|Home | About | Journals | Submit | Contact Us | Français|
Glucocorticoids (GCs) cause cell cycle arrest and apoptosis in lymphoid cells which is exploited to treat lymphoid malignancies. The mechanisms of these anti-leukemic GC effects are, however, poorly understood. We previously defined a list of GC-regulated genes by expression profiling in children with acute lymphoblastic leukemia (ALL) during systemic GC monotherapy and in experimental systems of GC-induced apoptosis. PLZF/ZBTB16, a transcriptional repressor, was one of the most promising candidates derived from this screen. To investigate its role in the anti-leukemic GC effects, we performed overexpression and knock-down experiments in CCRF-CEM childhood ALL cells. Transgenic PLZF/ZBTB16 alone had no detectable effect on cell proliferation or survival, but reduced sensitivity to GC-induced apoptosis but not apoptosis induced by antibodies against Fas/CD95 or 3 different chemotherapeutics. Knock-down of ZBTB16 entailed a small, but significant, increase in cell death induction by GC. Affymetrix Exon array-based whole genome expression profiling revealed that PLZF/ZBTB16 induction did not significantly alter the expression profile, however, it interfered with the regulation of numerous GC response genes, including BCL2L11/Bim, which has previously been shown to be responsible for cell death induction in CCRF-CEM cells. Thus, the protective effect of PLZF/ZBTB16 can be attributed to interference with transcriptional regulation by GC.
Glucocorticoids (GCs) induce apoptosis in certain cells of the lymphoid lineage and are therefore useful therapies in the treatment of lymphoid malignancies, most notably childhood acute lymphoblastic leukemia (ALL) . GCs mediate their effects via the GC receptor (GR), a ligand-activated transcription factor of the nuclear receptor super-family that resides in the cytoplasm and, upon ligand binding, translocates into the nucleus, where it modulates gene expression via binding to specific DNA response elements or by protein–protein interactions with other transcription factors [2–5]. A large number of genes have been identified that are regulated by GCs in lymphoid lineage cells in experimental systems  and related clinical samples [7,8], but the genes responsible for cell death induction and other effects of GCs on the immune system are not well understood (for recent reviews see [6,9–12]).
One of the most interesting GC-regulated candidate genes identified in the above screen was ZBTB16 (zinc finger and BTB domain containing gene 16), also referred to as promyelocytic leukemia zinc finger (PLZF, reviewed in [13,14]). It was induced by GC in the majority of children with precursor B-cell and T-cell ALL, in one patient with adult ALL and in peripheral blood lymphocytes from 2 healthy donors treated with GC, but not in 2 ALL cell lines . More recently, we re-analyzed these data using a normalization procedure that better resolves regulations of low abundance genes (GC-RMA)  and also performed additional expression profiling studies of 14 more ALL children (J. Rainer, in preparation) and found that PLZF/ZBTB16 was induced by GC more than 2-fold in 23 of 27 such cases.
PLZF/ZBTB16 was initially identified as a fusion partner of retinoic acid receptor α (RARα) in a variant chromosomal translocation t(11;17)(q23;q21) that occurred in a subset of acute promyelocytic leukemia patients [16,17]. PLZF/ZBTB16 is a transcriptional repressor of the POK (POZ and Krüppel) family of proteins. It contains one BTB (Broad complex, Tramtrack, and Bric à brac)/POZ (poxviruses and zinc finger and Krüppel) domain at the NH2-terminal moiety and 9C2H2 Krüppel-type zinc fingers at the carboxy-terminal end of the protein. The POZ/BTB domain mediates interactions with proteins such as transcriptional co-repressors entailing chromatin remodeling and transcriptional silencing. The Krüppel-type zinc fingers confer specificity of the repressor activity to particular promoters by interacting with corresponding response elements in regulatory regions of genes repressed by PLZF/ZBTB16 [18–21]. The hinge region of the protein contains a PEST domain with two consensus sites for CDK2-mediated phosphorylation that triggers ubiquitination and subsequent degradation of PLZF through the ubiquitin-proteasome pathway . The human PLZF/ZBTB16 gene maps to chromosome 11q22–q23 with seven exons  distributed over a region of approximately 200 kb. Although additional alternative transcripts encoding distinct proteins have been reported , most recent NCBI and Ensembl databases contain 2 and 3 transcripts, respectively, that differ only in their 5′ untranslated region and thus encode the same protein.
Regarding its function, a natural mutation (luxoid) in, and knock-out of, the mouse homologue Zfp145/ZBTB16 unraveled a crucial role in limb and skeletal patterning and spermatogonial stem cell maintenance [25–27]. PLZF/ZBTB16 has further been implicated in tumor suppression in melanoma  and prostate cancer , ascribed to its ability to cause cell cycle arrest and induce apoptosis in certain cell systems. The complex effects of PLZF/ZBTB16 have been associated with transcriptional repression of numerous genes such as members of the Hox family of transcription factors , kit , CRABPI , c-myc , CCNA2/Cyclin A, CDKN1B/p27/Kip1  and possibly others [26,28,29,33,34]. Thus, its suggested role in cell cycle arrest and apoptosis induction in some systems [28,31,34,35], further supported a possible role of PLZF/ZBTB16 induction in the anti-leukemic GC response.
In this study, we extended our analysis of GC regulation of PLZF/ZBTB16 to several leukemic cell lines and performed functional analyses on the role of PLZF/ZBTB16 in the anti-leukemic effects of GC. We found that transgenic PLZF/ZBTB16 alone had no effect on cell proliferation and survival, but reduced sensitivity to apoptosis induced by the GC analogue dexamethasone in the CCRF-CEM model for T-ALL. Similarly, knock-down of PLZF/ZBTB16 in CCRF-CEM cells entailed a small, but significant, increase in cell death induction by dexamethasone. We further exploited whole genome expression profiling to derive a molecular explanation for the protective effect of PLZF/ZBTB16 on GC-induced apoptosis.
The T-ALL cell lines CCRF-CEM-C7H2 , CEM-C7H2-2C8 , a CEM-C7H2 derivative with constitutive expression of the tetracycline-regulated reverse transactivator, rtTA , MOLT4 (CRL-1582, ATCC, Rockville, MD), and Jurkat (untransfected and a rat GR-transfected derivative ), the precursor B-cell lines 697/EU-3 (ACC 42, DSMZ, Braunschweig, Germany), NALM6 (ACC 128, DSMZ), RS4;11 (ACC 508, DSMZ) and AT-1 (kindly provided by R. Panzer Grümeier, Vienna), and the Burkitt's lymphoma Daudi (CCL-213, ATCC) were cultured in RPMI 1640 supplemented with 10% fetal calf serum and 2 mM l-glutamine at 37 °C, 5% carbon-dioxide and saturated humidity. HEK293T packaging cells (ATCC, Manassas, VA) were cultured in DMEM supplemented as above. The cells were free of mycoplasma infection and their authenticity was verified by DNA fingerprinting, as detailed previously . Taxol was dissolved in DMSO (10 mg/ml), doxycycline (100 μg/ml) in phosphate buffered saline, doxorubicin (10 mg/ml) and mitomycin C (500 μg/ml) in destilled water, and β-dexamethasone (10−4 M) in 100% ethanol. The final ethanol-concentration in the dexamethasone-treated and control cultures was maintained at 0.1%. All above reagents were from Sigma (Vienna, Austria).
Apoptosis was determined by FACS analysis of propidum iodide (PI)-treated permeabilized cells , as previously detailed . Even though this method may over-estimate the rate of apoptotic cells as compared to other methods, it allows simultaneous analysis of cell death induction and cell cycle progression in the same samples. Moreover, since this study addresses whether PLFZ/ZBTB16 reduces or increases cell death, the absolute extent of apoptosis does not affect its conclusions. Cells were analyzed with a FACScan cytometer (Becton Dickinson Biosciences, San Jose, CA) in combination with CellQuest Pro software (Becton Dickinson Biosciences) acquiring forward scatter/sideward scatter, FL-2 (log), and FL-3 (linear). In FL-2, the percentage of nuclei with reduced DNA-content (subG1 peak) was assessed. For cell cycle determination, the FACS data from PI-treated nuclei were analyzed using ModFit LT 2.0 (Verity, Topsham, ME) with the following quality parameters: minimal number of events analyzed: 10,000 and maximal %CV: 8. For the knock-down experiments, the percentage of viable cells in culture was determined by staining cell suspensions with 1 μg/ml 7-amino-actinomycin D (7-AAD; Sigma–Aldrich) plus APC-coupled Annexin-V (Becton Dickinson) and analyzing the samples in a FACScan (Becton Dickinson). Cells staining positive for Annexin-V and/or 7-AAD were defined as undergoing cell death.
Our immunoblotting procedure has been described in detail recently . Briefly, proteins were extracted from 5 × 106 cells in 100 μl RIPA-buffer, quantified by Bradford analysis, mixed with 40 μl loading buffer (4× SSB, 5% β-mercaptoethanol), denatured, fractionated on a 12.5% SDS-PAGE and electroblotted onto nitrocellulose. The membranes were incubated overnight with rabbit polyclonal antibodies against PLZF/ZBTB16 (HPA001499, Prestige antibodies, Sigma, Vienna, Austria) or BCL2L11/Bim (#559685, Becton Dickinson Biosciences, San Jose, CA), or mouse monoclonal antibodies against α-tubulin (DM1A, CalBiochem, Nottingham, UK) as a loading control. Specifically bound antibodies were detected with anti-rabbit or anti-mouse horseradish-peroxidase-conjugated secondary antibodies (Amersham Pharmacia Biotech, Uppsala, Sweden) and visualised by chemiluminescence (ECL, Amersham) and subsequent exposure to AGFA Curix X-ray films for 1 s to 30 min.
The lentiviral conditional expression construct pHR-tetCMV-ZBTB16 (U317) was generated using the GATEWAY™ technology (Invitrogen, Carlsbad, CA). The details of this procedure and the generation of stable clonal cell lines with tetracycline-regulated expression of cDNAs cloned into such constructs has been described previously . In brief, human PLZF/ZBTB16 mRNA was PCR-amplified from IRAT p970B0341D6 (RZPD, Germany) using forward primer 5′-CAAAAAAGCAGGCTCCGCCACCATGGATCTGACAAAA-3′ and reverse primer 5′-CAAGAAAGCTGGGTCTTACACATAGCACAGGTAGAGG-3′, and (after a second PCR reaction to complete the attB sites) recombined into pDONR207 (Invitrogen) resulting in pENTR207-ZBTB16 (U316). The construct was sequence-verified and subsequently recombined into the “destination vector” pHR-tetCMV-Dest-IRES-GFP (U192) to generate pHR-tetCMV-ZBTB16-ires-GFP (U317). The lentiviral shRNA-expressing constructs pHR-THT-ZBTB16-shRNA1-SFFV-GFP (U547), pHR-THT-ZBTB16-shRNA2-SFFV-GFP (U548) and pHR-THT-cocntrol-shRNA-SFFV-GFP (U288) were generated as described previously . In brief, complementary shRNA oligonucleotides directed against PLZF/ZBTB16 and a non-target nucleotide sequence (control shRNA) containing BamHI and HindIII sticky ends were synthesized (MWG Ebersberg, Germany) sequences:
The annealed oligonucleotides were cloned into the BglII–HindIII sites of pENTR-THT (U191) and the THT-shRNA cassette was recombined into the lentiviral RNAi “destination” vector pHR-Dest-SFFV-eGFP (U218) to generate pHR-THT-shRNA-1-SFFV-GFP (U547), pHR-THT-shRNA-2-SFFV-GFP (U548) and pHR-THT-shRNA-control-SFFV-GFP (U288).
The above described lentiviral constructs for gene overexpression or knock-down were transfected into 293T packaging cells along with the packaging plasmids pSPAX and pVSV-G (kindly provided by Didier Trono) and the lentivirus-containing supernatants used to transduce CCRF-CEM derivatives. For conditional PLZF/ZBTB16 overexpression, CEM-C7H2-2C8 (which constitutively express rtTA ) were transduced and two clonal cell lines, termed CEM-C7H2-2C8-ZBTB16#19 (ZBTB16#19) and CEM-C7H2-2C8-ZBTB16#58 (ZBTB16#58), derived by limiting dilution cloning. For knock-down experiments, the lentivirus-containing supernatants were concentrated using poly-ethylene glycol  and used to transduce CEM-C7H2 and ZBTB16#58 cells. Over 85% GFP-expressing bulk populations (if required, after enrichment by FACS-sorting) were used for further experiments.
For PLZF/ZBTB16 mRNA quantification by real-time RT-PCR, 50 μl of diluted cDNA (2 ng/μl) were added to 50 μl of TaqMan Universal MasterMix (Applied Biosystems, Foster City, CA) and introduced into microfluidic cards containing real-time RT-PCR mixes for human PLZF/ZBTB16 (Hs00232313_m1, Applied Biosystems) and TBP (Hs00427620_m1) as normalization control according to the manufacturer's guidelines: After equilibration at RT, the channels were filled with 100 μl of reaction mix and centrifuged two times for 1 min at 1000 rpm. Thereafter, the cards were sealed, loaded into the HT7900 real-time machine (Applied Biosystems), and run with a 2-step PCR thermo-protocol that included an initial 94.5 °C step for 10 min followed by 40 cycles of 97 °C for 15 s alternating with 60 °C for 1 min. Fluorescence signal intensities were read during the 60 °C temperature step. Primary real-time PCR data analysis was performed with SDS software version 2.2.1, and further analysis was performed in R (version 2.8). Data from 3 technically replicated measurements were averaged and normalized to the internal TBP control (Hs.00427620, Applied Biosystems). Log 2-fold change values (M values) were calculated for 3 biological replicates by comparing normalized real-time RT-PCR data from GC-treated samples against data from the corresponding control samples. M values were averaged for the 3 biological replicates and p-values were calculated (Student's t-test) to test against the null hypothesis of no differential expression (mean M = 0). In some instances (knock-down experiments, determination of Bim mRNA), individual real-time RT-PCRs were performed using the same kit as above for PLZF/ZBTB16 and Hs00197982_m1 for BCL2L11/Bim (Applied Biosystems). For individual assays, 20 ng transcribed cDNA was amplified in TaqMan Universal MasterMix (Applied Biosystems) in 96-well plates (94.5 °C for 5 min, 40 cycles of 94 °C for 15 s alternating with 60 °C for 1 min) and fluorescence signal intensities were read during the 60 °C temperature step (iQ5 Multicolor Real-Time PCR Detection System; Bio-Rad Laboratories Ges.m.b.H., Vienna, Austria).
To determine PLZF/ZBTB16 response genes, C7H2-2C8-ZBTB16#19 and #58 cells were cultured in duplicate in the absence (set A) or presence (set B) of 400 ng/ml doxycycline for 2, 6 or 24 h. Total RNA was prepared and 1.5 μg RNA subjected to expression profiling on Exon 1.0 microarrays (total of 24 arrays) as detailed previously . To assess the effect of PLZF/ZBTB16 on the GC response, the above cell lines were cultured for 24 h in the absence (set C) or presence (set D) of 200 ng/ml doxycycline and subsequently exposed to 10−8 M dexamethasone for 6 and 24 h and expression-profiled as above, resulting in a total of 15 arrays (one of the four 6 h replicates in set C was removed because it failed quality control).
The bioinformatic analysis of the microarrays was based on a custom CDF (CEL definition file) that allowed generating expression intensities of transcripts by summarizing the intensity signals of the individual oligonucleotide probes on the microarray targeting a particular transcript. In brief, the sequences of the >5 million 25 nt-long oligonucleotide probes on the Affymetrix Exon array (HuEx-1.0-st v2) were aligned to the genomic DNA sequence from Ensembl (version 52_36n). For all probes with a single complete alignment and no additional alignments with up to 1 mismatch, we determined whether the alignment was within the exon boundaries of a gene. Probes with sequences with a G/C content >18 were excluded from the analysis. Exon boundaries and transcript definitions were taken from the Ensembl core database for all known protein- and non-coding transcripts and from the ASTD database (version 1.1)  for all potential isoforms of a gene.
Based on the above annotation, a CDF file was compiled defining the probes targeting individual transcripts. Probes with no alignment (allowing 1 mismatch) to the human genome where defined as “background probes”. The resulting custom CDF contained 1,292,740 probes organized in 131,642 transcript probe sets corresponding to 28,532 Ensembl genes (49,701 and 88,789 transcript probe sets were based on annotations from the Ensembl core database and the ASTD database, respectively). 72,150 probes were defined as “background probes”. Pre-processing and all further data analyses were performed in R (version 2.8.0) using packages from Bioconductor (version 2.3)  and the custom CDF file described above. The raw signal intensities where preprocessed using the GC-RMA algorithm , where the “background probes” were used to predict the non-specific binding potential of the oligonucleotide probes. After pre-processing, the transcript with the highest variance across all samples was selected for further analysis as representative for a given gene. The moderated t-test  was then used to assess the significance for differential expression between the compared conditions. The resulting raw p-values were adjusted for multiple hypothesis testing using Benjamini and Hochberg's method for a strong control of the false discovery rate .
We have previously shown that PLZF/ZBTB16 is regulated by GC in numerous patients with ALL, but not in 2 ALL cell lines . To determine whether PLZF/ZBTB16 regulation is restricted to the in vivo situation or occurs in vitro as well and, if so, whether it relates to GC sensitivity, we determined PLZF/ZBTB16 regulation by GC in various in vitro leukemia models using real-time RT-PCR for PLZF/ZBTB16 and 9 leukemia cell lines treated with 10−7 M dexamethasone for 2, 6, 12 and 24 h in biological triplicates. As shown in Fig. 1, the extent of GC sensitivity in these cell lines was markedly different, i.e., untransfected Jurkat and MOLT4 T-ALL cell lines as well as AT-1 precursor B-ALL and Daudi Burkitt lymphoma cells were resistant to GC-induced apoptosis, while all others were GC-sensitive, although with varying kinetics. GR-transfected Jurkat cells died after 24 h exposure to 10−7 M dexamethasone, the response in CEM-C7H2 was delayed about 24 h, NALM6 and RS4;11 showed a somewhat slower death response than CEM-C7H2, while 697/EU-3 cells required 96 h to reach ~40% apoptosis. Effects of GC on cell cycle progression were also determined (Fig. 1) and showed that GC-induced apoptosis was frequently preceded or accompanied by an increase of cells in the G1 phase of the cell division cycle, whereas the cell lines resistant to GC-induced apoptosis were also resistant to the GC effects on the cell cycle. The notable exception was AT-1, in which G1 cell cycle arrest was observed in the complete absence of apoptosis.
Basal PLZF/ZBTB16 expression levels among these cell lines differed by close to 3 orders of magnitude (Fig. 2a), reflecting the situation in children with ALL . Expression ranged from undetectable (697, not shown in Fig. 2) to levels similar to those of the “housekeeping gene” TBP (MOLT4). PLZF/ZBTB16 expression showed no correlation with GC sensitivity or T- or precursor B-cell origin of the cell lines. Significant, as defined by an M value >1 (M values correspond to log 2 of fold induction) and a p-value of <0.05, and consistent (i.e., at all time points) GC-induction of PLZF/ZBTB16 was seen in NALM6, RS4;11, AT-1 and GR-transfected Jurkats (Fig. 2b). The latter finding showed that PLZF/ZBTB16 induction is critically dependent on sufficient amounts of functional GR in the investigated cell line (Jurkat cells lacking transgenic GR express low levels of wild type GR but fail to auto-induce the GR . Taken together, PLZF/ZBTB16 regulation by GC is not restricted to the in vivo situation, requires sufficient levels of functional GR and neither its regulation nor its expression appear to be clearly correlated with GC sensitivity.
To assess a possible contribution of PLZF/ZBTB16 to the anti-leukemic GC effects in ALL, we generated 2 stably-transduced derivatives of the CCRF-CEM T-ALL cell line with conditional expression of PLZF/ZBTB16, termed CEM-C7H2-2C8-ZBTB16#19 and #58 (ZBTB16#19 and #58), respectively. As revealed by quantitative RT-PCR and immunoblotting, PLZF/ZBTB16 mRNA and protein expression in these cell lines could be controlled by the addition of the tetracycline analogue doxycycline to the culture medium (Fig. 3). As little as 3.1 ng/ml doxycycline led to a detectable induction of PLZF/ZBTB16 mRNA that increased up to 50 ng/ml and reached a plateau thereafter. Interestingly, close to maximal induction was seen after only 2 h (Fig. 3a, the decrease in expression levels after 6 and 24 h at 3.1 and 12.5 ng/ml doxycycline is probably explained by the half-life of doxycycline in the culture). Similar observations were made at the protein level (Fig. 3b), although the detection limit was somewhat higher (starting from 12.5 ng/ml doxycycline).
The above clonal cell lines were then used to determine whether PLZF/ZBTB16 expression affects cell survival or cell cycle progression and/or modulates the anti-leukemic effects of GC. As shown in Fig. 4, transgenic expression of PLZF/ZBTB16 alone, even at levels that clearly exceeded those induced by GC, had no detectable effect on survival or cell cycle progression. However, it significantly reduced the extent of apoptosis induced by 10−7 M and, even more so, 10−8 M dexamethasone in this T-ALL in vitro model (Fig. 4a). GC-induced cell cycle arrest was also reduced, but this effect was less pronounced (Fig. 4b).
Since PLZF/ZBTB16 overexpression provided a protective effect, we investigated whether knock-down of PLZF/ZBTB16 might increase GC-induced cell death. To this end, we generated 2 lentiviral shRNA constructs targeting either exon 7 (shRNA-1) or 6 (shRNA-2) of PLZF/ZBTB16 and tested them in comparison with a negative control shRNA in C7H2-2C8-ZBTB16 cells. A clear, although incomplete, PLZF/ZBTB16 knock-down was observed both on the mRNA and protein levels (Fig. 5a). Next we infected GC-sensitive CEM-C7H2 cells (the parent of the C7H2-2C8-ZBTB16 cell lines) with recombinant lentiviruses containing these constructs. Since the PLZF/ZBTB16 levels in CEM-C7H2 cells are below the detection limit of the immunoblotting technique even after GC treatment, we could not verify the knock-down by Western technology. On the mRNA level, however, we observed a partial knock-down in this system as well (5 experiments resulting in a mean reduction of 1.6- and 2.1-fold for shRNA-1 and -2, respectively). Corresponding to the extent of knock-down observed with the 2 constructs, a small, but statistically significant increase in cell death induction was observed (Fig. 5b) further supporting the protective role of PLZF/ZBTB16 seen in the conditional overexpression experiments (Fig. 4a).
Next we explored whether the protective effect of PLZF/ZBTB16 was specific for GC-induced apoptosis or reflected a more general protective effect against cell death induction. To this end, we cultured ZBTB16#19 and #58 cells in the presence or absence of doxycycline (to induce transgenic PLZF/ZBTB16 expression) and determined the extent of apoptosis elicited by antibodies against CD95/FAS (representative of an apoptosis inducer via the extrinsic pathway) or the chemotherapeutics taxol, doxorubicin, or mitomycin C (representatives of apoptosis inducers via the intrinsic pathway). As depicted in Fig. 6, there was no significant protective effect against apoptosis induction by these 4 substances. Thus, within the limits of the investigated substances, the protective effect of PLZF/ZBTB16 appeared to be specific for GC-induced apoptosis.
To test whether the transcriptional repressor PLZF/ZBTB16 protects cells from GC-induced apoptosis by directly regulating expression of genes involved in cell death/survival decisions, we assessed the transcriptional response to transgenic PLZF/ZBTB16 in the above ALL in vitro model by whole genome expression profiling. We cultured 2 biological replicates of ZBTB16#19 and #58 (expressing PLZF/ZBTB16 in a tetracycline-dependent manner) in the absence (“set A”) or presence (“set B”) of 400 ng/ml doxycycline for 2, 6 and 24 h and subjected the corresponding RNAs to expression profiling on human Exon 1.0 microarrays (Affymetrix, Inc., total of 24 arrays) as indicated in Section 2. Unexpectedly, however, apart from PLZF/ZBTB16 itself, which served as an internal control, no significant regulations, as defined by a mean M value better than ±0.7 and a Benjamini–Hochberg  adjusted p-value (pBH) of <0.05, were observed after 2 and 6 h of doxycycline exposure. After 24 h, a single gene reached the above level of significance, i.e., HBEGF (heparin-binding EGF-like growth factor) was repressed with a mean M value of −2.3 and a pBH of 0.0299. These data strongly argued against the hypothesis that the inhibitory effect of PLZF/ZBTB16 on GC-induced apoptosis was caused by direct transcriptional regulation of death/survival genes by PLZF/ZBTB16 (the complete data can be viewed at the GEO web page, series GSE15820).
Since PLZF/ZBTB16 has previously been shown to interfere with GC-induction of a transiently transfected, GC-responsive reporter construct , we investigated whether PLZF/ZBTB16 interfered with the transcriptional response to GCs. We determined Exon array-based expression profiles from ZBTB#19 and #58 cells treated in duplicate with the GC analogue dexamethasone together with, or in the absence of, doxycycline. Together with the expression profiles generated further above, we generated 4 sets of experimental conditions that allowed us to study the effect of PLZF/ZBTB16 on the transcriptional response to GC. As explained above, sets “A” and “B” corresponded to data derived from ZBTB16#19 and #58 cells cultured in duplicates in the absence (set A) or presence (set B) of doxycycline for 6 or 24 h. Sets “C” and “D” derived from ZBTB16#19 and #58 cells that were first cultured for 24 h in the absence (set C) or presence (set D) of doxycycline to induce PLZF/ZBTB16 and subsequently exposed to 10−8 M Dex for 6 and 24 h. Comparison between sets A (no doxycycline, no dexamethasone) and C (no doxycycline, but plus dexamethasone) was used to identify GC-regulated genes (defined by mean M values better than ±0.7 and pBH < 0.05) in this system in the absence of transgenic PLZF/ZBTB16. After 6 h of GC exposure, 308 down- and 378 up-regulated transcripts fulfilled this criterion. After 24 h, 929 genes were down- and 717 up-regulated. The regulation of these genes was then studied after induction of PLZF/ZBTB16 by comparing set B (plus doxycycline, but no dexamethasone) with set D (plus doxycycline and plus dexamethasone). As shown in Fig. 7, the above GC response genes were significantly less regulated in the presence of doxycycline than in its absence, and this was true both for induced and repressed genes. The inhibitory effect included genes such as BCL2L11/Bim, the induction of which was previously shown to play a crucial role in GC-induced apoptosis in this model system . For this functionally relevant gene, the inhibitory effect of ZBTB16 overexpression on GC-induction was verified on the mRNA level by quantitative real-time RT-PCR (Fig. 8a) and on the protein level by immunoblotting (Fig. 8b). In conclusion, the data supported the hypothesis that PLZF/ZBTB16 reduces GC-induced apoptosis by interfering with the hormone's potential to regulate gene expression.
The transcriptional repressor PLZF/ZBTB16 was one of the most promising GC-regulated candidate genes for the anti-leukemic effects of GC: It was induced by GC during the in vivo response to systemic GC monotherapy in childhood ALL and, as shown in this study, in several leukemia in vitro cell lines. However, the functional analyses performed in this study revealed that, at least in the CCRF-CEM T-ALL model, PLZF/ZBTB16 exhibited a protective effect rather than copying, or contributing to, the anti-leukemic GC effects. Attempts to understand the molecular mechanism of this protective effect revealed that this bona fide transcriptional repressor failed to significantly affect gene expression in the investigated T-ALL model. This finding was unexpected because PLZF/ZBTB16 has been reported to cause changes in gene expression in several systems [26,28,29,33,34]. Even a specific search for regulation of some of the most prominent PLZF/ZBTB16 response genes, such as KIT, HOXB7, MYC, CCNA1/cyclin A, PBX1, CRABPI, BID or RUNX2, failed to provide evidence for transcriptional activity. This included a comparative analysis between the parental CEM-C7H2 cell line (assayed in biological triplicates on Exon arrays)  and ZBTB16#19 and #58 to assess possible effects of the small levels of PLZF/ZBTB16 expressed as a consequence of some leakiness of the tetracycline-regulated expression constructs. Several explanations can be offered for this phenomenon: In most instances [26,28,29,33], non-lymphoid tissues were analyzed and PLZF/ZBTB16 might require additional co-factors or post-transcriptional modifications, such as acetylation , that may not be present in lymphoid cells. The sole expression analysis in lymphoid cells (Jurkat T-ALL cells ) indicated that PLZF/ZBTB16 altered gene expression only at low cell density and after 72 h of PLZF/ZBTB16 induction. After 24 h and at “normal” cell densities, the findings in Jurkat cells were compatible with ours.
Although apparently inactive as a transcription factor when tested alone, PLZF/ZBTB16 significantly interfered with the transcriptional response to GC (Fig. 7). Although the transgenic PLZF/ZBTB16 levels were considerably higher than the endogenous levels in CCRF-CEM cells, similar levels were observed in lymphoblasts from some children with precursor B-ALL during systemic GC monotherapy  (and our unpublished data). Since both gene inductions and repressions were attenuated in the presence of PLZF/ZBTB16, and since regulations of many transcription units were affected, a general mechanism of action seems likely, such as sequestration of the activated GR by PLZF/ZBTB16. A more specific mechanism, such as binding of PLZF/ZBTB16 to the GR on the promoter and recruitment of transcriptional repressors, would be expected to interfere with transcriptional induction but not repressions and, hence, is less likely given that both repression and inductions were attenuated. A previous report suggested that PLZF/ZBTB16 might bind to, and inhibit DNA binding of, the GR . Mutational analyses revealed that this property maps to the C-terminal portion of PLZF/ZBTB16, which showed particularly strong GR binding ability. Current analyses in our lab suggest that in at least in one of our cell line systems, GC might specifically induce exons 3–7 (presumably via alternative promoter usage) that encode the above-mentioned particularly active portion of PLZF/ZBTB16.
A question of considerable interest is whether PLZF/ZBTB16 is a direct target of the GC receptor (GR). Previous studies using PLZF/ZBTB16 promoter-driven reporter assays in primary human endometrial stromal and myometrial smooth muscle cells, where PLZF/ZBTB16 is regulated by progesterone and GC, suggested that it is not . We have performed Affymetrix promoter tiling array-based Chip-on-CHIP analyses using CCRF-CEM and NALM6 cells and failed to find evidence of direct binding of the GR to the PLZF/ZBTB16 promoter (Rainer et al., in preparation). However, both negative evidences are limited by the sequences analyzed, i.e., the reporter study covered an ~5 kb region upstream and the promoter tiling array 7.5 kb upstream and 2.5 kb downstream of the putative start site.
In conclusion, we report that the transcriptional repressor PLZF/ZBTB16, one of the most promising GC-regulated candidate genes for the anti-leukemic effects of GC, did not contribute to, but rather protected CCRF-CEM leukemia cells from, the anti-leukemic GC effects. Although counterintuitive on first sight, induction of pro-survival signaling by GC in the course of an apoptotic response is not unheard of. Thus, we observed repression of the pro-apoptotic BH3-only molecule PMAIP1/Noxa in children and experimental systems of GC-induced apoptosis  and could show that this regulatory event is functionally relevant, since it interferes with the kinetics of, and sensitivity to, GC-induced apoptosis . Since GC are known to mediate pro-survival effects in certain systems, such as erythroblasts , neutrophils [59,60] or solid tumors , the GC response may inherently be complex, containing pro- and anti-apoptotic components. The cellular context may then determine which of the 2 opposing responses prevails. The growing understanding of this complexity might eventually allow specific interference with the pro-survival component of the response, thereby increasing the efficacy of GC in the therapy of lymphoid malignancies.
The authors would like to thank Dr. S. Geley for discussions, M. Brunner, K. Götsch, B. Gschirr, A. Kofler, S. Lobenwein, and C. Mantinger for technical assistance, and M. Kat Occhipinti-Bender for editing the manuscript. Supported by the Austrian Science Fund (SFB-F021, P18747), the Higher Education Commission of Pakistan, and ONCOTYROL, a COMET Center funded by the Austrian Research Promotion Agency (FFG), the Tiroler Zukunftsstiftung and the Styrian Business Promotion Agency (SFG). The Tyrolean Cancer Research Institute is supported by the “Tiroler Landeskrankenanstalten Ges.m.b.H. (TILAK)”, the “Tyrolean Cancer Aid Society”, various businesses, financial institutions and the People of Tyrol.