Volume 14, Issue 12 , Pages 1408-1416, December 2008
Vascular Endothelial Growth Factor Gene Polymorphisms May Predict the Risk of Acute Graft-versus-Host Disease following Allogeneic Transplantation: Preventive Effect of Vascular Endothelial Growth Factor Gene on Acute Graft-versus-Host Disease
Article Outline
Abstract
Microvessel injury is associated with the development of graft-versus-host disease (GVHD), whereas high levels of posttransplantation vascular endothelial growth factor (VEGF) have a protective effect on severe acute GVHD (aGVHD) and transplantation-related mortality. The current study aimed to determine the impact of VEGFA gene single-nucleotide polymorphisms (SNPs) on the risk of aGVHD after allogeneic stem cell transplantation (SCT). Using polymerase chain reaction and restriction fragment length polymorphism, 4 VEGFA SNPs— -2578 C>A (rs699947), -460 T>C (rs833061), +405 G>C (rs2010963), and +936 C>T (rs3025039)—were analyzed in 98 recipients. Strong linkage disequilibrium was noted among loci -2578, -460, and +405, but not among these loci and locus +936. Accordingly, 4 haplotypes were generated based on the genotypes of -2578, -460, and +405: CTC (47.9%), CTG (26.7%), ACG (24.2%), and CCC (1.0%). The group with low VEGF production (ie, +936CT genotype and 2 copies of the ACG haplotype) had a higher incidence of aGVHD. Significant associations were found between the risk of grade 2-4 aGVHD and the +936 CT (P = .006), -2578 AA (P = .003), and -460 CC (P = .002) genotypes and the ACG haplotype (P = .003). No association between the VEGFA SNPs and chronic GVHD was observed. The VEGFA SNPs might predict a lower risk of aGVHD. Our findings suggest that VEGF may have a protective role in the pathogenesis of aGVHD.
key words: Vascular endothelial growth factor, Single nucleotide polymorphism, Graft-versus-host disease, Allogeneic stem cell transplantation
Introduction
The pathogenesis of graft-versus-host disease (GVHD) has yet to be fully elucidated, although it is generally accepted that alloreactive T cell cytotoxicity is a central mediator. Alloreactive T cells recognize the recipients’ target tissues as nonself and evoke GVHD. The final step in the development of GVHD occurs in targeted tissues, in which inflammation develops due to interactions between these tissues and cytotoxic T cells. It An association between these inflammatory reactions and angiogenesis is well established.
Recently, endothelialitis and subsequent microvessel injury were found to be involved in the pathogenesis of GVHD. One study found strikingly higher microvessel densities in skin samples in healthy normal donors than in patients with acute GVHD (aGVHD) or chronic GVHD (cGVHD). Accordingly, it has been suggested that the endothelium is targeted during GVHD and that microvessel injury is a consequence of perivascular inflammation and endothelial cell death, which results in progressive microvessel loss and consequent tissue ischemia and stimulates the production of VEGF [1]. Accordingly, angiogenesis also is involved in the pathogenesis of GVHD.
VEGF, a soluble 34- to 46-kDa heparin-binding glycoprotein dimer, is a potent angiogenic peptide with diverse biological activities that include angiogenesis in both physiological and pathological situations [2]. VEGF gene (VEGFA) expression is regulated by various growth factors, cytokines, and hormones, as well as by hypoxia [3]. VEGF can be produced by numerous cells, including lymphocytes, macrophages, vascular smooth muscle cells, fibroblasts, keratinocytes, megakaryocytes, neutrophils, basophils, and mast cells. Moreover, previous investigations have suggested that type 2 cytokine stimulates VEGF production 4, 5.
Interestingly, higher VEGF levels at day 14 or 15 posttransplantation have been suggested to protect against the development of severe GVHD 4, 6. The first study to investigate this concept found an association between high VEGF levels and a lower incidence of nonrelapse-related mortality (NRM) (23% vs 4%), along with an inverse correlation between VEGF levels at day 14 posttransplantation with the severity of aGVHD [4]. Moreover, patients with severe grade 3-4 aGVHD had significantly lower log-transformed VEGF levels than those with or without grade 1-2 aGVHD [4]. Another study similarly reported improved survival in patients with higher VEGF levels at day 15 posttransplantation [6]. These 2 studies suggest that VEGF protects against severe aGVHD.
Recent investigations have demonstrated that VEGFA polymorphisms contribute to interindividual variations in VEGF expression. The VEGFA gene is located on chromosome 6p21 and consists of 8 exons and 7 introns 7, 8. Furthermore, polymorphisms in its promoter region (loci -2578C>A [rs699947] and -460T>C [rs833061]), its 5′-untranslated region (+405C>G [rs2010963]) and its 3′-untranslated region (+936C>T [rs3025039]) have been associated with different levels of VEGF expression 9, 10, 11, 12, 13, 14. Accordingly, in the present study, we investigated the impact of VEGFA polymorphisms on the development of aGVHD on outcome after allogeneic stem cell transplantation (SCT).
Materials and Methods
The objective of the present study was to investigate an association between VEGFA polymorphisms and the development of aGVHD or cGVHD after allogeneic SCT.
Patient Characteristics and Transplantation Procedure
Ninety-eight consecutive patients who had received an HLA-matched sibling transplant at the Kyungpook National University Hospital, Daegu, Korea between August 1998 and June 2005 were included in this retrospective study. Detailed information is provided in Table 1. The conditioning regimens consisted of busulfan/cyclophosphamide (n = 57; 58%), fludarabine-based regimens (n = 31; 32%), and cyclophosphamide/antithymocyte globulin (ATG) (n = 10; 10%). All 98 patients received peripheral blood stem cells (PBSCs), as described previously [15]. GVHD prophylaxis included cyclosporin A (CSA) plus methotrexate (MTX) in 86 patients (88%) and CSA alone or FK506/MTX in 6 patients each (6%/6%). Treatment for aGVHD and cGVHD was provided according to a standard protocol, as described previously [16].
Table 1. Patient Characteristics and Transplantation Procedures
| Variable | No. of pts (%) |
|---|---|
| Recipients | |
 Sex, female/male, n (%) | 34/64 (35/65) |
| Age, years, median (range) | 33 (16 to 58) |
| Diagnosis | |
| AML/ALL | 50/11 (51/11) |
| CML/MDS | 14/4 (14/4) |
| SAA/NHL | 10/8 (10/8) |
| Solid tumor∗ | 1 (1) |
| Advanced disease | 48 (49) |
| Donors | |
| Sex, female/male, n (%) | 34/64 (35/65) |
| Age, years, median (range) | 34 (15 to 65) |
| Conditioning, n (%) | |
| BuCy | 57 (58) |
| CyATG | 10 (10) |
| Fludarabine-based RIST | 31 (32) |
| Infused cell dose, median | |
| MNCs, × 108/kg | 6.75 |
| CD34+ cells, × 106/kg | 6.32 |
| CD3+ cells, × 108/kg | 1.97 |
| GVHD prophylaxis, n (%) | |
| CSA/MTX | 86 (88) |
| CSA | 6 (6) |
| FK506/sMTX | 6 (6) |
∗Metastatic colorectal carcinoma. |
Genotyping of VEGFA and Genotype Analysis
For VEGFA genotyping, genomic DNA was extracted from peripheral blood using the Wizard genomic DNA purification kit (Promega, Madison, WI). The VEGF -2578C>A (rs699947), -460T>C (rs833061), +405C>G (rs2010963), and +936C>T (rs3025039) genotypes were determined by polymerase chain reaction (PCR) and restriction fragment length polymorphism, as described previously 17, 18, 19, 20. To confirm genotyping results, selected PCR-amplified DNA samples (n = 2 for each genotype) were examined by DNA sequencing [17]. The study design was approved by the Kyungpook National University Hospital Institutional Research Board and conformed to the Helsinki Declaration. Each patient provided written informed consent.
Four genotypes were evaluated using the χ2 test to determine whether they conformed with the Hardy-Weinberg equilibrium (HWE). Genotype frequencies were determined using Haploview software (available at http://www.broad.mit.edu/mpg/haploview). Additive, dominant, and recessive models were used to investigate associations between each single-nucleotide polymorphism (SNP) and transplantation outcomes. Haplotype analysis for deviation from the HWE was conducted, and haplotype frequencies were estimated using linkage disequilibrium (LD) coefficients, D′. Individual haplotypes were determined with a Bayesian algorithm using the Phase program (available at http://www.stat.washington.edu/stephens/phase.html) [21].
As in our previous study [20], here a VEGF risk score model was generated based on the genotype at locus +936 and the copy number of the ACG haplotype at loci -2578/-405/+460. A score of 1 was assigned to risk alleles (ie, +936 CT or TT genotypes, or 2 copies of the ACG haplotype), and a score of 0 was assigned to other alleles (ie, the +936 CC genotype, or 0 or 1 copy of the ACG haplotype). The scores were summed, and 2 risk groups were defined: high risk (composite score, 2 or 1) and low risk (composite score, 0).
Definition and Endpoints
The day of stem cell infusion was defined as day 0. Engraftment was confirmed by peripheral blood counts, that is, a peripheral absolute neutrophil count of > 0.5 ×109/L and a peripheral platelet count of > 20 × 109/L for at least 3 consecutive days without requiring transfusion. Overall survival (OS) was defined as the time from transplantation until death from any cause. aGVHD and cGVHD were diagnosed and graded based on established criteria 22, 23.
Statistical Analysis
The data were analyzed according to information available on July 2005. The clinical characteristics and transplantation outcomes of patients were compared using the χ2 test, Fisher's exact test, or the Mann-Whitney U test for different VEGFA genotypes.
Probabilities of OS were calculated and plotted using the Kaplan-Meier method. The incidences of aGVHD, cGVHD, NRM, and recurrence were estimated using the cumulative incidence method considering competing risks [24]. During single-marker analyses, the OSs of VEGFA SNPs were compared using additive, dominant, and recessive models through the log-rank test, whereas the incidences of aGVHD, cGVHD, NRM, and recurrence for different VEGFA SNPs were compared using Gray's test.
During multivariate analyses using Cox proportional hazard models, clinical factors and significant genotypes were considered as covariates for each event. Because our analysis was confined to HLA-matched sibling PBSCT transplants, HLA disparity, donor relationship, and stem cell source were not included in the multivariate analysis. Before introducing potential time-dependent covariates, such as aGVHD or cGVHD, into the time-dependent Cox proportional hazard model, we investigated the appropriateness of using a non–time-dependent hazard model with either aGVHD or cGVHD. In univariate analyses, the P values of the omnibus test (which indicates a model's appropriateness in a non–time-dependent Cox model) were < .001 for aGVHD and .238 for cGVHD. Based on this result, we applied cGVHD only as time-dependent covariate in the model.
In model 1, the following covariates were included in transplantation outcome (ie, OS, NRM, or relapse) analysis: cGVHD (time-dependent covariate), age (< 40 years vs ≥ 40 years), the development of aGVHD (grade 0-2 vs grade 3-4), disease risk (high risk vs standard risk), conditioning regimen (myeloablative vs reduced intensity), VEGFA +936 C>T genotype (CT vs the CC genotype), and VEGFA haplotype (2 copies vs 0 or 1 copy of the ACG haplotype). In model 2, VEGF risk score was adopted instead of VEGFA +936 C>T genotype and VEGFA haplotype, to confirm the VEGF risk score as a surrogate for GVHD risk. The covariates for aGVHD grade 2-4 or grade 3-4 included age (< 40 years vs ≥ 40 years), disease risk (high risk vs standard risk), conditioning regimen (myeloablative vs reduced intensity), VEGFA +936 C>T genotype (CT vs the CC genotype), and VEGFA haplotype (2 copies vs 0 or 1 copy of the ACG haplotype), whereas those for cGVHD included age (< 40 years vs ≥ 40 years), development of aGVHD (grade 0-2 vs grade 3-4), disease risk (high risk vs standard risk), conditioning regimen (myeloablative vs reduced intensity), VEGFA +936 C>T genotype (CT vs CC genotype), and VEGFA haplotype (2 copies vs 0 or 1 copy of the ACG haplotype). Multivariate analyses using time-dependent or non–time-dependent Cox proportional hazard models were conducted using backward-stepwise modeling and a P value > .05 for the likelihood ratio test. Hazard ratios (HRs) and 95% confidence intervals (CIs) also were estimated.
Statistic significance was accepted for P values < .05. Statistical data were obtained using SPSS version 13.0 (SPSS Inc, Chicago, IL), NCSS version 4.0 (NCSS, Kaysville, UT), and the R package (version 2.4.1, available at http://CRAN.R-project.org).
Results
Overall Transplantation Outcomes
With a median follow-up of 29.5 months posttransplantation (range, 0.5 to 74.5 months), 33 patients (34%) progressed and 53 patients (54%) succumbed to primary disease progression (n = 15) or NRM (n = 38). The 2-year OS was 44.3% ± 5.3%; cumulative incidences of NRM were 10.8% ± 3.6% at 100 days and 24.3% ± 5.1% at 2 years, and the cumulative incidence of recurrence at 2 years was 33.4% ± 5.0%. The cumulative incidences of grade 1-4, grade 2-4, and grade 3-4 severe aGVHD were 81.6% ± 0.2%, 72.2% ± 0.3%, and 39.3% ± 0.5%, respectively, whereas those of cGVHD were 58.1% ± 5.8% at 6 months and 68.9% ± 5.5% at 2 years.
Genotype Frequencies of the VEGFA Polymorphisms
The genotype frequencies of the VEGFA polymorphisms are summarized in Table 2. All polymorphisms complied with the HWE (Table 2). The LDs of VEGFA polymorphisms at loci -2578/-460/+405/+936 are shown in Figure 1. Strong LDs can be seen between loci -460 and +405 (D′ = 0.94), between loci -2578 and -460 (D′ = 1.00), and between loci -2578 and +405 (D′ = 1.00); however, linkages of locus +936 with -2578 (D′ = 0.07), -460 (D′ = 0.09), or +405 (D′ = 0.13) are weak (D′ < 0.5). Accordingly, we generated haplotypes of the VEGFA polymorphisms based on 3 genotypes at loci -2578, -460, and +405. The frequencies of these haplotypes at loci -2578/-460/+405 were 47.9% for CTC, 26.7% for CTG, 24.2% for ACG, and 1.0% for CCC.
Table 2. Genotype Frequencies of VEGF Gene Polymorphisms at Loci -2578/-460/+405/+936
| Allele Ffrequency | ||||
|---|---|---|---|---|
| Locus | % Tested | Major | Minor | HWE P value |
| -2578 | 86.2% | C: 0.778 | A: 0.222 | 0.303 |
| -460 | 87.2% | T: 0.768 | C: 0.232 | 0.450 |
| +405 | 100% | G: 0.511 | C: 0.489 | 0.038 |
| +936 | 100% | C: 0.819 | G: 0.181 | 0.052 |
Figure 1
Pairwise linkage disequilibrium in the study population. SNPs selected for haplotype are shown in bold.
Univariate Analyses for Factors Associated with Transplantation Outcomes, Especially aGVHD
On single-marker analysis, the incidence of grade 2-4 aGVHD was higher in patients with the +936 CT genotype (88%) than in those with the CC genotype (64%; P = .006; Figure 2A), higher in patients with the -2578 AA genotype (100%) than in those with the CA (71%) or CC genotype (73%; P = .003), and higher in patients with the -460 CC genotype (100%) than in those with the CT (73%) or TT genotype (73%; P = .002). On haplotype analysis, the incidence of grade 2-4 aGVHD (100%) was higher in patients with 2 copies of the ACG haplotype than in those with 0 or 1 copy of this haplotype (68%; P = .003; Figure 2B). No association was found between VEGFA SNPs and the incidence of cGVHD.
Figure 2
VEGFA genotype/haplotype and its association with the risk of aGVHD. Higher incidences of grade 2-4 aGVHD were observed in patients with the + 936 CT genotype (A) or the ACG haplotype for loci -2578/-460/+405 (B).
With respect to other transplantation outcomes, patients with the +936 CT genotype were found to have a better OS at 2 years than those with the CC genotype (55% vs 38%; P = .04), but this did not hold for NRM or recurrence. No differences in terms of OS, NRM, or recurrence were noted for genotypes at loci -2578/-460/+405 (Table 3).
Table 3. Transplantation Outcome According to the VEGFA Genotype at Position +936 C>T, the VEGFA ACG Haplotype for Loci -2578/-460/+405, and the VEGFA Risk Score
| Overall Patients | VEGFA +936 C>T Genotype | VEGFA ACG Haplotype at -2578/-460/+405 | VEGFA Risk Score | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| (n = 98) | CT Genotype (n = 62) | CC Genotype (n = 36) | P Value | 0/1 Copy (n = 92) | 2 Copies (n = 6) | P Value | Score 0 (n = 59) | Score 1-2 (n = 39) | P Value | |
| Follow-up, months | 29.5 | 26 | 30.5 | 28.5 | 53 | 26.5 | 30.5 | |||
| aGVHD, n (%) | ||||||||||
| Evaluable patients | 93 | 59 | 29 | 87 | 6 | 56 | 37 | |||
| Overall | 75 (81) | 45 (76) | 30 (88) | .2 | 69 (79) | 6 (100) | .6 | 42 (75) | 33 (89) | 0.1 |
| Grade 2-4 | 63 (68) | 34 (58) | 29 (85) | .006 | 57 (66) | 6 (100) | .08 | 31 (55) | 32 (87) | 0.002 |
| Grade 3,4 | 26 (28) | 16 (27) | 10 (29) | .8 | 23 (26) | 3 (50) | .3 | 14 (25) | 12 (32) | 0.5 |
| cGVHD, n (%) | ||||||||||
| Evaluable patients | 74 | 43 | 31 | 69 | 5 | 41 | 33 | |||
| Overall cGVHD | 54 (73) | 30 (70) | 24 (77) | .6 | 51 (74) | 3 (60) | .6 | 28 (68) | 26 (79) | 0.4 |
| Extensive cGVHD | 30 (41) | 16 (37) | 14 (45) | .6 | 27 (39) | 3 (60) | .4 | 14 (34) | 16 (49) | 0.2 |
| Survival, n (%) | ||||||||||
| Relapse | 33 (34) | 22 (36) | 11 (31) | .7 | 32 (35) | 1 (17) | .7 | 22 (37) | 11 (28) | 0.4 |
| Death | 53 (54) | 37 (60) | 16 (45) | .1 | 50 (54) | 3 (50) | 1.0 | 37 (60) | 16 (45) | 0.1 |
| NRM | 38 (39) | 27 (44) | 11 (31) | .3 | 36 (39) | 2 (33) | 1.0 | 26 (44) | 12 (31) | 0.2 |
VEGF Risk Score Model Predicting the Risk of aGVHD
Based on the +936 C>T genotype and VEGFA haplotype information, we scored VEGFA SNPs based on the genotype at locus +936 and copy number of the ACG haplotype at loci -2578/-405/+460. A score of 1 was assigned to risk alleles (ie, +936 CT or TT genotypes, or 2 copies of the ACG haplotype), and a score of 0 was assigned to other alleles (ie, +936 CC genotype, or 0 or 1 copy of the ACG haplotype). After summing scores, 2 risk groups were defined: high risk (composite score 2 or 1; n = 37) and low risk (composite score 0; n = 56). Significant correlations were found between the incidence of grade 1-4 aGVHD or grade 2-4 aGVHD and the VEGFA SNP score model (Figure 3A and B). The incidence of grade 2-4 aGVHD was 89% in the high-risk patients (score 1-2, ie, +936 CT/TT genotype or 2 copies of ACG haplotype) and 62% in the low-risk patients (score 0; ie, +936 CC genotype and 0 or 1 copy of ACG haplotype) (P = .001). No differences were seen between the high-risk and low-risk patients in terms of OS, NRM, or recurrence (Table 3).
Figure 3
VEGFA risk score model and its association with the risk of aGVHD. Patients with a VEGFA risk score of 1 or 2 had higher incidences of aGVHD grade 1-4 (A) or grade 2-4 (B) than those with a VEGFA risk score of 0.
Organ-Specific Development of aGVHD According to VEGFA SNPs
We examined the relationships between organ-specific onset of aGVHD and the VEGFA SNPs. The VEGFA SNPs, especially the -2578 and -460 genotypes, were found to be closely associated with the development of gut GVHD (Figure 4), and a higher risk of gut GVHD was found to be associated with the -2578 AA genotype (P = .04, adjusted P = .08; Figure 4A) or a non-C carrier (P = .02, adjusted P = .05), as well as with the -460 TT genotype (P = .05, adjusted P = .07; Figure 4B) or a non-T carrier (P = .02, adjusted P = .05). Adjustment was performed for age, conditioning regimen, and disease risk.
Figure 4
Associations between VEGFA genotypes with the risk of gut GVHD. The VEGFA -2578C>A (A) and -460T>C genotypes (B) were found to be associated with the risk of gut GVHD.
Multivariate Analysis
Multivariate analysis for the risk of grade 2-4 aGVHD identified 2 risk factors in model 1 (Table 4): the +936 CT genotype (P = .03; HR = 1.79; 95% CI = 1.07 to 2.99) and the ACG VEGFA haplotype (P = .01; HR = 3.09; 95% CI = 1.27 to 7.52). In model 2, the high-risk group (VEGFA risk score 1 or 2) had a significantly greater risk of aGVHD grade 2-4 (P = .001; HR = 2.32; 95% CI = 1.41 to 3.81). The only independent risk factor for grade 3-4 severe aGVHD was the ACG haplotype (P = .02; HR = 4.46; 95% CI = 1.29 to 15.38). Other clinical factors were not found to be associated with the risk of aGVHD, perhaps because our cohort included only HLA-matched sibling PBSCT recipients.
Table 4. Multivariate Analysis of aGVHD or cGVHD and Transplantation Outcomes
| Risk Factor | HR (95% CI) | P value | |
|---|---|---|---|
| aGVHD, grade 2-4 | |||
| Model 1∗ | |||
| +936 C>T VEGFA genotype | CC genotype | 1.00 | .03 |
| CT genotype | 1.79 (1.07 to 2.99) | ||
| ACG VEGFA haplotype | 0 or 1 copy of ACG haplotype | 1.00 | .01 |
| 2 copies of ACG haplotype | 3.09 (1.27 to 7.52) | ||
| Model 2† | |||
| VEGFA risk score | Score 0 | 1.0 | .001 |
| Score 1 to 2 | 2.32 (1.41 to 3.81) | ||
| aGVHD, grade 3-4∗ | |||
| ACG VEGFA haplotype | 0 or 1 copy of ACG haplotype | 1.00 | .02 |
| 2 copies of ACG haplotype | 4.46 (1.29 to 15.38) | ||
| cGVHD‡ | |||
| aGVHD | aGVHD, grade 0-2 | 1.00 | < .001 |
| aGVHD, grade 3-4 | 3.70 (1.96 to 6.99) | ||
| NRM§ | |||
| aGVHD | aGVHD, grade 0-2 | 1.00 | .003 |
| aGVHD, grade 3-4 | 4.15 (1.65 to 10.53) | ||
| Relapse§ | |||
| aGVHD | aGVHD, grade 0-2 | 1.00 | .05 |
| aGVHD, grade 3-4 | 4.386 (1.027 to 18.729) | ||
| Disease risk | Standard risk | 1.00 | .002 |
| High risk | 3.46 (0.133 to 0.627) | ||
∗Age, disease risk, conditioning regimen, VEGFA +936 C>T genotype, and VEGFA ACG haplotype were included in the analysis. |
†Age, disease risk, conditioning regimen, and VEGFA risk score were included in the analysis. |
‡Age, development of aGVHD, disease risk, conditioning regimen, VEGFA +936 C>T genotype, and VEGFA ACG haplotype were included in the analysis. |
§Age, development of aGVHD or cGVHD, disease risk, conditioning regimen, VEGFA +936 C>T genotype, and VEGFA ACG haplotype were included in the analysis. |
VEGFA SNPs were not found to be associated with the risk of cGVHD, although a history of a previous episode of grade 3-4 aGVHD was found to be significantly associated with a greater risk of cGVHD (P < .001; HR = 3.70; 95% CI = 1.96 to 6.99). In terms of overall survival, grade 3-4 aGVHD (P < .001; HR = 40.31; 95% CI = 9.97 to 162.89) and high risk of disease (P = .001; HR = 6.86; 95% CI = 2.15 to 21.86) were found to be independent risk factors. With respect to other transplantation outcomes, such as NRM and relapse, grade 3-4 aGVHD was found to be significantly associated with a greater risk of NRM (P < .001; HR = 74.02; 95% CI = 16.58 to 826.21), as was VEGFA +936C>T genotype (P = .01; HR = 29.43; 95% CI = 1.99 to 435.24). High-risk status (P = .002; HR = 3.46; 95% CI = 1.59 to 7.53) and a previous episode of grade 3-4 aGVHD (P = .05; HR = 0.23; 95% CI = 0.05 to 0.97) were found to be independent predictors of relapse after allogeneic transplantation.
Discussion
The results of the present study suggest an association between VEGFA SNPs and a lower risk of aGVHD after allogeneic SCT (especially gut GVHD), as well as a protective role for VEGF in the pathogenesis of aGVHD. VEGF exerts 2 actions in this respect: a proinflammatory effect that may provoke inflammatory reactions in target tissues [25] and an angiogenic effect that may facilitate tissue reperfusion and regeneration in inflamed tissues [26]. Consequently, VEGF may have a protective or augmentative role in the pathogenesis of GVHD in opposite directions; elevated VEGF could either increase the severity of GVHD by promoting inflammation or reduce the severity of GVHD by stimulating tissue perfusion.
Our findings suggest that VEGF may protect against the development of severe GVHD, in agreement with previous reports associating elevated VEGF levels on day 14 or 15 posttransplantation with decreased risk of severe aGVHD and NRM 4, 6 Whether VEGF up-regulation is a consequence of tissue hypoxia in GVHD or whether it acts as a mediator of endothelial regeneration by inhibiting irreversible fibrosis caused by cytotoxic T lymphocytes is unclear, however. Regardless, it is apparent that high VEGF level may be a surrogate marker of a decreased risk for severe aGVHD or NRM after allogeneic SCT.
On the other hand, a study in a murine model concluded that VEGF has a proinflammatory function in the alloimmunity setting [27]. In the murine renal allograft model used in that study, anti-VEGF antibody was found to markedly inhibit T cell infiltration into allografts and acute rejection, and thus it was concluded that VEGF exerts robust proinflammatory activity in the alloimmunity setting [27]. But the renal allograft environment in a murine model likely differs greatly from that in a human GVHD setting. Accordingly, further studies are warranted to clarify this issue, and to determine whether VEGF is a driver or passenger in the development of GVHD.
Similarly, studies of VEGF in inflammatory bowel disease (IBD) suggest that VEGF has proinflammatory activity, with VEGF expression found to be higher in patients with IBD than in healthy controls and to be positively correlated with disease activity 28, 29. These results are at odds with our finding that lower VEGF production is associated with a higher risk of gut GVHD. But microvessel densities are attenuated in GVHD tissues [1] but elevated in IBD [30], and thus we postulate that although gut GVHD and IBD share similar pathogenetic mechanisms and pathological findings, VEGF's proinflammatory activity is more prominent during the pathogenesis of IBD, whereas its angiogenic activity is more significant in GVHD.
Genetic variations in the genes encoding these molecules may affect transcription and translation or may modulate the functions of the gene products. Growing evidence supports the notion that cytokine gene polymorphisms are important predictors of transplantation-related complications, including aGVHD and cGVHD, and transplantation outcomes 15, 31, 32, 33. For example, it is generally accepted that the SNPs of proinflammatory cytokines, such as interleukin (IL)-1, IL-2, IL-6, interferon-γ, or tumor necrosis factor-α, and of anti-inflammatory cytokines, such as IL-10, affect transplantation outcomes, including aGVHD and cGVHD and transplantation-related mortality [15].
In the present study, patients with the +936 CT genotype were found to have a higher risk of grade 2-4 aGVHD than patients with the CC genotype (88% vs 64%; P = .006; Figure 2A). In addition, patients with 2 copies of the ACG haplotype were found to have a higher risk of grade 2-4 aGVHD than patients with 0 or 1 copy of the ACG haplotype (100% vs 68%; P = .003; Figure 2B). Furthermore, according to our risk score model, the incidence of grade 2-4 aGVHD was 89% in the high-risk patients (score 1-2, ie, +936 CT genotype or any ACG haplotype), compared with 62% in the low-risk patients (P = .001; Figure 3B). However, a limitation of the current study is the very low number of high-risk patients (ie, 6). Accordingly, further studies are needed with larger numbers of patients to enable a clear conclusion on this issue.
In terms of the specific effects of the VEGFA genotypes, the +936 CT genotype and the ACG haplotype (for loci -2578/-460/+405) appeared to reduce VEGF production. In our previous study of patients with acute myeloid leukemia (AML), those patients with the +936 CC genotype and CTG haplotype had a higher risk of relapse, and this was associated with VEGF up-regulation. Conversely, the +936 CT genotype and the ACG haplotype were associated with VEGF down-regulation. Previous studies also have found an association between the +936T allele in the +936C>T genotype and VEGF down-regulation 10, 34. VEGF expression differences caused by VEGFA SNPs may be due to (1) a loss of a potential binding with the transcription factor AP-4 in the presence of the C-to-T transition; (2) LD of the polymorphism with another, as-yet unidentified polymorphism; or (3) modification of the mRNA structure [10]. In the present study, we did not perform promoter assays to evaluate the functional role of VEGFA haplotypes; however, a previous study in a Korean population with the TG haplotype at loci -460/+405 suggested an association between the TG haplotype at loci -460/+405 and VEGF up-regulation [17].
In conclusion, our findings support the notion that VEGF protects against aGVHD, especially against the development of severe aGVHD. They also suggest that VEGFA polymorphisms can be used to predict the risk of aGVHD.
Acknowledgments
Financial disclosure: The authors have nothing to disclose.
Reference
- Endothelial injury mediated by cytotoxic T lymphocytes and loss of microvessels in chronic graft versus host disease. Lancet. 2002;359:2078–2083
- . Vascular endothelial growth factor, a potent and selective angiogenic agent. The Journal of biological chemistry. 1996;271:603–606
- . Vascular endothelial growth factor and its receptors. Cytokine & growth factor reviews. 1996;7:259–270
- Vascular endothelial growth factor (VEGF) is associated with reduced severity of acute graft-versus-host disease and nonrelapse mortality after allogeneic stem cell transplantation. Bone marrow transplantation. 2006;38:149–156
- TH2 Cytokine-enhanced and TGF-beta-enhanced vascular endothelial growth factor production by cultured human airway smooth muscle cells is attenuated by IFN-gamma and corticosteroids. The Journal of allergy and clinical immunology. 2003;111:1307–1318
- . Vascular endothelial growth factor and activin-a serum levels following allogeneic hematopoietic stem cell transplantation. Biol Blood Marrow Transplant. 2007;13:942–947
- The human gene for vascular endothelial growth factor. Multiple protein forms are encoded through alternative exon splicing. The Journal of biological chemistry. 1991;266:11947–11954
- . Assignment of the vascular endothelial growth factor gene to human chromosome 6p21.3. Circulation. 1996;93:1493–1495
- . Novel polymorphisms in the promoter and 5' UTR regions of the human vascular endothelial growth factor gene. Human immunology. 1999;60:1245–1249
- . A common 936 C/T mutation in the gene for vascular endothelial growth factor is associated with vascular endothelial growth factor plasma levels. Journal of vascular research. 2000;37:443–448
- . Identification of polymorphisms within the vascular endothelial growth factor (VEGF) gene: correlation with variation in VEGF protein production. Cytokine. 2000;12:1232–1235
- . Haplotype analysis of the polymorphic human vascular endothelial growth factor gene promoter. Cancer research. 2003;63:812–816
- . VEGF gene sequence variation defines VEGF gene expression status and angiogenic activity in non-small cell lung cancer. Lung cancer (Amsterdam, Netherlands). 2004;46:293–298
- DNA sequence variation in the promoter region of the VEGF gene impacts VEGF gene expression and maximal oxygen consumption. American journal of physiology. 2006;290:H1848–H1855
- IL-10 promoter gene polymorphism associated with the occurrence of chronic GVHD and its clinical course during systemic immunosuppressive treatment for chronic GVHD after allogeneic peripheral blood stem cell transplantation. Transplantation. 2005;79:1615–1622
- Time to first flare-up episode of GVHD can stratify patients according to their prognosis during clinical course of progressive- or quiescent-type chronic GVHD. Bone marrow transplantation. 2007;40:779–784
- Vascular endothelial growth factor gene polymorphisms and risk of primary lung cancer. Cancer Epidemiol Biomarkers Prev. 2005;14:571–575
- Association of vascular endothelial growth factor gene polymorphisms with behcet disease in a Korean population. Hum Immunol. 2005;66:1068–1073
- VEGF gene polymorphisms and susceptibility to rheumatoid arthritis. Rheumatology (Oxford). 2004;43:1173–1177
- . Vascular endothelial growth factor (VEGF) gene (VEGFA) polymorphism can predict the prognosis in acute myeloid leukaemia patients. British journal of haematology. 2008;140:71–79
- . A new statistical method for haplotype reconstruction from population data. American journal of human genetics. 2001;68:978–989
- 1994 Consensus Conference on Acute GVHD Grading. Bone marrow transplantation. 1995;15:825–828
- Chronic cutaneous graft-versus-host disease in man. The American journal of pathology. 1978;91:545–570
- . Competing risk analysis using R: an easy guide for clinicians. Bone marrow transplantation. 2007;40:381–387
- . Vascular endothelial growth factor and its relationship to inflammatory mediators. Clin Cancer Res. 2007;13:2825–2830
- . Vascular permeability factor/vascular endothelial growth factor: an important mediator of angiogenesis in malignancy and inflammation. International archives of allergy and immunology. 1995;107:233–235
- Proinflammatory functions of vascular endothelial growth factor in alloimmunity. J Clin Invest. 2003;112:1655–1665
- . VEGF, basic-FGF, and TGF-beta in Crohn's disease and ulcerative colitis: a novel mechanism of chronic intestinal inflammation. The American journal of gastroenterology. 2001;96:822–828
- Vascular endothelial growth factor in inflammatory bowel disease. International journal of colorectal disease. 2003;18:418–422
- Angiogenesis as a novel component of inflammatory bowel disease pathogenesis. Gastroenterology. 2006;130:2060–2073
- . Genetic polymorphisms predicting the outcome of bone marrow transplants. Br J Haematol. 2004;127:479–490
- . Genomic screening and complications of hematopoietic stem cell transplantation: has the time come?. Bone Marrow Transplant. 2005;35:1–16
- . Beyond the HLA typing age: genetic polymorphisms predicting transplant outcome. Blood Rev. 2005;19:333–340
- A common 936 C/T gene polymorphism of vascular endothelial growth factor is associated with decreased breast cancer risk. Int J Cancer. 2003;106:468–471
Financial disclosure: See Acknowledgments on page 1415.
Presented at the European School of Hematology Euroconference on Translational Research in Transplantation, Integrating Immunity: From Genomics to Cell Therapy, Leiden, The Netherlands, November 2007.
PII: S1083-8791(08)00421-7
doi:10.1016/j.bbmt.2008.09.022
© 2008 American Society for Blood and Marrow Transplantation. Published by Elsevier Inc. All rights reserved.
Volume 14, Issue 12 , Pages 1408-1416, December 2008
