Influence of NOS3 rs2070744 genotypes on hepatocellular carcinoma patients treated with lenvatinib

We investigated whether or not nitric oxide synthase 3 (NOS3) rs2070744 genotypes can affect the response for lenvatinib treatment in patients with hepatocellular carcinoma (HCC). We evaluated the relation of the NOS3 rs2070744 genotypes to the tumor response, progression-free survival (PFS), and overall survival (OS) as the response for lenvatinib. We also examined the association between fibroblast growth factor receptor (FGFR) gene polymorphisms, a potential feature of lenvatinib, and the response. There were no significant differences between the studies for either PFS or OS, even though patients with the TT genotype had a longer mean PFS (hazard ratio [HR] 0.60; p = 0.069) and mean OS (HR 0.46; p = 0.075) than those with the TC/CC genotypes. However, patients with a single-nucleotide polymorphism (SNP) combination pattern of the NOS3 rs2070744 TC/CC and FGFR4 rs351855 CT/TT genotypes had a significantly shorter mean PFS (HR 2.56; p = 0.006) and mean OS (HR 3.36; p = 0.013) than those with the other genotypes. The NOS3 rs2070744 genotypes did not influence the clinical response. However, the SNP combination pattern of the NOS3 rs2070744 and FGFR4 rs351855 genotypes may be helpful as treatment effect predictors and prognostic factors for HCC patients treated with lenvatinib.

Correlation between the NOS3 rs2070744 genotypes and clinical characteristics. The demographic characteristics of patients are summarized in Table 1. The frequencies of TT, TC, and CC of the NOS3 rs2070744 genotypes were 76%, 23%, and 1%, respectively. The number of patients with the TT and TC/CC genotypes was 76 and 24, respectively. Comparing the clinical characteristics of patients with the TT genotype with those of patients with the TC/CC genotypes, there were no significant differences between the groups.
Effect of NOS3 rs2070744 on PFS and OS. Figure 1 and Table 2 show the effect of the NOS3 rs2070744 genotypes on clinical outcomes. The mean progression-free survival (PFS) of all enrolled patients was 10.0 (95% CI 8.5-11.5) months. The mean PFS of the NOS3 rs2070744 TC/CC genotypes and the TT genotype was 7.5 (95% CI 4.8-10.2) months and 10.6 (95% CI 8.9-12.4) months, respectively. There were no significant differences between the PFS of the two groups (HR 0.60; 95% CI 0. 34 Table 3 shows the effect of the FGFR gene polymorphisms on clinical outcomes. FGFR2 (rs2912791, rs2981429, rs2981582) and FGFR4 (rs351855) showed no significant differences in tumor response, PFS, or OS.

Effects of the FGFR gene polymorphisms on clinical response.
The SNP combination pattern and clinical outcomes. Figure 2 and Table 3 show the effect of FGFR alone and in combination with the NOS3 polymorphism on clinical outcomes. According to the effects of an SNP combination pattern of the NOS3 rs2070744 and FGFR4 rs351855 genotypes, patients with the NOS3 rs2070744 TC/CC and FGFR4 rs351855 CT/TT genotypes had a significantly shorter mean PFS of 6.0 (95% CI 2.7-9.3) months than did those with the other genotypes who had a mean PFS of 10.6 (95% CI 9.0-12.3) months (HR 2.56; 95% CI 1. 28-5.09; p = 0.006). Similarly, patients with the TC/CC and CT/TT genotypes had a significantly shorter mean OS of 12.0 (95% CI 8.1-15.8) months than did those with the other genotypes who had a mean OS of 22.0 (95% CI 20.1-24.0) months (HR 3.36; 95% CI 1.21-9.29; p = 0.013).
Adverse event assessment. The most common adverse events among patients who received lenvatinib were anorexia (59%), fatigue (55%), palmar-plantar erythrodysesthesia (47%), abdominal pain (41%), and diarrhea (41%) (see Supplementary Table S2  .0] months, p = 0.611); however, the incidence rate of adverse events was similar. Table 5 shows the impact on clinical outcomes of adverse events with an incidence rate of 30% or more. The mean PFS of the patients who developed grades 2 or 3 hypertension (n = 20) was significantly longer than that for those who developed grade 1 hypertension (n = 15) (

Discussion
To the best of our knowledge, this is the first study to report a correlation between the NOS3 rs2070744 genotypes and the clinical response to lenvatinib in patients with unresectable HCC. In a study investigating the utility of NOS3 SNPs, the NOS3 rs2070744 TT genotype was strongly associated with worse prognoses in patients with advanced HCC treated with sorafenib 12 . However, the NOS3 rs2070744 TC/CC genotypes tended to correspond to a worse prognosis after the administration of lenvatinib, although there were no statistically significant differences. That is, there was a difference in the responsiveness of NOS3 polymorphisms to HCC between the administration of lenvatinib and sorafenib. This can be attributed to the fact that lenvatinib is a multiple receptor tyrosine kinase inhibitor targeting not only VEGFR but also FGFR.
HCC is a hypervascularized tumor. Tumor angiogenesis can be effectively suppressed by simultaneously suppressing both VEGFR and FGFR signals. Lenvatinib inhibits angiogenic factor-mediated pathways, including VEGF and FGF, resulting in a subsequent further reduction in NO production. Reduced production of NO, a vasodilator, induces contraction of hepatic stellate cells 13 . Moreover, it increases intrahepatic vascular resistance Table 2. Effect of the NOS3 rs2070744 on tumor response, progression-free survival, and overall survival. CI confidence interval, DCR disease control rate (objective response plus stable disease), HR hazard ratio, NOS3 nitric oxide synthase 3, OR odds ratio, ORR objective response rate (complete and partial response).  www.nature.com/scientificreports/  www.nature.com/scientificreports/  www.nature.com/scientificreports/ and decreases intrahepatic blood flow as well. This infers that the portal hemodynamic effects of lenvatinib are different from those of sorafenib. Lenvatinib significantly aggravates the Congestion index (− 23.1% vs. + 16.2%), which reflects portal hemodynamics, and may exacerbate portal hypertension in patients with advanced HCC 14,15 .
In the present study, we also revealed that an SNP combination pattern of the NOS3 rs2070744 and FGFR4 rs351855 genotypes was associated with PFS and OS. In another study, it was suggested that the FGFR4 Gly388Arg polymorphism (rs351855) resulted in the transmembrane domain missense mutation and may affect protein expression 16 . Mutations in the transmembrane domain of the tyrosine kinase receptor can alter the pathological phenotype or signal transduction of FGFR4 17 . Because FGFR4 has the highest expression of transcripts in the liver, FGFR4 rs351855 may regulate HCC signaling [18][19][20] . Moreover, the FGFR4 rs351855 CT/TT genotypes sustained receptor activation and improved stability of the receptor more than did the CC genotype. This suggested influence on the response to chemotherapy in HCC patients treated with tyrosine kinase inhibitors [21][22][23][24] .
The NOS3 rs2070744 TC/CC genotypes may reduce the promoter activity of the NOS3 gene compared to the TT genotype, resulting in reduced NO production [25][26][27] . As is the case with the portal hemodynamic study 14,15 , the effects of NO with the NOS3 rs2070744 when treated with lenvatinib may differ from the results of a previous study 12 . Therefore, we presumed that the NOS3 rs2070744 and FGFR4 rs351855 genotypes could be useful predictors of the clinical response to lenvatinib treatment. The results revealed that the SNP combination pattern of the NOS3 rs2070744 and FGFR4 rs351855 genotypes was helpful as treatment effect predictors and prognostic factors for HCC patients treated with lenvatinib. However, there was no significant difference in the objective response rate. The objective response rate can not always be used to directly evaluate the clinical benefit nor can it be comprehensively used to measure drug efficacy. More research is needed to validate whether or not this SNP combination pattern can be a potential biomarker for chemotherapy.
There have been cases reported correlating patients' clinical response to treatment with drug-related adverse events [28][29][30] . In this study, patients who developed grades 2 or 3 hypertension during lenvatinib treatment had significantly longer PFS than those who developed grade 1 hypertension. Moreover, palmar-plantar erythrodysesthesia and diarrhea of grades 2 or 3 showed a prolonged PFS compared with grade 1. Previous studies have shown that patients who experienced hypertension and hand-foot skin reaction during lenvatinib treatment had significantly better outcomes than did those who did not develop these adverse events 31,32 . Our findings indicate that hypertension can be a positive predictive marker of prognosis in patients receiving lenvatinib treatment. VEGFR inhibitors such as sorafenib and lenvatinib have been reported to cause endothelial dysfunction associated with reduced NO production and activation of endothelin-1, a potent vasoconstrictor [33][34][35][36][37] . However, it is unclear why hypertension develops and why it is a marker of improved prognosis. On the other hand, grades 2 or 3 of anorexia and/or fatigue were poor prognostic factors for survival, although the differences were not significant. Anorexia was reported to be one of the common adverse events and was associated with time to discontinuation of lenvatinib treatment 32 . Anorexia and fatigue may be challenging conditions to alleviate and may affect the continuation of treatment and quality of daily life. The impact of genetic polymorphisms on the prevalence and severity of adverse events should be assessed, and the mechanisms underlying the occurrence of adverse events should be elucidated.
In  40,41 . Therefore, patients in this study with intermediate BCLC stage B and thus less likely to benefit from TACE were also treated with lenvatinib from the initial step at the discretion of their physicians. This study can help provide evidence for the importance of lenvatinib as a first-line treatment in patients with intermediate-stage HCC, who are very difficult to manage with TACE alone.
There were three limitations to this study. First, the selection bias for enrolled patients in this study cohort was undeniable because genomic DNA could not be obtained from all the patients receiving lenvatinib. Some patients, who had a short survival after lenvatinib treatment, could not be enrolled in this study. Therefore, individuals who had an inadequate response to lenvatinib were less likely to participate in the study. The median survival time of all the enrolled patients in this study was longer than that in the REFLECT trial with lenvatinib vs. sorafenib as a first-line treatment in 2018 5 . Moreover, the selection bias was correlated with the outcome. The multivariate analysis showed that an SNP combination pattern of the NOS3 rs2070744 and FGFR4 rs351855 genotypes was an independent factor of lenvatinib response, and a previous report showed that the factors significantly associated with poor OS included skeletal muscle mass index and ALBI grade 42 . Second, there was no control arm in the present study. Therefore, a clear distinction could not be made between the prognostic role of the NOS3 rs2070744 genotypes in patients treated with lenvatinib. Finally, the expression levels of NOS, which are crucial for NO synthesis, were not evaluated. Therefore, NOS expression and NO levels warrant further study and elucidation in patients with HCC.
In conclusion, the NOS3 rs2070744 genotypes did not influence the clinical response. However, we found that the SNP combination pattern of the NOS3 rs2070744 and FGFR4 rs351855 genotypes may be helpful as treatment effect predictors and prognostic factors for HCC patients treated with lenvatinib. Multicenter and prospective studies enrolling an appropriate number of patients will be required to confirm further the correlation of genetic polymorphisms with clinical outcomes and adverse events of lenvatinib.

Methods
Ethics. The Institutional Review Board (Ethics Committee) approved this study at all study the institutes involved, and the protocol of this study conforms to the provisions of the Declaration of Helsinki. This study is registered in the University Hospital Medical Information Network (UMIN) Clinical Trials Registry (UMIN000036625). The study was approved by the research ethics board at each participating center, and written informed consent was obtained from all participants. Patients were informed about the risk and benefits of the research and then chose to give their consent to participate.
Patients. This retrospective multicenter study was conducted in four medical institutions: Kitasato University Hospital, Yokohama City University Medical Center, Shonan Kamakura General Hospital, and Shonan Fujisawa Tokushukai Hospital in Japan.
Eligible patients (age, ≥ 20 years) with advanced or unresectable HCC caused by chronic liver disease, the Eastern Cooperative Oncology Group-performance status of 1 or less, and adequate organ function were included in this study. We diagnosed HCC by blood tests and imaging modalities and histopathological analyses of biopsy specimens for patients with atypical imaging findings. Critical exclusion criteria included no genomic DNA extracted from whole blood, treatment of cases discontinued within 14 days, severe liver failure (Child-Pugh class C), refractory ascites, end-stage kidney disease, or malignancies other than HCC.
Treatment. Lenvatinib was recommended for patients with advanced BCLC stage C such as portal vein invasion, metastases to distant organs, and for those who were unsuitable and/or refractory to transcatheter arterial therapies. Patients with intermediate BCLC stage B also received systemic lenvatinib therapy at the discretion of their physicians. For patients without risk factors, the recommended dose of lenvatinib was based on body weight: 12 mg/day for patients with a body weight of 60 kg or more and 8 mg/day for patients with a body weight of less than 60 kg. Dose reduction from the initial dose was approved according to the attending physician's discretion for patients who had risk factors. After the administration of lenvatinib, the dose was changed based on the clinical response and/or the occurrence of severe adverse events. Treatment continued with progressive disease (PD) or until the occurrence of unmanageable adverse events.
Evaluation for treatment response. We defined tumor response according to the modified Response Evaluation Criteria in Solid Tumors (mRECIST) 43 . Tumor response was visualized by a radiologist using either dynamic computed tomography or dynamic magnetic resonance imaging techniques and was assessed by the attending physician. The target lesion in the mRECIST assessment was defined as the contrast-enhanced portion of the lesion in the arterial phase. For the mRECIST assessment, a complete response (CR) was defined as the disappearance of the target lesion. Partial response (PR) was defined as a reduction of at least 30% of the target lesion size compared to the baseline diameter. PD was defined as an increase of at least 20% of the target lesion size. Stable disease (SD) was defined as any case that does not qualify for either PR or PD.
Objective response was defined as the proportion of patients with the best overall response of confirmed CR or PR. Disease control was defined as the best tumor response of CR, PR, or SD. Patients with an unknown response because of missing data could not be evaluated and were therefore excluded from our analyses. We also recorded the best variation (maximal decrease or minimal increase) in the sum of the most significant lesion dimensions for each patient.
End-point measurement. To investigate whether or not the NOS3 rs2070744 genotypes can affect the response for lenvatinib treatment in patients with HCC, we examined the influence of the genotypes on the tumor response, PFS, OS, and adverse events. Furthermore, we investigated the impact of the FGFR gene on the response in HCC patients treated with lenvatinib. PFS and OS defined as the time from the first dose of lenvatinib to progression or death (for PFS) or death alone (for OS) according to mRECIST. Adverse event assessments were done according to the Common Terminology Criteria for Adverse Events version 4.0 definitions. Genomic DNA isolation. Blood samples from the follow-up patients were exclusively collected in EDTA (ethylenediaminetetraacetic acid) tubes. Genomic DNA was extracted from whole blood using an automated MagNA Pure Compact Instrument (Roche Diagnostics GmbH, Mannheim, Germany) with the MagNA Pure Compact Nucleic Acid Isolation Kit I (Roche) following the manufacturer's instructions.
TaqMan SNP genotyping assays. Genotypes for the NOS3 rs2070744 were determined using the TaqMan SNP Genotyping Assay kits (Applied Biosystems, Foster City, CA, USA). Patients were classified as heterozygotic or homozygotic for each allele. They were categorized as the TT genotype (wild type homozygous mutant), the TC genotype (heterozygous mutant), and the CC genotype (homozygous mutant). As SNPs for the FGFR gene, FGFR2 (rs2912791, rs2981429, rs2981582) and FGFR4 (rs351855) were also determined. Chromosomes, positions, and biological effects of the SNPs studied can be found in the Supplementary Table S3 online.

Scientific Reports
| (2020) 10:17054 | https://doi.org/10.1038/s41598-020-73930-3 www.nature.com/scientificreports/ Statistical analysis. We estimated that this study would require a total of 100 patients by considering loss to follow-up. Variables were examined using Fisher's exact test or the Chi-square test. The differences between the PFS and OS in the two groups were analyzed by the Kaplan-Meier method and the Log-rank test. Variables identified using the univariate analysis to be significantly associated with PFS and OS were entered into the Cox proportional hazards regression model for multivariate analysis. Univariate and multivariate analyses with logistic regression models were used to calculate the OR at 95% CI to assess the correlation between the response and the NOS3 rs2070744 genotypes for lenvatinib treatment. A two-sided p value of less than 0.05 was considered to indicate statistical significance. All analyses were performed using SPSS version 24 Consent to participate. The study was conducted following the Declaration of Helsinki and after obtaining approval of the institutional research ethics committee and written informed consent obtained from all participants.