|
|
||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

,¶
,¶
,¶
,¶
,¶
,¶,1
* CHRU de Tours, Laboratoires dHématologie et dImmunologie et Service dHématologie et Thérapie Cellulaire;
Université François Rabelais de Tours, équipes EA-3853 et EA-3855; and
¶ Inserm, équipe ESPRI "Microenvironnement de lhématopoïèse et cellules souches," Tours, France
1Correspondence: Laboratoire dHématologie / Equipe Inserm ESPRI-EA3855, CHRU-Hôpital Bretonneau, 2 bd Tonnelle, Tours Cedex F-37044, France. E-mail: olivier.herault{at}med.univ-tours.fr
ABSTRACT
Stromal cell-derived factor 1 (SDF-1), a chemokine abundantly produced by the bone marrow microenvironment, and its receptor CXCR4 have crucial roles in malignant cell trafficking. In acute myeloid leukemia (AML), blasts invade the bloodstream and may localize in extramedullar sites, with variations from one patient to another. We hypothesized that a polymorphism in the SDF-1 coding gene (CXCL12 G801A) could influence blast dissemination and tissue infiltration in AML. CXCL12 G801A polymorphism was determined in 86 adult patients and 100 healthy volunteers. The allelic status and CXCR4 expression on bone marrow blasts were analyzed in relation to peripheral blood blast (PBB) counts and frequency of extramedullar tumor sites. 801A carrier status (801G/A, 801A/A) was found to be associated with a higher PBB count compared with 801G/G homozygous patients (P=0.031) and higher frequency of extramedullar tumor sites (odds ratio 2.92, 95% confidence interval 1.187.21, P=0.018). Moreover, the PBB count was correlated with CXCR4 expression (correlation coefficient 0.546, P=0.001) when considering 801A carriers. In conclusion, a polymorphism in the SDF-1 gene is shown for the first time to be associated with the clinical presentation of a malignant hematological disease and more generally with the risk of distant tissue infiltration by tumor cells.Dommange, F., Cartron, G., Espanel, C., Gallay, N., Domenech, J., Benboubker, L., Ohresser, M., Colombat, P., Binet, C., Watier, H., Herault, O., for the GOELAMS Study Group. CXCL12 polymorphism and malignant cell dissemination/tissue infiltration in acute myeloid leukemia.
Key Words: SDF-1 CXCR4 blast cells microenvironment
ACUTE MYELOID LEUKEMIA (AML) is characterized by uncontrolled proliferation within the bone marrow of myeloid progenitors arrested in their maturation process. In contrast to normal hematopoiesis, it is usually associated with egress of immature cells from the bone marrow into the circulation before anchoring in extramedullar locations. Peripheral blood blast (PBB) counts and the number of extramedullar tumor sites are extremely variable from one patient to another and depend in part on AML subtype. Chemokine stromal cell-derived factor-1 (SDF-1) and its receptor CXCR4 are involved in the trafficking of malignant cells (1)
. SDF-1/CXCR4 signaling is active in many cancer cells, including those of solid tumors and hematological malignancies. CXCR4 and its ligand SDF-1 were recently shown to have an important role in breast, prostate, ovarian, and sympathetic nervous system cancer metastasis (2
3
4
5
6)
, as well as in the in vitro migration assays of malignant cells from pancreatic cancers (7)
, non-Hodgkin B cell lymphomas (8)
, chronic lymphocytic leukemias, and acute leukemias (9
10
11
12)
. SDF-1 is constitutively produced in the bone marrow by immature osteoblasts lining the endosteum region and by stromal and endothelial cells. SDF-1 is also produced by different hematopoietic cells as well as by AML blast cells, which express varying amounts of functionally active CXCR4 (10
11
12
13)
.
The mechanisms of blast dissemination are poorly understood and could partly mimic those involved in the egress of normal hematopoietic progenitor cells (HPCs) from bone marrow during mobilization in which SDF-1/CXCR4 axis has a pivotal role (14)
. The granulocyte colony-stimulating factor-induced mobilization process is associated with a decrease in medullary levels of SDF-1 and up-regulation of CXCR4 (15)
, and our group has reported an association between the mobilizing capacity of HPCs and a single nucleotide polymorphism (SNP) in CXCL12 (16)
, the SDF-1-encoding gene. This polymorphism is located at nucleotide position 801 (G to A transition, counting from the ATG start codon) in the 3' untranslated region (3'UTR) of the SDF-1ß transcript (GenBank accession number L36033). The ability of blasts to exit from the bone marrow microenvironment, circulate in the peripheral blood, and anchor in extramedullar locations might thus depend on the CXCL12 genotype. The purpose of our study was to determine whether CXCL12 G801A polymorphism is critical for the dissemination of malignant cells in de novo AML.
MATERIALS AND METHODS
Cases
Eighty-six consecutive adult Caucasian patients (37 female and 49 male) with newly diagnosed de novo AML and 100 healthy volunteers were included in this study after obtaining informed consent and approval by the ethics committee of the University Hospital of Tours. The French-American-British (FAB) cooperative group classification was applied to define the AML subtype and because of the heterogeneity, statistical analyses were performed grouping myeloid (FAB M0/M1/M2) and (myelo)monocytic subtypes (FAB M4/M5). Cytogenetic risk groups were classically defined as favorable, unfavorable and intermediate according to Grimwade et al. (17)
. Since all patients were treated in the same hospital, clinical data were homogeneous. Blast dissemination and tissue infiltration were evaluated by PBB count and the presence of at least one extramedullar tumor site (lymph nodes, liver, spleen, skin, gums, testicles, Waldeyer ring, vertebrae, and nervous system).
CXCL12 genotype analysis
Genomic DNA was extracted from marrow or blood samples, and CXCL12 G801A polymorphism was determined with a polymerase chain reaction-restriction fragment length polymorphism assay [according to the method described by Winkler et al. (18)
]. The PCR primers were CAGTCAACCTGGGCAAAGCC(F) and AGCTTTGTGCCTGAGAGTCC(R). The PCR product was digested with the restriction endonuclease MspI (New England Biolabs, Beverly, MA), resulting in two fragments of 202 and 100 bp for the 801G allele and in one fragment of 302 bp for the 801A variant because of the elimination of the MspI restriction site. These fragments were identified by 6% PAGE (Fig. 1
a).
|
Flow cytometric analysis
The expression of CXCR4 on the surface of bone marrow blasts was determined by flow cytometry (FACSCalibur, Becton-Dickinson, San Jose, CA) using anti CD45-APC (HI30), anti CXCR4-PE (12G5), and anti CD34-FITC (HPCA-2) monoclonal antibodies (mAbs) from BD Biosciences (San Jose, CA). The cells were incubated with saturating concentrations of these mAbs or isotypic controls for 30 min at 4°C and then washed twice with phosphate buffer. At least 5000 events were collected for each sample. CXCR4 expression was measured on leucoblast populations after setting a gate on the SSC/FL4 (CD45) scatter graph, according to Lacombe et al. (19)
completed in certain cases with CD34 expression to optimize leucoblast gating (Fig. 1b
). CXCR4 expression on gated events was evaluated by signal/noise (S/N) ratio of geometric mean fluorescence intensities (MFI) obtained with anti CXCR4 monoclonal antibody (mAb) and its isotypic control (G155178, BD Biosciences), respectively.
Statistical analysis
The results were expressed as median (range), and number (percentage) of patients in 801A carrier (801A/A homozygous and 801A/G heterozygous patients or healthy volunteers) and 801G/G groups. Age, gender, diagnosis (FAB subtype), cytogenetics (risk group), bone marrow blast percentage, PPB count, extramedullar tumor sites, CXCR4 expression, and CXCL12 G801A polymorphism were considered as variables in univariate analysis performed using the Mann-Whitney U test,
2 contingency test, and Z test for non-zero correlation, as appropriate. To evaluate the influence of independent factors on the risk of extramedullar tumor site(s), multivariate analysis was performed using the multiple regression method, including the variables with P < 0.10 in the univariate tests. The level of significance was set at 0.05.
RESULTS
Patient characteristics, CXCL12 genotypes, and the expression of CXCR4 on blast cells are presented in Table 1
for each FAB subtype. The frequency of the 801A allele in our group of patients was the expected value for Caucasians (20)
and was not different from those among healthy volunteers. Moreover the frequencies of this allele were similar in FAB groups (41 and 46% in M0-M1-M2 and M4-M5 groups, respectively). CXCR4 expression was not statistically different between FAB groups [1.7 (128) and 2.9 (137.4) in M0-M1-M2 and M4-M5 groups, respectively] and between CXCL12 genotypes [1.9 (128) and 2.5 (1.037.4) in 801A carriers and 801G/G patients, respectively].
|
The PBB count was not statistically influenced by FAB subtype [4.2 (094.1) and 10.3 (0137.1) PBB/µl in M0-M1-M2 and M4-M5 groups, respectively]. It was moderately correlated with the percentage of blasts in the bone marrow compartment (correlation coefficient 0.356, P<0.001), which was not different in 801A carriers and 801G/G patients [80 (2398)% and 76 (1798)%, respectively]. As presented in Fig. 2
, the presence of the 801A allele was associated with an increased PBB count when comparing 801A carriers to 801G/G patients [10.4 (0.194.1) and 2.6 (0137.1) PBB/µl, respectively, P=0.031]. Moreover, PBB count was correlated with CXCR4 expression on bone marrow blasts (Fig. 3
) in 801A carriers (correlation coefficient 0.546, P=0.001), whereas such a correlation could not be evidenced in 801G/G patients (correlation coefficient 0.176).
|
|
Patients presenting extramedullar tumor site(s) were characterized by a lower mean age [39 (1878) vs. 53 (1675) yr, P=0.005], a higher PBB count [10.9 (0.1137.1) vs. 2.5 (099.3) PBB/µl, P=0.002], and a higher percentage of 801A carriers (59.4 vs. 33.3%, P=0.018). CXCL12 801A carrier status was indeed highly associated with extramedullar locations, which were found in 51.4% of 801A carriers (66.7 and 48.4% of A/A and A/G patients, respectively) and in 26.5% of 801G/G patients (Fig. 2)
, with an odds ratio of 2.92 (95% confidence interval 1.187.21). Interestingly, CXCR4 expression was not different in patients with and without extramedullar locations [S/N=2.2 (137.4) and 2.2 (111.6), respectively]. Age, PBB count, 801A carrier status, and bone marrow blast percentages [87% (1798) and 75% (2398) in patients with and without tissue infiltration, respectively, P=0.090] were included in a multivariate analysis, and the variables found to be independently associated with risk of extramedullar tumor site(s) were PBB count (P=0.012), age (P=0.029), and CXCL12 G801A polymorphism (P=0.042). These results were supported by an absence of difference between 801A carriers and 801G/G patients when considering age [44 (1672) and 51 (1778) yr, respectively].
DISCUSSION
The significance of SNPs in cancer is a recent finding. For example, SNPs are involved in the clinical presentation of malignant diseases, e.g., SNP of vascular endothelial growth factor (VEGF) in melanomas (21)
, and in the therapeutic response, e.g., SNP of FCGR3A in lymphomas (22)
. In this study, we report a genetic determinant associated with the risk of metastasis.
Malignant cell migration, which is recognized as a critical step in metastasis, is a complex process mainly involving metalloproteases, adhesion molecules, and chemokines such as SDF-1. Recent reports indicate that the SDF-1/CXCR4 interaction may be important for the metastasis of solid tumors that express CXCR4 (2
3
4
5
6)
. SDF-1 is the major chemokine released by the bone marrow microenvironment. It is encoded by the CXCL12 gene, and our study demonstrates that the CXCL12 801A allele is an independent risk factor for distant tissue infiltration by malignant cells in AML, concomitant with a higher circulating malignant cell count. These results are supported by our previous findings concerning mobilization of normal HPCs (16)
and by the recent description of a higher frequency of splenomegaly and hepatomegaly in patients with chronic lymphocytic leukemia carrying this variant than in patients homozygous for the common 801G/G genotype (23)
.
The functional significance of this polymorphism has not been characterized. Several polymorphisms have recently been found elsewhere in the CXCL12 gene region that are in linkage disequilibrium with CXCL12 G801A, and the authors suggested the presence of haplotypic variation in expression of SDF-1 transcripts (24)
. Nevertheless, this study was performed with Indonesian islanders and the results may be different in Caucasians. Direct evidence of an influence of CXCL12 G801A polymorphism on the production or transcript half-life of SDF-1 has not been obtained in vitro, because SDF-1 expression is limited to stromal cells and other tissues not easy to analyze. It was hypothesized in 1998 that this mutation could be associated with increased marrow stromal cell secretion of SDF-1 (18)
, without confirmation to date. On the other hand, it could be associated with lower secretion of SDF-1, an hypothesis supported by the lower SDF-1 level recently described in the plasma of normal homozygous 801A subjects (25)
and in stromal layer supernatants of long-term marrow cultures established from 801A carriers patients suffering from lymphoid malignancies (unpublished observation from J. Domenech).
This decreased production of SDF-1 might explain the increased capability of malignant cells to egress from the bone marrow microenvironment. In this context, the correlation observed between CXCR4 expression on bone marrow blasts and the PBB count observed in A carrier patients might result from weaker SDF-1-induced down-regulation of this receptor (26)
, which is not sufficiently effective to hold back the malignant cells in the marrow compartment. The lack of correlation between CXCR4 expression and PBB count in 801G/G patients might be explained by the existence of a critical intramedullary threshold of concentration of SDF-1 below which blasts leave the marrow, regardless of the influence of other factors involved in the migration process (adhesion molecules, etc.). Based on the hypothesis of the effect of CXCL12 polymorphism on the intramedullary production of SDF-1, 801A-carrier patients might have a concentration below this threshold and therefore present higher PPB count correlated with the expression of CXCR4. On the contrary, 801G/G patients might have a SDF-1 concentration above the threshold and the level of expression of CXCR4 might be a minor cofactor in the extramedullar dissemination of blasts, hence the lack of correlation.
In conclusion, we report that CXCL12 G801A polymorphism is a genetic determinant involved in the clinical presentation of leukemia. This description of increased release of blasts from the bone marrow to the blood and higher frequency of distal dissemination in 801A carriers is the first report of an association between this polymorphism and the risk of tissue infiltration by malignant cells. It would be interesting to determine whether 801A carriers had reduced overall survival and a greater risk of recurrence of AML. Moreover, as the SDF-1/CXCR4 axis is involved in the migration process of various types of cancer cells, it could be hypothesized that the 801A variant constitutes a general risk factor for metastasis development and that assessment of CXCL12 G801A polymorphism should help in identifying patients at risk of metastasis. Further studies should clarify this question.
ACKNOWLEDGMENTS
We thank Danielle Truglio, Anne-Françoise Pannier, Anne Mathieu, and Chloé Charroing for technical assistance. This study was supported by the "Comité dIndre-et-Loire de la Ligue Nationale Contre le Cancer," the Rotary Club of Blois (37), and the French "Les Sapins de lEspoir contre le Cancer" and "CANCEN" associations.
Received for publication March 19, 2006. Accepted for publication April 17, 2006.
REFERENCES
This article has been cited by other articles:
![]() |
S. Hassan, A. Baccarelli, O. Salvucci, and M. Basik Plasma Stromal Cell-Derived Factor-1: Host Derived Marker Predictive of Distant Metastasis in Breast Cancer Clin. Cancer Res., January 15, 2008; 14(2): 446 - 454. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. Gallay, L. Anani, A. Lopez, P. Colombat, C. Binet, J. Domenech, B. B. Weksler, F. Malavasi, and O. Herault The Role of Platelet/Endothelial Cell Adhesion Molecule 1 (CD31) and CD38 Antigens in Marrow Microenvironmental Retention of Acute Myelogenous Leukemia Cells Cancer Res., September 15, 2007; 67(18): 8624 - 8632. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. Hirata, Y. Hinoda, N. Kikuno, K. Kawamoto, A. V. Dahiya, Y. Suehiro, Y. Tanaka, and R. Dahiya CXCL12 G801A Polymorphism Is a Risk Factor for Sporadic Prostate Cancer Susceptibility Clin. Cancer Res., September 1, 2007; 13(17): 5056 - 5062. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. N. Kremer, A. Kumar, and K. E. Hedin Haplotype-Independent Costimulation of IL-10 Secretion by SDF-1/CXCL12 Proceeds via AP-1 Binding to the Human IL-10 Promoter J. Immunol., February 1, 2007; 178(3): 1581 - 1588. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |