Leukemia (2014) 28, 566–576 & 2014 Macmillan Publishers Limited All rights reserved 0887-6924/14 www.nature.com/leu

ORIGINAL ARTICLE CXCR4-SERINE339 regulates cellular adhesion, retention and mobilization, and is a marker for poor prognosis in acute myeloid leukemia

L Brault1, A Rovo´ 2, S Decker3, C Dierks3, A Tzankov4,5 and J Schwaller1,5

The CXCR4 receptor is a major regulator of hematopoietic cell migration. Overexpression of CXCR4 has been associated with poor prognosis in acute myelogenous leukemia (AML). We have previously shown that ligand-mediated phosphorylation of the Serine339 (CXCR4-S339) residue of the intracellular domain by PIM1 is implicated in surface re-expression of this receptor. Here, we report that phosphorylation of CXCR4-S339 in bone marrow (BM) biopsies correlated with poor prognosis in a cohort of AML patients. To functionally address the impact of CXCR4-S339 phosphorylation, we generated cell lines-expressing CXCR4 mutants that mimic constitutive phosphorylation (S339E) or abrogate phosphorylation (S339A). Whereas the expression of CXCR4 significantly increased, both CXCR4-S339E and the CXCR4-S339A mutants significantly reduced the BM homing and engraftment of Kasumi-1 AML cells in immunodeficient mice. In contrast, only expression of the CXCR4-S339E mutant increased the BM retention of the cells and resistance to cytarabine treatment, and impaired detachment capacity and AMD3100-induced mobilization of engrafted leukemic cells. These observations suggest that the poor prognosis in AML patients displaying CXCR4-S339 phosphorylation can be the consequence of an increased retention to the BM associated with an enhanced chemoresistance of leukemic cells. Therefore, CXCR4-S339 phosphorylation could serve as a novel prognostic marker in human AML.

Leukemia (2014) 28, 566–576; doi:10.1038/leu.2013.201 Keywords: CXCR4; serine339; AML; homing; retention; resistance

INTRODUCTION receptor signaling, CXCR4 endocytosis seems to control down- The G protein-coupled receptor CXCR4 is activated by CXCL12 stream signaling pathways, including ERK cascade activation, 8,10 (stromal cell-derived factor 1) and is involved in the control of intracellular calcium flux and chemotaxis. Despite the wide migration and homing of cells notably for engraftment of normal range of factors that are known to be activated after CXCL12 and neoplastic hematopoietic cells in the bone marrow (BM).1,2 binding to CXCR4, there is limited information about the critical CXCR4 receptor activation enhances the proliferation and/or intracellular residues of CXCR4 and the associated activation of 11,12 metastasis formation (that is, to the BM) of many tumors and downstream signal transduction pathways. also functions as an entry receptor for HIV-1. All CXCR4 functions Although CXCR4 signaling was shown as being controlled by critically depend on its cell-surface expression, which is regulated G protein-regulated protein kinases targeting the C-terminal tail, at the transcriptional level and at the protein level by endocytosis, the intracellular signals specifically evoked by phosphorylation intracellular trafficking and recycling.3,4 of distinct residues and their functional consequences remain 13 The surface expression level of CXCR4 on acute myelogenous unclear. We previously identified ligand-mediated phosphorylation leukemia (AML) blasts has been previously associated with a poor of CXCR4-S339 to have a key role for in vitro migration and in vivo 14 prognosis, and analogous to normal hematopoietic cells, the level homing and engraftment of murine BM cells. Here we investigated of CXCR4 expression also correlated with the CXCL12-induced the clinical impact and the functional consequence of the CXCR4- chemotaxis.5–7 In addition, interference with the CXCL12/CXCR4 S339 status in AML. Our results suggest that CXCR4-S339 axis through treatment with granulocyte colony-stimulating factor phosphorylation might be a novel prognostic marker in AML and/or CXCR4 antagonist like AMD3100 (Plerixafor) increased the patients and a critical regulator of migration, homing and retention cellular mobilization from the BM.8 CXCR4 inhibition is a novel of leukemic cells. strategy to not only mobilize hematopoietic stem cells for autologous transplantation but also tumor cells in order to increase the effectiveness of chemotherapy notably in AML.9 MATERIALS AND METHODS Several studies have demonstrated that internalization of the Patients and immunohistochemistry CXCR4 receptor involves endocytosis provoked by phospho- Patient’s characteristics are summarized in Table 1. All patients had given rylation of the CXCR4 C-terminal domain. Beside its role for general informed consent for studies to be performed using tissue receptor internalization leading to desensitization that terminates materials that remained after diagnostic procedures. Tissue biopsies were

1Department of Biomedicine, University Children’s Hospital (UKBB), University of Basel, Basel, Switzerland; 2Department of Hematology, University Hospital Basel, Basel, Switzerland; 3Department of Hematology/Oncology, University Medical Center Freiburg, Freiburg, Germany and 4Institute for Pathology, University Hospital Basel, Basel, Switzerland. Correspondence: Dr J Schwaller, Department of Biomedicine, University Children’s Hospital (UKBB), University of Basel, Hebelstrasse 20, Basel 4031, Switzerland. E-mail: [email protected] 5Shared senior authorship. Received 8 November 2012; revised 21 June 2013; accepted 24 June 2013; accepted article preview online 2 July 2013; advance online publication, 26 July 2013 Role of CXCR4-S339 in AML L Brault et al 567 were killed 24 h, 7 and 46 days post-transplantation. Single-cell suspen- Table 1. Patient characteristics sions of indicated tissue samples were prepared by passing the tissue through 100-mm nylon mesh strainers (Falcon, Becton-Dickinson, Heidel- pCXCR4-S339 pCXCR4-S339 P- berg, Germany). Red blood cells of peripheral blood (PB) were lysed before positive negative value the analysis. Dead cells were excluded using 4’,6-diamidino-2-phenylindole staining, and the percentage of GFP-positive cells was monitored using N 39 36 flow cytometry and expressed as the percentage of all gated mononuclear Age median (range) 65 (26–95) 55 (19–83) 0.032 cells of the indicated tissue. Tissue samples were fixed in 4% phosphate- Male/female 20/19 14/22 buffered saline-formaldehyde, paraffin-embedded and 4-mm sections were Risk group 0.416 stained with hematoxylin-eosin or anti-human CD34 (Dako, Glostrup, Low 6 5 Denmark, M7165). A ratio between PB and BM (PB/BM) is determined as Intermediate 14 15 (PB GFP þ cells (%) WBC body weight (g) 0.08)/(BM GFP þ cells 18 High 17 10 (%) BM cellularity/0.056). Not evaluated 2 6 To test the ability of AMD3100 to mobilize leukemic cells, a 5 mg/kg dose was administrated subcutaneously 5 weeks after transplantation. FLT3-ITD mutated 22 14 0.051 Three hours after the treatment, the mice were killed and analyzed. Curative therapy yes/no 26/13 31/5 To test the in vivo resistance to cytarabine (Ara-C), a dose (75 mg/kg) per CR after induction 17 19 0.113 day for 5 consecutive days was administrated subcutaneously 5 weeks chemotherapy after transplantation. On the day after the last treatment, the mice were Post induction failure 7 7 killed and analyzed.19 Early relapse 6 7 0.139 Abbreviations: CR, complete remission; ITD, internal tandem duplication. Cellular detachment assay Kasumi-1 cells were added to confluent human umbilical vein endothelial cell (HUVEC) cells in six-well plates with or without activation by tumor necrosis factora (20 ng/ml). After 1 h, non-adherent Kasumi-1 cells were processed using an automated immunostainer (Nexes, Ventana Medical carefully removed and each well was monitored (12 microscopic fields) for Systems, Tucson, AZ, USA). The immunohistochemical staining of pCXCR4- GFP-positive cells. Then, CXCL12 (50 nM) was added to the medium for S339 (Abcam, Cambridge, UK, no.74012) was performed in all specimens 30 min and the detached cells were carefully removed and each well taken at diagnosis as previously described.15 The immunohistochemical (12 microscopic fields) was monitored using a live cell imaging microscope staining of CXCR4 was performed following the same procedure with a (IX81, Olympus, Tokyo, Japan). The surface area coverage of GFP was primary rabbit polyclonal antibody (Abcam, no. 2074). Sections were analyzed by ImageJ (National Institutes of Health). incubated 30 min with anti-CXCR4 at dilution of 1:50. At least 200 cells were assessed in each trephine biopsy, and the percentage of positive cells Viability assay (that is, cells with distinct staining) was calculated. Cases were considered Kasumi-1 cells were plated at 3 104cells/well with or without the as pCXCR4-S339 positive when the percentage of stained cells per case support of MS-5 cells. Stromal cells (3.6 105) were plated in six-well was above 15%, on the basis of the prognostically relevant cutoff score as plates in a-minimum essential medium with 10% fetal calf serum determined by receiver operating characteristic, see Supplementary the previous day. After 1 day of coculture, Ara-C was added for 48 h. Methods (area under the receiver operating characteristic curve, Cells were harvested, washed and stained with the PO-PRO-1 dye AUROC ¼ 0.641, 95% confidence interval 0.512–0.768, P ¼ 0.037). Cases (Invitrogen, Carlsbad, CA, USA) and 7-aminoactinomycin D before the were considered as CXCR4 positive when the percentage of stained blasts analysis using flow cytometry. The specific apoptosis was calculated by per case was above 20% (that is, above the median expression value for the formula: percentage specific apoptosis ¼ (test—control) 100/ CXCR4 as this marker was not of prognostic significance, and thus receiver (100—control).20 operating characteristic could not be applied to it). Additional methods for quantitative reverse transcription-PCR, CXCR4 surface expression, internalization, recycling, confocal microscopy analysis, immunoblotting, ERK phosphorylation, intracellular Ca2 þ flux measure- Cell culture, DNA constructs and lentivirus preparation ment and statistical analysis are provided in a Supplementary Data File. Kasumi-1 and HEK-293 cells were grown, respectively, in Roswell Park Memorial Institute-1640 and Dulbecco’s Modified Eagle medium contai- ning 10% fetal calf serum in the presence of 5% CO2 in humidified RESULTS atmosphere. The human CXCR4 complementary DNA was kindly provided PhosphoCXCR4-S339 in human AML by M Biard-Piechaczyk (Montpellier, France) and point mutants were To determine the role of CXCR4-S339 phosphorylation in AML, we generated using PCR. The HA-hCXCR4-GFP fusion was expressed with the pEGFP vector (Clontech, Mountain View, CA, USA). The lentiviral vector was performed immunohistochemical staining of pCXCR4-S339 and constructed by cloning the CXCR4 complementary DNAs into the pCMV CXCR4 in BM core biopsies of 78 patients and found that 75 of lentiviral vector (304) coexpressing the green fluorescent protein (EGFP).16 them were evaluable for both markers. Three cases either negative Lentivirus preparation and transduction were performed as described for CXCR4 and pCXCR4-S339, positive for CXCR4 only or positive previously.17 for both are presented in Figure 1a. The majority of the cases (66/75) were CXCR4 positive, but only 38/66 were also found to be Chemotaxis/chemokinesis assay positive for pCXCR4-S339. We also analyzed pCXCR4-S339 and CXCR4 expression levels using Western blot in freshly processed Cells were kept for 2 h in starving medium at 0.5 104/ml before transfering in the upper well of a Transwell chamber (5.0-mm pore size; blasts from six AML patients from our cohort. As shown in Costar, Corning, Cambridge, MA, USA). For chemotaxis, a CXCL12 gradient Figure 1b, we observed that pCXCR4-S339 expression levels (0–100 nM) was applied by adding ligand in the lower chamber. For determined using immunohistochemistry correlated well with the chemokinesis, CXCL12 was added to the upper and lower chamber. After levels found using immunoblotting of the matched samples. We 4 h, the number of cells migrated to the lower chamber was measured. The next evaluated whether CXCR4 or pCXCR4-S339 expression levels migration potential was calculated as the percentage of the input cells. correlate with previously assessed parameters for the same cohort.21,22 We found that patients with positive staining for Transplantation, mobilization and drug resistance experiments CXCR4 showed a trend toward longer overall survival that was Kasumi-1 cells were resuspended in phosphate-buffered saline, and 107 statistically not significant (P ¼ 0.279) (Figure 1c); however, the cells were injected into the tail vein of sub-lethally irradiated (200 rad) presence of pCXCR4-S339 in these patients was associated with non-obese diabetic.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG)-recipient mice significantly shorter overall survival (P ¼ 0.005) (Figure 1d). (Jackson Laboratory, Bar Harbor, ME, USA). All procedures were reviewed Analysis of risk groups according to the protocols of the and approved by the University’s supervisory animal care committee. Mice Dutch-Belgian Cooperative Trial Group for Hematology and

& 2014 Macmillan Publishers Limited Leukemia (2014) 566 – 576 Role of CXCR4-S339 in AML L Brault et al 568 a b CXCR4/pCXCR4 immunostaining in 3 representative AML cases CXCR4/pCXCR4 immunoblot of AML samples AML#22 AML#5 AML#4 AML#3 AML#4 AML#6 AML#1 AML#2 AML#5 Markers (kDa) 55 pCXCR4 40 55 CXCR4 40 55 Actin 40 FLT3-ITD -+ -++ - c Overall survival of AML patients 1.0

0.8

0.6 n=66

0.4 n=9 pCXCR4 CXCR4 Cumulative OS 0.2 P = 0.279 0.0 0 12 24 36 48 60 Follow up (months) CXCR4(-) CXCR4(+)

def Overall survival of CXCR4(+) Overall survival of CXCR4(+) Overall survival of CXCR4(+) patients FLT3 wt patients FLT3-ITD patients 1.0 1.0 1.0 n=17 0.8 n=29 0.8 0.8 n=12

0.6 0.6 0.6 n=31 0.4 n=37 0.4 0.4 Cumulative OS Cumulative OS 0.2 0.2 Cumulative OS 0.2 n=6 P = 0.005 P = 0.016 P = 0.092 0.0 0.0 0.0 0 12 24 36 48 60 0 12 24 36 48 60 0 12 24 36 48 60 Follow up (months) Follow up (months) Follow up (months) pCXCR4(-) pCXCR4(+) pCXCR4(-) pCXCR4(+) pCXCR4(-) pCXCR4(+) Figure 1. Expression of CXCR4/pCXCR4-S339 in patients with AML. (a) Expressions of CXCR4 and pCXCR4 in three representative AML cases. Case no. 22 negative for CXCR4 and pCXCR4, case no. 5 positive for CXCR4 and negative for pCXCR4 and case no. 4 positive for CXCR4 and for pCXCR4. (b) pCXCR4 and CXCR4 protein levels in AML patients samples determined using immunoblotting. Actin served as a loading control. (c) Kaplan–Meyer estimates of overall survival according to CXCR4 expression in BM samples of AML patients. (d) Kaplan–Meyer estimates of overall survival according to pCXCR4 expression in BM samples of CXCR4-positive AML patients. (e) Kaplan–Meyer estimates of overall survival according to pCXCR4 expression in BM samples of CXCR4-positive AML patients with FLT3 WT. (f) Kaplan–Meyer estimates of overall survival according to pCXCR4 expression in BM samples of CXCR4-positive AML patients with FLT3-internal tandem duplication (ITD).

Oncology (HOVON, http://www.hovon.nl/trials/trials-by-type/aml. Modeling CXCR4-S339 phosphorylation in cell lines html) revealed that positive-pCXCR4-S339 staining correlated with To study the consequence of aberrant CXCR4-S339 phosphoryla- a significantly shorter overall survival for the intermediate risk tion for the biology of AML, we engineered receptor variants that group (P ¼ 0.003), whereas the differences in the other risk groups either mimic a constitutively phosphorylated state by did not reach significance (Supplementary Figure 1). We found exchanging serine339 to glutamic acid (S339E) or that abrogate that the highest levels of pCXCR4-S339 in immunoblots were phosphorylation by mutating serine to alanine (S339A) associated with the presence of an activating FLT3-internal (Figure 2a).11,12,23,24 To monitor receptor internalization and tandem duplication mutation (Figure 1b). Therefore, we compared recycling, we established HEK-293 cells stably expressing the survival in the FLT3 wild-type (WT) and FLT3-internal tandem WT, S339E or S339A variants that were C-terminally fused to duplication subgroups and found lower overall survival in both GFP. To study the role of S339 in CXCR4 function in leukemic cells, FLT3 WT and FLT3-internal tandem duplication subgroups for AML clones of human Kasumi-1 AML cells were established that stably cases expressing high levels of pCXCR4-S339 that reached express CXCR4 variants in a physiological range (in respect to statistical significance only in the FLT3 WT group (P ¼ 0.016 and Jurkat cells) (Figure 2b and Supplementary Figure 2a). These cell P ¼ 0.092, respectively) (Figures 1e and f). These observations lines were selected based on their undetectable baseline levels of indicate that the presence of pCXCR4-S339 in BM biopsies could CXCR4 at the mRNA and protein level.25,26 The expression of serve as a prognostic biomarker in AML and that increased CXCR4-WT and CXCR4-S339 variants did not change the CXCR4-S339 phosphorylation might enable the leukemic cells to proliferation or survival of Kasumi-1 cells (data not shown). escape chemotherapy. Likewise, comparable surface expression levels and localization of

Leukemia (2014) 566 – 576 & 2014 Macmillan Publishers Limited Role of CXCR4-S339 in AML L Brault et al 569

a b Amino acid sequence of the intracellular C-terminal tail of CXCR4 Kasumi-1 cells 100 CXCR4 80 321 331 341 351 II I I 60 SRGSSLKILS KGKRGGHSSV STESESSSFH SS (WT) 40 SRGSSLKILS KGKRGGHSAV STESESSSFH SS (S339A) 20 SRGSSLKILS KGKRGGHSEV STESESSSFH SS (S339E) Relative cell number 0 100 101 102 103 104 Surface CXCR4

cd Ligand-induced receptor Receptor recycling after wash out of the ligand internalization 100 100 CXCR4-WT CXCR4-WT WT + Wash out CXCR4-S339A CXCR4-S339A S339A + Wash out 75 CXCR4-S339E 75 CXCR4-S339E S339E + Wash out * * 50 50 ** Wash out

25 25 Surface CXCR4 (%) Surface CXCR4 (%)

0 0 012810 0 30 60 90 120 150 180 210 CXCL12 (µM) Time (min)

ef Ligand-induced chemotaxis Ligand-induced chemokinesis 10 2.0 Mock Mock CXCR4-WT CXCR4-WT CXCR4-S339A CXCR4-S339A # # CXCR4-S339E CXCR4-S339E # 8 # 1.5

# ** 6

* 1.0

4 Migrated cells (%) Migrated cells (%)

0.5 2

0 0.0 0 1 10 20 50 100 0 1 10 20 50 100 CXCL12 (nM) CXCL12 (nM) Figure 2. Mutations of Ser339 affect CXCR4 function in vitro.(a) Amino-acid sequence alignment of the C-tail of CXCR4-WT and the two variants for Serine339 (S339A and S339E). (b) Surface expression of CXCR4 on Kasumi-1 cells expressing empty vector (Mock, thin dotted black line), CXCR4-WT (black line), CXCR4-S339A (red line) or CXCR4-S339E (green line) measured using flow cytometry. Isotype is depicted as filled light gray histogram, and Jurkat cells were used as positive controls (dark gray histograms). One out of four representative experiment is shown. (c) CXCR4 surface expression after 30-min stimulation with increasing concentration of CXCL12 measured by flow cytometry and expressed as percentage of non-treated cells. Data represent means of three independent experiments performed in duplicate. (d) CXCR4 recycling determined after CXCL12 stimulation þ / wash out of the ligand after 30 min and expressed as percentage of non-treated cells measured by flow cytometry. Data represent means±s.e. to the mean (s.e.m.) of three independent experiments performed in duplicate. Significance (*Po0.05 and **Po0.01) was calculated by the analysis of variance (ANOVA) test. (e) CXCL12-induced chemotaxis of CXCR4 variants Kasumi-1 cells measured in Transwell assays and expressed as percentage of total input. Data represent means±s.e.m. of three independent experiments performed in duplicate. Significance (*Po0.05, **Po0.01 and #Po0.001) was calculated by ANOVA test. (f) CXCL12-induced chemokinesis of CXCR4 variants Kasumi-1 cells measured in Transwell assays, and expressed as percent of total input. Data represent means±s.e.m. of three independent experiments performed in duplicate.

& 2014 Macmillan Publishers Limited Leukemia (2014) 566 – 576 Role of CXCR4-S339 in AML L Brault et al 570 the receptor were found in HEK-293 cells expressing CXCR4 to the spleen, reaching significance for the cells expressing CXCR4 variants (Supplementary Figure 2b). mutants. The impaired homing was associated with a significantly increased number of CXCR4-S339E-expressing cells in the Impact of S339 mutations on CXCR4 function periphery. No increase was seen for CXCR4-S339A-expressing On the basis of previous observations that the phosphorylation cells possibly due to infiltration into other CXCL12-producing status of CXCR4-S339 affects recycling of the receptor in vitro,we organs like the liver (Figure 3a and Supplementary Figure 5a). At 7 first compared ligand-mediated internalization and recycling days post-transplantation, we found significantly less CXCR4- of the receptor variants in both cell lines.8 We found a similar S339E-expressing cells in the BM that was associated with an dynamics of receptor endocytosis upon CXCL12 stimulation of all increased number in the PB (P40.05) (Figure 3b). No significant CXCR4 variants (Figure 2c). Nevertheless, after exposure to CXCL12 differences in the expression of several adhesion molecules (CD49d, CD49e and CD62L) or apoptosis regulators (Bax, Bak, (50 nM) for 30 min followed by washing out, the CXCR4-S339E variant showed a significantly increased reappearance of the Bcl-2 and Bcl-xL) were found between the variants suggesting that the CXCR4-S339 state is predominantly responsible for the receptor compared with the WT control in Kasumi-1 (Po0.05) and HEK-293 cells (Figure 2d and Supplementary Figure 3). observed homing and engraftment effects (Supplementary Unexpectedly, the CXCR4-S339A variant was not defective in surface Figures 5b and c). We also analyzed mice 7 weeks post-transplant, re-exposure of the receptor. We next compared the migration when the first signs of disease developed. We found that Kasumi-1 capacity of Kasumi-1 cells by measuring chemotaxis (directed cells expressing all CXCR4 variants had expanded in the BM. migration) and chemokinesis (random migration). As expected, However, we found a significantly lowernumberofCXCR4-S339E- the expression of CXCR4 restored the chemotaxis capacity toward and CXCR4-S339A-expressing cells in the PB (Figure 3c). Immu- a CXCL12 gradient in a dose-dependent manner (Figure 2e). Cells nostaining for human CD34 revealed that the CXCR4-expressing expressing CXCR4-S339E and surprisingly also -S339A showed transplanted cells were predominantly present in the BM and only a two- to four-fold increase of chemotaxis capacity when sporadically found in the spleen or other CXCL12-producing organs like the liver, lung or the meningeal space (Figure 3d and compared with WT reaching significance for 10 nM CXCL12 and data not shown). Taken together, these data suggest that the above (Po0.05–0.001). This increased chemotaxis was associated, even though non-significantly (P40.05), with a 1.5- to 3.5-fold CXCR4-S339 state not only regulates homing and engraftment but increased chemokinesis (Figure 2f). These data suggest that also influences long-term distribution of leukemic cells. mimicking phosphorylation of CXCR4-S339 supports ligand- stimulated recycling of the receptor. Surprisingly, abrogation Mutations of CXCR4-S339 affect cell retention and sensitivity to of CXCR4-S339 phosphorylation by the S339A mutant also Ara-C increased chemotaxis and chemokinesis in vitro. As downstream signaling could occur independently of receptor As CXCR4 is a major regulator of retention of hematopoietic cells internalization, we determined ligand-induced ERK phosphorylation in the BM, we also compared in vivo retention to the BM of cells 26 expressing the CXCR4-S339 variants by calculating the ratio and calcium efflux. In contrast to mock-transfected cells, we found 27 a rapid increase of ERK phosphorylation in CXCL12-exposed between cells in the PB and the BM (PB/BM ratio) (Figure 3e). Kasumi-1 cells peaking after 3 min. This ligand-induced ERK Seven days post-transplant, CXCR4-WT-expressing cells displayed phosphorylation increased more than 1.2-fold in cells expressing a low PB/BM ratio of 0.0011, whereas CXCR4-S339E-expressing either the S339E or the S339A variant (Supplementary Figure 4a). cells displayed a significantly higher ratio with 0.0063 (Po0.01). These data indicate that CXCL12-induced signaling is restored in Interestingly, after 7 weeks, we found a significantly lower ratio for Kasumi-1 cells expressing CXCR4-WT and ERK signaling is enhanced cells expressing CXCR4-S339E as compared with cells expressing in cells expressing CXCR4-S339 variants. CXCR4-WT with 0.003 and 0.017, respectively (Po0.01). Together, these data show that expression of CXCR4-S339E mutation initially impairs homing, but later significantly increases the retention Mutations of CXCR4-S339 impair homing and engraftment of of Kasumi-1 AML cells in the BM. leukemic cells As the homing and engraftment capacity in vivo did not As phosphorylation of CXCR4-S339 in AML blasts was associated correlate with the observed in vitro migration capacity (Figures 2e, with poor outcome of chemotherapy-treated patients, we aimed 3a and b), we also determined the adhesion and detachment of to study how aberrant S339 phosphorylation might affect Kasumi-1 cells expressing the CXCR4-S339 variants to a monolayer CXCR4-regulated homing, lodging or BM retention of leukemic of HUVECs known to express CXCL12.28 We found increased cells. We transplanted Kasumi-1 AML cells stably expressing the adhesion to HUVEC’s of CXCR4-S339A and -S339E (1.5-fold, CXCR4-S339 variants into severe immunodeficient (NSG) mice and Po0.01) when compared with CXCR4-WT-expressing cells tracked the cells 24 h (reflecting homing) and 7 days (reflecting (Figures 4a–c). Likewise, we observed increased chemotaxis of engraftment/lodging) post-transplant. As shown in Figure 3a, the cells toward HUVEC’s conditioned medium (Figure 4b). Kasumi-1 cells expressing CXCR4-WT efficiently reached the BM, To compare cellular detachment, we added exogenous CXCL12 whereas mock-transduced cells homed very poorly. Unexpectedly, before removing the non-adherent cells and measured the expression of the CXCR4-S339A or -S339E variant significantly number of cells remaining. As shown in Figure 4d, we found reduced homing to 60% (Po0.05) and 20% (Po0.01) of cells a significantly lower number of CXCR4-WT (30%, Po0.05) than found in the BM, respectively, when compared with CXCR4- CXCR4-S339E (92%)- and CXCR4-S339A (55%)-expressing cells. WT-expressing cells. A similar reduction was observed for homing Activation of the HUVEC’s by tumor necrosis factor-a did not

Figure 3. CXCR4-S339 variants affect the BM homing, engraftment and expansion of Kasumi-1 cells in immunodeficient NSG mice. (a) Homing is shown by the percentage of GFP-positive Kasumi-1 cell variants found in the BM, spleen and PB of NSG mice after 24 h (n ¼ 5). (b) Engraftment is shown by the percentege of GFP-positive Kasumi-1 cell variants found in the BM and spleen of NSG mice and residual cells in the periphery 7 days after transplantation (n ¼ 6). (c) Expansion is shown by percentage of GFP-positive Kasumi-1 cell variants found in the BM, spleen and residual cells in the periphery of NSG mice 7 weeks after transplantation (n ¼ 6). Significance (*Po0.05, **Po0.01) was calculated using a Mann–Whitney test. (d) Infiltration of Kasumi-1 cell variants in the BM and spleen (hematoxylin-eosin and hCD34 stain). (e) Ratio between the number of Kasumi-1 variant cells in the periphery to the BM (calculated as indicated in Methods section) at 7 days (engraftment) and 7 weeks (expansion) after transplantation (n ¼ 6). Data represent means±s.e.m.. Significance (**Po0.01) was calculated using ANOVA test.

Leukemia (2014) 566 – 576 & 2014 Macmillan Publishers Limited Role of CXCR4-S339 in AML L Brault et al 571 abcHoming Engraftment Expansion PB PB PB 0.4 ** 0.020 30 * * ** 0.3 0.015 * 20 0.2 0.010 10 0.1 0.005 GFP+ cells (%) 0.0 0.000 0 BM BM BM 0.25 0.20 50 ** * * 0.20 ** 0.15 40 0.15 30 0.10 0.10 20 0.05 0.05 10 GFP+ cells (%) 0.00 0.00 0 Spleen Spleen Spleen 0.04 0.015 15 * 10 5 0.03 0.010 1.0 0.02 * 0.005 0.5 0.01 GFP+ cells (%) 0.00 0.000 0.0

d hCD34 immunostaining in transplanted mice e Peripheral blood to bone marrow ratio Mock CXCR4-WT CXCR4-S339A CXCR4-S339E 7days 0.008 **

0.006

0.004

0.002 BM (7weeks)

0.000 hCD34 H&E 7weeks 0.025

0.020

0.015

0.010

0.005 **

0.000 Spleen (7weeks) hCD34 H&E

& 2014 Macmillan Publishers Limited Leukemia (2014) 566 – 576 Role of CXCR4-S339 in AML L Brault et al 572 significantly alter the adhesion capacity but abrogated CXCL12 important for the activation of the receptor in solid cancer cells. mediated cellular detachment (Supplementary Figures 6a and b). We have previously shown that S339 is a target of the PIM1 The increased retention of CXCR4 mutant cells led us to serine/threonine protein kinase and that PIM1-mediated evaluate their sensitivity to chemotherapy-induced apoptosis.29 phosphorylation of CXCR4-S339 accelerated recycling of the We found that in the absence of stroma, Kasumi-1 cells expressing receptor.14 Subsequent work has proposed that CXCR4-S339 the different CXCR4-S339 variants displayed a similar sensitivity to phosphorylation is associated with poor survival in adults B-cell Ara-C (Figure 4d). However, upon coculture on MS-5 stromal acute lymphoblastic leukemia.34 By immunohistochemical cells, we observed a significantly increased resistance to 1 mM staining of BM biopsies, we now found significant CXCR4-S339 Ara-C for CXCR4-S339E (2.9-fold, Po0.01) and CXCR4-S339A phosphorylation in 38 of 75 evaluable AML cases (Figure 1). As (1.7-fold, Po0.05)-expressing cells that was maintained for reported for B-cell acute lymphoblastic leukemia, CXCR4-S339 CXCR4-S339E-expressing cells at 5 mM Ara-C (Figure 4e and phosphorylation was significantly associated with poor overall Supplementary Figure 6c). These data indicate that the CXCR4- survival. In addition, pCXCR4-S339-positive staining seemed to be S339 status significantly influences in vitro adhesion, retention and associated with a poor prognosis independent of the FLT3-internal stroma-mediated drug-resistance of leukemic cells. tandem duplication status (Figure 1). Unfortunately, several attempts failed to quantify pCXCR4-S339 by flow cytometry using the currently available antibody (data not shown) limiting the Mutations of CXCR4-S339 affect leukemic cell mobilization and drug resistance comparative quantification of surface CXCR4 and pCXCR4 to the limited availability of material for immunoblotting. The overall As cells expressing CXCR4 variants displayed different in vitro and impact of pCXCR4-S339 as a novel prognostic encouraging in vivo retention capacities, we wondered, whether alteration of biomarker in the AML diagnostic workup will need to be CXCR4-S339 would also affect mobilization of the cells from the prospectively validated in a larger collective of patients. BM in response to the competitive CXCR4 antagonist AMD3100. To functionally study the role of CXCR4-S339, we established We found in vitro, that CXCR4-S339E-expressing Kasumi-1 cell lines-expressing receptors with different variants of this cells were more resistant to AMD-3100-mediated inhibition of residue either mimicking (S339E) or ablating phosphorylation Transwell migration toward CXCL12 (Figure 5a). To evaluate in vivo (S339A). The widespread expression of CXCR4 made it rather mobilization, we applied AMD3100 (5 mg/kg) to mice 5 weeks 2 challenging to establish a cellular model with stable expression of after transplanting Kasumi-1 cells expressing the CXCR4 variants. the CXCR4 variants in absence of a functional wild-type receptor. As shown in Figure 5b, AMD3100 treatment mobilized normal BM Previous studies have shown that expression of CXCR4, for (GFP negative) resulting in a twofold increase of the white blood example, in the K562 leukemia cell line is not able to fully cell count in the periphery. Kasumi-1 cells (GFP positive) reconstitute its function.35 We established Kasumi-1 human AML expressing CXCR4-WT mobilized to the periphery (4.2-fold) more as well as HEK-293 (both expressing negligible levels of efficiently than cells expressing the CXCR4-S339A (2.5-fold) or the endogenous CXCR4) cell clones that stably exposed CXCR4-S339 CXCR4-S339E variant (2.3-fold, Po0.001) (Figure 5b). Increased variants at their surface within a physiological range in steady mobilization was associated with a reduction of Kasumi-1 cells state. Upon CXCL12 stimulation, the exogenous expressed remaining in the BM (Figure 5c). These data indicate that receptor was properly internalized and recycled after washing the CXCR4-S339 status also determines the BM retention and out the ligand. Receptor internalization was not affected by AMD-3100-mediated mobilization of AML cells. mutations either mimicking (S339E) or abrogating (S339A) As cells expressing CXCR4 variants displayed an increased phosphorylation (Figure 2). Interestingly, significantly increased in vitro resistance to Ara-C in presence of stroma, we also recycling was observed for CXCR4-S339E variant in Kasumi-1 cells, investigated the impact of CXCR4-S339 mutation on drug going along with our previous findings suggesting that phospho- resistance in vivo. Hereby, we applied Ara-C (75 mg/kg) for 5 rylation of S339 can positively regulate receptor recycling.14 The consecutive days to mice 5 weeks after transplant of Kasumi-1 enhanced recycling of the CXCR4-S339E variant was associated cells expressing the CXCR4-S339 variants. As expected, Ara-C with reduced calcium efflux, suggesting a mechanistic link treatment decreased the white blood cell count and the BM between intracellular calcium levels and receptor trafficking cellularity (data not shown). Interestingly, the number of Kasumi-1 (Supplementary Figure 4b). cells (GFP positive) expressing CXCR4-WT or CXCR4-S339A was Unlike other currently used models, stable expression of CXCR4 markedly reduced upon Ara-C treatment, whereas cells expressing in Kasumi-1 cells allowed studying intracellular downstream the CXCR4-SE variant were significantly less affected (Po0.05, signaling like activation of ERK or calcium efflux.35 Despite Figure 5d). This observation suggests that the CXCR4-S339 the marginal impact on receptor internalization/recycling, the status influences Ara-C resistance of AML cells through altered CXCR4-S339 variants were associated with a 1.2-fold enhanced interaction with the microenvironment. ERK activation that correlated with a two- to four-fold increased migration capacity when compared with CXCR4-WT. In regard of the calcium efflux discrepancies, these results suggest that in vitro DISCUSSION cell migration might be mainly driven through ERK-associated Several studies have demonstrated that elevated levels of CXCR4 signaling rather than through calcium-dependent signaling.36,37 on the surface of leukemic blasts are associated with an It is unclear to what extent the observed small increase of unfavorable AML prognosis in particular in patients with normal receptor recycling is responsible for the differential behavior karyotypes, but the functional consequence remained of cells expressing the CXCR4-S339E variant. The contribution of unclear.5,6,30,31 Increased CXCR4 levels were proposed to receptor internalization and recycling for the CXCR4 function correlate with engraftment of leukemic blasts in non-obese remains mostly unknown. However, cellular migration and diabetic/severe combined immunodeficiency mice but not adhesion can be independent from receptor internalization as confirmed by other studies.2,32 Biochemical studies suggested shown in cells from patients with WHIM (Warts, Hypogamma- that the function of CXCR4 is dynamically regulated by globulinemia, Infections and Myelokathexis) syndrome, with site-specific phosphorylation of multiple serine residues (S321, almost entire loss of the intracellular C-terminal tail of S324, S325, S330, S339, S346 and S352) in the intracellular CXCR4 that abolishes internalization of the receptor while C-terminal domain of the receptor.12,13 Rubin and co-workers33 in vitro chemotaxis and engraftment upon transplantation is developed antibodies detecting phosphorylated S339 that enhanced.35,38 Likewise, CXCR4-mediated activation of ERK seems allowed demonstrating that phosphorylation of this residue is to occur independently from receptor internalization.26 Therefore,

Leukemia (2014) 566 – 576 & 2014 Macmillan Publishers Limited Role of CXCR4-S339 in AML L Brault et al 573 a Adhesion and detachment on HUVECs Mock CXCR4-WT CXCR4-S339A CXCR4-S339E - CXCL12 + CXCL12

bc d MigrationAdhesion Detachment 25 2000 1.5 * * ** ** 20 1500 1.0 15 1000 10 0.5 * 500 Migrated cells (%) 5 GFP+ area (pixels)

0 0 Proportion of adherent cells 0.0

e Ara-C resistance on stroma **

30 * 25 20 10

5 Specific apoptosis (%)

0 MS-5- + - + - +

CXCR4-WT CXCR4-S339A CXCR4-S339E Figure 4. CXCR4-S339 variants determine adhesion, detachment and stroma-mediated chemoresistance in vitro.(a) Adhesion of CXCR4 variants-expressing Kasumi-1 cells to a HUVEC cells monolayer was determined using live cell microscopy: representative phase contrast pictures ( 20) for the indicated cells before ( CXCL12) and after addition of ligand ( þ CXCL12) are presented. (b) Migration capacity of CXCR4 variants-expressing Kasumi-1 cells toward HUVEC’s conditioned media (CM) measured in Transwell assays and expressed as the percentage of total input. Data represent means±s.e.m. of three independent experiments performed in duplicate. Significance (*Po0.05) was calculated using ANOVA test. (c) Proportion of adherent CXCR4 variants-expressing Kasumi-1 cells after 1-h incubation on HUVECs expressed as relative units of GFP-positive cell density. Data represent means±s.e.m. of three independent experiments performed in quadruplicate. Significance (**Po0.01) was calculated using ANOVA test. (d) Detachment of CXCR4 variants-expressing Kasumi-1 cells after 1-h incubation on HUVEC followed by 30-min incubation in the presence of CXCL12 (50 nM) was expressed as ratio of GFP-positive cell compared with mock-control cells. Data represent means±s.e.m. of three independent experiments performed in quadruplicate. Significance (*Po0.05) was calculated using ANOVA test. (e) Proportion of specific apoptosis induced by Ara-C treatment (1 mM) in Kasumi-1 cells expressing CXCR4 variants cultured with or without MS-5 stromal cells. Data represent means±s.e.m. of three independent experiments performed in duplicate. Significance (*Po0.05) was calculated using an unpaired t-test. the differential behavior of CXCR4-S339E-expressing cells might However, rather unexpected, expression of both the S339E and be the result of altered receptor recycling in conjunction of altered S339A mutants resulted in increased migration. The reason for downstream signals. this behavior remains unclear. It could be the consequence Our experiments demonstrate for the first time that alterations of a conformational change associated with the mutation of S339 are directly translated into altered signaling and function. (either S339A or S339E) that might also affect receptor trafficking

& 2014 Macmillan Publishers Limited Leukemia (2014) 566 – 576 Role of CXCR4-S339 in AML L Brault et al 574

a Migration b Retention # 20 10

15 *

5 10 #

** GFP+ cells (%) Migrated cells (%) * 5

0 0 1 0 0 1 0 1 0 1 0.1 0.1 0.1 AMD3100 0.1 AMD3100 -+-+ -+ 0.01 0.01 0.01 0.01 0.001 0.001 0.001 (µM) 0.001

Mock CXCR4 -WT CXCR4 -S339A CXCR4 -S339E

c Mobilization d Ara-C resistance GFP- cells GFP+ cells * ) 3 60000 6 # 4000 * 3000 40000 2

2000 WBC/µL 20000 WBC/µL 1 1000

0 0 GFP+ cells per Femur (x10 0 AMD3100 -+-+-+ -+-+-+ Ara-C -+-+-+

e Proposed model for CXCR4-S339 role in vivo

CXCR4-S339 Wildtype CXCR4-S339 mutant Blood Blood Adhesion Adhesion

Mobilization Mobilization Homing Homing Engraftment Engraftment

Bone Expansion Bone Expansion marrow marrow

Bone Bone Early phase Late phase Early phase Late phase Figure 5. CXCR4 variants impair AMD3100-induced mobilization and increase chemoresistance of Kasumi-1 cells in vivo.(a) The effect of AMD3100 (0–1 mM) on CXCL12-induced chemotaxis of CXCR4 variants Kasumi-1 cells was measured in Transwell assays and expressed as percentage of cells migrated to the lower chamber of the total input (data represent means±s.e.m. of three independent experiments performed in duplicates, significance (*Po0.05 and **Po0.01) was calculated using ANOVA test). (b) Total white blood cell (WBC) count per microliter of PB was determined using automated counting and the proportion of GFP-positive Kasumi-1 cells was determined by flow cytometry (n ¼ 6). Data represent means±s.e.m. Significance (*Po0.05 and #Po0.001) was calculated using a Mann–Whitney test. (c) Proportion of the residual GFP-positive Kasumi-1 cell variants in the BM after AMD3100 injection (n ¼ 6). Data represent means±s.e.m. Significance (#Po0.001) was calculated using ANOVA test. (d) Number of the residual GFP-positive Kasumi-1 cell variants in the BM after 5 Ara-C injections (n ¼ 4–5). Data represent means±s.e.m. Significance (*Po0.05) was calculated using ANOVA test. (e) The proposed model for the role of the CXCR4-S339 phosphorylation status for adhesion, homing, engraftment, expansion and mobilization of leukemic cells.

Leukemia (2014) 566 – 576 & 2014 Macmillan Publishers Limited Role of CXCR4-S339 in AML L Brault et al 575 (internalization/recycling) which was not observed (Figures 2 and ACKNOWLEDGEMENTS 3, and Supplementary Figures 2 and 3). It could rather result from We thank Sarah Thommen and Sabine Juge for technical help, Susan Treves for a compensatory mechanism implicating other potentially advice and Emmanuel Traunecker for cell sorting. We would like also to thank Oliver phosphorylated sites of the C-tail.12 Further studies will be Pertz for fruitful discussion, Fawzia Louache, Cristina Lo Celso and Radek Skoda for necessary to dissect the complex signaling network, including critically reading of the manuscript. This work was supported by the Gertrude Von S339 and the 16 additional serine/threonine residues located in Meissner Foundation (Basel), and grants from SNF (31003A-116587) and ONCOSUISSE the C-terminal intracellular tail of the receptor. (OCS-01830-02-2006) to JS. The Kasumi-1 AML cells expressing the CXCR4-S339 variants allowed to study the impact on homing, retention and mobili- zation to the BM. Expression of exogenous CXCR4 restored the REFERENCES poor homing and engraftment capacity of Kasumi-1 AML cells 1 Peled A, Petit I, Kollet O, Magid M, Ponomaryov T, Byk T et al. Dependence of specifically to the BM when transplanted into NSG mice human stem cell engraftment and repopulation of NOD/SCID mice on CXCR4. demonstrating its functionality (Figure 3). The enhanced homing Science 1999; 283: 845–848. and early cellular engraftment to the BM observed at days 1 and 7 2 Tavor S, Petit I, Porozov S, Avigdor A, Dar A, Leider-Trejo L et al. CXCR4 regulates for CXCR4-WT-expressing cells were not associated with increased migration and development of human acute myelogenous leukemia stem cells in long-term expansion and retention (7 weeks). On the contrary, transplanted NOD/SCID mice. Cancer Res 2004; 64: 2817–2824. CXCR4-S339 mutant-expressing cells displayed an increased 3 Alkhatib G. The biology of CCR5 and CXCR4. Curr Opin HIV AIDS 2009; 4: 96–103. accumulation in the BM between 1 and 7 weeks post-transplant 4 Busillo JM, Benovic JL. Regulation of CXCR4 signaling. Biochim Biophys Acta 2007; (fold increase of GFP-positive cells (%): WT: 120 , S339A: 410 , 1768: 952–963. 5 Rombouts EJ, Pavic B, Lowenberg B, Ploemacher RE. Relation between CXCR-4 S339E: 500 ). This was associated with the increased retention expression, Flt3 mutations, and unfavorable prognosis of adult acute myeloid of the cells to the BM as reflected by a smaller number leukemia. Blood 2004; 104: 550–557. of cells detected in the periphery (Figure 3e). These observations 6 Spoo AC, Lubbert M, Wierda WG, Burger JA. CXCR4 is a prognostic marker in acute suggest that phosphorylation of the CXCR4 C-terminal myelogenous leukemia. Blood 2007; 109: 786–791. tail determines the retention capacity of hematopoietic cells to 7 Voermans C, van Heese WP, de Jong I, Gerritsen WR, van Der Schoot CE. Migratory the BM. behavior of leukemic cells from acute myeloid leukemia patients. Leukemia 2002; Expression of the CXCR4 variants correlated with enhanced 16: 650–657. cellular adhesion, whereas the expression of CXCR4-WT led to 8 Lefkowitz RJ. G protein-coupled receptors. III. New roles for receptor kinases and a significantly increased ligand-mediated detachment of the cells. beta-arrestins in receptor signaling and desensitization. J Biol Chem 1998; 273: Interestingly, the cells that expressed CXCR4-S339E mutant 18677–18680. showed stronger retention that correlated with reduced apoptosis 9 Calandra G, Bridger G, Fricker S. CXCR4 in clinical hematology. Curr Top Microbiol Immunol 2010; 341: 173–191. in vitro and with an increased leukemia burden in vivo upon Ara-C 10 Guinamard R, Signoret N, Ishiai M, Marsh M, Kurosaki T, Ravetch JV. B cell antigen treatment (Figures 4 and 5). Similarly, functional interference with receptor engagement inhibits stromal cell-derived factor (SDF)-1alpha CXCR4 in xenografts of leukemic cells was associated with chemotaxis and promotes protein kinase C (PKC)-induced internalization a reduced burden that translated into a significantly increased of CXCR4. J Exp Med 1999; 189: 1461–1466. 39,40 overall survival upon Ara-C treatment. Taken together, these 11 Minina S, Reichman-Fried M, Raz E. Control of receptor internalization, signaling observations indicate that aberrant CXCR4-S339 phosphorylation level, and precise arrival at the target in guided cell migration. Curr Biol 2007; 17: might be a critical nominator of retention and resistance to 1164–1172. chemotherapy of leukemic cells and could account for the poor 12 Orsini MJ, Parent JL, Mundell SJ, Marchese A, Benovic JL. Trafficking of the HIV overall survival observed in pCXCR4-positive AML patients. coreceptor CXCR4. Role of arrestins and identification of residues in the c-terminal Cells expressing CXCR4-S339E variant were significantly tail that mediate receptor internalization. J Biol Chem 1999; 274: 31076–31086. impaired in AMD3100-induced mobilization (Figure 5). Disrupting 13 Busillo JM, Armando S, Sengupta R, Meucci O, Bouvier M, Benovic JL. Site-specific phosphorylation of CXCR4 is dynamically regulated by multiple kinases and the interaction of leukemic blasts with the microenvironment 40 results in differential modulation of CXCR4 signaling. J Biol Chem 2010; 285: increases sensitivity of AML to chemotherapy, and feasibility of 7805–7817. therapeutic chemosensitization by AMD3100 (CXCR4 antagonist) 14 Grundler R, Brault L, Gasser C, Bullock AN, Dechow T, Woetzel S et al. Dissection of in relapsed or refractory AML patients has been recently PIM serine/threonine kinases in FLT3-ITD-induced leukemogenesis reveals PIM1 41 reported. It will be interesting to explore whether the as regulator of CXCL12-CXCR4-mediated homing and migration. J Exp Med 2009; CXCR4-S339 phosphorylation status might influence the 206: 1957–1970. response to this novel therapeutic strategy. 15 Brault L, Menter T, Obermann EC, Knapp S, Thommen S, Schwaller J et al. PIM Taken together, our observations indicate that the integrity kinases are progression markers and emerging therapeutic targets in diffuse large of CXCR4-S339 is essential for migration and homing to the BM B-cell lymphoma. Br J Cancer 2012; 107: 491–500. of AML cells, and proper mobilization. Mimicking aberrant 16 Amendola M, Venneri MA, Biffi A, Vigna E, Naldini L. Coordinate dual-gene CXCR4-S339 phosphorylation (by expression of the S339E transgenesis by lentiviral vectors carrying synthetic bidirectional promoters. Nat Biotechnol 2005; 23: 108–116. mutation) resulted in increased expansion and reduced mobi- 17 Liu T, Jankovic D, Brault L, Ehret S, Baty F, Stavropoulou V et al. Functional lization associated with increased chemoresistance (Figure 5e). characterization of high levels of meningioma 1 as collaborating oncogene in Whether CXCR4-S339 phosphorylation mediated aberrant acute leukemia. Leukemia 2010; 24: 601–612. retention to the BM or to other compartments is directly 18 Shaposhnikov VL. [Distribution of the bone marrow cells in the skeleton of mice]. responsible for the disease progression and/or relapse, as Biull Eksp Biol Med 1979; 87: 483–485. suggested by the poor prognosis of pCXCR4-S339-positive AML, 19 Ng KP, Ebrahem Q, Negrotto S, Mahfouz RZ, Link KA, Hu Z et al. p53 independent needs to be experimentally addressed. Nevertheless, our work epigenetic-differentiation treatment in xenotransplant models of acute myeloid suggests that interference at CXCR4-S339 phosphorylation leukemia. Leukemia 2011; 25: 1739–1750. with small-molecule inhibitors targeting the involved kinases 20 Zeng Z, Shi YX, Samudio IJ, Wang RY, Ling X, Frolova O et al. Targeting the like PIM1, GRK6 or PKC might offer a strategy to increase the leukemia microenvironment by CXCR4 inhibition overcomes resistance to kinase mobilization of potentially relapse-initiating leukemic stem cells. inhibitors and chemotherapy in AML. Blood 2009; 113: 6215–6224. 21 Obermann EC, Arber C, Jotterand M, Tichelli A, Hirschmann P, Tzankov A. Expression of pSTAT5 predicts FLT3 internal tandem duplications in acute myeloid leukemia. Ann Hematol 2010; 89: 663–669. 22 Tzankov A, Strasser U, Dirnhofer S, Menter T, Arber C, Jotterand M et al. In situ CONFLICT OF INTEREST RHAMM protein expression in acute myeloid leukemia blasts suggests poor The authors declare no conflict of interest. overall survival. Ann Hematol 2011; 90: 901–909.

& 2014 Macmillan Publishers Limited Leukemia (2014) 566 – 576 Role of CXCR4-S339 in AML L Brault et al 576 23 Hicke L, Zanolari B, Riezman H. Cytoplasmic tail phosphorylation of the 33 Woerner BM, Warrington NM, Kung AL, Perry A, Rubin JB. Widespread CXCR4 alpha-factor receptor is required for its ubiquitination and internalization. J Cell activation in astrocytomas revealed by phospho-CXCR4-specific antibodies. Biol 1998; 141: 349–358. Cancer Res 2005; 65: 11392–11399. 24 Pitcher C, Honing S, Fingerhut A, Bowers K, Marsh M. Cluster of differentiation 34 Konoplev S, Jorgensen JL, Thomas DA, Lin E, Burger J, Kantarjian HM et al. antigen 4 (CD4) endocytosis and adaptor complex binding require activation of Phosphorylated CXCR4 is associated with poor survival in adults with B-acute the CD4 endocytosis signal by serine phosphorylation. Mol Biol Cell 1999; 10: lymphoblastic leukemia. Cancer 2011; 117:31. 677–691. 35 Kawai T, Choi U, Whiting-Theobald NL, Linton GF, Brenner S, Sechler JM et al. 25 Mohle R, Bautz F, Rafii S, Moore MA, Brugger W, Kanz L. The chemokine receptor Enhanced function with decreased internalization of carboxy-terminus CXCR-4 is expressed on CD34 þ hematopoietic progenitors and leukemic cells truncated CXCR4 responsible for WHIM syndrome. Exp Hematol 2005; 33: and mediates transendothelial migration induced by stromal cell-derived factor-1. 460–468. Blood 1998; 91: 4523–4530. 36 Delgado-Martin C, Escribano C, Pablos JL, Riol-Blanco L, Rodriguez-Fernandez JL. 26 Roland J, Murphy BJ, Ahr B, Robert-Hebmann V, Delauzun V, Nye KE et al. Role of Chemokine CXCL12 uses CXCR4 and a signaling core formed by bifunctional Akt, the intracellular domains of CXCR4 in SDF-1-mediated signaling. Blood 2003; 101: extracellular signal-regulated kinase (ERK)1/2, and mammalian target 399–406. of rapamycin complex 1 (mTORC1) proteins to control chemotaxis and 27 Rosu-Myles M, Gallacher L, Murdoch B, Hess DA, Keeney M, Kelvin D et al. survival simultaneously in mature dendritic cells. J Biol Chem 2011; 286: The human hematopoietic stem cell compartment is heterogeneous for CXCR4 37222–37236. 37 Ryu CH, Park SA, Kim SM, Lim JY, Jeong CH, Jun JA et al. Migration of human expression. Proc Natl Acad Sci USA 2000; 97: 14626–14631. umbilical cord blood mesenchymal stem cells mediated by stromal cell-derived 28 Zernecke A, Bidzhekov K, Noels H, Shagdarsuren E, Gan L, Denecke B et al. factor-1/CXCR4 axis via Akt, ERK, and p38 signal transduction pathways. Biochem Delivery of microRNA-126 by apoptotic bodies induces CXCL12-dependent Biophys Res Commun 2010; 398: 105–110. vascular protection. Sci Signal 2009; 2: ra81. 38 Kawai T, Choi U, Cardwell L, DeRavin SS, Naumann N, Whiting-Theobald NL et al. 29 Macanas-Pirard P, Leisewitz A, Broekhuizen R, Cautivo K, Barriga FM, Leisewitz F et al. WHIM syndrome myelokathexis reproduced in the NOD/SCID mouse Bone marrow stromal cells modulate mouse ENT1 activity and protect xenotransplant model engrafted with healthy human stem cells transduced with leukemia cells from cytarabine induced apoptosis. PLoS One 2012; 7: e37203. C-terminus-truncated CXCR4. Blood 2007; 109: 78–84. 30 Konoplev S, Rassidakis GZ, Estey E, Kantarjian H, Liakou CI, Huang X et al. 39 Chen Y, Jacamo R, Konopleva M, Garzon R, Croce C, Andreeff M. Overexpression of CXCR4 predicts adverse overall and event-free survival in CXCR4 downregulation of let-7a drives chemoresistance in acute myeloid patients with unmutated FLT3 acute myeloid leukemia with normal karyotype. leukemia. J Clin Invest 2013; 123: 2395–2407. Cancer 2007; 109: 1152–1156. 40 Nervi B, Ramirez P, Rettig MP, Uy GL, Holt MS, Ritchey JK et al. Chemosensitization 31 Tavernier-Tardy E, Cornillon J, Campos L, Flandrin P, Duval A, Nadal N et al. of acute myeloid leukemia (AML) following mobilization by the CXCR4 antagonist Prognostic value of CXCR4 and FAK expression in acute myelogenous leukemia. AMD3100. Blood 2009; 113: 6206–6214. Leuk Res 2009; 33: 764–768. 41 Uy GL, Rettig MP, Motabi IH, McFarland K, Trinkaus KM, Hladnik LM et al. 32 Monaco G, Konopleva M, Munsell M, Leysath C, Wang RY, Jackson CE et al. A phase 1/2 study of chemosensitization with the CXCR4 antagonist Engraftment of acute myeloid leukemia in NOD/SCID mice is independent of plerixafor in relapsed or refractory acute myeloid leukemia. Blood 2012; 119: CXCR4 and predicts poor patient survival. Stem Cells 2004; 22: 188–201. 3917–3924.

Supplementary Information accompanies this paper on the Leukemia website (http://www.nature.com/leu)

Leukemia (2014) 566 – 576 & 2014 Macmillan Publishers Limited