Stromal-Derived Factor 1 Expression in the Human Thymus

Immunohistochemical analysis

The following Abs were used throughout the study: goat polyclonal Ab specific to human SDF-1 (Santa Cruz Biotechnology, Santa Cruz, CA), rabbit polyclonal anti-S100 Abs (DAKO, Carpinteria, CA), mouse mAbs specific to cytokeratins A1/A3 and CD68 Ag (DAKO), biotinylated swine anti-rabbit Abs (F(ab′)₂; DAKO), biotinylated horse Abs specific to goat IgG, and biotinylated horse Abs specific to mouse IgG (Vector Laboratories, Burlingame, CA).

Fresh thymus fragments were obtained during cardiac surgery on children (aged 1 mo to 3 yr) with congenital valvular malformations. The research studies were approved by the institutional review board of the National Institutes of Health. Fresh thymic tissues were immersed in OCT compound (Stephens Scientific, Riverdale, NJ) and rapidly frozen in liquid nitrogen, and 6-μm sections were cut on a cryostat. In parallel, thymic tissues were fixed in 10% formalin (Sigma, St. Louis, MO) for 4 h, washed twice with PBS, and immersed in 70% ethanol. Serial sections of the formalin-fixed paraffin-embedded thymic tissue blocks were prepared by American HistoLabs (Gaithersburg, MD). Immunohistochemical staining was performed on cryostat sections fixed in acetone for 7 min or on paraffin-embedded sections after deparaffinization and rehydration. Sections were exposed to 3% hydrogen peroxide solution (Sigma) to neutralize endogenous peroxidases. After incubation with the blocking buffer (2% normal serum from the source of the secondary Ab, 1% BSA (Vector), and 2% Tween 20) for 45–60 min at room temperature, sections were treated with Abs against S100, A1/A3, CD68, or SDF-1 for the indicated times, followed by treatment with biotinylated anti-rabbit, anti-mouse, or anti-goat Abs for 1 h. Subsequently, sections were incubated using the Vectastain ABC kit or the Vectastain-ABC-AP kit, and color was developed using the following substrate solutions: diaminobenzidine (DAB), Vector VIP, or Vector Red according to the manufacturer’s instructions (Vector). Sections were counterstained with hematoxylin (Vector) and mounted with Permount (Fisher Scientific, Fairlawn, NJ).

In some experiments, anti-SDF-1 Abs (2 μg/ml) were incubated with 20 μg/ml of either SDF-1 or RANTES (PeproTech, Rocky Hill, NJ) overnight at 4°C on a rocking platform. Anti-SDF-1 Abs adsorbed with chemokines were then used in immunohistochemistry on cryostat sections.

Detection of apoptosis

Apoptosis was detected by Molecular Histology Laboratory (Montgomery Village, MD). In brief, paraffin-embedded slides were digested in proteinase K, exposed to 3% hydrogen peroxide solution, and subjected to TdT end labeling, followed by development with avidin-conjugated peroxidase and New Fuchsin substrate (red chromogen). Slides were counterstaining with hematoxylin.

Preparation of DC

DC were cultured from adult plastic-adherent PBMC as previously described (18), but with minor modifications. Briefly, PBMC from healthy blood donors were resuspended in RPMI 1640 (Life Technologies, Grand Island, NY) supplemented with 10% (v/v) heat-inactivated FBS (Biofluids, Rockville, MD), 100 U/ml penicillin (Life Technologies), 100 μg/ml streptomycin (Life Technologies), 2 mM l-glutamine (Life Technologies), 10 mM HEPES (Life Technologies), and 5 × 10⁻⁵ M 2-ME (Sigma; complete medium) at 5–8 × 10⁶ cells/ml and placed into 100- × 20-mm tissue culture plates (BD Biosciences, Lincoln Park, NJ) for 2 h at 37°C. Nonadherent cells were drawn off, fresh complete media were returned to culture wells supplemented with 1000 U/ml recombinant human GM-CSF (Immunex, Seattle, WA) and 1000 U/ml rhIL-4 (R&D Systems, Minneapolis, MN), and then the cells were placed in a humidified 5% CO₂ environment at 37°C. Half the total volume of medium with fresh complete medium and cytokines was replaced every other day. On day 7, DC were harvested and washed, and contaminating T cells, monocytes, NK cells, and B cells were removed from DC by immunomagnetic bead separation as previously described (18). The purity of DC was always >95%.

Induction of apoptosis in thymocytes

Thymocyte purification was performed as described previously (16). To induce apoptosis, thymocytes were first labeled with membrane linker PKH26 Red (Sigma) and then cultured in RPMI 1640 medium (Sigma) with 10% FCS (Sigma) and dexamethasone solution (Sigma) at a final concentration of 20 μM. Cells were harvested after 20 h of culture and analyzed for levels of apoptosis using the TACS annexin V-FITC kit (Trevigen, Gaithersburg, MD) according to the manufacturer’s instructions. Flow cytometry was performed using FACScan and CellQuest software (BD Biosciences).

Phagocytosis of apoptotic thymocytes

Phagocytosis of apoptotic thymocytes by immature DC was examined using an immunofluorescence microscope and by flow cytometry. For both assays, immature DC were labeled with PKH67 Green (Sigma) according to manufacturer’s instructions and mixed with PKH26-labeled untreated thymocytes or PKH26-labeled thymocytes pretreated with dexamethasone overnight (apoptotic) at a 1:3 ratio. After 4 h of coculture, an aliquot of 20,000 labeled cells was aspirated, cytocentrifuged onto glass slides using Cytospin 2 centrifuge (Shandon, Pittsburgh, PA), air-dried for 1 h, and then examined using an immunofluorescence microscope. Digital images of the cell monolayers containing phagocytic cells were generated using FITC and rhodamine filters. In parallel, after 24 h of cell culture cells were harvested, and flow cytometry was performed to determine the level of phagocytosis by enumerating the amount of double-positive DC.

Evaluation of CXCR4 expression and CXCR4 function on DC

Immature DC were cultured alone, in the presence of either untreated thymocytes or apoptotic thymocytes at a 1:3 ratio, or in the presence of soluble trimeric recombinant human CD40 ligand (CD40L; 1 μg/ml; Immunex) in medium supplemented with GM-CSF and IL-4. After 24-h culture, cells were harvested, and CXCR4 expression and function were evaluated by flow cytometry and chemotaxis assay, respectively. For CXCR4 surface staining, cell suspensions were incubated with PE-conjugated 12G5 Abs (BD PharMingen, San Diego, CA). During FACS analysis, gates were set up on DC using forward and side scatter parameters that allowed determination of CXCR4 expression on DC.

For the chemotaxis assay, DCs were counted using a hemocytometer, and 200,000 DC in 25 μl serum-free RPMI 1640 were loaded into the upper compartment of a microchemotaxis chamber (Neuroprobe, Cabin John, MD). Thirty-one microliters containing 10 nM SDF-1 (PeproTech) were added to the lower compartment. A polyvinylpyrollidone-free polycarbonate filter with 5-μm pores separated the two compartments. The chemotaxis chamber was incubated for 3 h at 37°C with 100% humidity and 5% CO₂. The filter was then removed, and the numbers of DCs migrating into each bottom compartment were counted from the lower chambers. In preliminary experiments we determined that DCs were not attached to the underside of the filter at the end of incubation period. All experimental groups were set up in triplicate. In addition, preliminary chemotaxis experiments with a range of SDF-1 concentrations from 1 nM to 1 μM demonstrated that 10 nM SDF-1 induced the largest number of migrating DC. Therefore, only this concentration was used in additional experiments.

After the filters of the chemotaxis chambers were removed, 20 μl of the suspensions from the lower compartments of 10 identical microchemotaxis chambers were pooled, diluted 1/1 with PBS, and subjected to flow cytometry to determine the numbers of migrated DC. The acquisition parameter of the FACScan was set up for 100 s, and the number of events collected within the DC gates was counted.

Infection of SCID-hu mice

SCID-hu mice were constructed by implanting second-trimester fetal tissues under both kidney capsules and were infected with macrophage-tropic, laboratory-adapted viruses (SF-162, JR-FL) or with primary, dual-tropic virus isolate 3284 by i.p. injection of 8000 50% tissue culture infective doses in a volume of 0.8 ml (19, 20). These SCID-hu mice were sacrificed by lethal ether inhalation 3 mo after infection. Thymic implants were removed, and sections were prepared from formalin-fixed paraffin-embedded thymic tissue blocks. The consent forms and procedures used in this study were reviewed and approved by the Albert Einstein College of Medicine committee on clinical investigation. All tissue was obtained from HIV-1-negative fetuses.

Location of SDF-1 protein expression in human thymic tissue

SDF-1 mRNA was shown by us and others to be expressed within the human thymus (10, 16). To decipher the role of SDF-1 in thymopoiesis, the location of SDF-1-expressing cells within the thymus was studied on cryostat sections of normal human thymic tissue. High levels of SDF-1 expression were detected in cells forming the outer walls of Hassall’s corpuscles (HC) in the medulla (Fig. 1⇓, a and b). In addition, a low level of SDF-1 staining was detected in the subcapsular area of the thymic cortex (data not shown). The specificity of staining was demonstrated by the absence of signal in the presence of normal goat IgG (Fig. 1⇓c) and the complete blocking of staining with recombinant SDF-1, but not with RANTES (Fig. 1⇓, d and e). To determine the identity of SDF-1-expressing cells in the medulla, serial sections from paraffin-embedded thymic blocks were stained with anti-SDF-1 and anti-cytokeratin A1/A3 Abs. In the paraffin-embedded sections, SDF-1-expressing cells were also detected in the outer walls of the HC, although with somewhat lesser staining intensity compared with cryostat sections (Fig. 1⇓f). Staining of serial sections with Abs against cytokeratins demonstrated that the SDF-1-positive cells in the thymic medulla were of epithelial origin (Fig. 1⇓g). A role of HC in the removal of dying cells from the thymus was previously suggested (21). To obtain direct evidence of this role of HC, sections derived from normal human thymus were subjected to in situ apoptosis detection (TUNEL assay) (Fig. 1⇓h). Strong positive staining was detected in the HC in the presence (Fig. 1⇓H), but not in the absence (Fig. 1⇓i), of TdT enzyme.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

In situ detection of SDF-1 production using cryostat sections and paraffin-embedded sections of the human thymus. Cryostat sections from normal human thymuses were stained with goat anti-SDF-1 Abs at 2 μg/ml (a, ×10; b, ×40) or with normal goat IgG (c, ×10) for 1 h, followed by biotinylated horse anti-goat Abs at 7.5 μg/ml. Slides were developed with the ABC-AP kit and Vector Red substrate. Stained cells are present in the outer walls of the HC in the medulla. Anti-SDF-1 Ab (2 μg/ml) were adsorbed with recombinant SDF-1 (d) or recombinant RANTES (e) at 20 μg/ml and were used for staining of cryostat sections as in a. Serial slides prepared from paraffin-embedded thymic blocks were incubated with anti-SDF-1 Abs at 5 μg/ml overnight, followed by biotinylated horse anti-goat Abs at 30 μg/ml (f, ×63) or with anti-A1/A3 Abs at 5 μg/ml for 1 h, followed by biotinylated swine anti-rabbit Abs at 5 μg/ml (g, ×63). Sections were developed with the ABC kit, followed by VIP (f, light purple) or DAB (g, dark brown) substrates. Sections from paraffin-embedded thymus were stained for apoptosis using TdT enzyme end labeling (h) or without TdT enzyme in the negative control (i) and were developed with New Fuchsin substrate (red; ×63).

Immunolocalization of DC in human thymus to HC

CXCR4, the only known receptor for SDF-1 (22), has been found to be expressed by many cell types that colonize the thymus, including thymocytes, macrophages, and mature DC. Mature thymocytes are localized in the thymic medulla; they express very low levels of CXCR4 (23, 24) and do not respond to SDF-1-mediated signaling (16). Macrophages express various levels of CXCR4 and do not migrate in response to SDF-1 (25) (data not shown). DCs were shown to be present at both the cortex-medulla junction and within the thymic medulla (26). In addition, mature DCs were shown to express high levels of CXCR4 and to migrate in response to SDF-1 (27). Therefore, we studied the distribution of DC within the thymic medulla in relation to SDF-1 staining. Immunohistochemistry studies using Abs against the DC marker S100 demonstrated that DC were located at the cortex-medulla junction and were scattered throughout the medulla region (Fig. 2⇓, a and c). The S100⁺ cells in thymic tissues exhibited morphology of DC, as confirmed by high power microscopy (Fig. 2⇓c, inset). No staining was detected in the presence of normal rabbit IgG (Fig. 2⇓b). Similar results were obtained using Abs against another DC marker, CD83, which is expressed intracellularly by immature and mature DC (28) (Fig. 2⇓d). In addition to S100 and CD83 markers for DCs, staining for the CD68 marker, common for macrophages and immature DC, was performed (29). CD68⁺ cells were scattered throughout the medullary and, to a lesser extent, cortical regions of the thymus and were also detected within HC (Fig. 2⇓e). Double immunostaining performed using anti-SDF-1 and anti-S100 Abs demonstrated that S100⁺ DC were found in locations overlapping that of the SDF-1⁺ epithelial cells in the thymic medulla or within the circle of SDF-1⁺ cells forming HC (Fig. 2⇓, e and f). Together, these immunohistochemistry studies demonstrated that both mature and immature DC could be found in situ in the human thymic medulla, and a fraction of these DC are located within close proximity to HC.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Close association of DC with SDF-1-expressing cells in the medullar region of human thymus. Serial slides were prepared from paraffin-embedded thymic tissues and were incubated with 2 μg/ml rabbit anti-S100 Ab (a, ×10; c, ×20), or normal rabbit IgG (b, ×10) or with mouse anti-CD83 mAb at 20 μg/ml (d, ×20) for 1 h. The slides were then incubated with biotinylated swine anti-rabbit IgG Ab (F(ab′)₂) at 2.25 μg/ml (a–c) or biotinylated horse anti-mouse IgG Ab at 5 μg/ml (d) for 1 h. Slides were developed with an ABC kit and DAB substrate. Inset, right upper corner, High power view (×100) of DCs stained with anti-S100 Ab (area located within black rectangle of c). Staining of thymic tissue was performed with mouse anti-CD68 mAb at 10 μg/ml, followed by biotinylated horse anti-mouse IgG Ab (F(ab′)₂) at 5 μg/ml (e, ×20). Double staining for SDF-1 and DC was performed on slides from paraffin-embedded normal thymic tissue (f, ×63). Slides were stained with anti-S100 Abs and developed with DAB as in a. The same slides were then incubated with the blocking buffer for 1 h and stained with anti-SDF-1 Ab at 5 μg/ml overnight, followed by biotinylated horse anti-goat Abs at 30 μg/ml for 1 h. Slides were developed with an ABC kit, followed by Vector VIP substrate. Note brown cells (DC) either attached to or enclosed by light purple cells (SDF-1-positive epithelial cells).

Phagocytosis of apoptotic thymocytes by immature DC

Immature DC generated from peripheral blood have been shown to engulf apoptotic cells (28, 30, 31, 32, 33, 34, 35). To determine whether immature DC can uptake apoptotic human thymocytes, thymocytes were labeled with PKH26 Red (28), treated with dexamethasone to induce apoptosis, and then cocultured with monocyte-derived immature DC. Initially, the number of apoptotic thymocytes was evaluated using FACS analysis by enumerating the number of cells double positive for FITC-annexin V and for PKH26 Red. Apoptosis was detected in 2–6% of untreated and in 60–72% of dexamethasone-treated thymocytes in five separate experiments (Fig. 3⇓, a and b). Next, the PKH26 Red-labeled apoptotic (or untreated) thymocytes were mixed with PKH67 Green-labeled immature, monocyte-derived DC. After 4 h of coculture, cells were placed on glass slides via cytospin and examined by immunofluorescence microscopy using individual filters for each color (Fig. 4⇓). In Fig. 4⇓a only DC can be seen using the FITC filter, and in Fig. 4⇓b only thymocytes are visualized with the rhodamine filter. The combining of images generated with individual filters allowed detection of DC that had engulfed thymocytes. These cells appeared yellow (Fig. 4⇓c). Unphagocytosed Red-labeled apoptotic thymocytes could also be detected on the same slides (Fig. 4⇓d). For a quantitative analysis of phagocytosis, dyed DC were cultured alone or in the presence of dyed thymocytes (either untreated or dexamethasone-treated), cultured for 24 h, and subjected to FACS analysis (Fig. 5⇓). Phagocytotic uptake was determined by the presence of double-positive cells. Phagocytosis was detected in DC cultured with apoptotic, but not untreated, thymocytes or when DC were cultured alone (Fig. 5⇓).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Induction of apoptosis in human thymocytes. Total human thymocytes were dyed red with PKH26 and cultured in the absence (a) or the presence of dexamethasone (b). FACS analysis was performed after 20 h of culture. Apoptotic, PKH26-labeled thymocytes, which are positive for annexin V, appear in the upper right quadrant. Results are representative of five individual experiments.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Immature DC phagocytose apoptotic thymocytes. Immature DC were labeled with PKH67 Green and incubated with PKH26 Red-labeled apoptotic thymocytes. After 4 h of coculture, cells were adhered to glass slides. Digital images of cell cocultures were obtained using FITC (a) or rhodamine (b) filters. Images derived with individual filters (a and b) were combined, and green DCs that engulfed red apoptotic thymocytes appear in yellow in c. Clusters of red thymocytes that have not been engulfed by DCs are also shown (d). Pictures are representative of two separate experiments.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Quantitative analysis of phagocytosis. Thymocytes were labeled with PKH26 Red and were cultured either alone (untreated thymocytes) or in the presence of dexamethasone (apoptotic thymocytes) for 20 h. Immature DC were labeled with PKH67 Green and cultured alone (a) or in the presence of PKH26-labeled apoptotic thymocytes (b) or PKH26-labeled untreated thymocytes (c). After 24 h of coculture, phagocytosis was evaluated by FACS analysis, and the amount of double-positive DC was enumerated. Arrows indicate a subset of PKH26 Red thymocytes that was not engulfed by DC. Data are representative of three separate experiments.

Analysis of CXCR4 expression and function on DC cultured with apoptotic thymocytes

To determine whether SDF-1 may play a role in attracting DC, we examined whether phagocytosis of apoptotic thymocytes by immature DC induced up-regulation of CXCR4 on DC and enhanced chemotaxis in response to SDF-1. Immature DC were cultured alone or with apoptotic thymocytes or in the presence of soluble CD40L (as a positive control). After 24 h of culture, CXCR4 expression was determined by flow cytometry (Fig. 6⇓a). A low level of CXCR4 was detected on DC cultured alone (ΔMFC 83). Addition of either apoptotic thymocytes or CD40L to the DC induced an increase in CXCR4 expression on DC; ΔMFC 148 and 191,respectively (Fig. 6⇓a). To determine whether the observed increase in CXCR4 surface expression correlated with increased migration of DCs in response to SDF-1, chemotaxis assays were performed. Incubation of DC with apoptotic thymocytes induced an increase in migration of DC in response to SDF-1 from 1200 ± 50 to 3400 ± 100 (Fig. 6⇓b). A similar increase in migration of DC in response to SDF-1 was observed when DCs were cultured with irradiated thymocytes or CD40L (data not shown). Only minimal migration of DC was observed in the presence of medium (Fig. 6⇓b). The identity of the migrating cells was confirmed by removing cells from the lower chambers of the chemotaxis plates and subjecting them to flow cytometry at fixed (100 s) acquisition intervals (Fig. 6⇓c). When DC were cultured alone, 1500 DC were acquired. A 2-fold increase in the numbers of migrated DC was detected after DC were cultured with apoptotic thymocytes (Fig. 6⇓c). Incubation of DC with untreated thymocytes induced an approximately 30% increase in the number of migrated DC (Fig. 6⇓c). Thus, our data demonstrate that uptake of apoptotic thymocytes by DC induces up-regulation of CXCR4 that correlates with increased chemotaxis in response to SDF-1.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Increase in CXCR4 expression and in SDF-1-induced migration in DC after uptake of apoptotic thymocytes. a, CXCR4 expression on DC after coculture with apoptotic thymocytes. Immature DC were cultured alone (thin line), with apoptotic thymocytes (thick line), or in the presence of CD40L (dotted line). After 24 h FACS analysis was performed using PE-12G5 Ab (anti-CXCR4) or isotype control IgG (shaded histogram) and gating on DC. Data are representative of three experiments. b, DCs were cultured alone or with apoptotic thymocytes. After 24 h DC were tested for chemotaxis in presence of SDF-1α or medium. Data represent the mean ± SEM of triplicate wells. c, FACS analysis of migrated DC. DCs were cultured alone (left), with apoptotic thymocytes (middle), or with untreated thymocytes (right) for 24 h, harvested, and subjected to SDF-1-induced chemotaxis as in b. At the end of chemotaxis, membrane filters of the chemotaxis plate were removed, 20 μl of cell suspensions were aspirated from the bottom chambers of 10 identical wells, pooled, and subjected to flow cytometry. Data are representative of three experiments.

Thymuses from HIV-1-infected SCID-hu mice display enlarged HC

It was of interest to determine whether the morphology of HC is altered in human thymuses with enhanced apoptosis of thymocytes due to viral infection. To test this hypothesis, SCID-hu mice were constructed by implanting human fetal thymic and liver tissues under the kidney capsule of SCID mice, and mice were either mock-infected or infected with macrophage-tropic, laboratory-adapted HIV-1 viruses or a dual-tropic HIV-1 primary isolate (36). Three months postinfection, thymic implants were isolated and examined by immunohistochemistry. In thymuses from mock-infected animals, HC were spread throughout the medulla, and their appearance was very similar to that of cells in normal thymic tissues (Fig. 7⇓a). SDF-1 staining in SCID-hu thymuses was more intense than in thymic explants from infants (Fig. 1⇑). SDF-1 was expressed throughout the medulla and also along the walls of HC. This difference in SDF-1 staining pattern may reflect the earlier developmental stages of the SCID-hu thymuses (fetal) or the xenogenic implantation environment. Importantly, in the thymuses derived from SCID-hu mice infected with macrophage-tropic and dual-tropic HIV-1, HC were dramatically enlarged (Fig. 7⇓, b–d). Staining with anti-S100 Ab revealed an intensive network of DC in the thymic medulla from both uninfected and infected tissues (Fig. 8⇓). DCs were also found in close proximity to and within HC in both uninfected and HIV-1-infected thymuses (Fig. 8⇓, a–c). TUNEL assay with TdT enzyme revealed numerous apoptotic cells (red stain) in infected (Fig. 8⇓d), but not in uninfected, thymuses (data not shown). When TdT labeling was performed on serial sections of an infected thymus, in parallel with staining with S100 Abs, S100⁺ DC could be detected in the areas overlapping the apoptotic cells (Fig. 8⇓, e and f). These data suggest that in HIV-infected thymuses, DCs may be involved in the uptake of apoptotic thymocytes. Furthermore, HC enlargement most likely reflects the need to remove large numbers of dying cells within the infected thymuses due to direct or indirect effects of HIV-1 infection.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Increased size of HC in SCID-hu thymuses infected with HIV-1. SCID-hu mice were uninfected (a) or infected with macrophage-tropic, laboratory-adapted viruses (SF-162 (b) and JR-FL (c)) or with primary, dual-tropic virus isolate (3284, d). Slides were prepared from formalin-fixed, paraffin-embedded thymic tissues and stained with anti-SDF-1 Abs at 5 μg/ml overnight, followed by horse anti-goat Ab at 30 μg/ml. Slides were developed with Vectastain ABC-AP, followed by Vector Red substrate.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

Effect of HIV-1 infection on the distribution of DC within the thymic medulla. Thymic tissues were derived from SCID-hu mice either uninfected (a) or infected with HIV-1 SF-162 (b) or JR-FL (c). Slides were stained with S100 Ab and developed as described in Fig. 2⇑. S100-positive DC stained brown and located close to or within HC in infected thymuses. Slides from HIV-1-infected thymus were subjected to apoptosis detection as in Fig. 1⇑h (×20, d). Serial slides from infected thymus were tested for apoptosis (×63, e) or were stained with S100 Ab as in Fig. 2⇑ (×63, f).