ± 0.2 38.4 ± 2.6 89.6 ± 0.8 24.9 ± 3.8 
After fibronectin-driven migration (%) 90.1 ± 2.2 84.2 ± 2.1b 89.6 ± 2.0 54.6 ± 0.7b 94.3 ± 0.5 61.3 ± 0.2b 89.0 ± 2.6 64.5 ± 1.2b 
Before migration (MFI) 28.8 ± 1.1 26.4 ± 2.4 16.6 ± 0.3 11.6 ± 0.3 24.0 ± 0.4 12.7 ± 0.3 21.3 ± 0.3 11.3 ± 0.3 
After fibronectin-driven migration (MFI) 27.4 ± 3.5 27.7 ± 3.1 16.0 ± 1.3 14.4 ± 0.4b 24.3 ± 2.5 16.6 ± 1.3b 19.4 ± 1.8 12.5 ± 0.9b 
CD4CD8CD4+CD8+CD4+CD8CD4CD8+
C57BL/6NODC57BL/6NODC57BL/6NODC57BL/6NOD
Before migration (%) 86.4 ± 2.0 53.8 ± 4.7 92.7 ± 0.8 48.2 ± 0.8 93.4 ± 0.2 38.4 ± 2.6 89.6 ± 0.8 24.9 ± 3.8 
After fibronectin-driven migration (%) 90.1 ± 2.2 84.2 ± 2.1b 89.6 ± 2.0 54.6 ± 0.7b 94.3 ± 0.5 61.3 ± 0.2b 89.0 ± 2.6 64.5 ± 1.2b 
Before migration (MFI) 28.8 ± 1.1 26.4 ± 2.4 16.6 ± 0.3 11.6 ± 0.3 24.0 ± 0.4 12.7 ± 0.3 21.3 ± 0.3 11.3 ± 0.3 
After fibronectin-driven migration (MFI) 27.4 ± 3.5 27.7 ± 3.1 16.0 ± 1.3 14.4 ± 0.4b 24.3 ± 2.5 16.6 ± 1.3b 19.4 ± 1.8 12.5 ± 0.9b 
a

Results are expressed as the mean ± SE of percentages or MFI of VLA-5+ cells.

b

Differences in NOD thymocytes were statistically significant when compared to VLA-5 expression before migration. No differences were seen concerning C57BL/6 thymocytes. Data were derived from three control and six NOD mice.

Interestingly, the VLA-5 enrichment seen after transendothelial NOD thymocyte migration was particularly seen in mature single-positive cells, as ascertained by relative cell numbers and MFI (Table II).

Differing from the enrichment in VLA-5-expressing cells, seen after NOD thymocyte migration across fibronectin plus endothelial cells, no alterations were seen in VLA-6 expression on NOD thymocytes before vs after migration, neither in terms of VLA-6+ relative cell numbers nor in the MFI (data not shown).

Although to a lesser extent, decreased numbers of migrating thymocytes were also seen when transendothelial cell migration was performed in the presence of fibronectin plus CXCL12. Again, the same trend was seen, not only when total thymocytes were evaluated, but also occurred in each CD4/CD8-defined subsets (Fig. 5 F).

Conjointly, these data reinforce the hypothesis that VLA-5/fibronectin interactions play a role in thymocyte emigration.

Previous data demonstrated the existence of an abnormal migration of NOD mouse thymocytes, with formation of giant PVS (5, 8, 9, 10). In this study, we showed that Foxp3+ cells are accumulated in giant PVS, indicating that regulatory T cells are also progressively retained within the NOD thymus. In fact, the percentages of CD4+CD8Foxp3+ cells were 4-fold reduced, as compared with what was found in normal mouse thymocytes.

We had previously demonstrated that the defect in VLA-5 expression in the NOD thymus was not seen in the epithelial component of the thymic microenvironment (9). In this study, we observed that NOD thymic dendritic cells and B lymphocytes also exhibit a VLA-5 defect. Nevertheless, such a defect was not seen in NOD thymic macrophages. The VLA-5 defect seen in B lymphocytes fits with our previous data showing clusters of B lymphocytes within NOD giant PVS (5) and raises the notion that the abnormal control of VLA-5 expression in the NOD thymus comprises distinct (but not all) hemopoietic-derived cell types.

Should we draw attention to the fact that the NOD thymus microenvironment is altered in terms of production of cell migration-related molecules, including fibronectin, laminin, and CXCL12. As seen by means of the morphometric analysis, NOD TEC should produce higher amounts of this chemokine, which in turn, is deposited onto the laminin and fibronectin-containing ECM fibrils throughout the NOD thymic lobules.

With regard to ECM ligands and receptors, we have previously examined various normal mouse strains, including the outbred Swiss mice. The results were conclusive in that the VLA-5 defect was restricted to NOD thymocytes (9). Functionally, only NOD thymocytes exhibited a decrease in fibronectin-driven migration. Moreover, giant PVS, bearing a fibronectin-containing network, were not seen in age-matched normal mice nor in various congenic mouse strains (23).

We further provided evidence strongly indicating that the VLA-5 defect seen in NOD thymocytes is part of a complex multivectorial mechanism, comprising hypo- and hyperresponsiveness to specific migratory stimuli. Accordingly, fixed amounts of a given stimulus applied ex vivo resulted in opposing responsiveness. As compared with C57BL/6 mice, NOD thymocytes exhibited a lower migratory response to fibronectin, but overresponded to laminin and CXCL12. Conceptually, these findings demonstrate an intrinsic altered migratory responsiveness of NOD thymocytes.

It is noteworthy that hypo- and hyperresponsiveness, respectively, to fibronectin and laminin are correlated with the membrane density of the respective ECM receptors on NOD thymocytes, namely, a decrease in VLA-5 and an increase in VLA-6. However, this was not the case regarding CXCR4, which was not overexpressed in NOD thymocytes, despite the fact that these cells exhibited a higher ex vivo migratory responsiveness, as compared with controls. This indicates that, in addition to the changes in membrane density, alterations in the activation levels of cell migration-related receptors also occur in NOD thymocytes.

In any case, when NOD thymocytes were allowed to migrate toward a combination of CXCL12 plus a given ECM ligand, the resulting migration was higher or lower than normal thymocytes, if the ECM protein was laminin or fibronectin, respectively.

A second aspect provided by these data is the synergy between the ECM ligands fibronectin and laminin and the chemokine CXCL12 in terms of inducing oriented thymocyte migration. This was seen in both C57/BL6 and NOD thymocytes. Interestingly, it has also been reported in growth hormone transgenic mice, as well as in BALB/c mice infected with the protozoan parasite Trypanosoma cruzi, the causative agent of Chagas’ disease (16, 17).

To correlate the migration-related abnormalities of NOD thymocytes with the progressive in vivo cell arrest and consequent appearance of giant PVS, we performed ex vivo experiments. We standardized a modified ECM/transendothelial migration assay, in which the contact with fibronectin preceded the migration across the endothelial cell layer. In these conditions, transendothelial migration of NOD thymocytes was significantly diminished as compared with controls. Such a decrease was seen in all CD4/CD8-defined subsets, but was more severe in mature CD4+ and CD8+ single-positive cells. Differently, when laminin or CXCL12 were added alone in this transendothelial migration system, migration values were similar with regard to NOD and normal mouse thymocytes.

Most importantly, we found a significant enrichment in VLA-5 expression in those NOD thymocytes that migrated across the endothelial cell layer covered with fibronectin. By contrast, such VLA-5 enrichment was not seen in C57BL/6 thymocytes.

Interestingly, although to a lesser extent, such a fibronectin-related decrease of transendothelial migration values was even observed when exogenous CXCL12 was applied in the bottom well of the transmigration chambers. This indicates that in NOD thymocytes, the migration vector driven by CXCL12 is not sufficient to counterbalance the defective migratory response though fibronectin, even when thymocytes migrate through a thymic endothelial cell layer.

Since in these ex vivo experiments we always used freshly isolated thymocytes, it is likely that the response reflects the in vivo behavior of these cells. In this respect, although in very short-term BrdU experiments differences in thymocyte emigration were not seen in NOD vs controls, long-term experiments revealed the accumulation of labeled thymocytes within giant PVS (8). Yet, more direct evidence in this regard could be attained with the generation of GFP-transgenic NOD mice, so as to allow a direct two-photon microscopy analysis of cell migration within the organ.

In any case, our data support the hypothesis in which those NOD thymocytes that succeed in exiting the organ have achieved a threshold of VLA-5 expression that allows them to migrate from the interior of the thymic parenchyma across the blood vessel wall.

The abnormalities of NOD thymocyte migratory capacity should be also placed in the context of the thymus microenvironment. In fact, through their migratory journey, NOD thymocytes likely encounter abnormally high amounts of migration-related stimuli, as revealed herein by fibronectin, laminin, and CXCL12. Although we have not analyzed this issue, it is likely that velocity of NOD thymocytes through the thymic microenvironmental network should be a resulting vector that corresponds to the sum of all of these ligand/receptor pairs. In this respect, each individual migrating vector should reflect the avidity of the corresponding ligand-receptor interaction, which is defined, not only by the affinity of the interaction itself, but also by the numbers of ligand/receptor pairs simultaneously interacting at a given moment (Fig. 6). Among others, these stimuli comprise ECM ligands, cytokines, and chemokines expressed in different concentrations depending on the thymic region. In parallel, thymocytes up- and down-regulate different membrane receptors that allow them to tune their responsiveness to ECM and chemokines. As a result of all stimuli together, thymocytes will respond migrating to a given direction, with a given velocity, which can change depending on the concentration and combination of each stimulus in each thymic region, as well as their capacity to respond via their specific receptors. This can be named multivectorial cell migration.

FIGURE 6.

Multivectorial thymocyte migration defects in NOD mice. The left panel schematically defines the hypothesis stating that the thymocyte migration results from a balance of several and simultaneous interactions between cell migration modulators and respective receptors. Accordingly, individual vectors form a resulting vector that leads thymocytes throughout their journey within the thymic lobules, from the cortex toward the medulla, with mature subsets ultimately being exported from the organ through the blood vessel walls. Right panel, Representing the situation in the NOD mouse, state that such multivectorial migration is altered in this particular pathological condition. Narrow arrows represent individual migration vectors induced by specific stimuli as fibronectin (green), laminin (blue), CXCL12 (violet), and other molecules (gray). Dashed arrows represent the resulting migration vector of thymocytes from C57BL/6 and NOD mice (yellow and pink, respectively). As compared with controls, thymocyte migration in NOD mice is altered concerning at least the three molecules analyzed herein. This leads to distinct resulting vectors; with NOD being slightly smaller then C57BL/6. Along with time, a proportion of mature VLA-5-negative thymocytes do not exit the organ to the blood vessels, and these cells are progressively accumulated within the PVS.

FIGURE 6.

Multivectorial thymocyte migration defects in NOD mice. The left panel schematically defines the hypothesis stating that the thymocyte migration results from a balance of several and simultaneous interactions between cell migration modulators and respective receptors. Accordingly, individual vectors form a resulting vector that leads thymocytes throughout their journey within the thymic lobules, from the cortex toward the medulla, with mature subsets ultimately being exported from the organ through the blood vessel walls. Right panel, Representing the situation in the NOD mouse, state that such multivectorial migration is altered in this particular pathological condition. Narrow arrows represent individual migration vectors induced by specific stimuli as fibronectin (green), laminin (blue), CXCL12 (violet), and other molecules (gray). Dashed arrows represent the resulting migration vector of thymocytes from C57BL/6 and NOD mice (yellow and pink, respectively). As compared with controls, thymocyte migration in NOD mice is altered concerning at least the three molecules analyzed herein. This leads to distinct resulting vectors; with NOD being slightly smaller then C57BL/6. Along with time, a proportion of mature VLA-5-negative thymocytes do not exit the organ to the blood vessels, and these cells are progressively accumulated within the PVS.

Close modal

It should be noted that in vivo, NOD thymocytes, once reaching the PVS, will likely encounter the fibronectin-containing intra-PVS network before seeing the endothelial cell layer, through which those bearing enough membrane VLA-5 will migrate so as to exit the organ, whereas thymocytes with no or low membrane expression of the receptor will preferentially be retained within the PVS, thus progressively promoting the enormous enlargement of these structures.

Interestingly, the enlargement of thymic PVS was reported in a few other autoimmune disorders such as myasthenia gravis in humans (24) and in BWF1 mice, a murine model for systemic lupus erythematosus (25). In these mice, B1 cells migrate aberrantly in the thymus during the development of lupus nephritis and their intra-PVS accumulation was correlated with the altered expression of the chemokine CXCL13 and adhesion molecules.

One might argue that in vitro/ex vivo approaches for phenotypic and functional analyses are not sufficient to ultimately clarify the in vivo significance of the results. Nevertheless, our data should also be placed in the context of the in situ findings previously published by our group, as well as those provided in the present manuscript. In particular, we have demonstrated that those thymocytes that are retained within giant PVS do not express significant levels of VLA-5 (9). This in situ finding is in keeping with the flow cytometry data showing that mature T cells (that correspond to the large majority of intra-PVS lymphocytes) are those in which the VLA-5 defect is more severe. Moreover, we showed that Foxp3+ cells are retained within NOD giant PVS. When evaluated by flow cytometry, CD4+CD825+Foxp3+ NOD thymocytes exhibited a large decrease in VLA-5 expression. Overall, these findings fit the data showing that in the periphery, there is a decrease in regulatory T cell numbers and that such a defect is related to the appearance of diabetes in this animal (26, 27, 28).

In conclusion, we demonstrate the existence of multiple cell migration-related abnormalities in NOD thymocytes, comprising both down- and up-regulation of specific responses, and unraveling the complexity of cell migration in the NOD mouse thymus. Although additional experiments are obviously necessary, our data indicate that the VLA-5 defect in NOD thymocytes may be of pathophysiological significance in terms of the type 1 diabetes seen in these animals.

We thank Dr. Thereza Christina Barja-Fidalgo (University of Rio de Janeiro State) for providing the tEnd-1 cell line, Dr. Luiza M. Araujo (Necker Hospital) for the anti-Foxp3 Ab, Meriem Garfa (René Descartes University Medical School, Paris, France) for confocal analysis assistance, and Dr. Ingo Riederer (Oswaldo Cruz Foundation, Rio de Janeiro) for helping with confocal image computer-based analysis. This work was developed in the context of the Centre National de la Recherche Scientifique-Fiocruz International Associated Laboratory of Immunology and Immunopathology.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by Fiocruz, Capes, Conselho Nacional de Pesquisas (Brazil) and Centre National de la Recherche Scientifique (France). A.C.K. was a recipient of a postdoctoral fellowship from Conselho Nacional de Pesquisas (Brazil).

3

Abbreviations used in this paper: ECM, extracellular matrix; PVS, perivascular space; TEC, thymic epithelial cell; MFI, mean fluorescence intensity.

1
Bach, J. F..
1994
. Insulin-dependent diabetes mellitus as an autoimmune disease.
Endocr. Rev.
15
:
516
-542.
2
Mandrup-Poulsen, T..
2003
. β cell death and protection.
Ann. NY Acad. Sci.
1005
:
32
-42.
3
Kwon, H., H. S. Jun, Y. Yang, C. Mora, S. Mariathasan, P. S. Ohashi, R. A. Flavell, J. W. Yoon.
2005
. Development of autoreactive diabetogenic T cells in the thymus of NOD mice.
J. Autoimmun.
24
:
11
-23.
4
Dardenne, M., F. Lepault, A. Bendelac, J. F. Bach.
1989
. Acceleration of the onset of diabetes in NOD mice by thymectomy at weaning.
Eur. J. Immunol.
19
:
889
-895.
5
Savino, W., C. Boitard, J. F. Bach, M. Dardenne.
1991
. Studies on the thymus in nonobese diabetic mouse: I. Changes in the microenvironmental compartments.
Lab. Invest.
64
:
405
-417.
6
Atlan-Gepner, C., M. Naspetti, R. Valero, M. Barad, F. Lepault, B. Vialettes, P. Naquet.
1999
. Disorganization of thymic medulla precedes evolution towards diabetes in female NOD mice.
Autoimmunity
31
:
249
-260.
7
Rosmalen, J. G., W. van Ewijk, P. J. Leenen.
2002
. T-cell education in autoimmune diabetes: teachers and students.
Trends Immunol.
23
:
40
-46.
8
Savino, W., C. Carnaud, J. J. Luan, J. F. Bach, M. Dardenne.
1993
. Characterization of the extracellular matrix-containing giant perivascular spaces in the NOD mouse thymus.
Diabetes
42
:
134
-140.
9
Cotta-de-Almeida, V., D. M. Villa-Verde, F. Lepault, J. M. Pleau, M. Dardenne, W. Savino.
2004
. Impaired migration of NOD mouse thymocytes: a fibronectin receptor-related defect.
Eur. J. Immunol.
34
:
1578
-1587.
10
Geutskens, S. B., D. A. Mendes-da-Cruz, M. Dardenne, W. Savino.
2004
. Fibronectin receptor defects in NOD mouse leucocytes: possible consequences for type 1 diabetes.
Scand. J. Immunol.
60
:
30
-38.
11
Campbell, J. J., J. Pan, E. C. Butcher.
1999
. Cutting edge: developmental switches in chemokine responses during T cell maturation.
J. Immunol.
163
:
2353
-2357.
12
Poznansky, M. C., I. T. Olszak, R. H. Evans, Z. Wang, R. B. Foxall, D. P. Olson, K. Weibrecht, A. D. Luster, D. T. Scadden.
2002
. Thymocyte emigration is mediated by active movement away from stroma-derived factors.
J. Clin. Invest.
109
:
1101
-1110.
13
Vianello, F., P. Kraft, Y. T. Mok, W. K. Hart, N. White, M. C. Poznansky.
2005
. A CXCR4-dependent chemorepellent signal contributes to the emigration of mature single-positive CD4 cells from the fetal thymus.
J. Immunol.
175
:
5115
-5125.
14
Pelletier, A. J., L. J. van der Laan, P. Hildbrand, M. A. Siani, D. A. Thompson, P. E. Dawson, B. E. Torbett, D. R. Salomon.
2000
. Presentation of chemokine SDF-1 α by fibronectin mediates directed migration of T cells.
Blood
96
:
2682
-2690.
15
Yanagawa, Y., K. Iwabuchi, K. Onoe.
2001
. Enhancement of stromal cell-derived factor-1α-induced chemotaxis for CD4/8 double-positive thymocytes by fibronectin and laminin in mice.
Immunology
104
:
43
-49.
16
Smaniotto, S., V. de Mello-Coelho, D. M. Villa-Verde, J. M. Pleau, M. C. Postel-Vinay, M. Dardenne, W. Savino.
2005
. Growth hormone modulates thymocyte development in vivo through a combined action of laminin and CXC chemokine ligand 12.
Endocrinology
146
:
3005
-3017.
17
Mendes-da-Cruz, D. A., J. S. Silva, V. Cotta-de-Almeida, W. Savino.
2006
. Altered thymocyte migration during experimental acute Trypanosoma cruzi infection: combined role of fibronectin and the chemokines CXCL12 and CCL4.
Eur. J. Immunol.
36
:
1486
-1493.
18
Johnson-Leger, C. A., M. Aurrand-Lions, N. Beltraminelli, N. Fasel, B. A. Imhof.
2002
. Junctional adhesion molecule-2 (JAM-2) promotes lymphocyte transendothelial migration.
Blood
100
:
2479
-2486.
19
D’Amico, G., G. Bianchi, S. Bernasconi, L. Bersani, L. Piemonti, S. Sozzani, A. Mantovani, P. Allavena.
1998
. Adhesion, transendothelial migration, and reverse transmigration of in vitro cultured dendritic cells.
Blood
92
:
207
-214.
20
Ziegler, S. F..
2006
. FOXP3: of mice and men.
Annu. Rev. Immunol.
24
:
209
-226.
21
Savino, W., D. A. Mendes-da-Cruz, J. S. Silva, M. Dardenne, V. Cotta-de-Almeida.
2002
. Intrathymic T-cell migration: a combinatorial interplay of extracellular matrix and chemokines?.
Trends Immunol.
23
:
305
-313.
22
Savino, W., D. A. Mendes-Da-Cruz, S. Smaniotto, E. Silva-Monteiro, D. M. Villa-Verde.
2004
. Molecular mechanisms governing thymocyte migration: combined role of chemokines and extracellular matrix.
J. Leukocyte Biol.
75
:
951
-961.
23
Colomb, E., W. Savino, L. Wicker, L. Peterson, M. Dardenne, C. Carnaud.
1996
. Genetic control of giant perivascular space formation in the thymus of NOD mice.
Diabetes
45
:
1535
-1540.
24
Sato, T., S. Ishikawa, K. Akadegawa, T. Ito, H. Yurino, M. Kitabatake, H. Yoneyama, K. Matsushima.
2004
. Aberrant B1 cell migration into the thymus results in activation of CD4 T cells through its potent antigen-presenting activity in the development of murine lupus.
Eur. J. Immunol.
34
:
3346
-3358.
25
Flores, K. G., J. Li, G. D. Sempowski, B. F. Haynes, L. P. Hale.
1999
. Analysis of the human thymic perivascular space during aging.
J. Clin. Invest.
104
:
1031
-1039.
26
Chatenoud, L., J. F. Bach.
2005
. Regulatory T cells in the control of autoimmune diabetes: the case of the NOD mouse.
Int. Rev. Immunol.
24
:
247
-267.
27
Piccirillo, C. A., M. Tritt, E. Sgouroudis, A. Albanese, M. Pyzik, V. Hay.
2005
. Control of type 1 autoimmune diabetes by naturally occurring CD4+CD25+ regulatory T lymphocytes in neonatal NOD mice.
Ann. NY Acad. Sci.
1051
:
72
-87.
28
Bluestone, J. A., Q. Tang.
2004
. Therapeutic vaccination using CD4+CD25+ antigen-specific regulatory T cells.
Proc. Natl. Acad. Sci. USA
101
: (Suppl. 2):
14622
-14626.