Distribution of SDF1-3’A, GNB3 C825T and MMP-9 C-1562T Polymorphisms in HSC CD34+ from Peripheral Blood of Patients with Hematological Malignancies

Mobilized peripheral blood stem cells (MPBSC) have nearly replaced bone marrow (BM). So, they become the primary source of hematopoietic grafts especially for patients with hematological malignancies undergoing aggressive myelosuppressive or myeloablative chemotherapy. It allows faster engraftment and equivalent disease-free survival compared with bone marrow cells [Siena S et al, 2000; To LB et al, 1997; Roberto M. Lemoli and Alessandra D’Addio, 2008].


Introduction
Mobilized peripheral blood stem cells (MPBSC) have nearly replaced bone marrow (BM). So, they become the primary source of hematopoietic grafts especially for patients with hematological malignancies undergoing aggressive myelosuppressive or myeloablative chemotherapy. It allows faster engraftment and equivalent disease-free survival compared with bone marrow cells [Siena S et al, 2000;To LB et al, 1997;Roberto M. Lemoli and Alessandra D'Addio, 2008].
Some reports suggested that hematopoietic stem cell mobilization involves a complex interplay between adhesion molecules, cytokines, proteolytic enzymes such as MMP-9 and MMP-2, stromal cells and chemokines among them (e.g,; SDF-1/CXCR4) play a central role [Roberto M. Lemoli and Alessandra D'Addio, 2008;Tsevee Lapidot and Isabelle Petit, 2002]. It has been reported that increased secretion of SDF-1 downmodulates CXCR4 on CD34+ cells, thus preventing the homing of hematopoietic progenitors to the bone marrow [Signoret N et al, 1997]. Moreover, Dlubek D et al, have observed a negative correlation between mobilization capacity and a reduced expression of CXCR4 on mobilized HPC CD34+ in the leukapheresis product [Dlubek D et al, 2006].
These data suggested a central role for CXCR4 and SDF-1 on mobilization of hematopoietic stem cell as well as their homing to the bone marrow [Dlubek D et al, 2006].
The reason for poor mobilization of hematopoietic stem cells that occur in many donors or patients is fully recognized and patients' characteristics (age, BMI, mobilization regimen, diagnosis and clinical status or ulterior therapy) did not explain the whole thing.
Benboubker and his colleagues identified an association of a polymorphism in the SDF-1 gene, designated as SDF1-3'A, with the rate of mobilization of HPCs CD34+ into peripheral blood [Benboubker L et al, 2001]. Hence, we hypothesized that individual genetic factors might explain, at least in part, this variability and that polymorphism analysis can be used to anticipate CD34+ cells mobilization.
So, identifying SNPs predictive of poor or good response to G-CSF or any mobilization regimen, in terms of number of CD34+ cells mobilized, might be useful in discussing the possibility of using a different mobilizing agent or a different source of CD34+ cells for auto-HSCT and allo-HSCT.
In this issue, we proposed to study the distribution of three genetic polymorphisms: SDF1-3'A, MMP-9 C-1562T and GNB3 C825T in Tunisian patients with malignant hematological diseases who underwent stem cell mobilization for autologous transplantation compared to a group of healthy allogenic PBPC donors. Group 4: composed of 15 patients with Acute Myeloid Leukemia (9 men, 6 women).

Materials and methods
Besides, a group composed of 41 subjects (24 men, 17 women) with mean age of 32 years (range 12-63 years) designated for peripheral blood stem cells (PBSC) mobilization. They were visiting the Cellular Immunology and Cytometry and Cellular therapy Laboratory of National Blood Transfusion Center of Tunis-Tunisia as allogenic donors for stem cell transplantation.
Then, a group of 165 healthy blood donors visiting the Blood Transfusion Service of National Blood Transfusion Center of Tunis -Tunisia served as a control group was enrolled in the study. Whole details concerning the subjects will be resumed in Table 1.
Written informed consent was obtained from all subjects according to a protocol approved by the ethical committee for scientific and medical research of the National Blood Transfusion Center and National Bone marrow transplantation center of Tunis (Tunisia) in accordance with the Declaration of Helsinki.
Circulating hematopoietic progenitors CD34+ were evaluated daily by flow cytometry and PBSC collections or apheresis were begun when peripheral CD34+ cells were ~20 cells/µl. Apheresis was usually performed daily using continuous flow blood cell separators COBE SPECTRA and MCS+.

DNA extraction and genotyping
Genomic DNA was prepared from EDTA anticoagulated peripheral blood by using a common salting-out procedure [Miller SA et al, 1988
The reaction conditions were: For SDF1-3'A an initial denaturation at 95°C for five minutes, then 35 cycles at 94°C for 30 seconds, at 58°C for 30 seconds, at 72°C for 1min, and finally extension at 72°C for 7 minutes.
All specimens were examined for the presence of amplifiable DNA. PCR products were digested with 10units HpaII/reaction (Fermentas) at 37°C for overnight [Benboubker L et al, 2001] (figure 1).

Statistical analysis
Allele and genotype frequencies of the studied polymorphisms in patients and healthy controls were formulated by direct counting. Statistical analysis was performed using SPSS software (SPSS 16.0 for windows; SPSS Inc., Chicago, IL.).
The allele frequencies of SDF1-3'A, GNB3 C825T and MMP-9C-1562T polymorphisms were tested for the Hardy-Weinberg equilibrium of the whole group or subgroups of patients and were compared to the respective frequencies of the control group using the Pearson chi-square test or Fisher's exact test when appropriate. The same test was applied to compare the genotype frequency between patients and controls. Association of the allelic frequencies with the clinico-pathologic parameters was evaluated by χ2 test. The odds ratios (OR) and 95% confidence intervals (CI) were calculated too. P<0.05 was required for statistical significance.

Patient's distributions according to their CD34+ cell yield and failure rates
Overall 83% of patients included in this study collected ≥2x10 6 CD34+ cells/kg after a maximum of 4 aphereses, among them 20% collected 2-5x10 6 CD34cells/kg, and 63% collected ≥5x10 6 CD34 cells/kg. Beside, 10% are remobilizers as they did not achieve the threshold of CD34+ cell yield of 2x10 6 CD34/kg within 4 apheresis days and are subjects to another mobilization protocol. Among them, the group of NHL represented the highest rate (40%), the lower ones, the group of MM and AML, which represented respectively 19% and 13%. By contrast, others are designed as first mobilizers (90%) since they have already collected ≥2x10 6 CD34+ cells/kg after a maximum of 4 aphereses days. Amongst them the group of multiple myeloma was the most frequent (40%), thereafter the group of Non-Hodgkin's lymphoma (34%) and Hodgkin's disease with 26%. For the patients included in this study, mobilization failure was defined as <2x10 6 CD34+ cells/kg obtained within 4 apheresis days. So, especially MM patients collected ≥5x10 6 CD34+ cells/kg and contained the highest CD34+ cell yield (8,89x10 6 CD34/kg for MM, and 5,51x10 6 CD34/kg for the others patients). Furthermore, the fact that MM patients had higher yield of CD34+ cells compared to NHL and HD is likely since that NHL and HL patients are frequently more heavily pretreated with cytotoxic chemotherapy than patients with MM [Iskra Pusic et al, 2008] (figure 4).

Analysis of the studied polymorphisms in the 4 subgroups of patients according to disease: A comparison between healthy donors of PBSC and patients
According to this study, SDF1-3'A and MMP-9 C-1562 T polymorphisms were significantly different between the patients and healthy controls (table 3). Particularly, we found significant differences in all the allelic and genotypic frequencies of the For the group of NHL, the distribution of the SDF1-3'A polymorphism was significantly different between patients and healthy controls especially for the A allele which seemed to be associated to this disease (p=0,019). Moreover, a decrease in GG genotype frequency compared to the control group was observed too reaching a statistically significance (p=0.029).
Concerning the MMP-9 C-1562T polymorphism, like the MM group, high significant differences were seen especially for the T allele ( So the presence of the MMP-9 C-1562T might be associated with this disease. When considering the GNB3 C825T polymorphism, we observed that the TT genotype was more frequent in patient with MM and NHL with respectively 20.69% and 15.15% compared to the Hodgkin's disease group (only 10.52%). Whereas, the CC genotype was more frequent in the NHL group (24.24%) ( Table 3).

Association of the SDF1-3'A allele with a good mobilizing capacity
As the clinicians have defined mobilization failure as <2x106 CD34+ cells/kg obtained within 4 apheresis days, two mainly group of patients emerged: the subjects with a good capacity of mobilization who collected ≥2x106 CD34+ cells/kg obtained within 4 apheresis days. Others with a poor mobilizing capacity and didn't collect 2x106 CD34+ cells/kg within 4 apheresis days. For the healthy allogenic PBSC donors, the mobilization failure was defined as <3x106 CD34+ cells/kg obtained within 4 apheresis days.
When considering the SDF1-3'A polymorphism, significant difference was observed in the SDF1-3'A allele carriers and GG carriers (p=0.023). A higher concentration of CD34+ cells in the leukapheresis products was detected in SDF1-3'A positive patients compared to GG homozygous subjects Besides, a lower increase in the GG genotypes was observed in the "poor" mobilizer group compared to the "good" ones reaching a statistical significance (p=0.023; OR =0.494; CI (95%) [0.268-0.912]) (Table 4).
Thus, the SDF1-3'A allele carriers, especially the SDF1-3'AA homozygous individuals in the group of healthy allogenic PBSC donors had a better mobilization potential (table 4).
OR designed as odd radio and p value of all genotypic and allelic frequencies are provided in the table by using statistical software (SPSS 16.0) as it was mentioned above in section materials and methods-statistical analysis.
However, when considering the group of remobilizers in our study population we have observed that 48% of subjects were GG, 12% were AA and 40% were GA. This led us to consider a probable association of the GG genotypes to mobilization failure.
For the GNB3 C825T polymorphism, we didn't observe any difference between the 2 groups of poor and good mobilizers.
This let us consider that there's no association between GNB3 C825T polymorphism and the capacity of mobilization of hematopoietic stem cells.
For the group of healthy PBSC donors, and with respect to our classification according to mobilization failure (<3x106 CD34/kg within 4 apheresis days), we have found an important association of SDF1-3'A distribution with higher mobilization yield of hematopoietic stem cells CD34+ reaching a higher statistical significance (p=0.001; OR=12.6;

Discussion
In the present study, we investigated the effect of polymorphisms in the genes SDF-1, GNB3 and MMP-9 on the outcome of mobilization of peripheral blood stem cells for autologous transplantation by using a PCR-RFLP analysis.
We observed a significant association for SDF-1 and MMP-9 polymorphisms exclusively in patients with MM, NHL and Hodgkin's disease suggesting that these polymorphisms are fair candidate gene variants to these 3 hematological diseases.
In fact, Association of these polymorphisms to cancer has been previously reported by many investigators [De Oliveira KB et al, 2009;Rabkin CS et al, 1999].
Our results were in agreement with other studies suggesting that SDF1-3 De Oliveira KB et al, when studying distribution of SDF1-3'A polymorphism have reported also a significant difference in genotype distribution between NHL patients (GG: 51.4%; GA: 47.1%; AA: 1.5%) compared to healthy controls (GG: 65.6%; GA: 28.9%; AA: 5.5%). Whereas, they didn't find any significant differences in genotypes distributions with breast cancer and Hodgkin's lymphoma [De Oliveira KB et al, 2009].
Moreover, previous reports on AIDS related non-Hodgkin's lymphoma (NHL) demonstrated that the CXCL12-3'A chemokine variant was associated with approximate doubling of the NHL risk in heterozygotes and an approximately fourfold increase in homozygotes [Rabkin CS et al, 1999;A Zafiropoulos et al, 2004]. Hence, this might let us suggest the possible role of such variant in the pathogenesis of NHL.
In this present work, we did not find a significant association between SDF1-3'A polymorphism and our group of patients with AML, this could be due to the lower number of patients (15 patients).
However, Dommange et al, have reported the implication of SDF1-3'A polymorphism in the clinical representation of acute myeloid leukemia in 86 patients with AML, as an association between this polymorphism and the risk of tissue infiltration by malignant cell was established by an increased release of the blast from the bone marrow in the blood in the SDF1-3'A carriers suggesting that this SDF-1 variant is associated with clinical representation of AML [A Zafiropoulos et al, 2004].
MMP-9 polymorphism -1567 C/T in the promoter region of the MMP-9 gene in group of patients with some haematological malignancies as well as in patients undergoing stem cell mobilization.
Then, we observed that the T allele was highly associated to the susceptibility to the four diseases studied (table 3). We have to investigate either this variant have major influence on the circulating levels of MMP-9.
Concerning the group of MM, we observed a significant association in all allelic and genotypic frequencies of SDF1-3'A polymorphism with statistical differences when compared to control. Hence, as increased angiogenesis was related to the pathogenesis of MM, and because SDF-1 chemokine induces increased VEGF production, which is responsible for an angiogenic activity [Florence Dommange et al, 2006], we hypothesize that the SDF1-3'A polymorphism might increase SDF-1 protein which would have a role in developing angiogenesis and in the pathogenesis of the disease.
On the other hand, frequent distribution of the SDF-1 3'A allele in multiple myeloma patients confirms the implication of SDF-1 in hematopoietic stem cells. This logical consequence of the widely distribution of SDF-1 3'A allele proving that multiple myeloma patient's could be considered as good mobilizers.
For the GNB3 polymorphism we've observed that the TT genotype and the T allele frequencies are more frequent especially in patients with MM (0.72 for Tallele frequency) and NHL (0.45 for Tallele frequency) compared to healthy donors of PBSC ( peripheral blood stem cells) (Table 3) which is far from the others populations [Maggie C.Y et al, 2004] . Then, suggesting the possible relation with these diseases.
Maggie et al when studying the ethnic differences in the linkage disequilibrium and distribution of single-nucleotide polymorphisms in 35 candidate genes for cardiovascular diseases have reported that the frequency of the T allele of GNB3 polymorphism in Chinese population is about 0.545. Then, such frequency is far from those of the French and of the Spanish population (0.329 and 0.359) and more closer to our result in Tunisian population [Yair Gazitt & Cagla Akai, 2004].
When interesting to the capacity of mobilization which was largely demonstrated to vary from a subject to another, several studies have focused on such phenomena and have reported that 10-30% of patients with hematological malignancies fail to mobilize PBSC [Ingrid G. Winkler & Jean-Pierre Levesque, 2006] and either a small proportion of normal donors (1-5%) fail to mobilize sufficient CD34+ cells.
Besides, many reports suggest that numerous factors are related to poorer mobilization including age, gender, type of growth factor, dose of the growth factor and in the autologous setting patient's diagnosis, chemotherapy regimen and number of previous chemotherapy cycles or radiation [Sugrue MW et al, 2001].
In our study we were interested in the possible implication of some genetic factors in mobilization and as we've found an association with the SDF-1 3'A variant only, then we supposed that this polymorphism is the only predictor of mobilization capacity of PBSC CD34+.

313
In fact, when analyzing the distribution of the two functional polymorphisms SDF-1 G801A and MMP-9 C-1562T considering the two groups of "good" and "poor" mobilizers, we've found an association only with SDF1-3'A polymorphism. While no association with capacity of mobilization was observed with GNB3 C825T and MMP-9 C-1562T polymorphisms.
When observing the distribution of the two polymorphisms not only when considering the mobilization capacity but also in relation to each studied disease enrolled in this work we've found that the good mobilizer group was mainly composed of MM patients. Whereas the poor mobilizer group contains Hodgkin's disease who are considered in previous studies as hard-to-mobilize patients [Benboubker L et al, 2001;Patrick J Stiff, 1999].
The fact that multiple myeloma patients mobilized better PBSC CD34+ (peripheral blood stem cells) than the others groups seem to be related to their ulterior chemotherapy (dexamethasone + thalidomide) and didn't receive any radiation therapy unlike the HD and NHL groups.
In the good mobilizer group composed of patients needing fewer apheresis than the other group, genotypes frequencies for the GG,GA, AA represented respectively 30.6%, 58% and 11.4%, and corresponded respectively to 45.5%, 47% and 7.6% in the poor mobilizer's group, and significant differences were found for GG genotype (p=0.007) and for A allele (p=0.009).
This confirms on the one hand that the SDF1-3'A allele was associated with good mobilizing capacity not only in the group of patients but for instance in the group of healthy allogenic PBSC donors (see table 5). Thus, our results regarding patients undergoing autologous transplantation of haematopoietic stem cells concur with those reported by Benboubker et al [Bogunia-Kubik K et al, 2009].
Moreover this deduction is already found in the group of healthy allogenic transplantation donors as it was reported in the present study and by Bogunia-Kubik K et al who have suggested that the SDF1-3'A allele was associated with a higher yield of CD34+ cells from healthy donors of PBPC for allogeneic haematopoietic SCT (stem cell transplantation) compared to GG homozygotes [Patrick J Stiff, 1999].
Recent studies by the same group underlined an association of the SDF1-3'A allele with faster granulocyte and platelet recovery after transplantation. Therefore they suggested that the SDF-1 gene polymorphism could be a useful tool of prognostic value for recipients of autologous haematopoietic stem cells [A. Gieryng et al, 2010]. The allelic variant SDF1-3'A is a result of the SNP rs1801157, which is located in a highly demethylated area of the 3'UTR region. This SNP confers a G to A transition in the nucleotide position 801, resulting in a loss of a methylation site, which could affect the methylating effect of G-CSF [Nagler A et al, 2004], and leading to a more decreased SDF-1 expression in healthy individuals carrying the polymorphism.
So, it's of interest to investigate either this variant have major influence on the circulating levels of SDF-1 and its mRNA expression, one of our future's interests.
Further studies examining how these three polymorphisms interact with disease risk factors are needed.
Interestingly, the possible implication of others genes involved of homing and migration process of CD34+ cells and for instance VCAM-1 to higher or lower mobilization yield of PBPC might emphasize new strategies for poor mobilizers subjects and lead to the identification of new biomarkers and/or therapeutic targets.

Conclusion
In the present study, we observed a significant association for CXCL12 and MMP-9 polymorphisms exclusively in patients with MM, NHL and Hodgkin's disease suggesting that these polymorphisms are fair candidate gene variants to these 3 hematological diseases.
Furthermore we've confirmed that the SDF1-3'A allele was highly associated to a good mobilizing capacity especially in the group of healthy allogenic PBSC donors where the analysis not biased by background disease or chemotherapy.
Besides, we suggested a possible association of GG genotypes to poorer mobilization is already deduced.

Acknowledgment
We thank all participant and all patients in this work