Portal de Periódicos da CAPES

SIRT1 is downregulated during neutrophil differentiation of acute promyelocytic leukaemia cells

2009; Wiley; Volume: 146; Issue: 3 Linguagem: Inglês

10.1111/j.1365-2141.2009.07749.x

1365-2141

Julian Wampfler, Mario P. Tschan, Deborah Shan, Alexander Laemmle, Elisabeth Oppliger Leibundgut, Gabriela M. Baerlocher, Deborah Stroka, Martin F. Fey, Christian Britschgi,

Autophagy in Disease and Therapy

Silent mating type information regulation 2 homolog 1 (SIRT1) is an NAD-dependent histone deacetylase with wide functions in metabolism, cancer and longevity (Brooks & Gu, 2008). SIRT1 also deacetylates non-histone targets, such as p53. Cellular regulation of SIRT1 is crucial, and this can be achieved at the transcriptional, translational or post-translational level. The tumour suppressor hypermethylated in cancer (HIC1) negatively regulates SIRT1 at the transcriptional level (Chen et al, 2005). HIC1 inactivation has been observed in a wide variety of malignancies and we reported HIC1 induction during myeloid differentiation and low HIC1 expression in acute myeloid leukaemia (AML) (Britschgi et al, 2008a). In agreement, SIRT1 overexpression was observed in human cancers including AML (Bradbury et al, 2005). However, SIRT1 expression may be reduced in cancer, and SIRT1 might have tumour suppressive as well as oncogenic functions (Brooks & Gu, 2008). Given that SIRT1 inactivates p53 and p53 in turn induces HIC1 transcription (Britschgi et al, 2006), this constitutes a feedback loop that might possibly be disturbed in leukaemia. We therefore set out to assess the role of SIRT1 in neutrophil differentiation of AML cells. SIRT1 mRNA expression levels were measured in haematopoietic CD34+ progenitor cells (n = 4), granulocytes of healthy donors (n = 5), and 79 primary AML using real-time quantitative reverse transcription polymerase chain reaction (RQ-PCR). The median relative expression ΔCt of SIRT1 mRNA was −0·36 in CD34+ and 0·03 in primary AML cells (not significantly different), whereas in granulocytes, SIRT1 mRNA was significantly higher expressed with a ΔCt of 4·38 (Fig 1A). Next, we assessed SIRT1 mRNA levels in five patients with t(15;17)-positive acute promyelocytic leukaemia (APL) at diagnosis and after therapy with all-trans retinoic acid (ATRA). In four out of five patients, SIRT1 mRNA upregulation was observed upon differentiating therapy, with ΔCt-values ranging from −0·09 to 4·23 (Fig 1B). These data are in line with our observations of elevated SIRT1 mRNA expression levels in granulocytes. We then aimed to confirm those findings in cell line models. ATRA was used to differentiate the t(15;17)-positive APL cell lines NB4 and HT93 towards granulocytes. SIRT1 mRNA was induced in NB4 and HT93 cells upon ATRA-induced granulocytic differentiation (1·5- and 3·5-fold respectively), whereas SIRT1 total protein levels consistently and strongly decreased (Fig 1C and D). The main compartment of expression and downregulation of SIRT1 protein was the nucleus (Fig 1D), whereas cytosolic expression was scarce. Disparate SIRT1 mRNA and SIRT1 protein regulation has been previously reported (Ford et al, 2008). Since the onset of SIRT1 protein downregulation preceded the onset of SIRT1 mRNA induction, a feedback loop might be involved. The tumour suppressor HIC1 forms a complex with SIRT1 to repress SIRT1 transcription (Chen et al, 2005). A reduction in SIRT1 protein levels might consequently lead to de-repression of SIRT1 transcription and increased mRNA levels. This feedback might also explain why SIRT1 mRNA expression is low in primary AML samples compared to granulocytes. As the reduction in SIRT1 mRNA levels is not responsible for the observed protein downregulation, low SIRT1 protein could be due to reduced protein stability during granulocytic differentiation. Studies in other cell line models strengthen the hypothesis that SIRT1 might be regulated post-translationally (Ford et al, 2008). SIRT1 mRNA and SIRT1 protein expression in models of neutrophil differentiation. (A) Total RNA was extracted from CD34+ selected progenitor cells from peripheral blood, peripheral blood granulocytes from healthy donors and primary AML samples of different FAB subtypes (peripheral blood or bone marrow), and reverse transcribed using pd(N)6 random primers. HMBS, ABL1 and SIRT1 cDNA expression was then measured in RQ-PCR using TaqMan® low-density arrays (Applied Biosystems, Rotkreuz, Switzerland. Assay IDs Hs00609297_m1, Hs00245445_m1 and Hs00202021_m1, respectively). SIRT1 mRNA Ct values were normalized to HMBS and ABL1 as housekeeping genes, and expression levels were calculated and expressed as ΔCt as described (Oberli et al, 2008). CD34+ progenitor cells show a SIRT1 mRNA expression level comparable to AML, whereas in granulocytes SIRT1 expression is significantly higher. (B) Five patients with newly diagnosed t(15;17)-positive APL were treated with orally administered ATRA at a dosage of 45 mg/m2 daily. Total RNA was extracted from blast cells isolated using a Ficoll gradient at day 0 (before treatment) and at follow-up. Expression levels of SIRT1 mRNA were assessed using RQ-PCR as in (A). (C) The t(15;17)-positive HT93 cell line was differentiated in vitro for the indicated time-points using 1 μmol/l ATRA. Successful differentiation was confirmed by fluorescence-activated cell sorting (FACS) analysis of surface differentiation markers (CD11b and CD18, data not shown). SIRT1 mRNA expression was assessed using RQ-PCR, normalized to HMBS as a housekeeping gene and untreated cells as the experimental starting point, expressed as n-fold change in regulation, as described (Britschgi et al, 2008a). There is induction of SIRT1 mRNA in ATRA treated cells. Similar results were obtained in the t(15;17)-positive NB4 cell line (data not shown). (D) In parallel to total RNA, we extracted proteins to assess changes in SIRT1 protein expression. Cells were lysed and Western blotting performed as described (Britschgi et al, 2006). The antibodies used were anti-SIRT1 (sc-153404; Santa Cruz, Labforce AG, Nunningen, Switzerland) and anti-β-actin (Sigma-Aldrich, Buchs, Switzerland). Upper: SIRT1 protein levels are decreased under ATRA induced differentiation, starting after 24 h of treatment. β-actin served as loading control. Lower panel: Western blot of nuclear protein extracts shows that the main compartment of SIRT1 expression and regulation is the nucleus. Nuclear SP-1 protein served as a loading control (anti-SP-1: sc-59; Santa Cruz). To assess whether SIRT1 protein downregulation during granulocytic differentiation is of functional significance, we inactivated SIRT1 in NB4 and HT93 cell lines. Efficient knock-down was confirmed in Western blots (Fig 2C). SIRT1 knock-down did not impact on cell viability, since the shSIRT1-cells could be maintained in culture over several weeks without loss of SIRT1 suppression. When treated with ATRA, both shSIRT1-NB4 and -HT93 cells differentiated less efficiently: expression of the myeloid surface marker CD11b was diminished in both cell lines upon SIRT1 inactivation. After 4 d of ATRA treatment, mean fluorescence intensity (MFI) was 8·9 in shSIRT1-NB4 (constructs sh1958 and sh3206) versus 43·5 in controls (construct SHC002, Fig 2A), and 8·7 vs. 19·5 in shSIRT1-HT93 and controls, respectively (data not shown). Those findings were corroborated by expression analyzes of genes known to be induced in granulocytic differentiation. CCAAT-enhancer binding protein ε (CEBPE) mRNA was induced 359·0-fold in HT93 control cells during differentiation, but only 282·0-fold in shSIRT1-HT93 cells (sh3206). In NB4 cells, CEBPE induction was 27·0- and 14·0-fold, respectively. Analysis of CSF3R mRNA produced similar results: 4·1-fold induction in HT93 controls compared to 3·5-fold in shSIRT1-HT93 and 16·0-fold induction in NB4 controls compared to 7·5- and 10·6-fold in shSIRT1-NB4 (sh3206 and sh1958, respectively) (Fig 2B). SIRT1 knock-down impairs granulocytic differentiation. (A) t(15;17)-positive NB4 cells were transduced to stably express either one of two small hairpin RNAs (shRNA) targeting SIRT1 (clones NM_012238.3-1958 and NM_012238.3-3206, named sh1958 and sh3206 respectively; Sigma-Aldrich), or a non-targeting shRNA (SHC002; Sigma-Aldrich) as described previously (Britschgi et al, 2008a). Puromycin selected subclones were differentiated with 1 μmol/l ATRA for 96 h and analyzed for CD11b surface expression. The grey area denotes the difference in CD11b expression between ATRA-treated shSIRT1-cells and controls, the two curves on the left correspond to untreated shSIRT1 and control cells. (B) In the same cells, we determined mRNA levels of genes known to be upregulated in differentiation. RNA was extracted and RQ-PCR performed as described above (Assay IDs Hs00167918_m1 for CSF3R and Hs00357657_m1 for CEBPE). mRNA levels are expressed as n-fold changes in regulation when compared to untreated cells using HMBS mRNA expression as a reference gene. NB4 control cells show a higher CSF3R induction than shSIRT1 cells (sh3206 and sh1958), confirming the impairment of granulocytic differentiation observed in (A). (C) Confirmation of SIRT1 knockdown efficiency. NB4 Cells were lysed and Western blotting performed as described in Fig 1D. β-actin served as a loading control. Results in HT93 were comparable (data not shown). These findings indicate that SIRT1 function might be important during induction of myeloid differentiation. This is further supported by recent findings: nicotinamide phosphoribosyltransferase (NAMPT) catalyzes the formation of the SIRT1-cofactor NAD+ and is essential in granulocyte colony-stimulating factor (G-CSF)-induced granulopoiesis via SIRT1-induction (Skokowa et al, 2009). We now show that a complete SIRT1 knock-out impairs terminal differentiation. This might be due to an interference with autophagy, since SIRT1 was shown to be an important in vivo regulator of autophagy, and SIRT1−/− mice resembled autophagy deficient ATG5−/− mice (Lee et al, 2008). We recently reported that ATG5 expression and functional autophagy are necessary for normal neutrophil development, and low ATG5 expression correlates with a leukaemic phenotype (Britschgi et al, 2008b). Consequently, a complete knock-down of SIRT1 might disturb autophagy, leading to impaired neutrophil differentiation. In normal late neutrophil development, however, SIRT1 protein might be dephosphorylated, leading to destabilization and downregulation, followed by de-repression of SIRT1 transcription. In summary, we found significantly lower SIRT1 mRNA expression in primary AML samples compared to granulocytes. Interestingly, we found that constitutive SIRT1 knock-out impairs pharmacologically induced differentiation of APL cells pointing to a possible involvement of SIRT1 in the initiation of granulocytic differentiation. Consequently, targeting SIRT1 might be of therapeutic relevance in AML, and our findings point to a new molecular avenue to be further explored in this disease. This work was supported by grants from the Bernese Cancer League (to CB and MPT), the Swiss National Foundation SNF 3100A0-112385 (to MFF), the Werner and Hedy Berger-Janser Foundation of Cancer Research (to MFF and MPT), the Bernese Foundation of Cancer Research and the Ursula-Hecht-Foundation for Leukaemia Research (to MFF). JW performed experiments, analyzed data and wrote the paper. MPT prepared lentivirus, established SIRT1 KO cells and analyzed data. DS performed experiments and analyzed data. AL and DSt helped in designing the project and analyzed data. GMB and EOL provided CD34+ and AML samples. MFF and CB designed the project, analyzed data and wrote the paper. The authors have no conflicts of interest to disclose.