Vascular endothelial growth factor (VEGF) is a potent proangiogenic factor. Several single nucleotide polymorphisms (SNPs) in the VEGF gene with influence on VEGF expression have been described. In multiple myeloma, VEGF stimulates angiogenesis which is correlated with disease progression and prognosis. In this study, we evaluated the association between genetic variations in the VEGF gene in patients with multiple myeloma and time to treatment failure (TTF) after high-dose melphalan and stem cell support (HDT), overall survival (OS) and efficacy of the anti-angiogenic drug thalidomide. Retrospectively, the SNPs −2,578C>A (rs699947), −460C>T (rs833061), +405G>C (rs2010963) and +936C>T (rs3025039) in the VEGF gene were examined in 348 patients with newly diagnosed multiple myeloma initially treated with HDT, where 176 patients were treated with thalidomide at relapse. None of the examined geno- or haplotypes was associated with differences in TTF after initial therapy or OS. A possible relation between the haplotype −2,578A/−460C/+405G (ACG) and effect of thalidomide was seen. Patients with no copies of the haplotype ACG had a longer time to next treatment than patients with one or two copies of the haplotype ACG, median 13.7 months vs. 9.2 months, p = 0.007. In conclusion, the haplotype ACG in the VEGF gene may influence the efficacy of thalidomide in multiple myeloma. Further analyses are needed to confirm these findings and get insight into the functional effect of these polymorphisms, so in the future we may be able to select multiple myeloma patients who especially will benefit from treatment with thalidomide.
Vascular endothelial growth factor (VEGF) is a homodimeric glycoprotein with a relative molecular mass of 45 kDa and the only mitogen that specifically acts on endothelial cells. The VEGF gene is located on chromosome 6p21.3 and consists of 8 exons exhibiting alternate splicing to form a family of proteins.1, 2 VEGF expression is modulated by different effectors including cytokines, lipo-polysaccharides, hormones and hypoxia.3–5 In some malignancies, increased VEGF expression resulting in inappropriate VEGF-induced angiogenesis is linked with tumour growth and metastasis.6, 7 In multiple myeloma, an incurable bone marrow cancer, increased angiogenesis is associated with diseased activity and survival.8, 9 Myeloma cell lines and bone marrow plasma cells from multiple myeloma patients express VEGF while VEGF receptors (FLT-1 and KDR/FLK-2) are present on endothelial cell and bone marrow stromal cell membranes. Activation of the VEGF receptors increases the production of various haematopoietic growth factors, i.e., interleukin-6 and thereby stimulates the myeloma cell in a paracrine pathway.10–12
The expression of VEGF in response to different stimuli is highly variable among individuals. Single nucleotide polymorphisms (SNPs) in the VEGF gene may contribute to this variation.4, 13 Some polymorphisms are related to increased VEGF expression (i.e., −2,578C allele and −634C allele) while others are shown to decrease VEGF expression (i.e., +936T allele and −460T allele).14–17 SNPs in the VEGF gene have been examined in different malignancies, both solid tumours and haematological neoplasms. In some malignancies, SNPs in the VEGF gene are associated with prognosis, survival or susceptibility of the disease.18–20
In multiple myeloma, high-dose melphalan and stem cell support (HDT) is the standard treatment of newly diagnosed patients under the age of 65 years. Though HDT is a more effective treatment than conventional chemotherapy concerning time to progression and overall survival (OS), some patients still have a poor prognosis with early relapse after HDT and short survival.21 Thalidomide is a well known anti-angiogenic and immunomodulatory drug and is included in relapse treatment and more recently primary treatment in elderly patients with multiple myeloma.22 Patients responding to thalidomide have decreasing plasma levels of VEGF and bone marrow micro vessel density, an estimate of bone marrow angiogenesis.23, 24 Unfortunately, thalidomide has several side effects as constipation, neuropathy, somnolence and venous thromboembolism and response is variable.25
In multiple myeloma, VEGF-induced angiogenesis seems to be important for disease activity, survival and effect of thalidomide treatment. The purpose of this study was to examine the potential association between different SNPs in the VEGF gene and time to treatment failure (TTF) and OS in newly diagnosed multiple myeloma patients treated with HDT. Furthermore, to evaluate whether SNPs in the VEGF gene influence the efficacy of thalidomide used as relapse treatment.
Material and Methods
Patient's clinical data, response criteria, eligibility criteria and treatment have previously been described.26 Briefly, patients diagnosed with multiple myeloma and treated with HDT from August 1994 to August 2004 were recruited from four participating hospitals in Denmark. A total of 348 patients were included in the study. Of these, 185 patients were included in the high-dose treatment protocols implemented by the Nordic Myeloma Study Group (NMSG no. 5/94, 7/98 and 11/00),21, 27, 28 whereas the remaining 163 patients were treated with similar treatment regimens off protocol.
Information on treatment at relapse was obtained from the medical reports from February to May 2008. Thalidomide was used in relapse treatment either alone or in combination with steroids and/or chemotherapy. In total, 176 patients were treated with thalidomide at some point of relapse. Some patients were treated more than once with thalidomide. In these cases, only the first treatment was included in the calculation of response and time to next treatment (TNT).
TTF and OS were calculated from date of stem cell infusion to date of progression after HDT or death, respectively.26 TNT was used as follow-up after relapse treatment and defined as the time from start of relapse treatment to start of new relapse treatment, progression without new treatment or death.
Definition of progressive disease has been described previously.26 Response to thalidomide treatment was catagorized as (i) partial response or better and (ii) less than partial response (stable disease and progressive disease). Partial response was defined as more than a 50% reduction in serum M-protein and a reduction in light chain proteinuria to less than 0.2 g/24 hr.
The study was approved by the Danish Ethical Committee (01-158/03).
Human tissue samples
Peripheral blood mononuclear cells (PBMCs) were purified from 292 leukapheresis products by buffy coat preparation. From 56 patients, 10 times 10-μm sections were collected from paraffin-embedded bone marrow samples.
DNA for analysis was purified from PBMCs by salting out method29 or from paraffin-embedded tissue by phenol extraction.30
Detection of single nucleotide polymorphisms
The polymorphisms in the VEGF gene at positions −2,578C>A (rs699947), −460C>T (rs833061), +405G>C (rs2010963) (also known as −634, relative to translation start) in the 5′ UTR region and +936C>T (rs3025039) in the 3′ UTR region were determined on a ABI 7900HT using allelic discrimination and predeveloped assays (Applied Biosystems, Birkerød, Denmark). Reactions of 5 μl contained aproximately 50 ng DNA, 2.5 μl mastermix (Applied Biosystems, Birkerød, Denmark) and the predesigned assays. Controls were included in each run, and a repeat analysis of 10% subset of samples yielded 100% identical genotypes. Moreover for 10 patients, DNA from both bone marrow and leukapheresis products was genotyped with identical results.
The statistical data were obtained using R statistical software (version 2.9.2, 2009, The R Foundation for Statistical Comouting, Vienna, Austria). All test were two-sided and p-values less than 0.05 were regarded as statistically significant. Fisher's exact test was used to compare categorical variables. The Kaplan–Meier method was used to draw survival curves and the Cox proportional hazards model and log-likelihood statistics to compare differences in TTF, TNT and OS between groups. The haplotype analysis and control for deviation from the Hardy–Weinberg equilibrium for the VEGF gene polymorphisms were conducted using Haploview software (available at http://www.broad.mit.edu/mgp/haploview). Subsequently, haplotype frequencies were estimated with linkage disequilibrium (LD) coefficient, D. Individual haplotypes were determined based on a Bayesian algoritm using Phase program31 (available at http://www.stat.washington.edu/stephens/phase.html).
In this study, the median follow-up of all 348 patients still alive was 93.4 months (range 54.6–174.2 months) and the median OS was 69.8 months. The charateristics of the patients are summarized in Table 1.
Table 1. Patient characteristics at the time of diagnosis
Association between VEGF genotypes, time to treatment failure after high dose treatment and overall survival
The genotypes frequencies for the polymorphisms at positions −2,578, −460, +405 and +936 in the VEGF gene are summarized in Table 2. The frequency of each allele and genotype is comparable to previous studies in caucasian populations.19, 32 No association was found between the examined polymorphisms and TTF after high-dose treatment and OS (Table 2).
Table 2. Univariate analysis of the association between different genotypes at the −2,578, −460, +405 and +936 loci in the VEGF gene and time to treatment failure (TTF) after high dose treatment and overall survival (OS)
Association between VEGF haplotypes, time to treatment failure after high dose treatment and overall survival
There was a strong LD between loci −2,578 and −460 (D′ = 0.96), −460 and +405 (D′ = 0.98) and −2,578 and +405 (D′ = 0.95), whereas their linkage with +936 was much weaker (D′ = 0.29, D′ = 0.23 and D′ = 0.16, respectively). Accordingly, eight different haplotypes of VEGF polymorphisms based on the −2,578, −460 and +405 loci were generated. The most common haplotypes were −2,578A/−460C/+405G (ACG), −2,578C/−460T/+405C (CTC) and −2,578C/−460T/+405G (CTG) (45.1%, 23.7% and 20.3%, respectively) while the frequencies of the haplotypes CCG, ATG, CCC, ACC and ATC were 5.7%, 4.2%, 0.4%, 0.3% and 0.3%, respectively.
No difference in TTF after high-dose treatment and OS was seen for the most common haplotypes ACG, CTC and CTG when divided into groups depending on the number of copies of the haplotypes (Table 3, Figs. 1a and 1b).
Table 3. The impact of the VEGF gene haplotype −2578/−460/+405 on time to treatment failure (TTF) after high dose treatment and overall survival (OS)
Efficacy of thalidomide in relation to VEGF geno- and haplotypes
Thalidomide was used as relapse treatment either alone or in combination with steroids and/or chemotherapy. The different genotypes for the examined VEGF gene polymorphisms had no influence on the response rates to thalidomide treatment. Neither were the three most common haplotypes ACG, CTC and CTG associated with response to thalidomide treatment (Table 4).
Table 4. The effect of genotypes at −2,578, −460, +405 and +936 loci and haplotypes −2,578/−460/+405 in the VEGF gene on response to thalidomide treatment
On the other hand, thalidomide treated patients with no copies of the haplotype ACG had a longer TNT than patients with one or two copies of the haplotype ACG, median 13.7 months vs. 9.2 months, respectively (p = 0.007) (Fig. 2a). No association was found between the haplotypes CTC and CTG and efficacy of thalidomide.
The efficacy of thalidomide as part of relapse treatment in patients with 0 copies of the haplotype ACG was not translated into a longer survival after relapse from initial treatment (Fig. 2b).
Multiple myeloma is a highly heterogenous disease concerning symptoms, response to treatment and survival. Several factors are responsible for this heterogenousity and SNPs may be one of them. A recent study in multiple myeloma has shown that patients who are homozygous carriers of the wild type ins-allele of the NFKB1 −94ins/delATTG polymorphism may benefit from maintenance treatment with interferon-α, in contrast to patients carrying the variant allele.33 Furthermore, a study has shown that polymorphisms in the DNA repair genes ERCC2, XRCC3 and CD3EAP can predict the outcome for patients treated with HDT.26
Polymorphisms in the VEGF gene have been associated with prognosis and survival in different types of cancer including haematologic malignancies and may alter the expression of VEGF. In our study, we found that the anti-angiogenic drug thalidomide was less effective in patients with 1 or 2 copies of the VEGF haplotype −2,578A/−460C/+405G (ACG). The haplotype ACG may lead to a lower VEGF expression, containing the variant allele −2,578A and the wildtype allele +405G which in previously studies have been associated with low VEGF expression. In a study concerning acute renal graft rejection, VEGF production was examined by stimulating PBMCs from 30 healthy individuals. Cells from homozygote carriers of the −2,578C allele had a significantly higher VEGF production.17 The haplotypes −2,578A/−1,154A/−634G and −2,578A/−1,154G/−634G have been associated with lowered circulation VEGF levels in vivo and reduced VEGF gene expression.16 Low plasma levels of VEGF in patients with acute respiratory distress syndrome was found in patients with the haplotype −460T/+405G/+936T.15 Theoretically, it is likely that multiple myeloma patients with 1 or 2 copies of the VEGF haplotype ACG have lower VEGF expression than patients with no copies of the haplotype and thereby reduced effect of thalidomide. In a previously study, thalidomide treatment was more effective in multiple myeloma patients with high plasma levels of VEGF and increased bone marrow angiogenesis.34
The improved efficacy of thalidomide in patients with no copies of the haplotype ACG was not translated into a better treatment response. Neither had any of the VEGF gene polymorphisms examined in this study including the haplotype ACG impact on TTF after HDT or OS. Thalidomide or other specific angiogenesis inhibitors were not used in the initial treatment of the patients in this study which may explain why no difference in TTF after HDT according to number of ACG copies was seen. The haplotype ACG had no influence on OS after HDT or relapse probably due to different treatment regimens at relapse. We are aware that this study has several limitations, including sample size, the retrospective design and the heterogeneous treatment with thalidomide concerning dose and time.
In conclusion, this study indicate that the VEGF haplotype ACG may influence the efficacy of thalidomide treatment in patients with multiple myeloma. It is important to evaluate the results in a prospective study including examination of VEGF expression either in blood samples or immunohistochemical in bone marrow samples together with examination of bone marrow angiogenesis to learn more about the functional effect of these polymorphisms. If the data are confirmed we will be able to select the multiple myeloma patients who will particularly benefit from treatment with thalidomide.
Lars Bentzen is thanked for excellent technical assistance.