Prospective validation of VEGF and eNOS polymorphisms as predictors of first-line bevacizumab efficacy in patients with metastatic colorectal cancer

Bevacizumab (Bev) plus chemotherapy is a standard first-line treatment in metastatic colorectal cancer (mCRC), however to date no predictive factors of response have been identified. Results of our previous analysis on patients enrolled in a randomized prospective phase III multicenter study (ITACa study) showed a predictive value of Vascular Endothelial Growth Factor (VEGF) polymorphism (VEGF + 936), a 27-nucleotide variable number tandem repeat (VNTR) of the endothelial nitric oxide synthase (eNOS) gene and eNOS + 894 polymorphism. mCRC patients, treated with Bev plus chemotherapy, were included in this prospective validation trial. eNOS + 894G > T was analyzed by Real time PCR, while the eNOS VNTR and VEGF + 936C > T were determined by standard PCR and direct sequencing analysis. These polymorphisms were assessed in relation to progression-free survival (PFS), overall survival (OS) and objective response rate (ORR). These three polymorphisms were not predictive of PFS (p 0.91, 0.59 and 0.09, respectively), and OS (p 0.95, 0.32 and 0.46, respectively). Moreover, the haplotype analyses did not confirm what was found in our previous study; patients bearing a specific haplotype of eNOS had not significantly improved outcomes. This prospective study failed to validate the predictive impact of eNOS and VEGF polymorphisms for response to Bev plus first-line chemotherapy in mCRC patients.

VEGF represents the main pro-angiogenic factor, and acts binding to specific recep-tors present on the membrane of endothelial cells, VEGFR-1 and VEGFR-2.This led to the variation of the structure of the cytoskeleton of the endothelial cells, with the consequent modification of the cell motility.VEGF expression is related to blood vessel formation and metastasization in various cancers.
Several researchers assessed the role of VEGF Single Nucleotide Polymorphisms (SNPs) in relation to Bev responsiveness with contradictory findings [4][5][6][7] .Our team analyzed VEGF and endothelial nitric oxide synthase (eNOS) SNPs in relation to clinical out-come (PFS, OS and overall response rate, ORR) in mCRC patients undergoing Bev-based first-line chemotherapy in the phase III prospective multicenter randomized "Italian Trial in Advanced Colorectal Cancer (ITACa)" study 8 .This study demonstrated that VEGF 936C/T, eNOS + 894 G/T, and VNTR were linked with prognosis only in Bev plus first-line chemotherapy patients.Moreover, patients with a particular haplotype combination of the two eNOS polymorphisms (defined as eNOS Haplo1/Haplo1 and eNOS Haplo 2/Haplo2), in comparison to those with the other genotypes, demonstrated significantly longer PFS and OS, as well as a higher ORR 8 .
The results of the abovementioned work needed to be confirmed in an independent case series, thus the present trial was designed to validate the role of these polymorphisms and haplotypes.If confirmed, these data could represent a valid selection criterion for candidates for treatment with Bev.

Results
From January 2016 to October 2019, 182 patients were enrolled into this trial.Thirteen patients were excluded from the analysis due to eligibility criteria violation.Table 1 summarizes the patient characteristics.Fifty-five percent of patients were men and the median age was 69 years.ECOG PS was equal to 0 in 75% of cases, 1 in 23% of cases and equal to 2 in a minority (2%).As regards the localization of the primary tumor, 69 patients (42%) had a neoplasm originating from the left colon, 64 patients (39%) from the right colon and 32 patients (19%) from the rectum.Analyzing the site of metastases, 58% of patients had me-tastases to the liver, 37% to the lungs and 24% to the peritoneum.Grade 3 adenocarcinomas were the most represented (38%).Prior surgery for primary tumor and adjuvant therapy occurred in 74% and 30% of cases, respectively.Most patients (88%) received a first-line chemotherapy treatment containing intravenous fluoropyrimidine and oxaliplatin or irinotecan (specifically, 107 patients received FOLFOX and 41 FOLFIRI).As regards mutational analyses performed on tissue, KRAS, NRAS e BRAF mutations were detected in 58%, 7% and 13% of cases, respectively.Table 2 shows the distribution of the polymorphisms VEGF + 936, eNOS + 894G > T and eNOS VNTR in the study population.At a median follow-up of 52.6 months, there had been 161 (95%) progressions and 134 (79%) deaths in the cohort under analysis.Median PFS and OS were 12.2 months (95% CI 11.2-13.4)and 22.6 months (95% 18.6-29.1),respectively.During treatment, 83 patients (72%) reached a CR or PR as best overall response.

Association between polymorphisms and clinical outcomes.
We evaluated the correlation between PFS and VEGF + 936, eNOS + 894 and eNOS VNTR polymorphisms.The analysis carried out did not confirm what was found in the previous study, as none of the polymorphisms correlated significantly with PFS (Table 3, Supplementary Fig. S1).
With regard to OS, no statistically significant difference was observed between polymorphism genotypes (Table 3, Supplementary Fig. S1).
We also investigated eNOS + 894 and eNOS VNTR polymorphisms in relation to ORR.VEGF + 936 was not considered in this analysis due to the small frequencies of CT and TT genotypes.eNOS VNTR was the only polymorphism significantly associated with ORR in univariate analysis, but this result was not confirmed in multivariate analysis (Supplemetary Table S1).eNOS haplotype.In our previous analysis, two haplotypes based on eNOS VNTR 4a/b and eNOS + 894 G/T were described: eNOS Haplo 1 (4b-G), eNOS Haplo 2 (4b-T).
The analyses performed did not confirm what was found in the previous study, in particular patients bearing eNOS Haplo1/Haplo1 or the combination of eNOS Haplo1/Haplo1 and eNOS Haplo2/Haplo2 had not significantly improved outcomes in terms of PFS, OS and ORR (Table 4, Fig. 1, Supplementary Table S2).A significant effect on PFS was instead observed in the group eNOS Haplo2/Haplo2.In particular eNOS Haplo2/Haplo2 showed worse PFS than other geno-types (HR 2.25, 95% CI 1.26-4.03,p = 0.006).The result obtained was confirmed in multivariate analysis (adjusted-HR 2.83, 95% CI 1.33-5.99,p = 0.006) (Table 4 and Fig. 1).

Discussion
The present study investigated VEGF and eNOS polymorphisms in relation to clinical outcome in patients with mCRC treated with first line Bev-based chemotherapy.
The inhibition of angiogenesis, through blockade of the VEGF/VEGFR pathway, is an effective strategy in the treatment of mCRC.Positive results from randomised phase III clinical trials are available not only with Bev 9 , but also with the multikinase inhibitor regorafenib 10 and aflibercept, a fully human fusion protein that highly specifically and potently blocks all VEGF isoforms and placental growth factor (PlGF) 11 .From a clinical perspective, the relatively small absolute benefit provided by these new agents, as well as the availability of an increasing number of therapeutic options, make the identification of predictive biomarkers an essential requirement for optimizing the use of antiangiogenic agents.Unfortunately, studies in this context, characterized by different designs, have shown contradictory results.In fact, although several studies have reported significant results regarding VEGF polymorphisms, others have failed to confirm their predictive role 12   www.nature.com/scientificreports/For example, the retrospective study by Loupakis and colleagues hypothesized a role of the VEGF-1498 C/T polymorphism as a predictive biomarker of Bev efficacy, showing that patients with this genotype had a worse prognosis 13 .However, the same authors did not confirm this hypothesis in a subsequent prospective validation study 7 .In another retrospective study, VEGF-1498 C/T and -2578 A/C were correlated with outcome 6 .In other prospective studies, VEGF-1154 G/A and VEGF-634 G/C were significantly corre-lated with patient response to treatment 14 .
One of the most compelling reasons for the failure of predictive marker studies is the biological complexity of tumor angiogenesis.Data in the literature suggest multiple pathways in the growth of new tumor vessels, leading to study the simultaneous inhibi-tion of multiple angiogenic targets as a potentially effective strategy 15 .The role of different cell types in the development of such an intricate network of signals has also emerged and the contribution of both stromal cells and bone marrow-derived vascular progenitors recruited and stimulated by hypoxic conditions has been highlighted 16 .The relevance of the tumor microenvironment as a crucial player in the growth and stabiliza-tion of new vessels and the critical implication of the extracellular matrix in supporting angiogenesis are now well established, thus confirming the contribution of both host and tumor factors to the so called "angiogenic balance".
Mechanisms of action of anti-VEGF therapy remain unclear since blockade of circu-lating VEGF can affect not only tumors, but also stroma and endothelial cell proliferation and maturation.While the antiangiogenic properties of Bev were initially attributed to the inhibition of endothelial cell proliferation, several biological effects are now recognized, such as the inhibition of bone marrow-derived progenitors, the normalization of vessel structure, the vascular "constriction", the destruction of the cancer stem cell niche, the di-rect effect on cancer cells and the interaction with the host immune system.Because of these multiple mechanisms of action, the translation of in vitro and in vivo results into the human model is not immediate, and several difficulties make the development of effective preclinical models extremely complicated.
The ITACA study secondary analysis has shown that specific polymorphisms of eNOS were significantly correlated with patient survival and response to treatment.Nota-bly, patients with a specific combination of eNOS Haplo1/Haplo1 and eNOS Haplo2/Haplo2 had significantly longer PFS and OS and a higher ORR than those with the other genotypes 8 .The correlation was significant in the Bev-treated group and not in the chemotherapy-only group, highlighting the potential role of eNOS polymorphisms as pre-dictive biomarkers of Bev.Nitric oxide (NO), which is produced by the endothelium through the constitutively expressed gene known as eNOS, is essential for preserving the functional integrity of endothelial cells, controlling hemodynamics, and developing collateral circulation 17 .For the prevention of thrombotic and atherogenic processes, appropriate eNOS expression and activity, which results in enough NO generation, are crucial 18 .It has been demonstrated that VEGF inhibition causes a reduction in eNOS expression, which in turn results in a reduction in NO production 19 , and that this phenomenon is connected to the induction of hypertension, one of the most frequent dose-limiting www.nature.com/scientificreports/toxicities of VEGF inhibitors 20 .Intron 4 eNOS VNTR polymorphism plays a role in regulating eNOS expression by acting as an enhancer/repressor and by coding for a 27-nt small RNA, which appears to inhibit eNOS expression at the transcriptional level [21][22][23][24] .The higher the number of 27-nt repeats, the more 27nt sir-RNA is produced, inhibiting eNOS expression.However, the association between eNOS VNTR in intron 4 and eNOS expression is still a much-debated issue [25][26][27] .
The purpose of the current study was to validate our previous findings, which, if verified, might have served as a reliable criterion for choosing individuals for Bev therapy.
Unfortunately, the study did not confirm the earlier findings and demonstrated the failure of both VEGF and eNOS polymorphisms as a potential predictor of benefit from Bev.The validation of haplotypes predictive role was thus compromised by the negative results on eNOS polymorphisms.
The lack of correlation evidence in the study could be due to the reduced power caused by fewer patients' enrolled respect to the planned sample size.Although eNOS VNTR showed a significant association with ORR in univariate analysis the result was not confirmed in multivariate analysis probably because of the further reduction of sample to non-missing values.
The present study confirms the absolute importance of prospective validation as an essential step on the road of biomarkers towards clinical application.Furthermore, based on the complexity of the biology of tumor angiogenesis and the involvement of multiple actors in this process, it seems rather unlikely that a single germline SNP alone could explain the efficacy of Bev.The failure of the "candidate SNP strategy" certainly opens the way to new questions on the possibility of effectively exploiting the pharmacogenetic ap-proach to identify predictors of benefit from Bev.
This prospective study failed to validate the hypothesized predictive impact of eNOS and VEGF polymorphisms for response to Bev plus first-line chemotherapy for mCRC.The results of the previous exploratory analysis performed on patients enrolled in a ran-domized and prospective phase III multicenter study (ITACa study) were not confirmed.While on the one hand the reduction in the sample size might have played a role in the negative result, we believe that the most probable cause is biological.Furthermore, since eNOS is not the direct target of Bev, other factors could definitely be involved in the relationship between eNOS activity and Bev efficacy.Therefore the main weakness of our trial is to have hypothesized that a single polymorphism could predict such a complex event as the inhibition of angiogenesis.We believe that future directions in this research field must necessarily include more comprehensive approaches, also considering the complexity of the mechanism of angiogenesis and the mecha-nisms of action of anti-VEGF therapy.

Materials and methods
Patients and treatment.This is a prospective non-pharmacological trial aimed at validating VEGF and eNOS polymorphisms as predictors of efficacy of Bev plus first-line chemotherapy in patients with mCRC.The primary objective of the study was to assess the correlation of the selected polymorphisms with PFS.Secondary objectives were the correlations with OS and ORR.The Local Ethics Committee (Comitato Etico Area Vasta Romagna and IRST) approved the study and informed consent was obtained from all patients before blood samples were obtained for genotype testing.
This study involved 182 patients with mCRC treated with first-line chemotherapy plus Bev at the IRCCS Istituto Romagnolo per lo Studio dei Tumori (IRST) "Dino Amadori" in Meldola.Patients ≥ 18 years of age, with histologically confirmed mCRC, one or more unidimensional measurable lesions according to Response Evaluation Criteria in Solid Tumours (RECIST) 1.0, Eastern Cooperative Oncology Group (ECOG) performance status ≤ 2 (≤ 1 if age ≥ 70 years), and an estimated life expectancy of at least 12 weeks were included.Patients were required to have adequate hematological, hepatic and renal func-tion.Prior adjuvant chemotherapy for CRC or neoadjuvant/adjuvant chemo-radiotherapy for rectal cancer were permitted, if completed at least 6 months before recurrence; conversely, prior chemotherapy or immunotherapy for advanced or metastatic disease were not permitted.Patients received one of the following chemotherapy regimens in combina-tion with Bev: FOL-FOX4, CAPOX, FOLFIRI, CAPIRI.
Treatment was to continue until disease progression (PD), withdrawal of consent, or unacceptable toxicity, whichever came first.If a patient became eligible for curative resection of metastatic disease, Bev was discontinued at least 6-8 weeks prior to the scheduled surgery date.After surgery, the choice of treatment was at the physician's discretion, and patients could resume treatment at least 28 days after surgery or complete wound healing, until PD.We analyzed KRAS gene (codons 12, 13, 59, 61, 117, 146), BRAF gene (codons 594, 600, 601) and NRAS gene (codons 12, 13, 18, 59, 61, 117, 146) as for clinical practice.
The analysis of the selected polymorphisms was performed on peripheral blood samples.Genomic DNA was extracted from 200 μl of whole blood using the QIAamp DNA Mini Kit (Qiagen SPA, Milan, Italy) and quantified with the Nanodrop instrument (Celbio Spa, Milan, Italy).
eNOS + 894 were analyzed by Real time PCR (7500 Applied Biosystems) with TaqMan SNP Genotyping assays (Assay ID C_3219460_20) (Applied Biosystems, Foster City, CA, USA) and starting from 10 ng of DNA.The eNOS VNTR and VEGF + 936 polymorphisms were instead determined by standard PCR and direct sequencing analysis on the ABI 3130 sequencer (Applied Biosystems).PCRs were performed starting from 50 ng of

Table 1 .
Patient characteristics.ECOG PS, Eastern Cooperative Oncology Group Performance Status.

Table 3 .
PFS and OS in relation to polymorphisms.