Abstract
Early VEGF-A reduction (EVR) by targeting abundant VEGF-A is a potential predictive marker of bevacizumab (BEV). The CCL5/ CCR5 axis modulates VEGF-A production via endothelial progenitor cells migration. We tested whether genetic polymorphisms in the CCL5/CCR5 pathway could predict efficacy of BEV in patients with metastatic colorectal cancer (mCRC) in a first-line setting. Genomic DNA was extracted from 215 samples from three independent cohorts: 61 patients receiving FOLFOX+BEV (evaluation cohort); 83 patients receiving FOLFOX (control cohort); 71 patients receiving FOLFOX/XELOX+BEV (exploratory cohort) for validation and serum biochemistry assay (n = 48). Single nucleotide polymorphisms of genes in the CCL5/CCR5 pathway were analyzed by PCR-based direct sequencing. Considering the unbalanced distribution of patient baseline characteristics between the evaluation and control cohorts, propensity score matching analysis was performed. Serum VEGF-A levels during treatment were measured using ELISA. Among the evaluation and control cohorts, patients with any CCL5 rs2280789 G allele had longer progression-free survival (PFS) and overall survival (OS) when receiving FOLFOX+BEV than FOLFOX (PFS: 19.8 vs. 11.0 months, HR 0.44, 95%CI: 0.24–0.83, p = 0.004; OS: 41.8 vs. 24.5 months, HR: 0.50, 95%CI: 0.26–0.95, p = 0.024). No significant difference was shown in patients with the A/A variant. In the exploratory cohort, CCL5 rs2280789 G alleles were associated with higher VEGF-A levels at baseline and a greater decrease in VEGF-A levels at day 14 compared to the A/A variant. CCL5 and CCR5 impact the angiogenic environment, and the genotypes in CCL5/CCR5 genes may identify specific populations who will benefit from BEV in first-line treatment for mCRC.
Keywords: CCL5, CCR5, VEGF-A, bevacizumab, metastatic colorectal cancer
Introduction
Early vascular endothelial growth factor (VEGF)-A reduction (EVR) by targeting abundant VEGF-A is a potential predictive marker of anti-VEGF therapy.1 The C-C motif chemokine ligand 5 (CCL5)/CC motif chemokine receptor 5 (CCR5) axis modulates VEGF-A production via endothelial progenitor cell migration in a CCR5-dependent manner.2 A recent study reported that pretreatment serum CCL5 levels and a decrease in serum VEGF-A levels during treatment predicted the efficacy of regorafenib in refractory metastatic colorectal cancer (mCRC).3 We previously reported that the homozygote in CCL5 rs2280789 and rs3817655 was significantly associated with lower serum CCL5 levels compared to other variants at baseline and day 21, which affected VEGF-A production in refractory mCRC patients receiving regorafenib.4 CCL5 is a CC chemokine and is characterized as late expression after T cell activation, and localizes with tumor-infiltrating leuko-cytes.5 CCL3 and CCL4 are also involved in EPCs migration via binding to their receptor CCR5, however an in vitro study showed that CCL5 is the most potent chemoattractant of EPCs.6 The CCL5/CCR5 signaling pathway positively activates protein kinase Cd (PKCδ), c-Src and hypoxia-inducible factor-1a (HIF-1α) to activate VEGF-A expression.2 CCL5 is also known as RANTES (regulated on activation normal T cell expressed and secreted). Krüppel-like transcription factor (KLF) 13 is a transcription factor located at upstream of CCL5/CCR5 signaling that regulates RANTES expression in T lymphocytes, and is known as RFLAT-1 (RANTES factor of late activated T lymphocytes 1).7
BEV is a recombinant, humanized monoclonal antibody that targets vascular endothelial growth factor and has been widely used in several cancer types including mCRC. However, its predictive and prognostic value is not yet fully understood. We therefore tested whether genetic polymorphisms in the CCL5/CCR5 signaling pathway predict outcomes in mCRC patients receiving BEV in a first-line setting. In addition, the angiogenic environment and its variation for each genotype were evaluated by measuring circulating angiogenic factors including VEGF-A throughout treatment with BEV.
Materials and Methods
Study design and patients
Our study investigated three independents cohorts composed of patients with histologically-confirmed mCRC; an evaluation cohort of 61 patients treated with FOLFOX plus BEV and a control cohort of 84 patients treated with FOLFOX alone from a retrospective study,8 and an exploratory cohort of 71 patients treated with BEV plus oxaliplatin-based treatment FOLFOX or XELOX with participating in a blood analysis research. The exploratory cohort was focused on the measurement of serum cytokine levels but was also used for validation of the results from the evaluation cohort. The control and evaluation cohorts started the treatment before and after approval of BEV in Japan, respectively, which could reduce bias by physician’s choice. Al patients were treated at The Cancer Institute Hospital (Tokyo, Japan). Eligible patients had a histologically confirmed diagnosis of mCRC, measurable or evaluable disease according to Response Evaluation Criteria in Solid Tumors (RECIST) v1.0, had no prior treatment for metastatic disease, and provided signed informed consent. FOLFOX ± BEV treatment (oxaliplatin 85 mg/m2, 5-fluorouracil [5-FU] bolus 400 mg/m2, 5-FU infusion 2,400 mg/m2, levofolinate calcium 200 mg/m2, with or without BEV 5 mg/kg) was administered every 2 weeks. XELOX + BEV treatment (oxaliplatin 130 mg/m2, capecitabine 1,000 mg/m2 given twice daily from the evening of day 1 to the morning of day 15, with or without BEV 5 mg/kg) was administered every 3 weeks. Doses were adjusted based on adverse events at the treating physician’s discretion after the manufacturer’s recommendations. Treatment was continued until any of the after occurred: disease progression, unmanageable toxicity, or patient refusal. We were fully compliant with the reporting recommendations for tumor marker prognostic studies (http://jnci.oxfordjournals.org/content/97/16/1180/T1.expansion.html). Tissue analysis was approved by the Institutional Review Board of each institute and was conducted at the University of Southern California/Norris Comprehensive Cancer Center in accordance with the Declaration of Helsinki and Good Clinical Prac Guidelines.
Selection of candidate single-nucleotide polymorphisms (SNPs)
The five candidate SNPs involved in the CCL5/CCR5 signaling pathway present in genes CCL3, CCL4, CCL5, CCR5, and KLF13 were selected using either of the after criteria: i) SNP with biological significance according to a published literature review, ii) tagging SNPs selected by the HapMap genotype data with r2threshold = 0.8: http://snpinfo.niehs.nih.gov/snpinfo/snptag.php, or iii) minor allele frequency ≥10% in both Caucasians and East Asians (in the Ensembl Genome Browser: http://uswest.ensembl.org/index.html). Functional significance was predicted based on the functional single-nucleotide polymorphism (F-SNP) database: http://compbio.cs.queensu.ca/F-SNP/ (Supporting Information Table S1). All selected SNPs met criteria ii) and iii) except CCL3 SNPs did not meet ii).
DNA extraction and genotyping
Genomic DNA was extracted from formalin-fixed paraffin-embedded tissues from the patients in all cohorts using the QIAmp Kit (Qiagen, Valencia, CA) according to the manufacturer’s protocol (www.qiagen.com). The candidate SNPs were tested using PCR-based direct DNA sequence analysis using an ABI 3100A Capillary Genetic Analyzer and Sequencing Scanner v1.0 (Applied Biosystems). The forward and reverse primers used for amplification of extracted DNA are listed in Supporting Information Table S1. For quality control purposes, a randomly selected 10% of the samples was analyzed by direct DNA sequencing for each SNP, and the genotype concordance rate was 99% or higher. The investigators analyzing SNPs were blinded to the clinical data.
Analysis of angiogenic factors levels
In the exploratory cohort, blood samples were obtained at baseline, day 14 and day 56 during the treatment. Separated serum was stored at −80 °C. The serum levels of VEGF-A, stromal cell derived factor-1 (SDF-1), angiopoietin-1 (Ang-1), angiopoietin-2 (Ang-2), VEGF-C, platelet-derived growth factor subunit B (PDGFB), interleukin-8 (IL-8), transforming growth factor beta (TGFB), placental growth factor (PlGF) and basic fibroblast growth factor (bFGF) were measured using Quantikine ELISA kits (R&D Systems). Blood analysis was finally available for 48 of 71 patients in the exploratory cohort, who received BEV plus FOLFOX or XELOX treatment at The Cancer Institute Hospital. Other 23 patients were excluded from the analysis due to the lack of samples. In our study, we investigated the association between circulating angiogenic factors and candidate SNPs under BEV treatment.
Statistical analysis
The primary endpoint of our study was progression-free survival (PFS), and the secondary endpoints were overall survival (OS) and overall response rate (ORR). PFS was defined as the interval between the date of starting treatment and the date of confirmed disease progression or death. Data of patients without disease progression or death were censored at the date of last follow up. OS was calculated from the date of starting treatment until the date of death from any cause. In patients who were lost to follow-up, data were censored at the date of last follow up. The ORR was based on the proportion of patients who had a complete response and partial response according to RECISTv1.0. Chi-square tests were used to examine the difference in baseline patient characteristics among the three cohorts. Considering the unbalanced distribution of patient baseline characteristics between the evaluation and control cohorts, propensity score (PS) matching analysis was performed. The logistic regression was conducted for calculating the PS on treatment including significantly unbalanced variables. The PS matching was then performed based on the nearest neighbor matching method.9 Allelic distribution of all SNPs by ethnicity for deviation from the Hardy–Weinberg equilibrium and the allelic frequencies of SNPs were tested using Chi-square tests. Fisher’s exact test was applied to examine the associations between SNPs and response. PFS and OS were estimated by the Kaplan–Meier method and were compared using the log-rank test, with predictive or prognostic clinical factors and candidate SNPs that were identified by univariate analysis using codominant, dominant, or recessive genetic models if appropriate. Multivariable analysis using the Cox proportional hazards model was conducted to identify factors influencing PFS and OS. The baseline demographic and clinical characteristics that remained statistically significantly associated with PFS and OS in multivariable analyses were included in the final models to reevaluate the independent effect of candidate SNPs. Serum cytokine levels at the three points and level changes during treatment were compared among the genetic variants of candidate SNPs using Student’s unpaired t test and one-way analysis of variance. The power would be 80% to detect a minimum hazard ratios ranged from 2.25 to 2.66 corresponding to the minor allele frequency of 0.1 to 0.4 in the association between a SNP and PFS among the evaluation cohort (N = 61, PFS events = 43) considering a dominant model and using a two-sided 0.05-level log-rank test. The power was greater than 80% and 85% when applying the same model and test to the control and exploratory cohorts, respectively. All analyses were carried out with SAS 9.4 (SAS Institute, Cary, NC) and R (version 3.4.3) package “twang”. All tests were 2-sided at a significance level of 0.05. p Values were not adjusted for multiple testing.
Results
Baseline patient and tumor characteristics
The baseline characteristics of the three cohorts before and after the propensity score matching were summarized in Supporting Information Table S2 and Table 1, respectively. Patient adjuvant history and KRAS status remained differently distributed after the PS matching. There was only one patient with adjuvant chemotherapy history in the control cohort. In addition, there was 82% of patients with KRAS status unavailable in the control cohort. The median follow-up time, median PFS and OS were 39.2 months, 18.4 months and 40.5 months in the evaluation cohort; 57.6 months, 12.3 months and 29.3 months in the control cohort; and 28.9 months, 13.3 months and 35.5 months in the exploratory cohort. Frequencies of genetic variants of the SNPs were within a Hardy–Weinberg equilibrium (p > 0.05) in each cohort except CCL4 rs1634517 in the control cohort (p > 0.01).
Table 1.
Baseline patients and tumor characteristics after propensity score matching
| Evaluation cohort (N = 61) |
Control cohort (N = 61) |
Exploratory cohort (N = 71) |
|||||
|---|---|---|---|---|---|---|---|
| Cohort | N | % | N | % | N | % | p Value1 |
| Sex | 0.39 | ||||||
| Male | 28 | 46 | 33 | 54 | 30 | 42 | |
| Female | 33 | 54 | 28 | 46 | 41 | 58 | |
| Age (year) | |||||||
| Median (range) | 60 (27–74) | 58 (28–74) | 60 (32–79) | 0.89 | |||
| ≤65 | 50 | 82 | 46 | 75 | 49 | 69 | |
| >65 | 11 | 18 | 15 | 25 | 22 | 31 | |
| Performance status | 0.34 | ||||||
| ECOG 0 | 61 | 100 | 60 | 98 | 71 | 100 | |
| ECOG 1–2 | 0 | 0 | 1 | 2 | 0 | 0 | |
| Primary tumor site | 0.48 | ||||||
| Right | 17 | 28 | 16 | 26 | 25 | 35 | |
| Left | 44 | 72 | 45 | 74 | 46 | 65 | |
| Liver metastasis | 0.19 | ||||||
| Yes | 37 | 61 | 27 | 44 | 37 | 52 | |
| No | 24 | 39 | 34 | 56 | 34 | 48 | |
| Lung metastasis | 0.82 | ||||||
| Yes | 23 | 38 | 23 | 38 | 30 | 42 | |
| No | 38 | 62 | 38 | 62 | 41 | 58 | |
| Lymph node metastasis | 0.66 | ||||||
| Yes | 26 | 43 | 31 | 51 | 33 | 46 | |
| No | 35 | 57 | 30 | 49 | 38 | 54 | |
| Peritoneal metastasis | 0.68 | ||||||
| Yes | 16 | 26 | 12 | 20 | 17 | 24 | |
| No | 45 | 74 | 49 | 80 | 54 | 76 | |
| Number of metastases | 0.43 | ||||||
| 1 | 23 | 38 | 21 | 34 | 30 | 42 | |
| 2 | 25 | 41 | 23 | 38 | 31 | 44 | |
| ≥3 | 13 | 21 | 17 | 28 | 10 | 14 | |
| Primary tumor resected | 0.36 | ||||||
| Yes | 52 | 85 | 46 | 75 | 55 | 77 | |
| No | 9 | 15 | 15 | 25 | 16 | 23 | |
| Adjuvant history | <0.001 | ||||||
| Yes | 23 | 38 | 1 | 2 | 17 | 24 | |
| No | 38 | 62 | 60 | 98 | 54 | 76 | |
| KRAS exon2 status | <0.001 | ||||||
| Wild-type | 29 | 48 | 8 | 13 | 29 | 41 | |
| Mutant | 15 | 25 | 3 | 5 | 40 | 56 | |
| Unknown | 17 | 28 | 50 | 82 | 2 | 3 | |
p Values <0.05 were shown in bold.
p Value was based on Chi-square test, or the Kruskal-Wallis test when appropriate.
Clinical outcome and SNPs tested in the cohorts receiving BEV plus chemotherapy
There were no significant differences between SNPs in genes, tumor response, PFS and OS in the evaluation cohort receiving BEV plus chemotherapy (Table 2). In addition, no statistical significance was observed in the exploratory cohort treated with BEV plus chemotherapy (Supporting Information Table S3).
Table 2.
Association between gene polymorphism and clinical outcome
| Tumor response |
Progression-free survival |
Overall survival |
||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| N | CR + PR | SD + PD | p Value* | Median, months (95% CI) | HR (95% CI)† | p Value* | HR (95% CI) ‡ | P Value* | Median, months (95%CI) | HR (95% CI) † | p Value* | HR (95% CI)‡ | p Value* | |
| Evaluation cohort | ||||||||||||||
| CCL5 rs2280789 | 0.67 | 0.24 | 0.79 | 0.30 | 0.26 | |||||||||
| A/A | 22 | 13 (59%) | 9 (41%) | 18.0 (7.8,20.3) | 1 (Reference) | 1 (Reference) | 36.3 (16.6,47.41) | 1 (Reference) | 1 (Reference) | |||||
| A/G2 | 28 | 17 (61%) | 11 (39%) | 19.8 (13.9,28.5) | 0.70 (0.38,1.29) | 1.14 (0.43,3.06) | 41.8 (29.0,45.11) | 0.67 (0.32,1.43) | 0.53 (0.18,1.61) | |||||
| G/G2 | 9 | 7 (78%) | 2 (22%) | |||||||||||
| CCR5 rs1799988 | 0.84 | 0.15 | 0.28 | 0.93 | 0.74 | |||||||||
| C/C | 20 | 12 (60%) | 8 (40%) | 18.0 (7.3,28.8) | 1 (Reference) | 1 (Reference) | 45.1 + (18.4,45.11) | 1 (Reference) | 1 (Reference) | |||||
| C/T | 26 | 18 (69%) | 8 (31%) | 13.3 (10.6,19.8) | 1.25 (0.60,2.60) | 1.23 (0.52,2.91) | 41.8 (27.7,47.41) | 0.92 (0.37,2.26) | 1.32 (0.45,3.91) | |||||
| T/T | 14 | 9 (64%) | 5 (36%) | 20.1 (9.0,41.6) | 0.62 (0.27,1.45) | 0.62 (0.23,1.63) | 39.3 (23.0,47.01) | 1.08 (0.41,2.85) | 1.59 (0.48,5.25) | |||||
| 0.58 | 0.76 | 0.86 | 0.97 | 0.51 | ||||||||||
| C/C | 20 | 12 (60%) | 8 (40%) | 18.0 (7.3,28.8) | 1 (Reference) | 1 (Reference) | 45.1 + (18.4,45.11) | 1 (Reference) | 1 (Reference) | |||||
| Any T | 40 | 27 (68%) | 13 (33%) | 18.4 (12.4,20.3) | 0.91 (0.46,1.78) | 0.93 (0.42,2.09) | 40.5 (28.6,47.41) | 0.98 (0.43,2.24) | 1.41 (0.51,3.94) | |||||
| 1.00 | 0.070 | 0.13 | 0.74 | 0.56 | ||||||||||
| Any C | 46 | 30 (65%) | 16 (35%) | 16.0 (12.4,23.1) | 1 (Reference) | 1 (Reference) | 41.8 (28.6,47.41) | 1 (Reference) | 1 (Reference) | |||||
| T/T | 14 | 9 (64%) | 5 (36%) | 20.1 (9.0,41.6) | 0.55 (0.27,1.13) | 0.54 (0.24,1.20) | 39.3 (23.0,47.01) | 1.14 (0.51,2.53) | 1.31 (0.53,3.23) | |||||
| KLF rs2241779 | 0.41 | 0.39 | 0.72 | 0.33 | 0.70 | |||||||||
| A/A | 24 | 15 (63%) | 9 (38%) | 19.9 (13.1,31.8) | 1 (Reference) | 1 (Reference) | 27.7,47.41) | 1 (Reference) | 1 (Reference) | |||||
| A/C2 | 23 | 17 (74%) | 6 (26%) | 14.7 (8.8,26.6) | 1.31 (0.69,2.48) | 1.15 (0.54,2.45) | 41.8 (25.1,44.01) | 1.46 (0.66,3.24) | 0.83 (0.32,2.15) | |||||
| C/C2 | 10 | 5 (50%) | 5 (50%) | |||||||||||
| CCL3 rs1130371 | 0.84 | 0.36 | 0.50 | 0.82 | 0.74 | |||||||||
| G/G | 31 | 20 (65%) | 11 (35%) | 18.0 (10.6,23.1) | 1 (Reference) | 1 (Reference) | 41.8 (25.1,47.01) | 1 (Reference) | 1 (Reference) | |||||
| G/A2 | 26 | 17 (65%) | 9 (35%) | 18.6 (12.4,27.9) | 0.76 (0.41,1.39) | 0.80 (0.42,1.52) | 40.5 (23.1,47.41) | 0.92 (0.45,1.89) | 1.14 (0.53,2.43) | |||||
| A/A2 | 4 | 2 (50%) | 2 (50%) | |||||||||||
| CCL4 rs1634517 | 0.50 | 0.50 | 0.68 | 0.38 | 0.83 | |||||||||
| C/C | 37 | 25 (68%) | 12 (32%) | 18.4 (11.8,23.1) | 1 (Reference) | 1 (Reference) | 36.3 (28.5,44.01) | 1 (Reference) | 1 (Reference) | |||||
| C/A2 | 19 | 12 (63%) | 7 (37%) | 18.6 (12.7,26.6) | 0.80 (0.42,1.52) | 1.17 (0.56,2.46) | 47.4 + (23.0,47.41) | 0.71 (0.32,1.55) | 1.10 (0.45,2.67) | |||||
| A/A2 | 5 | 2 (40%) | 3 (60%) | |||||||||||
| Control cohort | ||||||||||||||
| CCL5 rs2280789 | 0.94 | 0.43 | 0.77 | 0.006 | 0.022 | |||||||||
| A/A | 28 | 15 (54%) | 13 (46%) | 12.4 (7.1,27.8) | 1 (Reference) | 1 (Reference) | 33.0 (17.3,46.1) | 1 (Reference) | 1 (Reference) | |||||
| A/G | 21 | 11 (52%) | 10 (48%) | 13.0 (7.5,21.4) | 1.03 (0.49,2.19) | 1.07 (0.46,2.44) | 30.8 (20.5,69.1) | 0.78 (0.39,1.54) | 1.42 (0.64,3.14) | |||||
| G/G | 11 | 5 (45%) | 6 (55%) | 9.9 (2.7,18.2) | 1.64 (0.71,3.81) | 1.38 (0.57,3.39) | 16.6 (8.0,31.1) | 2.48 (1.15,5.33) | 3.24 (1.40,7.50) | |||||
| 0.80 | 0.52 | 0.63 | 0.68 | 0.048 | ||||||||||
| A/A | 28 | 15 (54%) | 13 (46%) | 12.4 (7.1,27.8) | 1 (Reference) | 1 (Reference) | 33.0 (17.3,46.1) | 1 (Reference) | 1 (Reference) | |||||
| Any G | 32 | 16 (50%) | 16 (50%) | 11.0 (7.5,14.8) | 1.22 (0.62,2.37) | 1.19 (0.59,2.42) | 24.5 (15.6,34.5) | 1.13 (0.63,2.04) | 1.97 (1.01,3.87) | |||||
| 0.74 | 0.19 | 0.48 | 0.002 | 0.008 | ||||||||||
| Any A | 49 | 26 (53%) | 23 (47%) | 12.4 (8.4,14.8) | 1 (Reference) | 1 (Reference) | 30.8 (22.3,46.1) | 1 (Reference) | 1 (Reference) | |||||
| G/G | 11 | 5 (45%) | 6 (55%) | 9.9 (2.7,18.2) | 1.62 (0.76,3.47) | 1.35 (0.59,3.11) | 16.6 (8.0,31.1) | 2.75 (1.34,5.66) | 2.86 (1.32,6.21) | |||||
| CCR5 rs1799988 | 0.94 | 0.78 | 0.72 | 0.041 | 0.009 | |||||||||
| C/C | 18 | 9 (50%) | 9 (50%) | 12.4 (6.8,19.4) | 1 (Reference) | 1 (Reference) | 35.3 (22.3,59.6) | 1 (Reference) | 1 (Reference) | |||||
| C/T2 | 31 | 17 (55%) | 14 (45%) | 12.9 (7.3,18.2) | 0.90 (0.41,1.98) | 0.85 (0.35,2.06) | 25.5 (16.0,34.5) | 2.04 (1.01,4.14) | 2.68 (1.27,5.62) | |||||
| T/T2 | 10 | 5 (50%) | 5 (50%) | |||||||||||
| KLF rs2241779 | 0.61 | 0.58 | 0.84 | 0.76 | 0.58 | |||||||||
| A/A | 29 | 17 (59%) | 12 (41%) | 12.9 (9.1,18.2) | 1 (Reference) | 1 (Reference) | 30.8 (20.5,44.6) | 1 (Reference) | 1 (Reference) | |||||
| A/C2 | 22 | 10 (45%) | 12 (55%) | 12.4 (6.3,21.4) | 1.19 (0.62,2.29) | 0.92 (0.43,1.98) | 25.5 (15.6,36.7) | 1.09 (0.61,1.97) | 0.83 (0.43,1.62) | |||||
| C/C2 | 9 | 4 (44%) | 5 (56%) | |||||||||||
| CCL3 rs1130371 | 0.21 | 0.45 | 0.91 | 0.59 | 0.95 | |||||||||
| G/G | 33 | 14 (42%) | 19 (58%) | 13.0 (7.3,21.4) | 1 (Reference) | 1 (Reference) | 29.2 (16.6,36.7) | 1 (Reference) | 1 (Reference) | |||||
| G/A2 | 25 | 16 (64%) | 9 (36%) | 11.0 (6.3,14.7) | 1.26 (0.65,2.47) | 0.95 (0.43,2.12) | 30.5 (17.3,44.6) | 0.85 (0.48,1.53) | 0.98 (0.51,1.88) | |||||
| A/A2 | 3 | 1 (33%) | 2 (67%) | |||||||||||
| CCL4 rs1634517 | 0.76 | 0.13 | 0.35 | 0.40 | 0.039 | |||||||||
| C/C | 31 | 15 (48%) | 16 (52%) | 14.4 (7.5,21.4) | 1 (Reference) | 1 (Reference) | 34.5 (20.5,46.1) | 1 (Reference) | 1 (Reference) | |||||
| C/A2 | 20 | 12 (60%) | 8 (40%) | 9.6 (6.2,14.7) | 1.60 (0.81,3.14) | 1.45 (0.66,3.18) | 25.5 (15.6,31.9) | 1.28 (0.71,2.33) | 2.11 (1.04,4.27) | |||||
| A/A2 | 7 | 4 (57%) | 3 (43%) | |||||||||||
Abbreviations: CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease, p Values <0.05 were shown in bold.
p Value was based on the Fisher’s exact test for response, log-rank test in the univariate analysis (†) and Wald test in the multivariable analysis within Cox regression model (‡) adjusted for sex, age, ECOG performance status, liver metastasis, number of metastases, primary tumor resection, and adjuvant chemotherapy.
Estimates did not reach yet.
Grouped together for estimate of HR.
Clinical outcome and SNPs tested in the control cohort receiving chemotherapy alone
Patients with the G/G variant in CCL5 rs2280789 had a significantly shorter OS compared to those with any A allele (16.6 vs. 30.8 months, HR 2.75, 95%CI, 1.34–5.66, p = 0.002); and patients with any T allele in CCR5 rs1799988 had a shorter OS compared to those with the C/C variant (25.5 vs. 35.3 months, HR 2.04, 95%CI, 1.01–4.14, p = 0.041) in univariate analysis. These effects remained in the multivariable model adjusted for sex, age, ECOG performance status, liver metastasis, number of metastases, primary tumor resection, and adjuvant chemotherapy (CCL5: HR 2.86, 95%CI; 1.32–6.21, adjusted p = 0.008; CCR5: HR 2.68, 95%CI: 1.27–5.62, adjusted p = 0.009). Concerning other CCR5 ligands, CCL4 rs1634516 A alleles were remarkably associated with shorter OS in the multivariable analysis (25.5 vs. 34.5 months; HR 2.11, 95%CI: 1.04–4.27, adjusted p =0.039)(Table 2).
Serum cytokine levels based on CCL5 and CCR5 SNPs in the exploratory cohort
The association between cytokine levels and SNPs is summarized in Table 3 and Supporting Information Table S4. VEGF-A levels were significantly higher in patients with any G allele in CCL5 rs2280789 than those with the A/A variant (mean 468.3 vs. 275.0 pg/ml, p = 0.024) at baseline. No significant differences were shown in other cytokines at baseline. Serum VEGF-A levels at baseline did not correlate with tumor response, PFS and OS (Supporting Information Table S5). In addition, serum Ang-2 levels were significantly lower in patients with any G allele in CCL5 rs2280789 than those with the A/A variant at day 56 (mean 1,678.3 vs. 2,225.0 pg/ml, p = 0.014) and a trend toward lower at day 14 (mean 1,779.8 vs. 2,134.3 pg/ml, p = 0.084) though there was no significant difference in change analyses. In change analysis, any G allele in CCL5 rs2280789 was significantly associated with a greater decrease of VEGF-A levels at day 14 (mean - 332.3 vs. −141.9 pg/ml, p = 0.028) and day 56 (mean - 228.7 vs. −110.5 pg/ml, p = 0.038) compared to the A/A variant (Figs. 1a and 1b). In addition, at day 56, Ang-1 and SDF-1 levels significantly decreased in patients with any A allele than the G/G variant (mean - 10,843.2 vs. −6,280.0 pg/ml, p = 0.008; mean - 165.4 vs. 102.0 pg/ml, p = 0.027) (Figs. 1e and 1f), while IL-8 significantly decreased in those with the G/G variant than any A allele (mean - 8.9 vs. −3.0 pg/ml, p = 0.049) (Fig. 1g).
Table 3.
Association between serum cytokine levels and SNPs in the exploratory cohort
| N | BL | D14 | D56 | Change (BL–D14) | Change (BL–D56) | |
|---|---|---|---|---|---|---|
| CCL5 rs2280789 | Mean serum VEGF-A levels (pg/ml) | |||||
| A/A | 21 | 275.0 | 133.1 | 173.9 | −141.9 | −110.5 |
| A/G | 17 | 492.0 | 139.6 | 196.4 | −352.4 | −295.6 |
| G/G | 10 | 427.9 | 129.7 | 150.9 | −298.2 | −277.0 |
| p Value1 | 0.099 | 0.833 | 0.169 | 0.112 | 0.141 | |
| A/A | 21 | 275.0 | 133.1 | 173.9 | −141.9 | −110.5 |
| Any G | 27 | 468.3 | 135.9 | 179.6 | −332.3 | −288.7 |
| p Value1 | 0.024 | 0.824 | 0.756 | 0.028 | 0.038 | |
| CCL5 rs2280789 | Mean serum Ang-1 levels (pg/ml) | |||||
| A/A | 21 | 34,900.0 | 28,209.5 | 24,990.0 | −6,690.5 | −10,240.0 |
| A/G | 17 | 35,900.0 | 28,905.9 | 24,347.1 | −6,994.1 | −11,552.9 |
| G/G | 10 | 33,180.0 | 26,720.0 | 26,900.0 | −6,460.0 | −6,280.0 |
| p Value1 | 0.793 | 0.845 | 0.781 | 0.970 | 0.017 | |
| Any A | 38 | 35,347.4 | 28,521.1 | 24,694.6 | −6,826.3 | −10,843.2 |
| G/G | 10 | 33,180.0 | 26,720.0 | 26,900.0 | −6,460.0 | −6,280.0 |
| p Value1 | 0.541 | 0.591 | 0.500 | 0.854 | 0.008 | |
| CCL5 rs2280789 | Mean serum Ang-2 levels (pg/ml) | |||||
| A/A | 21 | 2,558.6 | 2,134.3 | 2,225.0 | −424.3 | −335.5 |
| A/G | 17 | 2026.5 | 1,621.5 | 1,520.3 | −405.0 | −506.2 |
| G/G | 10 | 2,902.0 | 2049.0 | 1947.0 | −853.0 | −955.0 |
| p Value1 | 0.13 | 0.068 | 0.017 | 0.282 | 0.282 | |
| A/A | 21 | 2,558.6 | 2,134.3 | 2,225.0 | −424.3 | −335.5 |
| Any G | 27 | 2,359.7 | 1,779.8 | 1,678.3 | −570.9 | −672.4 |
| p Value1 | 0.538 | 0.084 | 0.014 | 0.518 | 0.258 | |
| CCL5 rs2280789 | Mean serum SDF-1 levels (pg/ml) | |||||
| A/A | 21 | 2,546.2 | 2,591.4 | 2,432.0 | 45.2 | −115.0 |
| A/G | 17 | 2,393.5 | 2,418.8 | 2,168.8 | 25.3 | −224.7 |
| G/G | 10 | 2,365.0 | 2,475.0 | 2,467.0 | 110.0 | 102.0 |
| p Value1 | 0.383 | 0.423 | 0.077 | 0.725 | 0.054 | |
| Any A | 38 | 2,477.9 | 2,514.2 | 2,311.1 | 36.3 | −165.4 |
| G/G | 10 | 2,365.0 | 2,475.0 | 2,467.0 | 110.0 | 102.0 |
| p Value1 | 0.437 | 0.790 | 0.284 | 0.440 | 0.027 | |
| CCL5 rs2280789 | Mean serum IL-8 levels (pg/ml) | |||||
| A/A | 21 | 14.2 | 10.2 | 9.3 | −3.9 | −4.3 |
| A/G | 17 | 10.9 | 10.3 | 9.4 | −0.6 | −1.5 |
| G/G | 10 | 17.4 | 10.8 | 8.5 | −6.6 | −8.9 |
| p Value1 | 0.111 | 0.979 | 0.914 | 0.111 | 0.083 | |
| Any A | 38 | 12.7 | 10.3 | 9.4 | −2.4 | −3.0 |
| G/G | 10 | 17.4 | 10.8 | 8.5 | −6.6 | −8.9 |
| p Value1 | 0.097 | 0.836 | 0.674 | 0.118 | 0.049 | |
| CCR5 rs1799988 | Mean serum VEGF-A levels (pg/ml) | |||||
| C/C | 11 | 266.0 | 135.6 | 180.2 | −130.4 | −85.8 |
| C/T | 27 | 430.4 | 131.7 | 180.8 | −298.8 | −262.8 |
| T/T | 10 | 387.0 | 141.9 | 164.1 | −245.1 | −222.9 |
| p Value1 | 0.366 | 0.821 | 0.758 | 0.355 | 0.318 | |
| C/C | 11 | 266.0 | 135.6 | 180.2 | −130.4 | −85.8 |
| Any T | 37 | 418.7 | 134,4 | 176.2 | −284.3 | −251.7 |
| p Value1 | 0.055 | 0.937 | 0.853 | 0.169 | 0.138 | |
| CCR5 rs1799988 | Mean serum PlGF levels (pg/ml) | |||||
| C/C | 11 | 16.4 | 25.4 | 33.4 | 9.1 | 17.0 |
| C/T | 27 | 17.5 | 24.1 | 29.3 | 6.7 | 11.8 |
| T/T | 10 | 17.3 | 21.4 | 27.8 | 4.1 | 10.5 |
| p Value1 | 0.565 | 0.295 | 0.280 | 0.045 | 0.071 | |
| C/C | 11 | 16.4 | 25.4 | 33.4 | 9.1 | 17.0 |
| Any T | 37 | 17.4 | 23.4 | 28.9 | 6.0 | 11.4 |
| p Value1 | 0.288 | 0.323 | 0.126 | 0.049 | 0.024 | |
Abbreviations: BL, baseline; D14, day 14; D56, day 56. p Values <0.05 were shown in bold.
P value was based on the Student’s unpaired t test and one-way analysis of variance (ANOVA) for the means of continuous measurements in each cytokines.
Figure 1.
Change of serum VEGF-A levels at day 14 and 56 based on CCL5 rs2280789 (a and b) and CCR5 rs1799988 (c and d) variant in patients receiving BEV-based chemotherapy. Change of serum cytokine levels based on CCL5 rs2280789 (e, f and g) or CCR5 rs1799988 (h and i) variant in patients receiving BEV-based chemotherapy: Ang-1 (e), SDF-1 (f) and IL-8 (g) at day 56; PlGF at day 14 (h) and day 56 (i).
Similarly, regarding CCR5 rs1799988, VEGF-A levels were higher in patients with any T allele than the C/C variant at baseline though not statistically significant (mean 418.7 vs. 266.0 pg/ml, p = 0.055). The change analysis in any T allele in CCR5 was compared to the change analysis in any G allele in CCL5. A numerical similar trend in change analysis for VEGF-A was observed though there was no statistical significance (Figs. 1c and 1d). In contrast, PlGF levels consistently increased in each variant especially in the C/C at day 14 and 56 than in any T allele (mean 9.1 vs. 6.0 pg/ml, p = 0.049; mean 17.0 vs. 11.4 pg/ml, p = 0.024) (Figs. 1h and 1i).
Clinical importance of BEV based on CCL5 and CCR5 SNPs
We compared the two cohorts by CCL5 and CCR5 SNPs based on our results mentioned above. Patients with any CCL5 rs2280789 G allele had longer PFS and OS when receiving FOLFOX plus BEV than FOLFOX (PFS: 19.8 vs. 11.0 months, HR: 0.44, 95%CI: 0.24–0.83, p = 0.004; OS: 41.8 vs. 24.5 months, HR: 0.50, 95%CI: 0.26–0.95, p = 0.024) (Figs. 2a and 2b). No significant difference was shown in patients with the A/A variant (Figs. 2c and 2d). Similarly, tumor response was marked in patients treated with FOLFOX plus BEV who carried any G allele, whereas there was no statistical significance (64.9% vs. 50.0%); in contrast, no difference was shown in the A/A variants (59.1% vs. 53.6%).
Figure 2.
Progression-free survival (PFS) and overall survival (OS) in the evaluation cohort receiving FOLFOX plus BEV (
FFBV) and control cohort receiving FOLFOX (
FF) according to any G or A/A CCL5 rs2280789 allele: PFS and OS in any G (a, b) and A/A (c, d).
Discussion
Our study shows the first evidence that SNPs of CCL5 or CCR5 genes are potential prognostic markers of first-line oxaliplatin-based treatment; and different VEGF-A status in each variant confers different prognosis and sensitivity to BEV in mCRC patients. We suggest that BEV-induced EVR within 2 months might improve outcomes, however the reason for higher VEGF-A levels at baseline in patients with any G allele than A/A in CCL5 rs2280789 or any T allele than C/C in CCR5 rs1799988 remains unclear. We thereby investigated the association between SNPs and other angiogenic factors.
CCR5 was originally studied as a receptor for human immunodeficiency virus (HIV) and its ligands CCL5, CCL3 and CCL4 confer protection from HIV infection.10 The CCL5/CCR5 axis has not been well investigated in cancer therapy, although its potential role in cancer progression was reported in some studies on varied tumor types including colorectal cancer.11,12 The CCL5/CCR5 axis is involved in the immune microenvironment and is exploited as a network for tumor progression with recruitment of specific immune cells and regulates either the host-derived anti-tumor immunity or tumor progression along with the concomitant chemo-kines.2,5,13 Genetic variants in the CCL5/CCR5-axis have been investigated in various cancers; however, the substantial role of SNPs of genes in the CCL5/CCR5 axis remains to be validated in terms of pro-oncogenic action or response to chemotherapy in cancer including colorectal cancer.14–19 In terms of genetic functionalities, transcriptional regulation of CCL5 is primarily regulated by CCL5 rs2280789 located in the promoter region, of which G alleles corresponded with a strong decrease in transcriptional activity of CCL5.20 Recent reports demonstrated that a low ability of VEGF-A production in any CCL5 rs2280789 G allele leads to a low risk of prostate cancer.21 However, there is limited evidence in respect of other SNPs including CCR5 rs1799988 tested in our study.
In our study, we focused on CCL5/VEGF-A signaling based on previous reports for refractory mCRC.4 Figure 3 illustrates the CCL5-VEGF-A signaling pathway in BEV or regorafenib treatment. According to the studies, low ability of VEGF-A production in patients with the CCL5 rs2280789 G/G variant was likely to achieve longer survival and severe hand-foot skin reaction in response to regorafenib treatment. Meanwhile, in our study, there was no significant difference in outcomes between genotypes in the cohorts receiving chemotherapy plus BEV, while G alleles in CCL5 rs2280789 and T alleles in CCR5 rs1799988 were associated with a poor outcome in the control cohort receiving chemotherapy alone. We speculate that this inconsistency derives from the large amount of VEGF-A at baseline followed by a remarkable EVR response to BEV in those genotypes, suggesting that patients harboring the G alleles or T alleles as biologically poor prognostic factors are likely to benefit from BEV. Taken together, the presence of quantitative VEGF-A reduction under low ability of VEGF-A production seems preferable in first-line treatment with BEV. An interesting study reported that neither baseline nor on-treatment VEGF-A levels in blood correlated with PFS; and moreover, the change in VEGF-A levels was revealed as independent of baseline levels during treatment with first-line BEV plus chemotherapy for mCRC though significant reduction of VEGF-A was observed at any point after 15 days until confirmed radiographic progression.22
Figure 3.
CCL5-VEGF-A signaling pathway in BEV treatment. RANTES, Regulated upon Activation Normal T cell Expressed and Secreted; MIP-1-alpha or -beta, macrophage inflammatory protein 1-alpha or beta; HRE, Hypoxia-response element; EPC, endothelial progenitor cell.
Thus we hereby underline that abundant circulating VEGF-A could be a positive predictive marker of BEV by testing genotypes, even though such an angiogenic environment is generally favorable for cancer cell progression and its predictive value remains unclear.23 Associations between SNPs and other angiogenic factors were analyzed to explore the reasons for increased VEGF-A levels in specific genotypes. Thereby, our exploratory analysis demonstrated that levels of other angiogenic factors at baseline did not correlate with outcome, while interestingly they could be changed under administration of BEV. In previous reports, blood concentrations of several angiogenic factors significantly increased prior to radiological disease progression24 or were higher in responders compared to non-responders at various points during BEV treatment.1 Such phenomena are consistent with our results: decrease in serum IL-8 levels is expected to achieve a better clinical outcome as shown in the CCL5 rs2280789 G/G variant, and a modest increase of PlGF marked in any CCR5 rs1799988 T allele than the C/C is also expected to provide a better outcome. Similarly, low Ang-2 levels in any G allele in CCL5 rs2280789 after day 56 might confer better clinical outcome. In a VEGF-A re-escalation in the early phase after the initial VEGF-A reduction was shown as a potential predictor of a poor response and early development of resistance to BEV.1 In contrast, activating other angiogenic cytokines was considered as alternative mechanisms of resistance to BEV.1,24
Taken together, high amounts of circulating VEGF-A might confer poor prognostic value of chemotherapy alone, and EVR followed by continuous low VEGF-A levels due to low ability of VEGF-A production may account for the positive predictive value of BEV in patients carrying minor alleles in the CCL5 and CCR5 SNPs. However, this does not guarantee resistance to BEV in patients with the other variants because there was no evidence of a detrimental effect of BEV on PFS and OS in the evaluation cohort. In addition, the true reason for high VEGF-A levels at baseline in specific genetic variants still remains unclear, suggesting VEGF-A as an independent prognostic or predictive marker, and alternative regulators of VEGF-A production might collaborate with the CCL5–VEGF-A signaling. Further pharmacogenetic studies are warranted to explore the role of SNPs in crosstalk between VEGF-A and other angiogenic cytokines such as PIGF for response and resistance to BEV.
Our study has limitations including: a retrospective study design with limited sample size; absence of correction for multiple testing; absence of validation of data from exploratory analysis for circulating cytokines; lack of preclinical data regarding the function of SNPs. In addition, regarding the patient characteristics, small number of adjuvant history, large number of KRAS exon2 status unknown and heterogenic tissue samples consisting primary tumor and metastatic site might affect the outcomes. Hence, our inferences from the results should be cautiously interpreted; however, the strength of our study is that all cohorts consisted of ethnically homogeneous populations and the potential role of genetic variants was demonstrated by after cytokine levels during treatment with BEV.
In conclusion, CCL5 and CCR5 impact the angiogenic environment and affect VEGF-A signaling. Furthermore, the genotypes may be surrogate markers of EVR and have differ-addition,ent ability of VEGF-A production, and could identifyspecific populations who may benefit from BEV-combined treatment in first-line treatment for mCRC.
Supplementary Material
What’s new?
Bevacizumab, the first anti-angiogenic agent targeting VEGF-A, has been widely used in several cancer types. However, an efficacy biomarker is still lacking. The CCL5/CCR5 axis modulates VEGF-A production via endothelial progenitor cells migration, and here the authors tested whether genetic polymorphisms could predict bevacizumab efficacy in patients with metastatic colorectal cancer. Gene polymorphisms in CCL5 and CCR5 genes were found to correlate with serum VEGF-A levels during bevacizumab treatment. Early VEGF-A reduction due to bevacizumab was differently observed between the genetic variants. Genetic variants in CCL5/CCR5 genes were demonstrated to vary sensitivity to bevacizumab due to different circulating VEGF-A levels.
Acknowledgements
This work was partly supported by the National Cancer Institute (grant number P30CA014089), the Gloria Borges WunderGlo Foundation-the Wunder Project, Dhont Family Foundation, San Pedro Peninsula Cancer Guild, Daniel Butler Research Fund and Call to Cure Fund. Mitsukuni Suenaga is the recipient of Takashi Tsuruo Memorial Fund. Martin D. Berger received a grant from the Swiss Cancer League (BIL KLS-3334–02-2014) and the Werner and Hedy Berger-Janser Foundation for cancer research. Yuji Miyamoto received a grant from Japan Society for the Promotion of Science (S2606).
Grant sponsor: Japan Society for the Promotion of Science; Grant number: S2606; Grant sponsor: Werner and Hedy Berger-Janser Foundation; Grant sponsor: Swiss Cancer League; Grant number: BIL KLS-3334–02-2014; Grant sponsor: Takashi Tsuruo Memorial Fund; Grant sponsor: Call to Cure Fund; Grant sponsor: Daniel Butler Research Fund; Grant sponsor: San Pedro Peninsula Cancer Guild; Grant sponsor: Dhont Family Foundation; Grant sponsor: Gloria Borges WunderGlo Foundation; Grant sponsor: National Cancer Institute; Grant number: P30CA014089
Footnotes
Conflict of interest: The authors have no conflicts of interest to declare in this work.
Additional Supporting Information may be found in the online version of this article.
References
- 1.Hayashi H, Arao T, Matsumoto K, et al. Biomarkers of reactive resistance and early disease progression during chemotherapy plus bevacizumab treatment for colorectal carcinoma. Oncotarget. 2014;5:2588–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Wang SW, Liu SC, Sun HL, et al. CCL5/CCR5 axis induces vascular endothelial growth factor mediated tumor angiogenesis in human osteosarcoma microenvironment. Carcinogenesis 2015;36:104–14. [DOI] [PubMed] [Google Scholar]
- 3.Suenaga M, Mashima T, Kawata N, et al. Serum VEGF-A and CCL5 levels as candidate biomarkers for efficacy and toxicity of regorafenib in patients with metastatic colorectal cancer. Oncotarget 2016;7: 34811–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Suenaga M, Schirripa M, Cao S, et al. Gene polymorphisms in the CCL5/CCR5 pathway as genetic biomarker for outcome and hand-foot skin reaction in metastatic colorectal cancer patients treated with regorafenib. Clin Colorectal Cancer 2018; 17:e395–e414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Song A, Nikolcheva T, Krensky AM. Transcriptional regulation of RANTES expression in T lymphocytes. Immunol Rev 2000;177:236–45. [DOI] [PubMed] [Google Scholar]
- 6.Ishida Y, Kimura A, Kuninaka Y, et al. Pivotal role of the CCL5/CCR5 interaction for recruitment of endothelial progenitor cells in mouse wound healing. J Clin Invest 2012;122: 711–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Song A, Patel A, Thamatrakoln K, et al. Functional domains and DNA-binding sequences of RFLAT-1/KLF13, a Krüppel-like transcription factor of activated T lymphocytes. J Biol Chem 2002;277:30055–65. [DOI] [PubMed] [Google Scholar]
- 8.Suenaga M, Mizunuma N, Matsusaka S, et al. Retrospective analysis on the efficacy of bevacizumab with FOLFOX as a first-line treatment in Japanese patients with metastatic colorectal cancer. Asia Pac J Clin Oncol 2014;10:322–9. [DOI] [PubMed] [Google Scholar]
- 9.Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivar Behav Res 2011; 46:399–424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Cocchi F, Devico AL, Garzino-Demo A, et al. Identification of RANTES, MIP-1 alpha, and MIP-1 beta as the major HIV-suppressive factors produced by CD8+ T cells. Science 1995;270: 1811–5. [DOI] [PubMed] [Google Scholar]
- 11.Barashi N, Weiss ID, Wald O, et al. Inflammation-induced hepatocellular carcinoma is dependent on CCR5 in mice. Hepatology 2013;58:1021–30. [DOI] [PubMed] [Google Scholar]
- 12.Musha H, Ohtani H, Mizoi T, et al. Selective infiltration of CCR5(+)CXCR3(+) T lymphocytes in human colorectal carcinoma. Int J Cancer 2005; 116:949–56. [DOI] [PubMed] [Google Scholar]
- 13.Balkwill F. Cancer and the chemokine network. Nat Rev Cancer 2004;4:540–50. [DOI] [PubMed] [Google Scholar]
- 14.Sáenz-López P, Carretero R, Cózar JM, et al. Genetic polymorphisms of RANTES, IL1-a, MCP-1 and TNF-A genes in patients with prostate cancer. BMC Cancer 2008;8:382. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Bodelon C, Malone KE, Johnson LG, et al. Common sequence variants in chemokine-related genes and risk of breast cancer in postmenopausal women. Int J Mol Epidemiol Genet 2013;4:218–27. [PMC free article] [PubMed] [Google Scholar]
- 16.Tahara T, Shibata T, Nakamura M, et al. RANTES promoter genotype and gastric cancer risk in a Japanese population. Anticancer Res 2009;29:4265–9. [PubMed] [Google Scholar]
- 17.Gawron AJ, Fought AJ, Lissowska J, et al. Polymorphisms in chemokine and receptor genes and gastric cancer risk and survival in a high risk polish population. Scand J Gastroenterol 2011;46:333–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Duell EJ, Casella DP, Burk RD, et al. Inflammation, genetic polymorphisms in proinflammatory genes TNF-A, RANTES, and CCR5, and risk of pancreatic adenocarcinoma. Cancer Epidemiol Biomarkers Prev 2006;15:726–31. [DOI] [PubMed] [Google Scholar]
- 19.Ying H, Wang J, Gao X. CCL5–403, CCR5–59029, and Delta32 polymorphisms and cancer risk: a meta-analysis based on 20,625 subjects. Tumour Biol 2014;35:5895–904. [DOI] [PubMed] [Google Scholar]
- 20.An P, Nelson GW, Wang L, et al. Modulating influence on HIV/AIDS by interacting RANTES gene variants. Proc Natl Acad Sci USA 2002;99: 10002–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kidd LR, Jones DZ, Rogers EN, et al. Chemokine ligand 5 (CCL5) and chemokine receptor (CCR5) genetic variants and prostate cancer risk among men of African descent: a case-control study. Hered Cancer Clin Pract 2012;10:16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Loupakis F, Cremolini C, Fioravanti A, et al. Pharmacodynamic and pharmacogenetic angiogenesis-related markers of first-line FOL-FOXIRI plus bevacizumab schedule in metastatic colorectal cancer. Br J Cancer 2011;104:1262–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Hegde PS, Jubb AM, Chen D, et al. Predictive impact of circulating vascular endothelial growth factor in four phase III trials evaluating bevacizumab. Clin Cancer Res 2013;19:929–37. [DOI] [PubMed] [Google Scholar]
- 24.Kopetz S, Hoff PM, Morris JS, et al. Phase II trial of infusional fluorouracil, irinotecan, and bevacizumab for metastatic colorectal cancer: efficacy and circulating angiogenic biomarkers associated with therapeutic resistance. J Clin Oncol 2010;28:453–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.