Growth Hormone Modulates Thymocyte Development in Vivo through a Combined Action of Laminin and CXC Chemokine Ligand 12
http://www.100md.com
内分泌学杂志 2005年第7期
Institut National de la Santé et de la Recherche Médicale/Fiocruz Associate Laboratory of Immunology, Laboratory on Thymus Research (S.S., D.M.S.V.-V., W.S.), Department of Immunology, Oswaldo Cruz Institute, Oswaldo Cruz Foundation, Rio de Janeiro 21045-900, Brazil; Department of Morphology, Federal University of Alagoas (S.S.), Maceió 57072-970, Brazil; H?pital Necker, Unité Mixte de Recherche 8147 (S.S., J.-M.P., M.D., W.S.), and Institut National de la Santé et de la Recherche Médicale, Unité 344 (M.-C.P.-V., W.S.), 75743 Paris Cedex 15, France; and National Institute of Aging, National Institutes of Health (V.d.M.-C.), Baltimore, Maryland 21224
Address all correspondence and requests for reprints to: Dr. Wilson Savino, Laboratory on Thymus Research, Department of Immunology, Oswaldo Cruz Institute, Oswaldo Cruz Foundation, Avenue Brasil 4365, Manguinhos, 21045-900 Rio de Janeiro, Brazil. E-mail: savino@fiocruz.br.
Abstract
Previous evidence indicates that GH modulates thymic cell migration. In this study we approached this issue in vivo, studying thymocyte migration in GH transgenic animals and in normal mice treated intrathymically with GH. Extracellular matrix and chemokines are involved in thymocyte migration. In this respect, thymocyte adhesion to laminin was higher in GH-treated animals than controls, and the numbers of migrating cells in laminin-coated Transwells was higher in GH-transgenic and GH-injected mice. Additionally, CXC chemokine ligand 12 (CXCL12)-driven migration was higher in GH-Tg and GH-treated animals compared with controls. Interestingly, although CXCR4 expression on thymocytes did not change in GH-Tg mice, the CXCL12 intrathymic contents were higher. We found that CXCL12, in conjunction with laminin, would additionally enhance the migration of thymocytes previously exposed to high concentrations of GH in vivo. Lastly, there was an augmentation of recent thymic emigrants in lymph nodes from GH-Tg and GH-injected animals. In conclusion, enhanced thymocyte migration in GH transgenic mice as well as GH-injected mice results at least partially from a combined action of laminin and CXCL12. Considering that GH is presently being used as an adjuvant therapeutic agent in immunodeficiencies, including AIDS, the concepts defined herein provide important background knowledge for future GH-based immune interventions.
Introduction
THE CROSS-TALK between the neuroendocrine and immune systems is well demonstrated, with common ligands and receptors being used in both systems, allowing a physiological communication circuitry that seems to play a relevant role in homeostasis. In this context, hormones and neuropeptides are potent immunomodulators, and one of their target organs is the thymus (1). More particularly, a series of evidence indicates that GH act upon the microenvironmental and lymphoid compartments of the organ (2).
The thymus is a primary lymphoid organ in which bone marrow-derived T cell precursors undergo differentiation, ultimately leading to the migration of positively selected thymocytes to the T cell-dependent areas of peripheral lymphoid organs (3). From the entrance of T cell precursors into the thymus to the exit of mature cells from the organ, a vast body of interactions promotes the complex process of T cell differentiation (1) that occurs as cells migrate within the thymic lobules. Such a process involves sequential expression of various proteins and rearrangements of the T cell receptor (TCR) genes. Most of the immature thymocytes, including those bearing the phenotypes TCR–CD3–CD4–CD8– and TCRlowCD3lowCD4+CD8+, are cortically located, whereas mature TCRhighCD3highCD4+CD8– and TCRhighCD3highCD4– CD8+ cells (that will normally leave the organ toward peripheral lymphoid organs) are found in the medulla. Along with such a journey, developing thymocytes encounter cortical and medullary nonlymphoid microenvironments through distinct cell-cell and cell-matrix interactions (1, 4). Cell migration is thus a crucial event for intrathymic T cell differentiation, and we recently proposed that chemokines act in concert with the extracellular matrix (ECM), resulting in the migration of a given cell subset, either within the thymus or at the entrance into and/or exit from the organ (5). Interestingly, microenvironmental cells produce both types of molecules (4, 5, 6, 7).
Few studies suggest that GH modulates intrathymic T cell migration: 1) recombinant GH increases human T cell engraftment into the thymus of severe combined immune deficiency mice, an effect that seems to be mediated by adhesion molecules and ECM, because it can be abrogated with anti-?1 integrin antibodies (8); 2) in vitro GH treatment increases the adhesion of thymocytes to thymic epithelial cells (TEC; the major component of the thymic microenvironment) as well as the production of ECM proteins and expression of ECM receptors by TEC (9). In this study, we approached this issue by studying thymocyte migration in GH-transgenic (GH-Tg) mice as well as in normal mice intrathymically injected with GH.
Materials and Methods
Animals
Female and male BALB/c, C57BL/6, and GH-Tg mice were bred and maintained under specific-pathogen free conditions at Necker Hospital animal facilities. GH-Tg mice were derived from animals provided by Drs. A. Bartke (Department of Physiology, Southern Illinois University School of Medicine, Carbondale, IL) and Thomas R. Wagner (Oncology Research Institute, Greenville Hospital System, Greenville, SC). These mice were generated by introducing the bovine GH transgene linked to the metallothionein promoter (10) and bred in a C57BL/6 genetic background. High levels of GH are maintained in these GH-Tg animals (11).
In most cases, 8- to 10-wk-old animals were used, whereas some experiments using GH-Tg mice were run using different ages, from 6 wk to 7 months. The animal facilities and care followed rules precluded by the European Union ethics committee for animal research.
Chemicals
Recombinant bovine GH was provided from Dr. A. F. Parlow (Pituitary Hormones and Antisera Center, National Institute of Diabetes and Digestive and Kidney Diseases, Torrance, CA). Stock solution (1 mg/ml) was prepared in PBS and kept at –20 C until use. Cellular fibronectin (from human foreskin fibroblasts), laminin (from murine Engelbreth-Holm-Swarm sarcoma), BSA, fluorescein isothiocyanate (FITC), and aminoethylcarbazole were purchased from Sigma-Aldrich Corp. (St. Louis, Mo), whereas the anesthetic 2,2,2-tribromoethanol was purchased from Aldrich Chemical Co. (Milwaukee, WI). The chemokines CXCL12/SDF-1 and CCL25/TECK were obtained from R&D Systems, Inc. (Minneapolis, MN). Pertussis toxin (PTX) and cholera toxin (CTX) were obtained from Sigma-Aldrich Corp. (St. Louis, MO).
Antibodies
For immunohistochemistry, rabbit polyclonal antisera specific for laminin or fibronectin were obtained from Novotec (St. Martin-la-Garenne, France); rabbit antihuman cytokeratin immune serum as well as peroxidase-labeled goat antirabbit Ig serum were purchased from DakoCytomation (Carpinteria, CA). Anti-CXCL12 monoclonal antibody (mAb) was purchased from R&D Systems, Inc. Rhodamine-labeled goat antirabbit Ig and FITC-coupled goat antirat Ig were purchased from Biosys (Compiegne, France). In all cases, optimal dilutions of each primary or secondary antibody were predetermined in our laboratory.
For cytofluorometric analyses, appropriate dilutions of the following mAb were used: anti-CD3/FITC, anti-CD4/phycoerythrin (PE), anti-CD4/peridinin chlorophyll protein (PercP), anti-CD4/allophycocyanin (APC), anti-CD8/CyChrome, anti-CD8/APC, anti-CD8/PercP, anti-CD44/FITC, anti-CD25/APC, anti-CD49d/PE, anti-CD49e/PE, anti-CD49f/PE, and anti-CXCR4/PE as well as isotype-matched negative controls for each fluorochrome applied in specific antibodies, thus including unrelated rat Igs labeled with FITC, PE, APC, or PercP (all purchased from BD Pharmingen, San Diego, CA). In some blocking experiments, purified anti-CD49f was applied (10 μg/ml) to inhibit laminin-very late antigen-6 (VLA-6)-mediated interactions.
Immunohistochemistry
Thymus frozen sections were submitted to indirect immunofluorescence or immunoperoxidase assay as previously described (12). Specimens were incubated with a given primary antibody for 1 h, washed with PBS, and treated with rhodamine- or peroxidase-coupled second antibody. In the case of immunofluorescence, specimens were analyzed under a confocal fluorescence microscope (MS 510, Zeiss, Oberkochen, Germany). When peroxidase-linked second antibody was applied, enzyme activity was revealed with aminoethylcarbazole in the presence of H2O2. Controls comprised specimens in which primary antibodies were replaced by unrelated Igs and generated no significant labeling (not shown).
Cytofluorometry
Thymus, spleen, as well as sc and mesenteric lymph node cell suspensions from each animal group were prepared in PBS as previously described (13). Three- or four-color cytofluorometry was performed by incubating cells with a mixture of appropriately diluted antibodies in 2% fetal calf serum in PBS for 20 min at 4 C. After washings, cells were fixed and analyzed by flow cytometry in a FACSCalibur device (BD Pharmingen) equipped with CellQuest software (BD Biosciences, Mountain View, CA). A gate excluding cell debris and nonviable cells was determined using forward vs. side scatter parameters and was confirmed in some experiments with the use of propidium iodide staining and immediate analysis of unfixed cells. Analyses were made after recording 10,000–100,000 events for each sample.
We also phenotyped the CD4–CD8– double-negative (DN) subset with the CD44 and CD25 markers. For that, we used anti-CD4/PercP and anti-CD8/PercP in conjunction with anti-CD25/allophycocyanin and anti-CD44/FITC. In these experiments, we recorded 200,000 events at the flow cytometer.
Real-time PCR
Mouse thymuses were homogenized in 1 ml TRIzol reagent (Invitrogen Life Technologies, Inc., Carlsbad, CA) using a Polytron instrument (Brinkmann Instruments, Westbury, NY). The total RNA was then isolated according to the instructions of the manufacturer. RT of 2 μg total RNA was performed using the Stratagene RT kit (La Jolla, CA), and the resulting cDNA was submitted to real-time PCR, to measure the expression of mRNA encoding for CXCL12. Amplification of ?2-microglobulin was performed as an endogenous control. The following primers were applied: CXCL12: forward: ccagtcagcctgacctaccg; reverse, cgggtcaatgcacacttgtc; and ?2-microglobulin: forward, tgaccgcttgtatgctatccagt; reverse, cagtgtgagccaggatatag.
The PCR was performed on an ABI PRISM 7700 sequence detector using the quantitative PCR Master Mix for SYBR Green (Eurogentec, Seraing, Belgium). Amplification was carried out in a total volume of 20 μl containing 0.5 μM gene-specific primers, 10 μl Master Mix SYBR Green, and 5 μl cDNA at different dilutions (three then fold dilutions). After initial enzyme activation (5 min at 55 C) and DNA denaturation, (5 min at 95 C) the PCRs were cycled 50 times with the following parameters: denaturation at 95 C for 30 sec, annealing at 59 C for 30 sec, and extension at 72 C for 30 sec. RNA expression was determined by the comparative cycle threshold (Ct) method (14). Four independent assays were performed, each one in triplicate.
Cell adhesion assay
To assess the thymocyte adhesive capacity on ECM substrata, 60 x 15-mm culture dishes (Nunc, Copenhagen, Denmark) were coated for 1 h at room temperature (plus 30 min for air drying) with 10 μg/ml fibronectin, laminin, or BSA as a control. In these assays, 1 x 107 thymocytes from three pooled thymuses from GH-Tg, GH intrathymically treated BALB/c mice, or their corresponding controls were allowed to adhere on the precoated dishes. After 1-h incubation, the adherent thymocytes were harvested, counted, and phenotyped for the expression of CD3, CD4, and CD8 molecules. The relative adhesion was calculated as the ratio of adhesion between GH-Tg or GH-treated over the respective control, which was, by definition, fixed at 1.
Isolation and culture of thymic nurse cells (TNCs)
Migration of immature thymocytes can be investigated in vitro by studying TNCs. These are cortically located lymphoepithelial complexes formed by one TEC, which harbors various numbers of thymocytes (mostly immature cells), being located in the cortical region of thymic lobules. In culture, TNCs spontaneously release thymocytes, and TNC-derived epithelial cells can reconstitute lymphoepithelial complexes after being cocultured with fetal thymocytes, thus corresponding to an in vitro model of thymocyte migration within the TEC context (5). TNCs were isolated from pools of five to 10 C57BL/6 wild-type or GH-transgenic female mice at 11–15 wk of age, as currently done in our laboratory (9), and the number of TNC per mouse was then calculated. Once obtained, such lymphoepithelial complexes were settled in culture, so as to obtain, 5 d later, highly purified TNC-derived epithelial cultures. These cells were used for immunocytochemical detection of laminin and CXCL12 or to evaluate reconstitution of thymic nurse cell complexes. In this kind of experiment, 5-d TNC-derived TEC were trypsinized and cocultured with fresh 16- to 17-d-old fetal thymocytes (105 thymocytes/104 TEC/well) in inverted Terasaki plates for 6 h in complete RPMI 1640 medium (9). Alternatively, cocultures were performed in the presence of purified antilaminin antibodies (10 μg/ml). Control cultures were incubated with control rabbit IgG at equivalent protein concentrations. Percentages of newly formed complexes were evaluated in blind by direct counting under the light microscope.
Cell chemotaxis assay
Cell chemotaxis is another strategy to evaluate immature as well as mature thymocyte migration in vitro, under distinct stimuli (5). Thymocyte migratory activity was assessed ex vivo in 5-μm pore size Transwell plates (Corning Costar, Cambridge, MA) using cell suspensions after pooling three thymuses from each group of mice. In this cell chemotaxis assay, membranes were coated on both sides with 10 μg/ml fibronectin, laminin, or BSA for 1 h at 37 C, followed by 1 h of blocking with 1% BSA. Thymocytes (2.5 x 106) were then added to the upper chamber in 100 μl 0.5% BSA/RPMI 1640, and 600 μl 0.5% BSA/RPMI 1640 were added to the lower chamber. After 3 h of incubation at 37 C in 5% CO2-containing air, cells migrated into lower chambers were removed, counted, and analyzed by flow cytometry. In additional experiments, we tested the ability of the chemokines, CXCL12 and CCL25, applied at 100 ng/ml, alone or in combination with laminin.
GH injection and intrathymic fluorescein labeling
One direct strategy to evaluate thymocyte exit in vivo is analysis of the so-called recent thymic emigrants. It is established that intrathymic injection of FITC labels large numbers of thymocytes, allowing the recovery of FITC+ cells that have recently emigrated from the thymus (15, 16). BALB/c as well as GH-Tg and corresponding age-matched control mice were anesthetized with 2,2,2-tribromoethanol, and 10 μl FITC solution (1 mg/ml)/thymic lobe were injected into the thymus in the open chest cavity. Controls for the injection procedure consisted of dropping 10 μl FITC solution into the mediastinal cavity (over the thymus) or intrathymic injection of PBS alone. In the case of BALB/c mice, the volume of 10 μl injected per thymic lobe contained FITC (1 mg/ml) plus GH (10–5 M, final concentration). In addition, in this model an additional control consisted of injecting heat-denatured GH, applied at the same concentration as the native hormone. We previously ascertained that FITC plus GH or GH alone had similar effects, stimulating TEC proliferation, indicating that FITC does not interfere in the GH activity (data not shown). We also ascertained that TEC proliferation was not enhanced by GH denatured by boiling (2). In additional experiments in which FITC was not injected, intrathymic treatment with GH (or its denatured form) followed the same protocol as that described above.
In all animals the chest cavity was closed with surgical clips; 16 h later, thymus, spleen, and lymph node cells were collected from all mice. In additional experiments, the organs were collected 2 wk after GH injection and 16 h after FITC injection. Mice whose thymuses contained less than 50% FITC-labeled cells were discarded.
Statistical analyses
Data are shown as the mean ± SE. Results were statistically analyzed by unpaired Student’s test; data were considered statistically different at P 0.05.
Results
Thymocyte numbers in GH-Tg and GH-injected mice
Relative cell numbers (per gram of body weight) in GH-Tg mice were higher than those in age-matched controls; the same pattern was observed in both young and middle-aged animals, in which differences were statistically significant (P < 0.05). Similar findings were seen in the total numbers of thymocytes from mice injected intrathymically with GH compared with controls, either untreated or treated with PBS or heat-denatured GH (dGH). Interestingly, this hormonal effect was still observed 2 wk after a single GH injection (Fig. 1, A and B). Such an increase in absolute numbers was seen in both immature and mature thymocyte subsets (Fig. 1, C and D), although the profiles regarding the expression of CD3, CD4, and CD8 in various thymocyte subsets were similar in the various experimental groups, as shown in Fig. 1E. We also evaluated the DN compartment with CD44 and CD25 markers. No differences were seen between GH-Tg and controls for any of the four CD44/CD25-defined DN subsets (Fig. 1F).
FIG. 1. Thymocyte numbers and CD3/CD4/CD8-defined phenotypes in GH-Tg and GH-injected mice. A, Comparison of the numbers of thymocytes per gram of body weight in GH-Tg mice () vs. corresponding wild-type controls (Wt; ), in young (6–10 wk) and middle-aged (26–32 wk) mice. B, Total thymocyte numbers in BALB/c mice intrathymically injected with native GH () compared with those in controls injected with denatured GH (), either 16 h or 2 wk after a single GH injection. In both GH-Tg and GH-injected animals, the numbers of thymocytes were higher than in the corresponding controls. C and D, Such an increase in absolute cell numbers comprises both immature and mature CD4/CD8 thymocyte subsets, respectively, in GH-Tg and GH-injected animals. E, Flow cytometry-derived three-color profiles for defining CD4 vs. CD8 as well as CD3 expression in both 8-wk-old wild-type and GH-Tg animals. The percentages of positive cells are seen in each profile. With regard to CD3 labeling, total CD3+- as well as CD3high-expressing cells are shown separately. No significant difference was seen between the two groups. Similarly, no difference was seen between GH-Tg and wild-type thymus in terms of the relative number of CD25/CD44-defined subsets within the CD4–CD8– DN compartment (F). Representative relative numbers of each subset are shown in each quadrant. With regard to GH-Tg vs. C57BL/6, 28–30 animals were used per group. In BALB/c mice treated with GH for 16 h, 46–49 animals were evaluated per group, whereas in the experiments performed 14 d after a single GH intrathymic injection, five to six mice per group were analyzed. *, P < 0.05; **, P < 0.02.
GH enhances ex vivo thymocyte adhesion to and migration through laminin
Knowing that cell migration depends on the adhesion of the migrating cell to the substrate, we also checked whether the adhesion capacity of thymocytes onto ECM-defined substrata was modified by in vivo GH treatment. Experiments performed in both GH-Tg and GH-injected BALB/c mice revealed that adhesion of thymocytes to laminin (but not to fibronectin) was higher than that in the corresponding controls: C57BL/6 and dGH-injected BALB/c, respectively (Fig. 2A). Although the total number of adherent cells was changed, their CD4/CD8-defined phenotype remained similar in both GH-Tg and GH-treated and the corresponding control groups, namely, untreated C57BL/6 and dGH-treated BALB/c mice (Fig. 2B).
FIG. 2. GH enhances laminin-driven thymocyte adhesion and migration. A, Relative adhesion to laminin or fibronectin of GH-Tg thymocytes (GH-Tg; ) compared with wild-type controls (Wt; ). Relative adhesion was calculated by dividing the numbers of adhered thymocytes from the GH-Tg group by the corresponding control number (whose relative adhesion was then defined as 1). The panel depicts two independent experiments. The cytofluorometric profiles shown in B illustrate that the percentages of CD4/CD8-defined thymocyte subsets of laminin-adherent cells are similar in Wt and GH-Tg mice. Similar data were obtained when thymocytes from GH-injected BALB/c and corresponding controls were allowed to adhere to laminin or fibronectin (data not shown). C and D, Thymocyte migration driven by laminin (but not fibronectin) is enhanced in both GH-Tg (C) and GH-injected (D) mice compared with that in controls. Mice used in these experiments were 8–10 wk of age, and at least six animals were used in each group. *, P < 0.05.
Taking into account the above findings, we examined whether high GH contents in vivo could also modulate ex vivo thymocyte migration through ECM-coated Transwells. Examining the total number of migrating cells, we observed that migration through laminin was higher in GH-Tg and GH-injected mice than in the corresponding controls (Fig. 2, C and D). In contrast, when fibronectin was applied as a haptotatic stimulus, the numbers of migrating cells were similar in GH-Tg and wild-type groups. No differences between control and GH-Tg animals were seen in BSA-driven cell migration as well. In GH-injected BALB/c mice, the differences observed were the same as those described for GH transgenic animals. In both cases, GH enhanced the adhesion and migration of thymocytes, but did not change the phenotypic patterns of adhered or migrating cells (data not shown).
Does GH modulate intrathymic ECM ligands and receptors in vivo?
Because we had shown that in vitro treatment of TEC with GH enhanced ECM deposition (9), we evaluated the distribution of laminin and fibronectin in the thymus of GH-Tg mice, compared with respective controls: GH induced an increase in ECM deposition in both cortical and medullary regions of the thymic lobules, as illustrated in Fig. 3. Interestingly, in situ labeling also revealed an increase in the laminin receptor VLA-6 in the microenvironmental compartment of GH-Tg thymuses compared with controls (Fig. 4A). Nevertheless, the membrane levels of ECM receptors (VLA-4, VLA-5, and VLA-6) on thymocytes did not change (Fig. 4B and Table 1). Similar findings were seen in GH-injected BALB/c mice (data not shown).
FIG. 3. Enhancement of ECM proteins in the thymic microenvironment of GH-Tg mice. Upper panels, Laminin staining by immunofluorescence; an increase in cortical laminin-containing fibrils in GH-Tg 10-wk-old mouse thymus (A) compared with the age-matched wild-type control (B). Laminin deposition is also increased in the medullary region of GH-Tg mouse thymus (C), compared with wild-type control (D). Bottom panels, Enhancement of fibronectin in the thymus of GH-Tg mice (F) compared with control thymus from a wild-type mouse. Magnifications, x400. C, Cortex; M, medulla; S, septum. The single arrow shows the capsule, whereas the double arrow indicates a blood vessel.
FIG. 4. Enhancement of VLA-6 in the thymic microenvironment of GH-Tg mice. Upper panels, CD49f (-chain integrin of the laminin receptor VLA-6) staining by immunoperoxidase. Compared with the age-matched wild-type control (C), an increase in immunostaining is seen in GH-Tg 10-wk-old mouse thymus (B). C, By contrast, VLA-6 membrane expression is similar in GH-Tg- and wild-type-derived CD4/CD8-defined thymocyte subsets. C, Cortex; M, medulla; S, septum; BM, blood vessel.
TABLE 1. Expression of fibronectin and laminin receptors in CD4/CD8-defined thymocyte subsets from GH-Tg and wild-type C57BL/6 mice
GH modulates laminin-mediated interactions in TNC complexes
We next analyzed the TNC complex, a model for intrathymic T cell migration (5), which is targeted by in vitro GH treatment (9) and is partially dependent on laminin-mediated interactions (5). We first noticed that the numbers of freshly isolated TNCs in both GH-Tg and GH-injected mice were higher than the corresponding control values (Fig. 5A). Additionally, as ascertained in the transgenic model, the percentages of larger and more granular TNC complexes were higher in GH-Tg than wild-type counterparts (Fig. 5B). Moreover, when we established epithelial cultures derived from GH-Tg TNC complexes, the amount of laminin detected was greater than that in epithelial cultures derived from wild-type mouse thymuses (Fig. 5C). These findings prompted us to attempt to reconstitute TNC lymphoepithelial complexes using thymocytes from normal animals that were cocultured with GH-Tg or wild-type-derived TEC. This provided a model to ascertain whether the GH-Tg laminin-enriched thymic microenvironment was also playing a role in the thymocyte migration, herein exemplified by the entrance of thymic lymphocytes into a well-defined thymic epithelial niche. In fact, formation of lymphoepithelial complexes was increased when freshly isolated fetal thymocytes obtained from C57BL/6 mice were cocultured with GH-Tg-derived TEC compared with TEC obtained from TNC cultures of age-matched, wild-type animals, and this effect was partially, yet statistically significantly, abrogated by antilaminin treatment (Fig. 5, D and E).
FIG. 5. Changes in TNCs from GH-Tg mice. A, Both the relative TNC numbers in GH-Tg mice and the total TNC numbers in animals injected with GH are higher than those in corresponding controls. B, The percentage of large granular TNCs from GH-Tg mice is higher compared with that in controls. Moreover, laminin labeling in TNC-derived epithelial cells from GH-Tg is stronger than that in TEC cultures from wild-type mice (C). Finally, reconstitution of TNC lymphoepithelial complexes is higher when the epithelial cells were cocultured with normal thymocytes and were derived from GH-Tg mice (D). This effect is partially abrogated with the use of antilaminin (LN) antibody, but not with the unrelated rabbit Igs (Rab Ig), used at the same concentration (E, right). Mice were 8 wk old, and pools of at least 10 animals were used per group in each experiment depicted in A and B, which are representative of three similar independent experiments. D and E, At least three independent cultures of triplicates were used.
GH enhances CXCL12-triggered thymocyte migration: synergy with laminin
Considering that thymocyte migration is also influenced by chemokines (6, 7), and that the chemokine CXCL12 was reported as having a combined effect with laminin in modulating thymocyte migration (17), we addressed the question of whether GH could modulate intrathymic CXCL12-driven migration. In fact, when CXCL12 was applied, thymocyte migration was consistently higher in GH-Tg mice (Fig. 6A). Similar results were obtained with GH-treated BALB/c animals (not shown). Interestingly, this effect seems to be rather specific, because it was not seen when another chemokine, CCL25, which is also known to preferentially attract immature thymocytes (5, 7), was applied (Fig. 6A).
FIG. 6. Enhanced thymocyte migration induced by laminin in conjunction, or not, with CXCL12. A, Experiments with 10-wk-old GH-Tg mice () and corresponding C57BL/6 wild-type controls (), with data expressed as the total number of cells that migrated through Transwell chambers. Nonspecific thymocyte migration, seen when BSA was applied in the Transwells, was similar in all groups. Thymocyte migration induced by the chemokine CCL25 was also similar in wild-type and GH-Tg animals. In contrast, laminin (LN)-driven or CXCL12-driven migration of thymocytes was significantly enhanced in GH-Tg mice compared with respective controls. Moreover, a synergy was seen when both stimuli were applied simultaneously, and this was due to the enhancement of CD4+CD8+ cell migration (B), although absolute numbers of all CD4/CD8-defined subsets from GH-Tg mice were also significantly increased after CXCL12-driven migration, in combination, or not, with laminin. Each column represents the mean ± SE of three independent experiments. *, P < 0.05; **, P < 0.01.
Concerning CXCL12 stimulation, GH-Tg and GH-injected CD4/CD8-defined thymocyte subsets migrated significantly more than the corresponding controls.
In conjunction with these ex vivo functional data, we found increased levels of CXCL12 in GH-Tg thymic stroma (Fig. 7, A and B) and GH-Tg-derived TNC cultures compared with corresponding controls; the latter was ascertained by real-time PCR and immunocytochemistry (Fig. 7, C and D). By contrast, the membrane levels of CXCR4 in thymocytes were similar in wild-type and transgenic animals, independent of the CD4/CD8-defined subset analyzed (Fig. 7E).
FIG. 7. Enhanced CXCL12 production by thymic cells of GH-Tg mice. Upper panels, Thymus sections from wild-type (A) and GH-Tg (B) mice, depicting stronger immunohistochemical labeling for CXCL12 in the GH-Tg thymus. Also, epithelial cells from GH-Tg animals produce larger amounts of CXCL12, as revealed by quantitative real-time PCR (C); the relative increase in specific mRNA was in three dilutions, and values were obtained with wild-type-derived cells (s) normalized to 1. GH-Tg TEC produced 4- to 5-fold more CXCL12 mRNA than controls. An increase in CXCL12 was confirmed by immunocytochemistry (D), showing an increased labeling in Tg mice-derived TEC (right) compared with controls (left). E, A series of cytofluorometric profiles, showing that the membrane levels of CXCR4 are similar in GH-Tg and wild-type thymocytes regardless of the CD4/CD8-defined thymocyte subset analyzed.
We then investigated whether CXCL12, in conjunction with laminin, could also enhance migration of thymocytes previously exposed to high concentrations of GH in vivo. We noticed that in control groups the simultaneous use of both stimuli resulted in a degree of migration that was higher than the summation of each stimulus alone. Such an additive effect was also seen in GH-Tg and GH-injected mice and essentially targeted the CD4+CD8+ subpopulation (Fig. 6).
Thymocyte migration elicited by laminin, with or without CXCL12, was also investigated with the use of specific blockers, anti-VLA-6 mAb and Bordetella PTX, respectively. As expected, migration induced by laminin was completely prevented by incubation of thymocytes with anti-VLA-6 mAb, whereas CXCL12-induced migration was abrogated by PTX (Fig. 8B). Control experiments revealed that an isotype-matched mAb did not change thymocyte migration triggered by laminin (not shown), and that CTX, which is known to be inefficient in blocking chemokine receptors (18), did not exert any effect on CXCL12-induced migration of thymocytes from wild-type or GH-Tg mice. Moreover, PTX did not change the profiles of laminin-triggered thymocyte migration in either wild-type or GH-Tg animals. However, when thymocytes from GH-Tg mice were preincubated with anti-VLA-6 and then subjected to CXCL12 stimulation, we found a consistent and statistically significant 20% decrease in the numbers of migrating cells (Fig. 8A). In control animals, such a decrease was much less and was not statistically significant.
FIG. 8. Blockage of laminin receptor interferes with CXCL12-driven migration of thymocytes from GH-Tg mice. A, Preincubation of thymocytes with the anti-VLA-6 antibody completely abrogated laminin (LN)-triggered thymocyte migration (which decreased to values obtained with BSA, used as a negative control) in both wild-type and GH-Tg animals. The numbers of migrating thymocyte CXCL12 were enhanced in GH-Tg mice (compared with wild type), and such an enhancing effect was abrogated with the anti-VLA-6 antibody to levels seen in CXCL12-induced migration of control thymocytes. Such a blocking effect was also seen when laminin and CXCL12 were used simultaneously. B, CXCL12-driven migration of thymocytes from wild-type and GH-Tg mice is prevented by PTX, but not by the corresponding control, CTX. PTX, however, was not able to interfere with the LN-driven migration, independently of laminin being used alone or together with CXCL12. *, P < 0.05; **, P < 0.01.
We also evaluated the expression of CXCR4 and VLA-6 in CD4/CD8-defined subsets of those thymocytes that migrated after laminin and/or CXCL12 stimulation. VLA-6 membrane levels did not change in migrating cells, whereas a down-regulation of CXCR4 was seen on thymocytes exposed to CXCL12, in conjunction, or not, with laminin, although no differences were found between GH-Tg and wild-type mice (Fig. 9).
FIG. 9. Expression of CD49f and CXCR4 in CD4/CD8-defined thymocyte subsets before and after migration induced by laminin and/or CXCL12. Representative cytofluorometric profiles are shown in various situations, with the relative number of each VLA-6/CXCR4-defined subpopulation shown within each rectangle. Before migration, both wild-type and GH-Tg mouse thymocytes demonstrated that the simultaneous expression of CD49f (the 6-chain of the laminin receptor VLA-6) and CXCR4 is more frequent in immature thymocyte subsets (CD4–CD8– and CD4+CD8+ cells) than in mature CD4+ or CD8+ single-positive cells. Such a pattern was maintained in those cells that migrated under an unrelated BSA stimulus as well as after stimulation by laminin. By contrast, when CXCL12 was applied as chemoattractant (alone or in combination with laminin), the relative numbers of VLA-6+CXCR4+ migrating cells were diminished in both wild-type and GH-Tg mice, an effect that was seen in all CD4/CD8-defined thymocyte subsets and was still more pronounced in immature CD4–CD8– and CD4+CD8+ cells. Similar results were seen in four animals from each experimental group.
GH modulates thymocyte export to peripheral lymphoid organs
In a last series of experiments we tested whether GH could modulate the exit of thymocytes in vivo. The analysis of recent thymic emigrants (RTEs) in GH-Tg and GH-injected animals revealed that compared with controls, absolute RTE numbers (i.e. FITC+ cells) in both GH-Tg and GH-injected animals were augmented in mesenteric and, to a lesser extent, in sc lymph nodes, but not in spleen. This augmentation was essentially due to the differences in the CD4+ subset, as seen by the CD4/CD8 ratios in wild-type compared with GH-Tg mice (Table 2). In contrast, no changes were observed in the rare CD4+CD8+ cells seen in peripheral lymphoid organs. These findings suggest that high levels of GH differentially modulate the relative distribution of these cells, at least with regard to the peripheral lymphoid organs evaluated in this study.
TABLE 2. Modulation of homing of thymic-derived CD4+ T lymphocytes in GH-Tg mice
Discussion
Previous studies revealed that distinct intrathymic cellular interactions can be influenced by GH in vitro, as exemplified by the secretion of thymulin, a TEC-derived thymic hormone, TEC proliferation, as well as the expression of ECM ligands and receptors with consequent modulation of ECM-mediated TEC/thymocyte interactions (1). By contrast, studies using knockout mice for the GH receptor revealed that this hormone is not essential for thymus development (19). Yet, in vivo addition of GH does appear to be effective; systemic GH injections in aging rats increased the total thymocyte number and the percentage of CD3-bearing cells (20, 21) as well as concanavalin A-induced mitogenic response and IL-6 production (22). In keeping with this finding, AIDS patients who received adjuvant GH therapy exhibited thymus growth, as demonstrated by computer tomography (23). Moreover, recombinant GH stimulated human peripheral blood lymphocyte engraftment and migration of T cells into the thymus of severe combined immune deficiency mice (8).
In this study we evaluated in vivo the effects of high levels of GH in the thymus using two distinct approaches: intrathymic injection of GH in BALB/c mice and evaluation of GH-Tg mice. Accordingly, the first protocol allowed us to evaluate the acute effect of the hormone within the organ; in the second, we could study the effects due to prolonged exposition of the thymus to high circulating levels of GH. Interestingly, in both conditions the numbers of thymocytes were slightly (yet significantly) higher than the corresponding control value. This is in keeping with preliminary data obtained with GH-injected mice, which presented 10–15% more cells in the S and G2M phases compared with denatured GH-injected animals, whereas no differences were seen in terms of membrane Fas expression (Smaniotto, S., M. Dardenne, and W. Savino, unpublished observations).
We focused herein on those GH-triggered effects related to thymocyte development. ECM ligands and receptors correspond to one group of players in this process (4). We observed that GH induced an increase in laminin deposition throughout the thymic lobules. This is in keeping with our in vitro data, which showed that treatment of cultured TEC with GH enhanced ECM deposition (9). Additionally, in both GH-Tg and GH-injected mouse thymus sections, we found that the VLA-6 labeling in the microenvironmental compartment was denser than that in the respective controls. This may result in a stronger attachment of laminin to the microenvironmental network, thus favoring its presentation to thymocytes. Such an idea is also supported by the fact that TNC-derived epithelial cells from GH-Tg mice apparently produce more laminin than those from wild-type animals, and that the degree of reconstitution of lymphoepithelial complexes, formed after coculturing GH-Tg-derived TEC with normal thymocytes, was higher than that seen when TEC from wild-type animals were used.
In contrast, however, membrane expression of VLA-6 on thymocytes did not change significantly. Yet, with regard to the laminin receptor, ex vivo binding to laminin was higher when thymocytes from GH-treated mice were applied, suggesting that high GH contents also activate VLA-6 in thymocytes, favoring their migrating progress. Furthermore, such an effect was found in both immature and mature CD4/CD8-defined thymocyte subsets. In keeping with this, thymocyte migration in laminin-containing Transwell chambers was higher in GH-Tg and GH-treated mice compared with controls. Overall, these data tell us that GH promotes a multifaceted change in laminin-mediated interactions; one consequence being an increased migratory response of thymocytes, which probably means a change in the developmental behavior. At present, we cannot explain why the expression of VLA-6 is enhanced in the microenvironment, but not in thymocytes. Yet it is conceivable that the mechanisms controlling the expression of this molecule in thymocytes and TEC are different. This hypothesis is actually supported by our recent data showing a defect in the expression of the fibronectin receptor VLA-5 on nonobese diabetic thymocytes, which is not found in the TEC of the same animals (24).
A nonmutually excluding hypothesis we raised is that chemokine-driven thymocyte migration could also be under the influence of GH. In fact, the thymocyte migratory response triggered by CXCL12 was higher in GH-Tg and GH-treated mice than in controls, although the membrane levels of CXCR4 remained similar in Tg and wild-type animals, suggesting that (similar to what was seen in relation to VLA-6) CXCR4 in GH-Tg thymocytes is spontaneously more activated. Moreover, the enhancement of CXCL12 seen in the thymus and TNC cultures derived from GH-Tg mice, ascertained at both mRNA and protein levels, suggest that in vivo, both microenvironmental and thymocyte sides contribute to the resulting enhancement of CXCL12-mediated interactions revealed in the GH-Tg mouse thymus.
Interestingly, it was shown that in vivo injection of CXCL12 in bovine GH-Tg mice inhibited ex vivo T cell migratory activity of peritoneal cells (25), suggesting that the GH-enhancing, CXCL12-triggered migration shown in this study might be restricted to early T cell developmental stages. In any case, the present data represent, to our knowledge, the first demonstration that chemokine-driven thymocyte migration can be hormonally controlled. This is a relevant issue when designing procedures involving GH-based therapy in both experimental animals and humans.
An additional aspect that we analyzed was the possibility that GH could enhance the combined effects of laminin plus CXCL12. When both molecules were applied, the number of migrating cells was higher than the sum of the two stimuli alone, an effect seen in control, GH-Tg, and wild-type animals and that essentially targeted the CD4+CD8+ cells. The fact that such a synergic effect was seen in both wild-type and GH-Tg animals is evidence that GH enhances a biological circuitry that already exists in the physiological situation. These data fit with the idea that ECM and chemokines may act in combination to drive thymocyte migration (5) and reinforce the idea of a cross-talk between integrin and chemokine receptors, a concept that has been established for CXCL12 and laminin or fibronectin (26, 27, 28). Accordingly it is conceivable that CXCL12/CXCR4-induced signaling also uses the Janus kinase/signal transducer and activator of transcription pathway, similar to integrins of the VLA family. In this respect, it is noteworthy that an antilaminin receptor mAb could significantly block approximately 20% of the migratory activity elicited by CXCL12 in GH-Tg thymocytes. However, because we did not detect a significant reciprocal blocking activity of PTX on laminin-driven thymocyte migration even in GH-Tg animals, it is conceivable that the relative influence of VLA-6 on the CXCR4 signaling pathway is more relevant than the reverse situation. Accordingly, disruption of VLA-6-triggered intracellular signaling, secondary to the binding of the anti-VLA-6 mAb, could somewhat alter the Janus kinase/signal transducer and activator of transcription pathway, which is also partially triggered by CXCR4 stimulation, as recently demonstrated (18). Although additional studies of this issue are obviously necessary, our results place GH-Tg derived thymocytes as an interesting model to study putative CXCR4/VLA-6 functional connections.
The fact that ex vivo migration of mature thymocytes was enhanced in high GH content conditions prompted us to check whether the hormone could also modulate the exit of thymocytes. We found in both GH-Tg and GH-injected animals an augmentation of CD4+FITC+ cells in mesenteric and sc lymph nodes (but not in the spleen), revealing that GH modulates the relative distribution of these cells among peripheral lymphoid organs. The mechanisms governing such differential RTE homing in GH-Tg mice have not been explored. Yet it is likely that CD62L (a well-known homing receptor for lymph nodes) is involved, because in BALB/c mice intrathymically injected with bovine GH, we found an up-regulation of this molecule in lymph node RTEs (29).
The biological circuit(s) triggered by high GH contents has not been addressed in this study. Nevertheless, an IGF-I/IGF-I receptor loop is likely to be involved. GH-Tg mice have high levels of circulating IGF-I (30). Moreover, we defined an intrathymic GH-controlled IGF-I/IGF-I receptor circuitry in human and mouse TEC and thymocytes (31). Lastly, we showed that in vitro GH-enhanced TEC/thymocyte interactions relevant to thymocyte migration could be abrogated not only by anti-ECM and anti-ECM receptor antibodies, but also by anti-IGF-I and anti-IGF-I receptor reagents (9).
Another issue to be discussed refers to whether the effects reported herein are induced via paracrine or endocrine pathways. This is particularly interesting considering that the results seen in GH-Tg mice were similar to those seen in normal animals injected intrathymically with GH. Although this issue has not been approached in a precise way, it is likely that both pathways occur; in vitro GH treatment of thymic cells does promote a multifaceted biological response, comprising, among other effects, the modulation of ECM proteins (9). Additionally, systemic injection of GH into mice enhances thymic hormone production by TEC (22).
Taken together, our data demonstrate that GH-Tg mice as well as GH-injected animals present altered thymocyte migration, with changes in the distribution of exiting thymocytes in the periphery of the immune system. Moreover, such a GH-related thymocyte migration results at least partially from a combined action of laminin and CXCL12. It is not known whether the opposite situation occurs in mouse GH receptor knockout mice, yet such a hypothesis seems plausible considering that these animals do exhibit alterations in the thymic epithelium, as revealed by increased density of the TEC network, the appearance of large epithelial cysts, and a decrease in thymic hormone secretion (32). Nevertheless, one should take into account that these animals as well as IGF-I-null mice (33) do develop negative and positive selection, with rather normal relative numbers of CD4/CD8-defined thymocyte subsets. This fits with the hypothesis recently postulated that the main role of GH and IGF-I in the immune system is a general antistress effect (34). In this context, and if we take into account that the thymus is one of the most stress-sensitive organs in the body, promoting massive thymocyte death as one of the main responses to acute stress, a putative antistress role of GH-IGF-I-related circuits would be crucial for the homeostasis in this organ, with likely consequences for the maintenance of peripheral T cells. In this context, and considering that GH is presently being used as an adjuvant therapeutic agent in immunodeficiencies, including AIDS (23), the concepts defined herein provide important background knowledge for future GH-based immune interventions.
Acknowledgments
We thank Dr. Tsvee Lapidot for reading the manuscript, and Veronique Alves and Jo?o Hermínio Silva for technical assistance.
References
Savino W, Dardenne M 2000 Neuroendocrine control of the thymus. Endocr Rev 21:412–443
Savino W, Postel-Vinay MC, Smaniotto S, Dardenne M 2002 The thymus gland: a target organ for growth hormone. Scand J Immunol 55:442–452
Benoit C, Mathis D 1999 T-Lymphocyte differentiation and biology. In: Paul W, ed. Fundamental immunology. 4th ed. Philadelphia: Lippincott-Raven; 367–409
Savino W, Dalmau SR, Cotta-de-Almeida V 2000 Role of extracellular matrix mediated interactions in thymocyte migration. Dev Immunol 7:19–28
Savino W, Mendes da Cruz DA, Silva JS, Dardenne M, Cotta de Almeida V 2002 Intrathymic T cell migration: a combinatorial interplay of extracellular matrix and chemokines? Trends Immunol 23:305–313
Annunziato F, Romagnani P, Cosmi L, Lazzeri E, Romagnani S 2001 Chemokines and lymphopoiesis in the thymus. Trends Immunol 22:277–281
Ansel KM, Cyster JG 2001 Chemokines in lymphopoiesis and lymphoid organ development. Curr Opin Immunol 13:172–179
Taub DD, Tsarfaty G, Llyoyd AR, Durum SK, Murphy WJ 1994 Growth hormone promotes human T cell adhesion and migration to both human and murine matrix proteins in vitro and directly promotes xenogeneic engraftment. J Clin Invest 94:293–300
Mello-Coelho V, Villa Verde DMS, Dardenne M, Savino W 1997 Pituitary hormones modulate extracellular matrix-mediated interactions between thymocyte and thymic epithelial cells. J Neuroimmunol 76:39–49
Chen WY, Wight DC, Mehta BV, Wagner TE, Kopchick JJ 1991 Glycine 119 of bovine growth hormone is critical for growth promoting activity. Mol Endocrinol 5:1845–1852
Knapp JR, Chen WY, Turner ND, Byers FM, Kopchick JJ 1994 Growth patterns and body composition of transgenic mice expressing mutated somatotropin genes. J Anim Sci 72:2812–2819
Villa Verde DMS, Chammas R, Lagrota Candido JM, Brentani RR, Savino W 1994 Extracellular matrix components of the mouse thymic microenvironment. IV. Thymic nurse cells express extracellular matrix ligands and receptors. Eur J Immunol 24:659–664
Dalmau SR, Freitas CS, Savino W 1999 High expression of fibronectin receptors and L-selectin as a hallmark of early steps of thymocyte differentiation: lessons from sublethally irradiated mice. Blood 93:974–990
Livak KJ, Schittgen TD 2001 Analyses of relative gene expression data using real-time quantitative PCR and the 2-C(T) method. Methods 25:402–408
Scollay R, Butcher EC, Weissmann L 1980 Thymus cell migration. Quantitative aspects of cellular traffic from the thymus to the periphery in mice. Eur J Immunol 10:210–218
Gabor MJ, Godfrey DI, Scollay R 1997 Recent thymic emigrants are distinct from most medullary thymocytes. Eur J Immunol 27:2010–2015
Yanagawa Y, Iwabuchi K, Onoe K 2001 Enhancement of stromal cell-derived factor-1-induced chemotaxis for CD4/8 double-positive thymocytes by fibronectin and laminin in mice. Immunology 104:43–49
Vila-Coro AJ, Rodriguez-Frade JM, Martin de Ana AM, Moreno-Ortiz MC, Martinez-AC, Mellado M 1999 The chemokine SDF-1 triggers CXCR4 receptor dimerization and activates the JAK/STAT pathway. FASEB J 13:1699–1710
Dorskind K, Horseman ND 2000 The roles of prolactin, growth hormone, insulin-like growth factor-I, and thyroid hormones in lymphocyte development and function: insights from genetic models of hormone and hormone receptor deficiency. Endocr Rev 21:292–312
Kelley KW, Brief S, Westly HJ, Novakofski J, Bechtel PJ, Simon J, Walker EB 1986 GH3 pituitary adenoma cells can reverse thymic aging in rats. Proc Natl Acad Sci USA 83:5663–5667
Li YM, Brunke DL, Dantzer R, Kelley KW 1992 Pituitary epithelial cell implants reverse the accumulation of CD4–CD8– lymphocytes in thymus glands of aged rats. Endocrinology 130:2703–2711
Goya RG, Gagnerault MC, Leite de Moraes MC, Savino W, Dardenne M 1992 In vivo effects of growth hormone on thymus function in aging mice. Brain Behav Immun 6:341–350
Napolitano LA, Lo JC, Gotway MB, Mulligan K, Barbour JD, Schmidt D, Grant RM, Halvorsen RA, Schambelan M, McCune JM 2002 Increased thymic mass and circulating naive CD4 T cells in HIV-1-infected adults treated with growth hormone. AIDS 16:1103–1111
Cotta de Almeida V, Villa-Verde DMS, Lepault F, Dardenne M, Savino W 2004 Impaired migration of NOD mouse thymocytes: a fibronectin receptor-related defect. Eur J Immunol 34:1578–1587
Soriano SF, Hernanz-Falcón P, Rodriguez-Frade JM, Martin da Ana AM, Garzon R, Carvalho-Pinto C, Vila-Coro AJ, Zaballos A, Balomenos D, Martinez-AC, Mellado M 2002 Functional inactivation of CXC chemokine receptor 4-mediated responses through SOCS3 up-regulation. J Exp Med 196:311–321
Weber C, Alon R, Moser B, Springer TA 1996 Sequential regulation of 4?1 and 5?1 integrin avidity by CC chemokines in monocytes: implications for transendothelial chemotaxis. J Cell Biol 134:1063–1073
Shen W, Bendall LJ, Gottlieb DJ, Bradstock KF 2001 The chemokine receptor CXCR4 enhances integrin-mediated in vitro adhesion and facilitates engraftment of leukemic precursor-B cells in the bone marrow. Exp Hematol 29:1439–1447
Wright N, Hidalgo A, Rodrígues-Frade JM, Soriano SF, Mellado M, Parmo-Cabanas, Briskin MJ, Teixidó J 2002 The chemokine stromal cell-derived factor-1 modulates 4?7 integrin-mediated lymphocyte adhesion to mucosal addressin cell adhesion molecule-1 and fibronectin. J Immunol 168:5268–5277
Smaniotto S, Ribeiro-Carvalho MM, Dardenne M, Savino W, Mello-Coelho V 2004 Growth hormone stimulates the selective trafficking of thymic CD4+CD8– emigrants to peripheral lymphoid organs. NeuroImmunoModulation 11:299–306
Sotelo AI, Bartke A, Kopchick JJ, Knapp JR, Turyn D 1998 Growth hormone (GH) receptors, binding proteins and IGF-I concentrations in the serum of transgenic mice expressing bovine GH agonist or antagonist. J Endocrinol 158:53–59
Mello-Coelho V, Villa-Verde DMS, Farias-de-Oliveira DA, Brito JM, Dardenne M, Savino W 2002 Functional IGF-1-IGF-1 receptor-mediated circuit in human and murine thymic epithelial cells. Neuroendocrinology 75:139–150
Savino W, Smaniotto S, Binart N, Postel-Vinay MC, Dardenne W 2003 In vivo effects of growth hormone upon the thymus. Ann NY Acad Sci 992:179–185
Montecino-Rodrigues E, Clarck RG, Power-Braxton L, Dorshkind K 1997 Primary B cell development is impaired in mice with defects of the pituitary/thyroid axis. J Immunol 159:2712–2719
Dorshkind K, Horseman ND 2001 Anterior pituitary hormones, stress, and immune system homeostasis. Bioessays 23:288–294(Salete Smaniotto, Valéria)
Address all correspondence and requests for reprints to: Dr. Wilson Savino, Laboratory on Thymus Research, Department of Immunology, Oswaldo Cruz Institute, Oswaldo Cruz Foundation, Avenue Brasil 4365, Manguinhos, 21045-900 Rio de Janeiro, Brazil. E-mail: savino@fiocruz.br.
Abstract
Previous evidence indicates that GH modulates thymic cell migration. In this study we approached this issue in vivo, studying thymocyte migration in GH transgenic animals and in normal mice treated intrathymically with GH. Extracellular matrix and chemokines are involved in thymocyte migration. In this respect, thymocyte adhesion to laminin was higher in GH-treated animals than controls, and the numbers of migrating cells in laminin-coated Transwells was higher in GH-transgenic and GH-injected mice. Additionally, CXC chemokine ligand 12 (CXCL12)-driven migration was higher in GH-Tg and GH-treated animals compared with controls. Interestingly, although CXCR4 expression on thymocytes did not change in GH-Tg mice, the CXCL12 intrathymic contents were higher. We found that CXCL12, in conjunction with laminin, would additionally enhance the migration of thymocytes previously exposed to high concentrations of GH in vivo. Lastly, there was an augmentation of recent thymic emigrants in lymph nodes from GH-Tg and GH-injected animals. In conclusion, enhanced thymocyte migration in GH transgenic mice as well as GH-injected mice results at least partially from a combined action of laminin and CXCL12. Considering that GH is presently being used as an adjuvant therapeutic agent in immunodeficiencies, including AIDS, the concepts defined herein provide important background knowledge for future GH-based immune interventions.
Introduction
THE CROSS-TALK between the neuroendocrine and immune systems is well demonstrated, with common ligands and receptors being used in both systems, allowing a physiological communication circuitry that seems to play a relevant role in homeostasis. In this context, hormones and neuropeptides are potent immunomodulators, and one of their target organs is the thymus (1). More particularly, a series of evidence indicates that GH act upon the microenvironmental and lymphoid compartments of the organ (2).
The thymus is a primary lymphoid organ in which bone marrow-derived T cell precursors undergo differentiation, ultimately leading to the migration of positively selected thymocytes to the T cell-dependent areas of peripheral lymphoid organs (3). From the entrance of T cell precursors into the thymus to the exit of mature cells from the organ, a vast body of interactions promotes the complex process of T cell differentiation (1) that occurs as cells migrate within the thymic lobules. Such a process involves sequential expression of various proteins and rearrangements of the T cell receptor (TCR) genes. Most of the immature thymocytes, including those bearing the phenotypes TCR–CD3–CD4–CD8– and TCRlowCD3lowCD4+CD8+, are cortically located, whereas mature TCRhighCD3highCD4+CD8– and TCRhighCD3highCD4– CD8+ cells (that will normally leave the organ toward peripheral lymphoid organs) are found in the medulla. Along with such a journey, developing thymocytes encounter cortical and medullary nonlymphoid microenvironments through distinct cell-cell and cell-matrix interactions (1, 4). Cell migration is thus a crucial event for intrathymic T cell differentiation, and we recently proposed that chemokines act in concert with the extracellular matrix (ECM), resulting in the migration of a given cell subset, either within the thymus or at the entrance into and/or exit from the organ (5). Interestingly, microenvironmental cells produce both types of molecules (4, 5, 6, 7).
Few studies suggest that GH modulates intrathymic T cell migration: 1) recombinant GH increases human T cell engraftment into the thymus of severe combined immune deficiency mice, an effect that seems to be mediated by adhesion molecules and ECM, because it can be abrogated with anti-?1 integrin antibodies (8); 2) in vitro GH treatment increases the adhesion of thymocytes to thymic epithelial cells (TEC; the major component of the thymic microenvironment) as well as the production of ECM proteins and expression of ECM receptors by TEC (9). In this study, we approached this issue by studying thymocyte migration in GH-transgenic (GH-Tg) mice as well as in normal mice intrathymically injected with GH.
Materials and Methods
Animals
Female and male BALB/c, C57BL/6, and GH-Tg mice were bred and maintained under specific-pathogen free conditions at Necker Hospital animal facilities. GH-Tg mice were derived from animals provided by Drs. A. Bartke (Department of Physiology, Southern Illinois University School of Medicine, Carbondale, IL) and Thomas R. Wagner (Oncology Research Institute, Greenville Hospital System, Greenville, SC). These mice were generated by introducing the bovine GH transgene linked to the metallothionein promoter (10) and bred in a C57BL/6 genetic background. High levels of GH are maintained in these GH-Tg animals (11).
In most cases, 8- to 10-wk-old animals were used, whereas some experiments using GH-Tg mice were run using different ages, from 6 wk to 7 months. The animal facilities and care followed rules precluded by the European Union ethics committee for animal research.
Chemicals
Recombinant bovine GH was provided from Dr. A. F. Parlow (Pituitary Hormones and Antisera Center, National Institute of Diabetes and Digestive and Kidney Diseases, Torrance, CA). Stock solution (1 mg/ml) was prepared in PBS and kept at –20 C until use. Cellular fibronectin (from human foreskin fibroblasts), laminin (from murine Engelbreth-Holm-Swarm sarcoma), BSA, fluorescein isothiocyanate (FITC), and aminoethylcarbazole were purchased from Sigma-Aldrich Corp. (St. Louis, Mo), whereas the anesthetic 2,2,2-tribromoethanol was purchased from Aldrich Chemical Co. (Milwaukee, WI). The chemokines CXCL12/SDF-1 and CCL25/TECK were obtained from R&D Systems, Inc. (Minneapolis, MN). Pertussis toxin (PTX) and cholera toxin (CTX) were obtained from Sigma-Aldrich Corp. (St. Louis, MO).
Antibodies
For immunohistochemistry, rabbit polyclonal antisera specific for laminin or fibronectin were obtained from Novotec (St. Martin-la-Garenne, France); rabbit antihuman cytokeratin immune serum as well as peroxidase-labeled goat antirabbit Ig serum were purchased from DakoCytomation (Carpinteria, CA). Anti-CXCL12 monoclonal antibody (mAb) was purchased from R&D Systems, Inc. Rhodamine-labeled goat antirabbit Ig and FITC-coupled goat antirat Ig were purchased from Biosys (Compiegne, France). In all cases, optimal dilutions of each primary or secondary antibody were predetermined in our laboratory.
For cytofluorometric analyses, appropriate dilutions of the following mAb were used: anti-CD3/FITC, anti-CD4/phycoerythrin (PE), anti-CD4/peridinin chlorophyll protein (PercP), anti-CD4/allophycocyanin (APC), anti-CD8/CyChrome, anti-CD8/APC, anti-CD8/PercP, anti-CD44/FITC, anti-CD25/APC, anti-CD49d/PE, anti-CD49e/PE, anti-CD49f/PE, and anti-CXCR4/PE as well as isotype-matched negative controls for each fluorochrome applied in specific antibodies, thus including unrelated rat Igs labeled with FITC, PE, APC, or PercP (all purchased from BD Pharmingen, San Diego, CA). In some blocking experiments, purified anti-CD49f was applied (10 μg/ml) to inhibit laminin-very late antigen-6 (VLA-6)-mediated interactions.
Immunohistochemistry
Thymus frozen sections were submitted to indirect immunofluorescence or immunoperoxidase assay as previously described (12). Specimens were incubated with a given primary antibody for 1 h, washed with PBS, and treated with rhodamine- or peroxidase-coupled second antibody. In the case of immunofluorescence, specimens were analyzed under a confocal fluorescence microscope (MS 510, Zeiss, Oberkochen, Germany). When peroxidase-linked second antibody was applied, enzyme activity was revealed with aminoethylcarbazole in the presence of H2O2. Controls comprised specimens in which primary antibodies were replaced by unrelated Igs and generated no significant labeling (not shown).
Cytofluorometry
Thymus, spleen, as well as sc and mesenteric lymph node cell suspensions from each animal group were prepared in PBS as previously described (13). Three- or four-color cytofluorometry was performed by incubating cells with a mixture of appropriately diluted antibodies in 2% fetal calf serum in PBS for 20 min at 4 C. After washings, cells were fixed and analyzed by flow cytometry in a FACSCalibur device (BD Pharmingen) equipped with CellQuest software (BD Biosciences, Mountain View, CA). A gate excluding cell debris and nonviable cells was determined using forward vs. side scatter parameters and was confirmed in some experiments with the use of propidium iodide staining and immediate analysis of unfixed cells. Analyses were made after recording 10,000–100,000 events for each sample.
We also phenotyped the CD4–CD8– double-negative (DN) subset with the CD44 and CD25 markers. For that, we used anti-CD4/PercP and anti-CD8/PercP in conjunction with anti-CD25/allophycocyanin and anti-CD44/FITC. In these experiments, we recorded 200,000 events at the flow cytometer.
Real-time PCR
Mouse thymuses were homogenized in 1 ml TRIzol reagent (Invitrogen Life Technologies, Inc., Carlsbad, CA) using a Polytron instrument (Brinkmann Instruments, Westbury, NY). The total RNA was then isolated according to the instructions of the manufacturer. RT of 2 μg total RNA was performed using the Stratagene RT kit (La Jolla, CA), and the resulting cDNA was submitted to real-time PCR, to measure the expression of mRNA encoding for CXCL12. Amplification of ?2-microglobulin was performed as an endogenous control. The following primers were applied: CXCL12: forward: ccagtcagcctgacctaccg; reverse, cgggtcaatgcacacttgtc; and ?2-microglobulin: forward, tgaccgcttgtatgctatccagt; reverse, cagtgtgagccaggatatag.
The PCR was performed on an ABI PRISM 7700 sequence detector using the quantitative PCR Master Mix for SYBR Green (Eurogentec, Seraing, Belgium). Amplification was carried out in a total volume of 20 μl containing 0.5 μM gene-specific primers, 10 μl Master Mix SYBR Green, and 5 μl cDNA at different dilutions (three then fold dilutions). After initial enzyme activation (5 min at 55 C) and DNA denaturation, (5 min at 95 C) the PCRs were cycled 50 times with the following parameters: denaturation at 95 C for 30 sec, annealing at 59 C for 30 sec, and extension at 72 C for 30 sec. RNA expression was determined by the comparative cycle threshold (Ct) method (14). Four independent assays were performed, each one in triplicate.
Cell adhesion assay
To assess the thymocyte adhesive capacity on ECM substrata, 60 x 15-mm culture dishes (Nunc, Copenhagen, Denmark) were coated for 1 h at room temperature (plus 30 min for air drying) with 10 μg/ml fibronectin, laminin, or BSA as a control. In these assays, 1 x 107 thymocytes from three pooled thymuses from GH-Tg, GH intrathymically treated BALB/c mice, or their corresponding controls were allowed to adhere on the precoated dishes. After 1-h incubation, the adherent thymocytes were harvested, counted, and phenotyped for the expression of CD3, CD4, and CD8 molecules. The relative adhesion was calculated as the ratio of adhesion between GH-Tg or GH-treated over the respective control, which was, by definition, fixed at 1.
Isolation and culture of thymic nurse cells (TNCs)
Migration of immature thymocytes can be investigated in vitro by studying TNCs. These are cortically located lymphoepithelial complexes formed by one TEC, which harbors various numbers of thymocytes (mostly immature cells), being located in the cortical region of thymic lobules. In culture, TNCs spontaneously release thymocytes, and TNC-derived epithelial cells can reconstitute lymphoepithelial complexes after being cocultured with fetal thymocytes, thus corresponding to an in vitro model of thymocyte migration within the TEC context (5). TNCs were isolated from pools of five to 10 C57BL/6 wild-type or GH-transgenic female mice at 11–15 wk of age, as currently done in our laboratory (9), and the number of TNC per mouse was then calculated. Once obtained, such lymphoepithelial complexes were settled in culture, so as to obtain, 5 d later, highly purified TNC-derived epithelial cultures. These cells were used for immunocytochemical detection of laminin and CXCL12 or to evaluate reconstitution of thymic nurse cell complexes. In this kind of experiment, 5-d TNC-derived TEC were trypsinized and cocultured with fresh 16- to 17-d-old fetal thymocytes (105 thymocytes/104 TEC/well) in inverted Terasaki plates for 6 h in complete RPMI 1640 medium (9). Alternatively, cocultures were performed in the presence of purified antilaminin antibodies (10 μg/ml). Control cultures were incubated with control rabbit IgG at equivalent protein concentrations. Percentages of newly formed complexes were evaluated in blind by direct counting under the light microscope.
Cell chemotaxis assay
Cell chemotaxis is another strategy to evaluate immature as well as mature thymocyte migration in vitro, under distinct stimuli (5). Thymocyte migratory activity was assessed ex vivo in 5-μm pore size Transwell plates (Corning Costar, Cambridge, MA) using cell suspensions after pooling three thymuses from each group of mice. In this cell chemotaxis assay, membranes were coated on both sides with 10 μg/ml fibronectin, laminin, or BSA for 1 h at 37 C, followed by 1 h of blocking with 1% BSA. Thymocytes (2.5 x 106) were then added to the upper chamber in 100 μl 0.5% BSA/RPMI 1640, and 600 μl 0.5% BSA/RPMI 1640 were added to the lower chamber. After 3 h of incubation at 37 C in 5% CO2-containing air, cells migrated into lower chambers were removed, counted, and analyzed by flow cytometry. In additional experiments, we tested the ability of the chemokines, CXCL12 and CCL25, applied at 100 ng/ml, alone or in combination with laminin.
GH injection and intrathymic fluorescein labeling
One direct strategy to evaluate thymocyte exit in vivo is analysis of the so-called recent thymic emigrants. It is established that intrathymic injection of FITC labels large numbers of thymocytes, allowing the recovery of FITC+ cells that have recently emigrated from the thymus (15, 16). BALB/c as well as GH-Tg and corresponding age-matched control mice were anesthetized with 2,2,2-tribromoethanol, and 10 μl FITC solution (1 mg/ml)/thymic lobe were injected into the thymus in the open chest cavity. Controls for the injection procedure consisted of dropping 10 μl FITC solution into the mediastinal cavity (over the thymus) or intrathymic injection of PBS alone. In the case of BALB/c mice, the volume of 10 μl injected per thymic lobe contained FITC (1 mg/ml) plus GH (10–5 M, final concentration). In addition, in this model an additional control consisted of injecting heat-denatured GH, applied at the same concentration as the native hormone. We previously ascertained that FITC plus GH or GH alone had similar effects, stimulating TEC proliferation, indicating that FITC does not interfere in the GH activity (data not shown). We also ascertained that TEC proliferation was not enhanced by GH denatured by boiling (2). In additional experiments in which FITC was not injected, intrathymic treatment with GH (or its denatured form) followed the same protocol as that described above.
In all animals the chest cavity was closed with surgical clips; 16 h later, thymus, spleen, and lymph node cells were collected from all mice. In additional experiments, the organs were collected 2 wk after GH injection and 16 h after FITC injection. Mice whose thymuses contained less than 50% FITC-labeled cells were discarded.
Statistical analyses
Data are shown as the mean ± SE. Results were statistically analyzed by unpaired Student’s test; data were considered statistically different at P 0.05.
Results
Thymocyte numbers in GH-Tg and GH-injected mice
Relative cell numbers (per gram of body weight) in GH-Tg mice were higher than those in age-matched controls; the same pattern was observed in both young and middle-aged animals, in which differences were statistically significant (P < 0.05). Similar findings were seen in the total numbers of thymocytes from mice injected intrathymically with GH compared with controls, either untreated or treated with PBS or heat-denatured GH (dGH). Interestingly, this hormonal effect was still observed 2 wk after a single GH injection (Fig. 1, A and B). Such an increase in absolute numbers was seen in both immature and mature thymocyte subsets (Fig. 1, C and D), although the profiles regarding the expression of CD3, CD4, and CD8 in various thymocyte subsets were similar in the various experimental groups, as shown in Fig. 1E. We also evaluated the DN compartment with CD44 and CD25 markers. No differences were seen between GH-Tg and controls for any of the four CD44/CD25-defined DN subsets (Fig. 1F).
FIG. 1. Thymocyte numbers and CD3/CD4/CD8-defined phenotypes in GH-Tg and GH-injected mice. A, Comparison of the numbers of thymocytes per gram of body weight in GH-Tg mice () vs. corresponding wild-type controls (Wt; ), in young (6–10 wk) and middle-aged (26–32 wk) mice. B, Total thymocyte numbers in BALB/c mice intrathymically injected with native GH () compared with those in controls injected with denatured GH (), either 16 h or 2 wk after a single GH injection. In both GH-Tg and GH-injected animals, the numbers of thymocytes were higher than in the corresponding controls. C and D, Such an increase in absolute cell numbers comprises both immature and mature CD4/CD8 thymocyte subsets, respectively, in GH-Tg and GH-injected animals. E, Flow cytometry-derived three-color profiles for defining CD4 vs. CD8 as well as CD3 expression in both 8-wk-old wild-type and GH-Tg animals. The percentages of positive cells are seen in each profile. With regard to CD3 labeling, total CD3+- as well as CD3high-expressing cells are shown separately. No significant difference was seen between the two groups. Similarly, no difference was seen between GH-Tg and wild-type thymus in terms of the relative number of CD25/CD44-defined subsets within the CD4–CD8– DN compartment (F). Representative relative numbers of each subset are shown in each quadrant. With regard to GH-Tg vs. C57BL/6, 28–30 animals were used per group. In BALB/c mice treated with GH for 16 h, 46–49 animals were evaluated per group, whereas in the experiments performed 14 d after a single GH intrathymic injection, five to six mice per group were analyzed. *, P < 0.05; **, P < 0.02.
GH enhances ex vivo thymocyte adhesion to and migration through laminin
Knowing that cell migration depends on the adhesion of the migrating cell to the substrate, we also checked whether the adhesion capacity of thymocytes onto ECM-defined substrata was modified by in vivo GH treatment. Experiments performed in both GH-Tg and GH-injected BALB/c mice revealed that adhesion of thymocytes to laminin (but not to fibronectin) was higher than that in the corresponding controls: C57BL/6 and dGH-injected BALB/c, respectively (Fig. 2A). Although the total number of adherent cells was changed, their CD4/CD8-defined phenotype remained similar in both GH-Tg and GH-treated and the corresponding control groups, namely, untreated C57BL/6 and dGH-treated BALB/c mice (Fig. 2B).
FIG. 2. GH enhances laminin-driven thymocyte adhesion and migration. A, Relative adhesion to laminin or fibronectin of GH-Tg thymocytes (GH-Tg; ) compared with wild-type controls (Wt; ). Relative adhesion was calculated by dividing the numbers of adhered thymocytes from the GH-Tg group by the corresponding control number (whose relative adhesion was then defined as 1). The panel depicts two independent experiments. The cytofluorometric profiles shown in B illustrate that the percentages of CD4/CD8-defined thymocyte subsets of laminin-adherent cells are similar in Wt and GH-Tg mice. Similar data were obtained when thymocytes from GH-injected BALB/c and corresponding controls were allowed to adhere to laminin or fibronectin (data not shown). C and D, Thymocyte migration driven by laminin (but not fibronectin) is enhanced in both GH-Tg (C) and GH-injected (D) mice compared with that in controls. Mice used in these experiments were 8–10 wk of age, and at least six animals were used in each group. *, P < 0.05.
Taking into account the above findings, we examined whether high GH contents in vivo could also modulate ex vivo thymocyte migration through ECM-coated Transwells. Examining the total number of migrating cells, we observed that migration through laminin was higher in GH-Tg and GH-injected mice than in the corresponding controls (Fig. 2, C and D). In contrast, when fibronectin was applied as a haptotatic stimulus, the numbers of migrating cells were similar in GH-Tg and wild-type groups. No differences between control and GH-Tg animals were seen in BSA-driven cell migration as well. In GH-injected BALB/c mice, the differences observed were the same as those described for GH transgenic animals. In both cases, GH enhanced the adhesion and migration of thymocytes, but did not change the phenotypic patterns of adhered or migrating cells (data not shown).
Does GH modulate intrathymic ECM ligands and receptors in vivo?
Because we had shown that in vitro treatment of TEC with GH enhanced ECM deposition (9), we evaluated the distribution of laminin and fibronectin in the thymus of GH-Tg mice, compared with respective controls: GH induced an increase in ECM deposition in both cortical and medullary regions of the thymic lobules, as illustrated in Fig. 3. Interestingly, in situ labeling also revealed an increase in the laminin receptor VLA-6 in the microenvironmental compartment of GH-Tg thymuses compared with controls (Fig. 4A). Nevertheless, the membrane levels of ECM receptors (VLA-4, VLA-5, and VLA-6) on thymocytes did not change (Fig. 4B and Table 1). Similar findings were seen in GH-injected BALB/c mice (data not shown).
FIG. 3. Enhancement of ECM proteins in the thymic microenvironment of GH-Tg mice. Upper panels, Laminin staining by immunofluorescence; an increase in cortical laminin-containing fibrils in GH-Tg 10-wk-old mouse thymus (A) compared with the age-matched wild-type control (B). Laminin deposition is also increased in the medullary region of GH-Tg mouse thymus (C), compared with wild-type control (D). Bottom panels, Enhancement of fibronectin in the thymus of GH-Tg mice (F) compared with control thymus from a wild-type mouse. Magnifications, x400. C, Cortex; M, medulla; S, septum. The single arrow shows the capsule, whereas the double arrow indicates a blood vessel.
FIG. 4. Enhancement of VLA-6 in the thymic microenvironment of GH-Tg mice. Upper panels, CD49f (-chain integrin of the laminin receptor VLA-6) staining by immunoperoxidase. Compared with the age-matched wild-type control (C), an increase in immunostaining is seen in GH-Tg 10-wk-old mouse thymus (B). C, By contrast, VLA-6 membrane expression is similar in GH-Tg- and wild-type-derived CD4/CD8-defined thymocyte subsets. C, Cortex; M, medulla; S, septum; BM, blood vessel.
TABLE 1. Expression of fibronectin and laminin receptors in CD4/CD8-defined thymocyte subsets from GH-Tg and wild-type C57BL/6 mice
GH modulates laminin-mediated interactions in TNC complexes
We next analyzed the TNC complex, a model for intrathymic T cell migration (5), which is targeted by in vitro GH treatment (9) and is partially dependent on laminin-mediated interactions (5). We first noticed that the numbers of freshly isolated TNCs in both GH-Tg and GH-injected mice were higher than the corresponding control values (Fig. 5A). Additionally, as ascertained in the transgenic model, the percentages of larger and more granular TNC complexes were higher in GH-Tg than wild-type counterparts (Fig. 5B). Moreover, when we established epithelial cultures derived from GH-Tg TNC complexes, the amount of laminin detected was greater than that in epithelial cultures derived from wild-type mouse thymuses (Fig. 5C). These findings prompted us to attempt to reconstitute TNC lymphoepithelial complexes using thymocytes from normal animals that were cocultured with GH-Tg or wild-type-derived TEC. This provided a model to ascertain whether the GH-Tg laminin-enriched thymic microenvironment was also playing a role in the thymocyte migration, herein exemplified by the entrance of thymic lymphocytes into a well-defined thymic epithelial niche. In fact, formation of lymphoepithelial complexes was increased when freshly isolated fetal thymocytes obtained from C57BL/6 mice were cocultured with GH-Tg-derived TEC compared with TEC obtained from TNC cultures of age-matched, wild-type animals, and this effect was partially, yet statistically significantly, abrogated by antilaminin treatment (Fig. 5, D and E).
FIG. 5. Changes in TNCs from GH-Tg mice. A, Both the relative TNC numbers in GH-Tg mice and the total TNC numbers in animals injected with GH are higher than those in corresponding controls. B, The percentage of large granular TNCs from GH-Tg mice is higher compared with that in controls. Moreover, laminin labeling in TNC-derived epithelial cells from GH-Tg is stronger than that in TEC cultures from wild-type mice (C). Finally, reconstitution of TNC lymphoepithelial complexes is higher when the epithelial cells were cocultured with normal thymocytes and were derived from GH-Tg mice (D). This effect is partially abrogated with the use of antilaminin (LN) antibody, but not with the unrelated rabbit Igs (Rab Ig), used at the same concentration (E, right). Mice were 8 wk old, and pools of at least 10 animals were used per group in each experiment depicted in A and B, which are representative of three similar independent experiments. D and E, At least three independent cultures of triplicates were used.
GH enhances CXCL12-triggered thymocyte migration: synergy with laminin
Considering that thymocyte migration is also influenced by chemokines (6, 7), and that the chemokine CXCL12 was reported as having a combined effect with laminin in modulating thymocyte migration (17), we addressed the question of whether GH could modulate intrathymic CXCL12-driven migration. In fact, when CXCL12 was applied, thymocyte migration was consistently higher in GH-Tg mice (Fig. 6A). Similar results were obtained with GH-treated BALB/c animals (not shown). Interestingly, this effect seems to be rather specific, because it was not seen when another chemokine, CCL25, which is also known to preferentially attract immature thymocytes (5, 7), was applied (Fig. 6A).
FIG. 6. Enhanced thymocyte migration induced by laminin in conjunction, or not, with CXCL12. A, Experiments with 10-wk-old GH-Tg mice () and corresponding C57BL/6 wild-type controls (), with data expressed as the total number of cells that migrated through Transwell chambers. Nonspecific thymocyte migration, seen when BSA was applied in the Transwells, was similar in all groups. Thymocyte migration induced by the chemokine CCL25 was also similar in wild-type and GH-Tg animals. In contrast, laminin (LN)-driven or CXCL12-driven migration of thymocytes was significantly enhanced in GH-Tg mice compared with respective controls. Moreover, a synergy was seen when both stimuli were applied simultaneously, and this was due to the enhancement of CD4+CD8+ cell migration (B), although absolute numbers of all CD4/CD8-defined subsets from GH-Tg mice were also significantly increased after CXCL12-driven migration, in combination, or not, with laminin. Each column represents the mean ± SE of three independent experiments. *, P < 0.05; **, P < 0.01.
Concerning CXCL12 stimulation, GH-Tg and GH-injected CD4/CD8-defined thymocyte subsets migrated significantly more than the corresponding controls.
In conjunction with these ex vivo functional data, we found increased levels of CXCL12 in GH-Tg thymic stroma (Fig. 7, A and B) and GH-Tg-derived TNC cultures compared with corresponding controls; the latter was ascertained by real-time PCR and immunocytochemistry (Fig. 7, C and D). By contrast, the membrane levels of CXCR4 in thymocytes were similar in wild-type and transgenic animals, independent of the CD4/CD8-defined subset analyzed (Fig. 7E).
FIG. 7. Enhanced CXCL12 production by thymic cells of GH-Tg mice. Upper panels, Thymus sections from wild-type (A) and GH-Tg (B) mice, depicting stronger immunohistochemical labeling for CXCL12 in the GH-Tg thymus. Also, epithelial cells from GH-Tg animals produce larger amounts of CXCL12, as revealed by quantitative real-time PCR (C); the relative increase in specific mRNA was in three dilutions, and values were obtained with wild-type-derived cells (s) normalized to 1. GH-Tg TEC produced 4- to 5-fold more CXCL12 mRNA than controls. An increase in CXCL12 was confirmed by immunocytochemistry (D), showing an increased labeling in Tg mice-derived TEC (right) compared with controls (left). E, A series of cytofluorometric profiles, showing that the membrane levels of CXCR4 are similar in GH-Tg and wild-type thymocytes regardless of the CD4/CD8-defined thymocyte subset analyzed.
We then investigated whether CXCL12, in conjunction with laminin, could also enhance migration of thymocytes previously exposed to high concentrations of GH in vivo. We noticed that in control groups the simultaneous use of both stimuli resulted in a degree of migration that was higher than the summation of each stimulus alone. Such an additive effect was also seen in GH-Tg and GH-injected mice and essentially targeted the CD4+CD8+ subpopulation (Fig. 6).
Thymocyte migration elicited by laminin, with or without CXCL12, was also investigated with the use of specific blockers, anti-VLA-6 mAb and Bordetella PTX, respectively. As expected, migration induced by laminin was completely prevented by incubation of thymocytes with anti-VLA-6 mAb, whereas CXCL12-induced migration was abrogated by PTX (Fig. 8B). Control experiments revealed that an isotype-matched mAb did not change thymocyte migration triggered by laminin (not shown), and that CTX, which is known to be inefficient in blocking chemokine receptors (18), did not exert any effect on CXCL12-induced migration of thymocytes from wild-type or GH-Tg mice. Moreover, PTX did not change the profiles of laminin-triggered thymocyte migration in either wild-type or GH-Tg animals. However, when thymocytes from GH-Tg mice were preincubated with anti-VLA-6 and then subjected to CXCL12 stimulation, we found a consistent and statistically significant 20% decrease in the numbers of migrating cells (Fig. 8A). In control animals, such a decrease was much less and was not statistically significant.
FIG. 8. Blockage of laminin receptor interferes with CXCL12-driven migration of thymocytes from GH-Tg mice. A, Preincubation of thymocytes with the anti-VLA-6 antibody completely abrogated laminin (LN)-triggered thymocyte migration (which decreased to values obtained with BSA, used as a negative control) in both wild-type and GH-Tg animals. The numbers of migrating thymocyte CXCL12 were enhanced in GH-Tg mice (compared with wild type), and such an enhancing effect was abrogated with the anti-VLA-6 antibody to levels seen in CXCL12-induced migration of control thymocytes. Such a blocking effect was also seen when laminin and CXCL12 were used simultaneously. B, CXCL12-driven migration of thymocytes from wild-type and GH-Tg mice is prevented by PTX, but not by the corresponding control, CTX. PTX, however, was not able to interfere with the LN-driven migration, independently of laminin being used alone or together with CXCL12. *, P < 0.05; **, P < 0.01.
We also evaluated the expression of CXCR4 and VLA-6 in CD4/CD8-defined subsets of those thymocytes that migrated after laminin and/or CXCL12 stimulation. VLA-6 membrane levels did not change in migrating cells, whereas a down-regulation of CXCR4 was seen on thymocytes exposed to CXCL12, in conjunction, or not, with laminin, although no differences were found between GH-Tg and wild-type mice (Fig. 9).
FIG. 9. Expression of CD49f and CXCR4 in CD4/CD8-defined thymocyte subsets before and after migration induced by laminin and/or CXCL12. Representative cytofluorometric profiles are shown in various situations, with the relative number of each VLA-6/CXCR4-defined subpopulation shown within each rectangle. Before migration, both wild-type and GH-Tg mouse thymocytes demonstrated that the simultaneous expression of CD49f (the 6-chain of the laminin receptor VLA-6) and CXCR4 is more frequent in immature thymocyte subsets (CD4–CD8– and CD4+CD8+ cells) than in mature CD4+ or CD8+ single-positive cells. Such a pattern was maintained in those cells that migrated under an unrelated BSA stimulus as well as after stimulation by laminin. By contrast, when CXCL12 was applied as chemoattractant (alone or in combination with laminin), the relative numbers of VLA-6+CXCR4+ migrating cells were diminished in both wild-type and GH-Tg mice, an effect that was seen in all CD4/CD8-defined thymocyte subsets and was still more pronounced in immature CD4–CD8– and CD4+CD8+ cells. Similar results were seen in four animals from each experimental group.
GH modulates thymocyte export to peripheral lymphoid organs
In a last series of experiments we tested whether GH could modulate the exit of thymocytes in vivo. The analysis of recent thymic emigrants (RTEs) in GH-Tg and GH-injected animals revealed that compared with controls, absolute RTE numbers (i.e. FITC+ cells) in both GH-Tg and GH-injected animals were augmented in mesenteric and, to a lesser extent, in sc lymph nodes, but not in spleen. This augmentation was essentially due to the differences in the CD4+ subset, as seen by the CD4/CD8 ratios in wild-type compared with GH-Tg mice (Table 2). In contrast, no changes were observed in the rare CD4+CD8+ cells seen in peripheral lymphoid organs. These findings suggest that high levels of GH differentially modulate the relative distribution of these cells, at least with regard to the peripheral lymphoid organs evaluated in this study.
TABLE 2. Modulation of homing of thymic-derived CD4+ T lymphocytes in GH-Tg mice
Discussion
Previous studies revealed that distinct intrathymic cellular interactions can be influenced by GH in vitro, as exemplified by the secretion of thymulin, a TEC-derived thymic hormone, TEC proliferation, as well as the expression of ECM ligands and receptors with consequent modulation of ECM-mediated TEC/thymocyte interactions (1). By contrast, studies using knockout mice for the GH receptor revealed that this hormone is not essential for thymus development (19). Yet, in vivo addition of GH does appear to be effective; systemic GH injections in aging rats increased the total thymocyte number and the percentage of CD3-bearing cells (20, 21) as well as concanavalin A-induced mitogenic response and IL-6 production (22). In keeping with this finding, AIDS patients who received adjuvant GH therapy exhibited thymus growth, as demonstrated by computer tomography (23). Moreover, recombinant GH stimulated human peripheral blood lymphocyte engraftment and migration of T cells into the thymus of severe combined immune deficiency mice (8).
In this study we evaluated in vivo the effects of high levels of GH in the thymus using two distinct approaches: intrathymic injection of GH in BALB/c mice and evaluation of GH-Tg mice. Accordingly, the first protocol allowed us to evaluate the acute effect of the hormone within the organ; in the second, we could study the effects due to prolonged exposition of the thymus to high circulating levels of GH. Interestingly, in both conditions the numbers of thymocytes were slightly (yet significantly) higher than the corresponding control value. This is in keeping with preliminary data obtained with GH-injected mice, which presented 10–15% more cells in the S and G2M phases compared with denatured GH-injected animals, whereas no differences were seen in terms of membrane Fas expression (Smaniotto, S., M. Dardenne, and W. Savino, unpublished observations).
We focused herein on those GH-triggered effects related to thymocyte development. ECM ligands and receptors correspond to one group of players in this process (4). We observed that GH induced an increase in laminin deposition throughout the thymic lobules. This is in keeping with our in vitro data, which showed that treatment of cultured TEC with GH enhanced ECM deposition (9). Additionally, in both GH-Tg and GH-injected mouse thymus sections, we found that the VLA-6 labeling in the microenvironmental compartment was denser than that in the respective controls. This may result in a stronger attachment of laminin to the microenvironmental network, thus favoring its presentation to thymocytes. Such an idea is also supported by the fact that TNC-derived epithelial cells from GH-Tg mice apparently produce more laminin than those from wild-type animals, and that the degree of reconstitution of lymphoepithelial complexes, formed after coculturing GH-Tg-derived TEC with normal thymocytes, was higher than that seen when TEC from wild-type animals were used.
In contrast, however, membrane expression of VLA-6 on thymocytes did not change significantly. Yet, with regard to the laminin receptor, ex vivo binding to laminin was higher when thymocytes from GH-treated mice were applied, suggesting that high GH contents also activate VLA-6 in thymocytes, favoring their migrating progress. Furthermore, such an effect was found in both immature and mature CD4/CD8-defined thymocyte subsets. In keeping with this, thymocyte migration in laminin-containing Transwell chambers was higher in GH-Tg and GH-treated mice compared with controls. Overall, these data tell us that GH promotes a multifaceted change in laminin-mediated interactions; one consequence being an increased migratory response of thymocytes, which probably means a change in the developmental behavior. At present, we cannot explain why the expression of VLA-6 is enhanced in the microenvironment, but not in thymocytes. Yet it is conceivable that the mechanisms controlling the expression of this molecule in thymocytes and TEC are different. This hypothesis is actually supported by our recent data showing a defect in the expression of the fibronectin receptor VLA-5 on nonobese diabetic thymocytes, which is not found in the TEC of the same animals (24).
A nonmutually excluding hypothesis we raised is that chemokine-driven thymocyte migration could also be under the influence of GH. In fact, the thymocyte migratory response triggered by CXCL12 was higher in GH-Tg and GH-treated mice than in controls, although the membrane levels of CXCR4 remained similar in Tg and wild-type animals, suggesting that (similar to what was seen in relation to VLA-6) CXCR4 in GH-Tg thymocytes is spontaneously more activated. Moreover, the enhancement of CXCL12 seen in the thymus and TNC cultures derived from GH-Tg mice, ascertained at both mRNA and protein levels, suggest that in vivo, both microenvironmental and thymocyte sides contribute to the resulting enhancement of CXCL12-mediated interactions revealed in the GH-Tg mouse thymus.
Interestingly, it was shown that in vivo injection of CXCL12 in bovine GH-Tg mice inhibited ex vivo T cell migratory activity of peritoneal cells (25), suggesting that the GH-enhancing, CXCL12-triggered migration shown in this study might be restricted to early T cell developmental stages. In any case, the present data represent, to our knowledge, the first demonstration that chemokine-driven thymocyte migration can be hormonally controlled. This is a relevant issue when designing procedures involving GH-based therapy in both experimental animals and humans.
An additional aspect that we analyzed was the possibility that GH could enhance the combined effects of laminin plus CXCL12. When both molecules were applied, the number of migrating cells was higher than the sum of the two stimuli alone, an effect seen in control, GH-Tg, and wild-type animals and that essentially targeted the CD4+CD8+ cells. The fact that such a synergic effect was seen in both wild-type and GH-Tg animals is evidence that GH enhances a biological circuitry that already exists in the physiological situation. These data fit with the idea that ECM and chemokines may act in combination to drive thymocyte migration (5) and reinforce the idea of a cross-talk between integrin and chemokine receptors, a concept that has been established for CXCL12 and laminin or fibronectin (26, 27, 28). Accordingly it is conceivable that CXCL12/CXCR4-induced signaling also uses the Janus kinase/signal transducer and activator of transcription pathway, similar to integrins of the VLA family. In this respect, it is noteworthy that an antilaminin receptor mAb could significantly block approximately 20% of the migratory activity elicited by CXCL12 in GH-Tg thymocytes. However, because we did not detect a significant reciprocal blocking activity of PTX on laminin-driven thymocyte migration even in GH-Tg animals, it is conceivable that the relative influence of VLA-6 on the CXCR4 signaling pathway is more relevant than the reverse situation. Accordingly, disruption of VLA-6-triggered intracellular signaling, secondary to the binding of the anti-VLA-6 mAb, could somewhat alter the Janus kinase/signal transducer and activator of transcription pathway, which is also partially triggered by CXCR4 stimulation, as recently demonstrated (18). Although additional studies of this issue are obviously necessary, our results place GH-Tg derived thymocytes as an interesting model to study putative CXCR4/VLA-6 functional connections.
The fact that ex vivo migration of mature thymocytes was enhanced in high GH content conditions prompted us to check whether the hormone could also modulate the exit of thymocytes. We found in both GH-Tg and GH-injected animals an augmentation of CD4+FITC+ cells in mesenteric and sc lymph nodes (but not in the spleen), revealing that GH modulates the relative distribution of these cells among peripheral lymphoid organs. The mechanisms governing such differential RTE homing in GH-Tg mice have not been explored. Yet it is likely that CD62L (a well-known homing receptor for lymph nodes) is involved, because in BALB/c mice intrathymically injected with bovine GH, we found an up-regulation of this molecule in lymph node RTEs (29).
The biological circuit(s) triggered by high GH contents has not been addressed in this study. Nevertheless, an IGF-I/IGF-I receptor loop is likely to be involved. GH-Tg mice have high levels of circulating IGF-I (30). Moreover, we defined an intrathymic GH-controlled IGF-I/IGF-I receptor circuitry in human and mouse TEC and thymocytes (31). Lastly, we showed that in vitro GH-enhanced TEC/thymocyte interactions relevant to thymocyte migration could be abrogated not only by anti-ECM and anti-ECM receptor antibodies, but also by anti-IGF-I and anti-IGF-I receptor reagents (9).
Another issue to be discussed refers to whether the effects reported herein are induced via paracrine or endocrine pathways. This is particularly interesting considering that the results seen in GH-Tg mice were similar to those seen in normal animals injected intrathymically with GH. Although this issue has not been approached in a precise way, it is likely that both pathways occur; in vitro GH treatment of thymic cells does promote a multifaceted biological response, comprising, among other effects, the modulation of ECM proteins (9). Additionally, systemic injection of GH into mice enhances thymic hormone production by TEC (22).
Taken together, our data demonstrate that GH-Tg mice as well as GH-injected animals present altered thymocyte migration, with changes in the distribution of exiting thymocytes in the periphery of the immune system. Moreover, such a GH-related thymocyte migration results at least partially from a combined action of laminin and CXCL12. It is not known whether the opposite situation occurs in mouse GH receptor knockout mice, yet such a hypothesis seems plausible considering that these animals do exhibit alterations in the thymic epithelium, as revealed by increased density of the TEC network, the appearance of large epithelial cysts, and a decrease in thymic hormone secretion (32). Nevertheless, one should take into account that these animals as well as IGF-I-null mice (33) do develop negative and positive selection, with rather normal relative numbers of CD4/CD8-defined thymocyte subsets. This fits with the hypothesis recently postulated that the main role of GH and IGF-I in the immune system is a general antistress effect (34). In this context, and if we take into account that the thymus is one of the most stress-sensitive organs in the body, promoting massive thymocyte death as one of the main responses to acute stress, a putative antistress role of GH-IGF-I-related circuits would be crucial for the homeostasis in this organ, with likely consequences for the maintenance of peripheral T cells. In this context, and considering that GH is presently being used as an adjuvant therapeutic agent in immunodeficiencies, including AIDS (23), the concepts defined herein provide important background knowledge for future GH-based immune interventions.
Acknowledgments
We thank Dr. Tsvee Lapidot for reading the manuscript, and Veronique Alves and Jo?o Hermínio Silva for technical assistance.
References
Savino W, Dardenne M 2000 Neuroendocrine control of the thymus. Endocr Rev 21:412–443
Savino W, Postel-Vinay MC, Smaniotto S, Dardenne M 2002 The thymus gland: a target organ for growth hormone. Scand J Immunol 55:442–452
Benoit C, Mathis D 1999 T-Lymphocyte differentiation and biology. In: Paul W, ed. Fundamental immunology. 4th ed. Philadelphia: Lippincott-Raven; 367–409
Savino W, Dalmau SR, Cotta-de-Almeida V 2000 Role of extracellular matrix mediated interactions in thymocyte migration. Dev Immunol 7:19–28
Savino W, Mendes da Cruz DA, Silva JS, Dardenne M, Cotta de Almeida V 2002 Intrathymic T cell migration: a combinatorial interplay of extracellular matrix and chemokines? Trends Immunol 23:305–313
Annunziato F, Romagnani P, Cosmi L, Lazzeri E, Romagnani S 2001 Chemokines and lymphopoiesis in the thymus. Trends Immunol 22:277–281
Ansel KM, Cyster JG 2001 Chemokines in lymphopoiesis and lymphoid organ development. Curr Opin Immunol 13:172–179
Taub DD, Tsarfaty G, Llyoyd AR, Durum SK, Murphy WJ 1994 Growth hormone promotes human T cell adhesion and migration to both human and murine matrix proteins in vitro and directly promotes xenogeneic engraftment. J Clin Invest 94:293–300
Mello-Coelho V, Villa Verde DMS, Dardenne M, Savino W 1997 Pituitary hormones modulate extracellular matrix-mediated interactions between thymocyte and thymic epithelial cells. J Neuroimmunol 76:39–49
Chen WY, Wight DC, Mehta BV, Wagner TE, Kopchick JJ 1991 Glycine 119 of bovine growth hormone is critical for growth promoting activity. Mol Endocrinol 5:1845–1852
Knapp JR, Chen WY, Turner ND, Byers FM, Kopchick JJ 1994 Growth patterns and body composition of transgenic mice expressing mutated somatotropin genes. J Anim Sci 72:2812–2819
Villa Verde DMS, Chammas R, Lagrota Candido JM, Brentani RR, Savino W 1994 Extracellular matrix components of the mouse thymic microenvironment. IV. Thymic nurse cells express extracellular matrix ligands and receptors. Eur J Immunol 24:659–664
Dalmau SR, Freitas CS, Savino W 1999 High expression of fibronectin receptors and L-selectin as a hallmark of early steps of thymocyte differentiation: lessons from sublethally irradiated mice. Blood 93:974–990
Livak KJ, Schittgen TD 2001 Analyses of relative gene expression data using real-time quantitative PCR and the 2-C(T) method. Methods 25:402–408
Scollay R, Butcher EC, Weissmann L 1980 Thymus cell migration. Quantitative aspects of cellular traffic from the thymus to the periphery in mice. Eur J Immunol 10:210–218
Gabor MJ, Godfrey DI, Scollay R 1997 Recent thymic emigrants are distinct from most medullary thymocytes. Eur J Immunol 27:2010–2015
Yanagawa Y, Iwabuchi K, Onoe K 2001 Enhancement of stromal cell-derived factor-1-induced chemotaxis for CD4/8 double-positive thymocytes by fibronectin and laminin in mice. Immunology 104:43–49
Vila-Coro AJ, Rodriguez-Frade JM, Martin de Ana AM, Moreno-Ortiz MC, Martinez-AC, Mellado M 1999 The chemokine SDF-1 triggers CXCR4 receptor dimerization and activates the JAK/STAT pathway. FASEB J 13:1699–1710
Dorskind K, Horseman ND 2000 The roles of prolactin, growth hormone, insulin-like growth factor-I, and thyroid hormones in lymphocyte development and function: insights from genetic models of hormone and hormone receptor deficiency. Endocr Rev 21:292–312
Kelley KW, Brief S, Westly HJ, Novakofski J, Bechtel PJ, Simon J, Walker EB 1986 GH3 pituitary adenoma cells can reverse thymic aging in rats. Proc Natl Acad Sci USA 83:5663–5667
Li YM, Brunke DL, Dantzer R, Kelley KW 1992 Pituitary epithelial cell implants reverse the accumulation of CD4–CD8– lymphocytes in thymus glands of aged rats. Endocrinology 130:2703–2711
Goya RG, Gagnerault MC, Leite de Moraes MC, Savino W, Dardenne M 1992 In vivo effects of growth hormone on thymus function in aging mice. Brain Behav Immun 6:341–350
Napolitano LA, Lo JC, Gotway MB, Mulligan K, Barbour JD, Schmidt D, Grant RM, Halvorsen RA, Schambelan M, McCune JM 2002 Increased thymic mass and circulating naive CD4 T cells in HIV-1-infected adults treated with growth hormone. AIDS 16:1103–1111
Cotta de Almeida V, Villa-Verde DMS, Lepault F, Dardenne M, Savino W 2004 Impaired migration of NOD mouse thymocytes: a fibronectin receptor-related defect. Eur J Immunol 34:1578–1587
Soriano SF, Hernanz-Falcón P, Rodriguez-Frade JM, Martin da Ana AM, Garzon R, Carvalho-Pinto C, Vila-Coro AJ, Zaballos A, Balomenos D, Martinez-AC, Mellado M 2002 Functional inactivation of CXC chemokine receptor 4-mediated responses through SOCS3 up-regulation. J Exp Med 196:311–321
Weber C, Alon R, Moser B, Springer TA 1996 Sequential regulation of 4?1 and 5?1 integrin avidity by CC chemokines in monocytes: implications for transendothelial chemotaxis. J Cell Biol 134:1063–1073
Shen W, Bendall LJ, Gottlieb DJ, Bradstock KF 2001 The chemokine receptor CXCR4 enhances integrin-mediated in vitro adhesion and facilitates engraftment of leukemic precursor-B cells in the bone marrow. Exp Hematol 29:1439–1447
Wright N, Hidalgo A, Rodrígues-Frade JM, Soriano SF, Mellado M, Parmo-Cabanas, Briskin MJ, Teixidó J 2002 The chemokine stromal cell-derived factor-1 modulates 4?7 integrin-mediated lymphocyte adhesion to mucosal addressin cell adhesion molecule-1 and fibronectin. J Immunol 168:5268–5277
Smaniotto S, Ribeiro-Carvalho MM, Dardenne M, Savino W, Mello-Coelho V 2004 Growth hormone stimulates the selective trafficking of thymic CD4+CD8– emigrants to peripheral lymphoid organs. NeuroImmunoModulation 11:299–306
Sotelo AI, Bartke A, Kopchick JJ, Knapp JR, Turyn D 1998 Growth hormone (GH) receptors, binding proteins and IGF-I concentrations in the serum of transgenic mice expressing bovine GH agonist or antagonist. J Endocrinol 158:53–59
Mello-Coelho V, Villa-Verde DMS, Farias-de-Oliveira DA, Brito JM, Dardenne M, Savino W 2002 Functional IGF-1-IGF-1 receptor-mediated circuit in human and murine thymic epithelial cells. Neuroendocrinology 75:139–150
Savino W, Smaniotto S, Binart N, Postel-Vinay MC, Dardenne W 2003 In vivo effects of growth hormone upon the thymus. Ann NY Acad Sci 992:179–185
Montecino-Rodrigues E, Clarck RG, Power-Braxton L, Dorshkind K 1997 Primary B cell development is impaired in mice with defects of the pituitary/thyroid axis. J Immunol 159:2712–2719
Dorshkind K, Horseman ND 2001 Anterior pituitary hormones, stress, and immune system homeostasis. Bioessays 23:288–294(Salete Smaniotto, Valéria)