Decrease of Bcl-xL and augmentation of thymocyte apoptosis in GILZ overexpressing transgenic mice

Glucocorticoids (GCs) are hormones and drugs that express their functions by binding and activating their specific intracellular receptor (GR), thus acting through genomic and nongenomic mechanisms. The thymus, one of their target organs, promotes thymocyte maturation to functional CD4⁺ or CD8⁺ single positive (SP) T lymphocytes that are ready to migrate to the periphery. T-cell development in the thymus is ordered by sequential steps that involve waves of cell proliferation and apoptosis. Apoptosis is a key process in thymus physiology and is triggered in 3 well-known ways: (1) negative selection, involving 5% of all thymocytes, eliminates autoreactive T-cell clones at the level of CD4⁺CD8⁺ double-positive (DP) cells; (2) death by neglect involves 90% of DP thymocytes that are neither positively nor negatively selected; and (3) stress-induced cell death.¹ 

GCs promote apoptosis of DP thymocytes²  but are not involved in apoptosis by negative selection.³  Their role in apoptosis by neglect is uncertain, but they certainly mediate stress-induced apoptosis.¹  The apoptotic signaling pathway triggered by the synthetic GC dexamethasone (DEX) in thymocytes has recently been clarified to some extent. The GC receptor coordinates activation of gene transcription and caspase-8, -9, and -3 in a sequence that leads to thymocyte apoptosis.^4-7 

Additionally, GC-mediated thymic apoptosis is regulated by many different molecules, and the B-cell leukemia 2 (Bcl-2) family of proteins is a critical regulator of apoptosis.⁸  The family is subdivided into antiapoptotic members such as Bcl-2 and Bcl-xL, and proapoptotic members such as Bcl-2-associated X protein (Bax) and BCL-2 homologous antagonist/killer (Bak) and the group of Bcl-2 homology 3 (BH3)-only death proteins.⁹  Specifically, Bcl-xL is expressed in the thymus only in DP cells¹⁰  and is essential for their function.¹¹  In the current model of GC-induced cell death, therefore, the GR, by binding to DNA, up-regulates expression of unknown proteins, which, in turn, influences the balance of Bcl-2 family members resulting in activation of different caspases.² 

GILZ (glucocorticoid-induced leucine zipper), one of several genes that are transcriptionally induced by GC-GR binding, was discovered in thymocytes committed to die as a consequence of DEX treatment. Its expression is up-regulated prominently in lymphoid organs.¹²  Interestingly, GILZ expression is reported to be induced in the kidney,¹³  in pluripotent mesenchymal cells,^14,15  and by GCs in erythroid progenitors,¹⁶  thus suggesting GILZ plays a role not only in the immune system but also in several other physiologic systems.

Within the immune system GILZ is strongly up-regulated by GC in the thymus, particularly, and in the bone marrow, spleen, and lymph nodes.¹²  Expressed in resting B lymphocytes, it is down-regulated when they are activated¹⁷ ; in macrophages GC and interleukin 10 (IL-10) up-regulate it.^18,19 

In thymic hybridoma cells, GILZ expression selectively protects from activation-induced cell death, as triggered by anti-CD3 monoclonal antibody (mAb) but not from apoptosis induced by other apoptotic stimuli.¹²  Moreover, GILZ inhibits T-cell receptor-induced IL-2/IL-2 receptor expression and nuclear factor κB (NF-κB) activity, by blocking nuclear translocation and DNA binding through a direct protein-to-protein GILZ/NF-κB interaction.²⁰  GILZ-mediated inhibition of Fas/FasL-dependent apoptosis may be explained in part by its inhibition of NF-κB. The role of GILZ overexpression in cell death in the absence of apoptogenic stimuli has not yet been evaluated. GILZ is hypothesized to play a role in regulating T-cell activation because its expression is down-regulated by T-cell receptor (TCR) triggering.²⁰  It inhibits the mitogen-activated protein kinase (MAPK) pathway by binding to RNA polymerase-activating factor-1 (Raf-1),²¹  and directly binds and inhibits activator protein 1 (AP-1).²² 

Up to now, as the function of GILZ in the thymus has been analyzed only in in vitro experimental systems, there is a lack of information about its physiologic and/or pathologic role that can only be obtained from in vivo studies. Transgenic (TG) mice, with transgene expression limited to the T-cell lineage, constitute an excellent in vivo model for investigating the influence of a molecule in the thymus.

In this study, we produced TG mice in which murine GILZ is controlled by the T-cell-specific human CD2 (hCD2) promoter and enhancer that program T-cell-specific transgene expression and are expressed at high levels in all thymocyte subsets, including double negative (CD4^-CD8^-, DN), DP, and CD4⁺ or CD8⁺ SP cells.²³  Results indicate that GILZ increased the apoptosis rate of thymocytes and decreased Bcl-xL expression. Increased apoptosis correlates with changes in thymocytes and is associated with altered T-lymphocyte subpopulation counts.

The transgene construct was made by cloning 874-base pair (bp) cDNA fragment containing the full open reading frame (ORF) of GILZ into the hCD2 expression vector (Figure 1A). The construct was injected into fertilized mouse embryos, and the generation of TG mice (DBA × C57BL/6) was tested by Southern blot analysis of their tail genomic DNA (Figure 1B). TG mice were backcrossed to C57BL/6 mice 8 generations, and TG mice were compared with WT littermates in the same experiments. Two independent transgenic lines were produced.

Thymi were weighed, and the cells were counted with a hemocytometer. Single thymocyte suspensions were cultured in RPMI 1640 medium containing 5% fetal calf serum (FCS), 100 U/mL penicillin/streptomycin, 0.1 mM nonessential amino acids, 5.5 × 10² μM β-mercaptoethanol (β-ME; GIBCO Invitrogen, San Giuliano Milanese, Italy) in flat-bottomed, 96-well plates (Costar, Cambridge, MA).

Total RNA was extracted using TRIzol (GIBCO Invitrogen). The assay was performed using the RPA III kit (Ambion, Austin, TX). Total RNA was hybridized with GILZ and the β-actin antisense probe. Samples were processed according to the manufacturer's instructions and loaded on to a 6% polyacrylamide/8 M urea denaturing gel. Autoradiographic exposure was carried out using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). To evaluate Bcl-2 family member mRNA expression, an mAPO-2 multiprobe template set from Pharmingen (San Diego, CA) was used.

Total proteins were separated on a 12% or 15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Primary Abs were rabbit polyclonal antiserum anti-GILZ, anti-caspase-8 (Santa Cruz Biotechnology, Santa Cruz, CA), anti-caspase-3, anti-caspase-9 (Cell Signaling Technology, Beverly, MA), anti-Bcl-xL (R&D Systems, Minneapolis, MN), anti-β-tubulin mAbs (Sigma, St Louis, MO) and anti-glutathione S-transferase (GST) mAb (Amersham, Buckinghamshire, United Kingdom). Secondary Abs were horseradish peroxidase-labeled goat anti-rabbit or anti-mouse (Pierce, Rockford, IL).

The following mAbs were used: fluorescein isothiocyanate (FITC)-conjugated rat anti-mouse CD8a (clone 53-6.7), Cy-Chrome anti-mouse CD8a (clone 53-6.7), phycoerythrin (PE)-conjugated rat anti-mouse CD4 (clone H129.19), and FITC-conjugated hamster anti-CD3 (clone 145-2C11). All mAbs and their isotype controls were purchased from Pharmingen (San Diego, CA).

Apoptosis was evaluated by TUNEL (transferase-mediated deoxyuridine triphosphate [dUTP] nick end labeling) technology using an In Situ Cell Death Detection Kit (Roche Diagnostics, Mannheim, Germany). Briefly, cells (2 × 10⁶/mL) were washed in phosphate-buffered saline (PBS) and fixed in 4% formaldehyde for 60 minutes at room temperature (RT), washed again, and incubated on ice for 2 minutes with the permeabilization solution (0.1% Triton X-100 in 0.1% sodium citrate). Cells were then stained by incubating them with 50 μL TUNEL reaction mixture for 60 minutes at 37°C. Cells were analyzed by flow cytometry.

Apoptosis was also measured by flow cytometry. After 24 hours of culture, cells were centrifuged, and the pellets were gently resuspended in 1.5 mL hypotonic propidium iodide (PI) solution (50 μg/mL in 0.1% sodium citrate plus 0.1% Triton X-100). The tubes were kept in the dark at 4°C for 1 hour. The PI fluorescence of individual nuclei was measured by flow cytometry using standard FACScan equipment (Becton Dickinson, Franklin Lakes, NJ). The nuclei traversed a 488-nm Argon laser light beam. A 560-nm dichroid mirror and a 600-nm band pass filter (band width 35 nm) were used to collect the red fluorescence because of PI DNA staining. The data were recorded in logarithmic scale in a Hewlett Packard (HP 9000, model 310; Palo Alto, CA) computer. The percentage of apoptotic cell nuclei (subdiploid DNA peak in the DNA fluorescence histogram) was calculated with specific FACScan research software (Lysis II).

The thymocyte suspension was incubated in complete medium with 10 μg/mL JC-1 for 10 minutes in the dark at RT. Cells were then washed, resuspended in 400 μL, and analyzed by flow cytometry. JC-1 was from Molecular Probes (Eugene, OR)

The GILZ-GST-TAT expression vector was constructed by inserting GILZ cDNA in the TAT-C vector to produce an in-frame fusion protein. Fusion protein were purified as previously described.¹² 

[3H]-thymidine (2.5 μCi [0.0925 MBq]) per well was added 15 hours before harvesting cells. Cells were collected with a multiple suction-filtration apparatus (Mash II) on a fiberglass filter (Whittaker, Walkersville, MD) and counted in a β counter (Packard, Milan, Italy).

Data are presented as the mean ± 1 standard error of the mean (SEM). Student t test was calculated for all suitable experiments. Values with P greater than .05 were not considered significantly different (^*P < .05, ^**P < .005, ^***P < .0005).

Generation of TG mice was tested and confirmed by Southern blot of DNA extracted from their tail (Figure 1A,-B and “Materials and methods”).

GILZ expression in the thymus was determined by using the RPA and Western blot analysis on TG and WT thymocytes to assess the amount of GILZ mRNA and protein, respectively. Basal levels of GILZ mRNA were about 5-fold higher in TG than in WT thymocytes, as shown by the quantitative analysis (Figure 1D) of a representative RPA (Figure 1C).

Western blots showed protein levels (Figure 1E) were higher in the thymus of TG than in WT mice. GILZ overexpression was compared in TG and in WT thymocytes from mice that were left untreated or treated with different doses of DEX in dose-response experiments. As shown in Figure 1E, thymic GILZ protein levels in 3 TG mice were higher than in untreated WT control mice (in which the protein was barely detectable) and were similar to levels in WT mice treated with 0.1 to 1 mg/kg DEX. To prevent confounding effects which could be due to genetic differences (other than the transgene) between the TG and WT mice, TG mice were backcrossed to C57BL/6 mice 8 generations, and single experiments were performed comparing TG and WT littermates.

To determine whether T-cell development was influenced by constitutive GILZ overexpression, thymi from TG and WT animals were weighed, total thymocytes were counted by hemocytometer, and thymocyte subsets were analyzed by flow cytometry using mAbs specific for CD4 and CD8 T-cell surface antigens. As shown in Figure 2A, in young adult mice, thymus weight was not significantly different in TG or WT mice. Figure 2B shows the total thymocyte count from young adult TG mice was not significantly different than in controls.

The relative percentage of thymocyte subsets was analyzed by 2-color flow cytometry. Figure 2C shows there was a significant (P < .05) decrease in DP (from 79.4% ± 0.9% to 74% ± 1.8%), and a parallel consistent (P < .05) increase in DN (from 4.6% ± 0.5% to 6% ± 0.3%) and CD8⁺ SP (from 4.3% ± 0.4% to 6.2% ± 0.8%) cells. No significant differences were detected in the relative number of the CD4⁺ SP subset. Thus, basal GILZ overexpression appears to perturb thymic subsets significantly in young adult TG mice.

To see whether increased cell death contributed to the decrease in DP thymocytes, apoptosis was evaluated by TUNEL technology on fresh thymocytes harvested immediately after mice were killed. TUNEL technology detects apoptosis in vivo in the thymus (usually ranging from 0.5% to 1%²⁴) before apoptotic cells are cleared by macrophages. Figure 3A shows that apoptosis of fresh thymocytes from TG was notably higher than WT control mice (from 0.7 ± 0.1 to 5.1 ± 0.1; P < .0005).

Increased apoptosis of TG thymocytes prompted us to investigate the role of the caspase cascade. Analysis was performed on cultured thymocytes because the apoptosis signaling pathway in fresh thymocytes is difficult to detect because of the in vivo apoptotic cell clearance by macrophages.²⁴  We first examined caspase-3, which is considered one of the terminal caspases involved in DEX-induced apoptosis.^5,7  After cells from TG or WT mice were harvested, 1 aliquot was taken for immediate total protein extraction, and the others were cultured for 30 minutes, 1 hour, or 2 hours before extracting the proteins to be used in Western blot experiments. Figure 3B shows that activated cleaved caspase-3 was absent in proteins from freshly isolated WT thymocytes (0 h) and after 30 minutes; its activation was detectable after 1 hour of culture but was fully evident after 2 hours of culture. In TG thymocytes caspase-3 activation, after 1 hour or 2 hours, was stronger than in WT thymocytes. Analysis of β-tubulin showed a similar protein content in WT and TG samples. So increased apoptosis in TG thymocytes correlated with augmented caspase-3 activation.

The executioner caspase-3 is activated by at least 2 different pathways. The mitochondrial pathway leads to sequential activation of caspase-8, BID, release of cytochrome c from mitochondria, and activation of caspase-9 which directly cleaves and activates caspase-3. The second pathway involves activation of caspase-8 that directly activates caspase-3.²⁵  We analyzed both pathways to see which accelerates activation of caspase-3 in TG mice. The filters we had used to analyze caspase-3 activation by Western blot were again adopted to probe caspase-8. It was activated after 30 minutes (Figure 3B) and activation increased after 1 or 2 hours of culture. In WT thymocytes caspase-8 was weakly active only after 1 or 2 hours of culture. Western blots were performed to determine whether caspase-8 activated caspase-3 directly or through the so-called “mitochondrial pathway.” Caspase-9 activation was not detectable at any time either in TG or WT thymocytes, suggesting caspase-3 is directly activated by caspase-8.

The change in mitochondrial transmembrane potential (Δψ_m) is widely accepted as an index of mitochondrial activity because Δψ_m is involved in release of cytochrome c and the ensuing caspase-9 activation in the mitochondrial pathway. Figure 3C shows a representative experiment. Δψ_m, as evaluated by JC-1 flow cytometry analysis, was at the same levels in TG thymocytes and WT thymocytes after 0, 1, or 2 hours of culture, providing further evidence that in our culture system caspase-3 is not activated through the mitochondrial/caspase-9 pathway.

In vitro pull-down experiments with the fusion protein GILZ-GST demonstrated that GILZ activated caspase-8 neither by directly binding to it nor by binding to Fas-associated death domain protein (FADD), the caspase-8 recruiting molecule (not shown). However, we cannot exclude that the interactions happen in vivo.

Whether GILZ inhibits anti-CD3-induced apoptosis in TG thymocytes, as has been reported in hybridoma cell lines,^12,20  was also tested. Figure 3D shows anti-CD3 triggering significantly increased apoptosis in WT but not in TG thymocytes, concurring with previously published data.

Caspase activation is regulated by balancing expression of antiapoptotic (Bcl-2, Bcl-xL) and proapoptotic (Bax, Bak) Bcl-2 family molecules. Bcl-xL is almost exclusively expressed in, and selectively regulates, DP but not DN or SP thymocyte apoptosis.¹⁰  With the aim of dissecting the role of Bcl-2 family, and particularly of Bcl-xL, RPA experiments with a multiprobe template kit and Western blotting were performed in WT or TG thymocytes to assess the mRNA and protein levels, respectively. Figure 4A shows a representative experiment (left) and quantitative analysis of all experiments (right). The antiapoptotic Bcl-xL mRNA levels were significantly lower in fresh thymocytes from TG mice than in controls, whereas mRNA levels of the proapoptotic Bax or Bak were unchanged. Expression of the other family molecules was too low for quantification. Low mRNA levels correlated with a parallel significant decrease in the amount of protein as shown in a representative experiment (Figure 4B, left side) and quantitative protein analysis from a series of experiments (right side). Thus, thymocytes from TG mice have a reduced Bcl-xL mRNA and protein expression which may account for increased thymocyte apoptosis.

During the course of this study, we observed that cultured thymocytes induced GILZ expression even in WT cells with their very low or undetectable basal GILZ expression. Expression induction by cell culture is a known property of some molecules because also serum- and glucocorticoid-regulated kinase 1 (SGK-1) is induced by glucocorticoids and by culture serum.²⁶  We exploited this feature to correlate GILZ and Bcl-xL expression and caspase-8 and -3 activation in time-course experiments by culturing aliquots of WT or TG thymocytes for 30 minutes, 1 hour, or 2 hours before extracting the RNA or proteins to be used in RPA or Western blot experiments, respectively. Figure 5A shows very low GILZ expression in proteins from freshly isolated WT thymocytes (0 h), which increased after 30 minutes, 1 hour, or 2 hours of culture; TG thymocytes showed the same phenomenon, although GILZ expression was higher. The same filters, reprobed for expression of Bcl-xL, showed a marked decrease in Bcl-xL expression in WT thymocytes starting 30 minutes after the increase in GILZ (1-hour culture). Bcl-xL expression in TG thymocytes was very low (as shown in Figure 4) in freshly isolated thymocytes and remained unchanged at all time points. In RPA (Figure 5B), Bcl-xL mRNA decreased after 1 hour of culture; the decrease was significantly stronger after 2 hours in WT thymocytes, reaching levels similar to the TG thymocytes. Bax or Bak mRNA from WT or TG thymocytes did not show any significant change with culture, thus demonstrating a specific effect on Bcl-xL expression.

The decrease in Bcl-xL correlated with apoptosis of WT or TG thymocytes cultured for 24 or 48 hours. As shown in Figure 5C, after 24 hours, apoptosis in TG thymocytes was still significantly higher than in WT cells, despite their Bcl-xL decreased expression, although the differences were smaller when compared with apoptosis in fresh thymocytes (compare with Figure 3). After 48 hours of culture, differences in apoptosis rates were further decreased and became nonsignificant.

To further support the correlation between GILZ, Bcl-xL, and thymocyte apoptosis, we produced a fusion protein GST-TAT-GILZ in which the transactivator of transcription (TAT) peptide pulls the fusion protein across the cell membrane, inside cells²⁷  (Figure 6A). WT thymocytes, treated with GST-TAT-GILZ (0.2 μg/mL) showed a marked decrease in Bcl-xL protein expression after 2 hours of culture when compared with controls (Figure 6B). This was followed (after 24 hours) by a dramatic increase in apoptosis compared with the group treated with the GST-TAT control fusion protein (Figure 6C, left panel). As a control of specificity, the GST-TAT-GILZ fusion protein did not increase apoptosis of spleen cells (Figure 6C, right panel). Our controls, in agreement with previous reports,²⁷  demonstrated that TAT fusion proteins are not toxic.

In young adult mice, functional phenomena (namely proliferation versus apoptosis) balance thymus weight, total cell number, and small changes in the relative number of thymocyte subsets. As we wondered whether differences are amplified as mice age, elderly animals (between 18 and 20 months of age) were killed, their thymocytes counted and analyzed by flow cytometry to determine the relative number of cell subsets. Figure 7A shows no differences were seen in total thymocyte number, whereas strongly significant differences which were not seen in young adult mice appeared in subset cell numbers (Figure 7B). Specifically, the number of CD4⁺ SP TG thymocytes was significantly greater than WT thymocytes (9 ± 0.7 and 12.7 ± 0.6 in WT and TG, respectively); DP cells from aged TG mice were very low (81.5 ± 1.3 WT versus 62.8 ± 3.6 TG); more DN (5.7 ± 0.7 WT versus 9.9 ± 0.8) and CD8⁺ SP cells (3.8 ± 0.3 WT versus 14.7 ± 3.7 TG) were found in aged mice. In conclusion, phenotypic experiments in aged mice suggest age-related impairments in the thymus subsets in elderly TG mice.

Finally, the proliferation potential of thymocytes from young adult mice was tested. Aliquots of each isolated thymic subset were cultured alone or with concanavalin A (ConA) mitogen for thymidine uptake experiments and analyzed after 48 hours of culture. As shown in Figure 7C, despite the appearance of thymidine uptake greater in TG than in WT thymocytes (mean values ± 1 SEM) in the CD4⁺ SP cells, there was no statistical difference between the 2 samples (P > .05).

In this study, we demonstrated that GILZ is involved in thymic apoptosis. Three experimental systems were used: (1) overexpressing-GILZ transgenic mice, (2) cultured WT thymocytes, and (3) a GILZ-TAT fusion protein. Thanks to these systems, it has been shown that GILZ promotes spontaneous thymic apoptosis (by stress and/or neglect) and inhibits apoptosis by negative selection. Moreover, analysis of the biochemical events determining spontaneous apoptosis possibly delineates an unknown pathway based on a decrease in Bcl-xL and activation of caspase-8 and -3 without involving the mitochondria. Loss of GILZ experimental systems will tell us whether the reported observations are important events in the physiology of thymus and in development of normal and autoreactive T lymphocytes.

Transgenic mice with the transgene specifically expressed in the T-cell lineage appeared to be a suitable experimental system. We chose the vector based on the hCD2 minigene,²³  with the gene of interest under the control of the hCD2 promoter and locus control region. This expression vector confers 3 important features on the TG mice: (1) tissue specificity, because the gene is expressed only in the T-cell lineage (thymocytes and peripheral T lymphocytes); (2) copy dependence; and (3) position-independent expression of the gene.^23,28,29  Expression studies performed by RPA and by Western blot confirmed the thymus in TG mice expressed about 5-fold more mRNA and protein than in WT mice. GILZ is strongly up-regulated by GC treatment. In our TG mice GILZ overexpression is constitutive, matching physiologic levels of expression achieved with 0.1 to 1 mg/kg DEX, a dose that causes low 15% to 20% thymocyte apoptosis (data not shown).

No significant differences were detected in thymus weight or total thymocyte number. A significant decrease in DP and increases in DN and CD8⁺ SP thymocytes in young adult mice were amplified with aging.

The functional analysis focused on thymocyte apoptosis. Although 95% thymocytes undergo cell death,³⁰  apoptosis detected in vivo on fresh thymocytes is usually very low (0.5%-1%) because of continuous clearance of apoptotic cells.^24,31  Our results indicate a strong increase, of up to 6%, in ex vivo thymic apoptosis, as evaluated by TUNEL, in TG mice, which was associated with an augmented activation of the caspase-8/caspase-3 pathway. This rise in apoptosis apparently contrasts with reports, indicating that GILZ inhibits apoptosis induced by anti-CD3 stimulation^13,20  or IL-2 withdrawal.³²  However, some molecules are known to play opposing roles in apoptosis. GCs promote thymic apoptosis but inhibit ovary apoptosis.³³  Moreover, although they determine thymocyte apoptosis, they also protect thymocytes from anti-CD3-induced apoptosis, a mechanism known as mutual exclusion. Our data suggest that GILZ behaves like GCs.

The GILZ-activated caspase-8/-3 apoptotic pathway we observed strengthened the similarities with GC-induced apoptotic mechanisms. This raises the question of whether GILZ is one of the molecules that are part of the apoptotic pathway triggered by GCs in thymocytes. GILZ is transcriptionally induced by DEX and DEX triggers apoptosis by achieving gene transcription and protein synthesis.^5,7  In addition, like GCs, GILZ determines apoptosis by sequential activation of caspase-8/caspase-3.⁷  Opinions are divergent on caspase-9 activation. Some reports claim it is crucial in DEX-induced apoptosis,^6,34,35  whereas others claim it amplifies DEX-induced apoptosis after triggering by caspase-8/caspase-3 activation.^7,36  We therefore speculate that GILZ is involved in the caspase-9-independent part of GC-induced thymocyte apoptosis.

Of interest, increased thymocyte apoptosis was a late event in cultured cells (24 hours), following the Bcl-xL decrease (2 hours), thus suggesting that the former event was possibly a consequence of the latter. This strongly suggests that decreased expression of Bcl-xL is limited to the DP cells, and that the increased apoptosis may be confined inside the DP subset because Bcl-xL is expressed almost exclusively in DP cells and selectively regulates their survival but does not influence the survival of SP cells.¹⁰  However, concurring with the small changes observed in our young adult TG mice, Bcl-xL down-regulation in Bcl-xL^+/- and Bcl-xL^-/- mice is associated with normal thymocyte and peripheral T-cell counts.¹⁰  It seems improbable that diminished Bcl-xL expression was due to fewer DPs because the DP decrease was only about 5%, whereas 50% less Bcl-xL was expressed in TG than in WT thymocytes. Furthermore, the unchanged proapoptotic Bax and Bak expression confirms the specificity of the Bcl-xL decrease. It is hardly surprising that a strong percentage decrease in Bcl-xL expression did not determine apoptosis of a similar percentage of thymocytes. In fact, it has previously been shown that in Bcl-xL^+/- mice with a 50% Bcl-xL decrease, “TUNEL-positive apoptotic cells increased drastically from 1.2% to 4.3% in male fetal germ cells.”^37,p210 Administration of GILZ protein in fusion with the TAT peptide strengthened the correlation with the Bcl-xL down-regulation and with the increase of apoptosis. We do not know what mechanism underlines the GILZ inhibition of Bcl-xL. Because GILZ is able to bind to the gene promoter, thus acting as a transcriptional repressor,¹⁵  it may directly repress Bcl-xL transcription. However, Bcl-xL could also be repressed indirectly as the NF-κB family directly activates Bcl-xL³⁸  and protects from apoptosis.³⁹  As GILZ binds and inhibits NF-κB,²⁰  one might presume it counters both expression of Bcl-xL and its antiapoptotic effect, thus promoting apoptosis. The decrease in Bcl-xL usually triggers the mitochondrial apoptotic pathway, with release of cytochrome c in the cytoplasm and activation of caspase-9 “apoptosome.”²⁵  In our model, decreased Bcl-xL did not correlate with caspase-9 activation, suggesting alternative “non-mitochondrial” mechanisms. For example, like Bcl-2 the antiapoptotic prototype of the family, Bcl-xL may regulate caspase activation independently of the cytochrome c/caspase-9 apoptosome.³⁶  Alternatively, Bcl-xL may regulate activation of caspase-8 which is bound to the outer mitochondrial membrane⁴⁰  or present inside the endoplasmic reticulum.⁴¹  Altered Bcl-xL expression features in many diseases such as tumors,⁴²  disorders in the central nervous system,⁴³  and vascular diseases.⁴⁴  Specifically, Bcl-xL is overexpressed in many cancers, including different lymphomas in which it (1) increases survival of cancer cells, thus augmenting their invasive and metastatic abilities, and (2) determines resistance of cancer cells to chemotherapeutic agents. The GILZ-dependent down-regulation of Bcl-xL expression indicates GILZ may play a role in protecting cells from the tumorigenic potential of Bcl-xL overexpression⁴⁵  and/or in rendering cancer cells responsive to antitumor therapies.

As we also wondered whether the dysfunctions could disrupt thymus balance over time, we analyzed thymic cells from aged mice. Although the total thymocyte count was normal, the thymic perturbation that we had observed in young TG mice was amplified in older animals: 20% fewer DP cells than in WT mice, the differences in CD8⁺ SP and DN counts were even greater, and an increase appeared in CD4⁺ SP cells. This pattern of amplified thymocyte number perturbation with aging concurs with previous reports in which a 2-fold increase Bcl-xL in thymocytes of Bcl-xL TG mice is associated with age-related abnormalities in thymic subset counts.⁴⁶  Further studies of in vivo challenge of TG mice will show whether thymic impairments generate an abnormal immune response and, possibly, pathologic alterations.

In conclusion, constitutive GILZ overexpression in the thymus of TG mice determined diverse functional and phenotypic abnormalities. Of importance, this in vivo model will be used to investigate the role of GILZ in the anti-inflammatory and immunosuppressive actions of glucocorticoids, with the aim of selecting specific therapeutic effects without the wide array of side effects that limit the pharmacologic usage of GCs.

We thank Dr Geraldine A. Boyd for her help in writing this paper in English, D. Kioussis for providing the hCD2 vector, and L. D'Adamio for providing the TAT-C vector.