Valproic acid exerts differential effects on CXCR4 expression in leukemic cells

Leukemia Research 34 (2010) 235–242

Contents lists available at ScienceDirect

Leukemia Research journal homepage: www.elsevier.com/locate/leukres

Valproic acid exerts differential effects on CXCR4 expression in leukemic cells Hilal Gul a,b , Leah A. Marquez-Curtis a , Nadia Jahroudi b , Loree M. Larratt b , Anna Janowska-Wieczorek a,b,∗ a b

Canadian Blood Services R & D, Edmonton, Alberta, Canada Department of Medicine, University of Alberta, Edmonton, Alberta, Canada

a r t i c l e

i n f o

Article history: Received 16 March 2009 Received in revised form 30 March 2009 Accepted 21 May 2009 Available online 18 June 2009 Keywords: Valproic acid SDF-1/CXCR4 axis Chemotaxis AML HDAC inhibitors

a b s t r a c t We recently reported that the histone deacetylase inhibitor, valproic acid (VPA), increases CXCR4 receptor expression and function in cord blood hematopoietic stem/progenitor cells (HSPC) and the immature, highly CD34-positive AML cell lines KG-1a and KG-1. In this study, we investigated whether VPA influences CXCR4 in CD34-negative AML cell lines (promyelocytic HL-60 and monocytic THP-1), as well as both CD34positive and CD34-negative primary AML cells. We found that VPA (i) diminishes CXCR4 expression and chemotaxis in HL-60 cells and in the CD34-negative subtypes of primary AML cells and (ii) increases CXCR4 expression and function in the highly CD34-positive subtypes of primary AML cells. Hence, we suggest that VPA exerts different effects on CXCR4 depending on cell maturation status, and this novel finding may have important implications for AML therapy. © 2009 Elsevier Ltd. All rights reserved.

1. Introduction The G protein-coupled chemokine receptor CXCR4 is expressed on primitive hematopoietic stem/progenitor progenitor cells (HSPC), and its ligand SDF-1␣ (CXCL12) is produced by cells, including stromal cells and osteoblasts, within the bone marrow (BM) microenvironment [1]. Interaction between SDF-1␣ and CXCR4 is known to play an essential role in mobilization, homing, retention and survival of HSPC [1–3]. HSPC expressing CXCR4 are attracted to the BM microenvironment by SDF-1␣, and disruption of the SDF1␣/CXCR4 axis facilitates their mobilization to the peripheral blood. SDF-1␣ has been also reported to activate adhesion receptors such as CD44, very late antigen-4 (VLA-4) and lymphocyte functionassociated antigen-1 (LFA-1) on HSPC, which may contribute to the homing process [4]. Thus, SDF-1␣/CXCR4-mediated migration of HSPC is believed to be one of the major mechanisms for homing and, aside from its importance in this respect, this axis has also been reported to play crucial roles in the migration, abnormal proliferation and anchorage of human acute myelogenous leukemic (AML) cells in the BM [5].

∗ Corresponding author at: Department of Medicine, University of Alberta & Canadian Blood Services, CBS Building, 8249-114 Street NW, Edmonton, Alberta, Canada T6G 2R8. Tel.: +1 780 431 8761; fax: +1 780 702 8622. E-mail address: [email protected] (A. Janowska-Wieczorek). 0145-2126/$ – see front matter © 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.leukres.2009.05.014

AML, a heterogeneous group of diseases characterised by uncontrolled clonal proliferation of myeloid blasts with reduced capacity to differentiate into mature cells, has been reported to express CXCR4 differently in its various subtypes: lower expression level in the case of undifferentiated (M0), myeloid (M1/M2), and erythroid (M6) AML, and higher expression in more differentiated myelomonocytic (M4/M5) and promyelocytic (APL; M3) AML [6]. The expression level of CXCR4 on AML blasts was postulated to be a major prognostic factor for disease progression [7,8]. Enhanced expression promotes cell retention, survival and growth within the BM microenvironment resulting in resistance to conventional chemotherapy and subsequent relapses [9,10]. This indicates that modulation of CXCR4 gene expression could be clinically relevant not only in the trafficking of leukemic cells but also in their response to chemotherapeutic agents. Gene expression can be regulated by chromatin remodelling [11]. Two enzyme classes, histone acetyltransferase (HAT) and histone deacetylase (HDAC) regulate chromatin structure in normal cells [12] and alterations in these molecules can lead to aberrant transcriptional networks in AML. For example, abnormal recruitment of HDACs by the PML/RARa and AML1/ETO fusion proteins in APL and AML-M2, respectively, results in repression of genes relevant for differentiation by reducing the core histone acetylation level [13]. Therefore, differentiation or transcription therapy using histone deacetylase inhibitors (HDI) through chromatin modelling is under investigation as a new approach to treat hematological malignancies [14–16].

236

H. Gul et al. / Leukemia Research 34 (2010) 235–242

Table 1 Clinical characteristics of patients and effects of VPA on CXCR4 mRNA, protein and chemotaxis. Age/sex

FAB type

Cytogenetics

FISH probe

%BM/PB

CD34 antigen

WBC (×109 /L)

Gene

Protein

Chemotaxis

1

53/F

M2

AML1/ETO

72/12

+

2.8

↔

↔

↔

2 3

87/F 80/M

M4 ND

– –

nd/65 28/48

+ +

206 35

↔ ↔

↔

ND ND

4 5

69/M 48/M

ND M2

– AML1/ETO

35/36 44/59

+ +

93 9.2

↑ ↔

↑ ↔

↑ ND

6

42/F

M3

PML/RARa

87/66

−

5.1

↓

↓

7

46/M

M3

–

87/91

−

9.2

↓

↓

8 9

57/M 49/F

ND ND

– DEK/CAN

18/38 70/47

− −

9.5 3.9

ND ↔

↓ ↔

↓ ND

10

46/M

M4

MLL

72/97

−

211

↓

↓

↓

11

73/F

M4

46XX t(8;21) q(22:22) 46XX 46XY Complex 46XY 46XX t(8;21) q(22:22) 46XX t(15;17) q(22:21) 46XY Complex 46XY 46XX t(6;9) (p23:q34) 46XX t(9;11) q(22:23) 46XX

–

28/0

−

5.8

ND

↔

ND

Patient #

WBC: white blood cell count at diagnosis; PB: peripheral blood; BM: bone marrow. ND: not determined; ↔: no effect, decrease

A specific and potent short-chain fatty acid HDI, valproic acid (VPA), has gained considerable attention among other HDI due to its safe and well-tolerated nature [17]. Recently, it was demonstrated that VPA induces differentiation of PML/RARa-transformed HSPC and primary AML blasts in vitro regardless of the primary genetic alteration [18]. Moreover, VPA has been reported to increase the cytotoxicity or favorable response of AML blasts in recent clinical trials [19–24]. On the other hand, we and others have shown that VPA can exert opposite effects on HSPC, including the enhancement of proliferation and self-renewal [25,26]. Furthermore, we found that VPA increases CXCR4 expression and chemotaxis of cord blood primitive HSPC towards SDF-1␣ and we suggested its use to improve homing and engraftment [27]. Moreover, a similar effect was observed in the CD34-positive human AML cell lines KG-1a and KG-1 [27] and this might explain both of the favorable and the adverse effects observed after treatment of AML with HDI. Here we have extended our investigation and examined the effect of VPA on CXCR4 (at the protein, transcriptional and functional levels) in CD34-negative promyelocytic HL-60 and monocytic THP-1 AML cell lines as well as CD34-positive and CD34-negative leukemic cells obtained from patients diagnosed with AML. 2. Materials and methods 2.1. Cells and viability testing The human promyelocytic HL-60 and monocytic THP-1 leukemic cell lines were purchased from American Type Culture Collection (Rockville, MD, USA). Blood samples were obtained from patients diagnosed with AML at the University of Alberta Hospital (Edmonton, AB, Canada) with the appropriate patients’ consents and the approval of the Health Research Ethics Board. AML subtypes were classified according to the criteria of the French American British (FAB) system as shown in Table 1. Light density mononuclear cells (MNC) were isolated by Ficoll/Hypaque density centrifugation as described previously [28]. All cells were then maintained in IMDM medium (GibcoBRL, Long Island, NY) supplemented with 20% bovine growth serum (BGS; Hyclone, Logan, UT) and treated with VPA (1 mM; Sigma, Oakville, ON) for 24 h and 48 h as described below. Cell viability was measured by the trypan blue exclusion assay. 2.2. Chemotaxis assay Chemotaxis was assayed in modified Boyden chambers (Neuro Probe Inc., Gaithersburg, MD) as described previously [27]. Briefly, pre-warmed serum-free medium containing SDF-1␣ (200 ng/ml; Biochemical Research Centre, University of British Columbia, Vancouver, BC) was added to the lower chambers. Aliquots of the cell suspension (1 × 105 cells/100 ␮l) were loaded onto the upper chambers and incubated for 3 h (37 ◦ C, 5% CO2 ). Cells from the lower chambers were recovered

Effect of VPA

: slight increase, ↑: increase,

: slight decrease, ↓:

for determination of cell numbers and percentage cell migration was calculated as the ratio of the number of cells recovered from the lower compartment to the total number of cells loaded in the upper compartment. Each experiment was performed at least twice using four or more chambers for each cell sample. 2.3. FACS analysis The surface expression of CXCR4 in cells was evaluated by FACS analysis, with detection of the CXCR4 antigen by PE-anti-CXCR4 monoclonal antibody (BD Biosciences, Oakville, ON). Briefly, the cells were stained in PBS (Ca- and Mg-free) supplemented with 5% BGS. After the final wash, cells were fixed in 2% paraformaldehyde prior to FACS analysis, which was performed by FACscan (Becton-Dickinson, San Jose, CA) using PE-goat-anti-mouse immunoglobulin (IgG) as the isotype control. To eliminate any nonspecific binding, the same ratio of fluorochrome/protein for the isotype control and specific antibody was used. 2.4. Quantitative real-time RT-PCR (qRT-PCR) Total RNA extraction using TRIzol reagent (Invitrogen, Burlington, ON) and reverse transcription was carried out as described previously [27]. The TaqMan Gene Expression Assay (Applied Biosystems, Streetsville, ON) was used to analyze CXCR4 mRNA expression according to the manufacturer’s instructions. Briefly, the thermal cycling conditions were 95 ◦ C for 2 min, followed by 40 cycles at 95 ◦ C for 15 s and 60 ◦ C for 1 min. The real-time polymerase chain reaction was performed in triplicate in 96-well plates using the ABI Prism 7700 Sequence Detector (Applied Biosystems). Results were analyzed with the ABI Prism 7700 Sequence Detection System software (Applied Biosystems). Each reaction was normalized by the cycle threshold (Ct) of GAPDH cDNA expression. Relative expression was calculated with the Ct method. 2.5. Chromatin immunoprecipitation assay (X-ChIP) The ChIP experiment was performed using commercially available reagents (Upstate Biotechnology, Lake Placid, NY) according to the protocol, with modifications as described [29]. In brief, cells were treated with 1% formaldehyde (Sigma) to obtain cross-linked chromatin. The reaction was stopped by the addition of 125 mM glycine. Soluble chromatin with an average size of ∼200–500 bp was prepared by sonication using VibraCellTM (Sonics and Material Inc., Danbury, CT). An aliquot of the chromatin (10 ␮l) was saved for analysis of total input, and the remaining samples were immunoprecipitated with 5 ␮g of acetylated histone H4 antibody (Upstate Biotechnology) or with no antibody. The immunoprecipitated chromatin was washed sequentially with wash buffers before elution by a 5-min incubation in 150 ␮l of fresh buffer E (1% SDS, 50 mM NaHCO3 ). To reverse cross-linking, the samples were incubated at 65 ◦ C for 4 h in a buffer containing 200 mM NaCl and 1 ␮g of RNase A followed by treatment with proteinase K and ethanol precipitation. The pellets were collected by microcentrifugation, resuspended in 20 ␮l of H2 O, and subjected to qRT-PCR analysis as described above using the appropriate primer pairs, namely the human CXCR4 forward primer: 5 -GGCCGACCTCCTCTTTGTC-3 , reverse primer: 5 -TTGCCACGGCATCAACTG-3 , TaqMan probe: 5 -FAM-CACGCTTCCCTTCTGTAMRA-3 ; GAPDH forward primer: 5 -CGCTCTCTGCTCCTCCTGTT-3 , reverse primer: 5 -CACCTGGCGACGCAAAA-3 , TaqMan probe: 5 -FAM-AGTCAGCCGCATCTT-TAMRA3 . The amplification of GAPDH promoter was used as an internal control.

H. Gul et al. / Leukemia Research 34 (2010) 235–242 2.6. Statistical analysis Arithmetic means and standard deviations were calculated and statistical significance was defined as p < .05 using the Student’s t-test.

3. Results 3.1. VPA downregulates CXCR4 expression in the CD34-negative HL-60 cells Recently we showed that VPA enhances CXCR4 expression in normal HSPC and the immature, highly CD34-positive leukemic cell lines KG-1 and KG-1a [27]. However, contrary to its effect in immature cells, VPA was reported to induce growth arrest and differentiation in cells of the more mature CD34-negative, promyelocytic AML cell line HL-60 [30]. We speculate that the effect of VPA on leukemic blasts differs according to their degree of maturation and that VPA may also reduce CXCR4 expression, enhancing cytotoxicity of more differentiated leukemic blasts. Hence, we evaluated the effect of VPA (1 mM) on surface CXCR4 expression by flow cytometry in HL-60 cells and the CD34-negative monocytic leukemia cell line THP-1. We found no significant effect of VPA on CXCR4 level in HL-60 or THP-1 cells after 24-h incubation but a

237

decrease (about 3-fold) in the percentage of CXCR4-positive cells after 48-h incubation (Fig. 1A). Using qRT-PCR we confirmed that the effect of VPA on surface CXCR4 levels was reflected at the mRNA level in HL-60 cells (2fold downregulation), but not in more differentiated THP-1 cells (Fig. 1B). These findings suggest that reduction of surface CXCR4 level is mainly a consequence of decreased CXCR4 gene expression in HL-60 but not in THP-1 cells. The viabilities of cells exposed to VPA were not significantly altered (data not shown). Silent genes appear either to lack or to have only minimal levels of histone acetylation at their promoter regions, whereas acetylated histones are normally associated with promoters of actively transcribed genes [11]. We have shown that modulation of CXCR4 gene transcription correlates with the acetylation status of histone H4 in HSPC [27]. To investigate whether modulation of CXCR4 gene transcription correlated with the acetylation status of histone H4 in HL-60 cells, we analyzed the levels of acetylated histones associated with the CXCR4 gene promoter in these cells by X-ChIP assay. We found that VPA treatment of HL-60 cells after 48 h did not significantly change the levels of acetylated histone H4 that were associated with the CXCR4 promoter compared to untreated cells (Fig. 1C), indicating that transcription of the CXCR4 gene in HL-60 cells is regulated differently from HSPC.

Fig. 1. VPA reduces CXCR4 level in HL-60 cells. Surface and mRNA expression of CXCR4 in HL-60 and THP-1 cells following 24- and 48-h exposure to VPA (1 mM). The results are shown as fluorescence histograms. (A) The percentage of CXCR4-expressing cells in CD34-negative HL-60 and THP-1 cells; representative of three experiments yielding similar results. (B) Expression of CXCR4 mRNA in HL-60 and THP-1 cells following 48-h exposure to VPA (1 mM). The relative expression levels of CXCR4 mRNA were determined by real-time quantitative PCR (qRT-PCR). GAPDH was used as an internal calibrator (control gene); representative of two experiments yielding similar results. (C) Modulation of CXCR4 gene transcription by VPA does not correlate with acetylation status of histone H4 in HL-60 cells as shown by X-ChIP assay; representative of two experiments yielding similar results.

238

H. Gul et al. / Leukemia Research 34 (2010) 235–242

the SDF-1-driven chemotaxis of these CD34-negative AML cells (Pts #10, #8, #6 and #7; Fig. 3A–C, right panels; and Table 1). However, no change was seen in CXCR4 expression or chemotaxis in two other CD34-negative AML samples (Pts #9 and #11) (Table 1). On the other hand, VPA stimulation led to an increase at surface and mRNA CXCR4 levels after 24 h in a highly CD34-positive (97%) AML patient #4 (Fig. 4A). In fact, these effects of VPA were also confirmed at the functional level as migration towards SDF-1␣ (Fig. 4A) was significantly enhanced, similar to results in our previous studies of highly CD34-positive KG-1a and KG-1 cells [27]. Moreover, we found a slight increase in surface CXCR4 levels with no effect on CXCR4 mRNA in Pt #3 (Fig. 4B) who had both CD34-positive and CD34-negative populations in about equal proportions. However, no effects on CXCR4 level were observed in other patients (Pts #1, #2 and #5), whose cells were low in CD34-antigen (Table 1). These data suggest that the effect of VPA on CXCR4 expression in AML blasts differs according to their maturation stage and/or type of cell population.

4. Discussion Fig. 2. VPA dimishes SDF-1␣ mediated migration of HL-60 cells, but not that of THP1 cells. Percentage cell migration was calculated as the ratio of the number of cells that had migrated towards SDF-1␣ to the total number of cells loaded. Migration towards an SDF-1␣ gradient (200 ng/ml) in HL-60 and THP-1 cells. Results represent the mean ± SEM of three independent experiments.*p < 0.05, statistically significant compared with untreated cells (Student’s t-test).

3.2. VPA reduces chemotaxis in HL-60 cells towards an SDF-1˛ gradient To assess the in vitro functional responses of HL-60 and THP-1 cells, we examined the migration of VPA-treated cells in response to an SDF-1␣ gradient in comparison to untreated cells. Consistent with the expression of CXCR4, VPA significantly decreased (3-fold) the migration of HL-60 cells towards a high SDF-1␣ gradient after 48-h incubation, but not of THP-1 cells (Fig. 2). However, no effect was observed on the chemotaxis of cells treated with VPA for 24 h (data not shown). 3.3. Effect of VPA on CXCR4 expression and function in primary AML cells Based on the hypothesis that the favorable or adverse responses of some subtypes of AML blasts may be related to the effect of VPA on CXCR4 expression, we evaluated it in CD34-positive and CD34-negative primary AML cells. We analyzed the effect of VPA on CXCR4 level and function after 24 h and 48 h in peripheral blood mononuclear cells (PB-MNC) from 11 AML patients. The clinical characteristics of patients and results are summarized in Table 1. After 48-h incubation, VPA treatment resulted in 1.5- and 2.6-fold decreases in the percentage of CXCR4-expressing cells in PB-MNC from CD34-negative patients Pt #10 (AML-M4) and Pt #8 (secondary AML), respectively (Fig. 3A and B, left panels). However, VPA treatment led to only slight decreases in surface CXCR4 level in CD34-negative APL (M3 subtype) cells from patients Pt #6 (Fig. 3C, left panel) or Pt #7 (Table 1). A similar effect was also observed in the CXCR4 mRNA level in Pt #10 (Fig. 3A, middle panel). However, VPA treatment resulted in 3- and 1.6-fold downregulation at the transcriptional level in Pts #6 and #7, respectively (Fig. 3C, middle panel; Table 1), which is more significant than surface CXCR4 level. Gene expression analysis for Pt #8 was not performed due to an insufficient number of cells. The favorable effect of VPA was also observed at the functional level as VPA significantly repressed

A number of studies have reported the pivotal role of the SDF-1/CXCR4 axis in mobilization, homing and repopulation of normal HSPC as well as trafficking and survival of leukemic blasts [1–6]. Therefore, the CXCR4 receptor has become an important target for improving both the transplantation of normal HSPC and the response of leukemic cells to chemotherapeutic agents. We have recently reported the regulation of CXCR4 expression by the potent HDI VPA, and the more potent but structurally different HDI Trichostatin A (TSA), through chromatin remodelling [27,31]. We showed that VPA enhances CXCR4 expression and homing responses towards an SDF-1␣ gradient and suggested employing HDI as a priming agent for improving HSPC engraftment [27]. In the present study, we focused on VPA because it is a safe, relatively well-tested drug and is under extensive evaluation for the treatment of myeloid malignancies [19]. We report here that VPA differentially regulates CXCR4 expression in various subtypes of AML blasts. VPA repressed CXCR4 and chemotaxis in more differentiated CD34-negative AML cells, whereas highly CD34-positive, immature AML cells and normal HSPC were found to respond to VPA with increased CXCR4 expression and chemotaxis. In fact, consistent with our observations it was reported that in melanoma cells the highly potent HDI TSA upregulated CXCR4 expression after 24 h incubation but downregulated it after 48 h, indicating that the effect is transient [32]. In addition, they showed that in lymphocytes TSA differentially regulates CXCR4 expression, suggesting that the effect of HDI may be cell-dependent as well. Furthermore, repression of CXCR4 expression and migration in acute lymphoblastic leukemia cells by HDIs such as SAHA and butyrate have been demonstrated [33]. Our results with VPA are consistent with the findings that the effects of HDI on cells differ according to cell type and/or degree of maturation [19,25,27,31]. Here we propose a novel regulation of CXCR4 receptor in AML cells which could explain why some patients were more and others less responsive to chemotherapeutic agents. Clinical data that we obtained previously indicated that exposure to VPA increased the response to chemotherapeutic agents, which resulted in partial remission in two patients with secondary AML arising from myeloproliferative disorder (MPD) [23,24]. In recent clinical trials, the best responses to VPA were observed in patients who had sAML/MDS (low-risk MDS;