Introduction
Neuroblastoma (NB) is a solid embryonal tumor of the autonomic nervous system, derived from primitive sympathetic neurons. In >50% of all cases of NB, tumors arise in the adrenal medulla, while the rest originate in the paraspinal sympathetic ganglia, with a highly variable clinical presentation (1). Between 1975 and 2011, NB occurred in 1 in 7,000 live births worldwide and accounted for 7–10% of all childhood cancers (2). The annual age-standardized incidence rate (ASR) of NB in the Children Oncology Group between 1986 and 2001 was estimated to be 10.53 cases per million individuals in North American populations (3). In Lebanon, the National Cancer Registry (NCR) reports NB as a very rare malignancy, with an ASR of 2 cases per million individuals in 2007 (4).
According to the Surveillance Epidemiology and End Results (SEER) databases, the clinical outcome for patients with NB has improved in the last 40 years (5). However, this improvement is primarily attributed to an increase in cure rates among patients with the benign form of the disease. The cure rates among children with malignant high-risk tumors have shown only modest improvement, despite marked escalations in the intensity of therapy provided (6). Despite recent therapeutic advances, no salvage treatment modalities exist to date, and high-risk NB is still characterized by a low survival rate, with 50–60% of patients showing a relapse (7).
A number of previous studies indicated that genetic factors may be involved in the clinical outcome and the response to treatment of NB patients (8). Tumor-derived genomic information has been used since the 1980s to predict prognosis, particularly since the MYCN oncogene was first discovered to be the target of high-level amplifications at chromosome band 2p24. Due to the influence of MYCN amplification on the clinical outcome of NB, it is currently used as a standard biomarker for treatment stratification (9). In 1985, NB patients with an MYCN gene amplification of 1, 3 and ≥10 gene copies were estimated to have an 18-month survival rate of 70, 30 and 5%, respectively (10). In addition to MYCN amplification, NB stage III or IV, abnormal chromosomal deletion/translocation and an older age at diagnosis have all been identified as being associated with poor prognosis (11).
Current treatments for patients with high-risk NB include induction of remission, consolidation of the remission, and eradication of the minimal residual disease. The most commonly used induction chemotherapy includes dose-intensive cycles of cisplatin and etoposide, alternated with vincristine, doxorubicin, cyclophosphamide or ifosfamide and topotecan (12,13). A number of drug-metabolizing enzymes (DMEs) appear to be involved in the activation or inactivation of these drugs, including cytochrome P450 (CYP), N-acetyltransferase (NAT) and glutathione S-transferase (GST) (14). Therefore, individuals possessing a modified ability to metabolize these drugs may have an increased risk of relapse or mortality. In a large cohort of Australian individuals with NB studied over 15 years (n=209), carriers of the NAT1*11 allele variant were significantly less likely to relapse or succumb to the disease, while children who were GSTM1 null were more likely to relapse or succumb to the disease, after adjusting for MYCN amplification, stage of the disease and age at diagnosis [hazard ratio (HR), 1.6, P=0.04] (15). The phase I oxidation enzymes CYP3A4 and CYP3A5 were also found to metabolize a number of the drugs used in NB treatment (16–22) (Table I). Single nucleotide polymorphisms (SNPs) in these CYP genes may alter their enzymatic activity, hence affecting clinical outcome. However, despite these observations, there are no studies on the role of cytochrome P450 in NB clinical outcome to date. In the present study, the potential association between the genotype and expression levels of CYP3A4 and CYP3A5 and mortality were examined in a group of Lebanese patients with NB.
Materials and methods
Study design and participants
The current study employed a retrospective cohort design. Three major medical centers in the Lebanese capital city of Beirut participated. Specimens from children with histologically confirmed stage III or IV NB, diagnosed between 1993 and 2012, were obtained from the medical archives of Saint George Hospital University Medical Center, Hotel Dieu De France Hospital and the University Medical Center - Rizk Hospital (all Beirut, Lebanon) and provided to the current study anonymously. Patients excluded from the study were those with stage I or II NB (n=2) and those with missing archival tumors (n=9). Of the 38 patients identified, 27 were included in the study.
Approval was obtained from the Institutional Review Board of the University of Balamand prior to conducting the study. A review of the medical records of all patients was performed anonymously to obtain the following information: Age at diagnosis, gender, NB stage, date of diagnosis, relapse state, date of mortality and administered chemotherapeutic drugs.
Nucleic acid extraction
Sections (10 μm thick) of identified paraffin-embedded NB tumors were obtained. These were prepared by a trained pathologist, avoiding necrotic areas to ensure effective nucleic acid extraction. Sections were subjected to deparaffinization by xylene and protein digestion by proteinase K (Qiagen, Valencia, CA, USA). Tumor DNA was extracted from 50% of the sections using a QIAamp DNA FFPE Tissue kit (Qiagen), while RNA was extracted from the remaining 50% of sections using an RNeasy FFPE kit (Qiagen), both according to the manufacturer’s instructions.
CYPs genotyping analysis
CYP3A4*1B and CYP3A4*1A alleles were detected by polymerase chain reaction and restriction fragment length polymorphism (PCR-RFLP) using 100–150 ng of extracted DNA and the appropriate primers (Table II) at an annealing temperature of 55°C as previously described (23). The resulting 334-bp PCR product was digested by the PstI restriction enzyme (Thermo Scientific, Waltham, MA, USA) overnight at 37°C. CYP3A5*1 and CYP3A5*3 alleles were detected using 100–150 ng of tumor DNA and the appropriate primers (Table II) at an annealing temperature of 55°C as previously described (24). The resulting 293-bp PCR product was digested by the Ssp1 restriction enzyme (Thermo Scientific) overnight at 37°C. Digested fragments were then analyzed using ethidium bromide staining and agarose gel electrophoresis, and examined under UV light.
MYCN amplification
MYCN amplification was assessed for each NB tumor sample using extracted DNA in a Real-Time PCR TaqMan Detection assay (Applied Biosystems, Foster City, CA, USA). The degree of MYCN amplification was derived from the ratio of the number of copies of the MYCN oncogene to the number of copies of the reference gene β-actin (MYCN:β-actin ratio). The 50-μl PCR mixture contained 10 mmol/l Tris-HCl (pH 8.3), 50 mmol/l KCl, 10 mmol/l EDTA, 60 nmol/l passive reference dye ROX, 3.5 mmol MgCl2, 0.2 mmol/l of each dNTP, 300 nmol/l of each primer, 200 nmol/l of each probe, 0.5 U of AmpliTaq Gold, and 1 U of AmpErase UNG (Applied Biosystems) (Table II). The two genes were amplified as follows: 120 sec at 95°C followed by 45 cycles of: 30 sec at 95°C, 30 sec at 60°C, and 30 sec at 72°C.
The concentrations of MYCN and β-actin were estimated in the same assay via extrapolation on external calibration curves. The threshold cycle value (CT) for each sample was used to estimate the target concentration on the curves. Calibration curves were constructed with reference DNA extracted from the peripheral blood of two healthy male donors (aged >50) with no chronic diseases. The DNA concentration was determined by spectrophotometric measurement (Nanodrop 1000; Thermo Scientific) and expressed as number of molecules per microliter as previously described (25). The DNA was serially diluted to plot the calibration curves and the degree of amplification of each sample was derived from the MYCN:β-actin ratio. Only samples in which the MYCN:β-actin ratio was >2.0 were considered amplified. In each assay, positive controls and non-amplified negative controls were used.
cDNA synthesis and quantitative PCR (qPCR) for CYPs
cDNA synthesis was performed by reverse transcription of extracted tumor RNA using a RevertAid First Strand cDNA Synthesis kit (Thermo Scientific). Briefly, the reaction consisted of RNA, hexamer primers (provided in the RevertAid kit), reaction buffer, RNase inhibitor, dNTP mix and reverse transcriptase which were incubated for 5 min at 25°C and then the reaction was terminated by heating at 70°C for 5 min. qPCR analysis was performed using a TaqMan system (Applied Biosystems) according to the manufacturer’s instructions. β-actin was used as the reference gene. The 50-μl PCR mixture consisted of 20 ng synthesized cDNA, primers for CYP3A4, CYP3A5 and β-actin (600 nmol/l each), 250 nmol/l TaqMan probes and PCR Master mix (Applied Biosystems) (Table II). Amplification and detection were performed with the MxPro qPCR Agilent system (Agilent Technologies, Santa Clara, CA, USA) using the following PCR reaction profile: 95°C for 10 min followed by 40 cycles of 95°C for 20 sec and 62°C for 1 min. In every assay, liver DNA extracted from paraffin-embedded healthy liver tissue, archived at St George Hospital University Medical Center (Beirut, Paris) and a non-DNA template were used as positive and negative controls, respectively.
CYP mRNA expression calculations
Gene expression was calculated by absolute quantification as previously described (26). Standard curves for the target genes CYP3A4 and CYP3A5, as well as the reference gene β-actin, were constructed by serial dilutions of a synthesized cDNA obtained from archival liver tissue of a healthy donor. The threshold cycle values obtained for each gene from the tumor samples were extrapolated on corresponding constructed standard curves to obtain the number of copies per gene per sample. CYP3A4 and CYP3A5 expression levels were then calculated as a ratio of the number of copies of the target gene to the number of copies of the reference gene obtained for each sample.
Statistical analysis
Variant genotypes were grouped for statistical analyses based on previous studies of enzymatic function or phenotypic consequence (27–30). Homozygote carriers of the CYP3A5*3/*3 mutant genotype were compared with CYP3A5*1/*3 heterozygotes or wild-type CYP3A5*1/*1 homozygotes. Similarly, homozygote and heterozygote carriers of CYP3A4*1B mutant allele (CYP3A4*1B/1B or CYP3A4*1A/1B) were compared with homozygote wild-type CYP3A4*1A/1A carriers. CYP3A4 and CYP3A5 gene expression levels were analyzed as dichotomous variables (above or below the median expression level). Overall Survival (OS) is a standard measurement of survival defined as the duration from the time of diagnosis of disease to either death or date of last contact (31). A Kaplan-Meier survival analysis was used to estimate the OS in the total group of patients, and the MYCN amplification subgroups. The difference in survival between the MYCN amplified and the MYCN non-amplified groups was compared using the Log rank test (25). A Cox proportional hazard regression model was used to assess associations between mortality and CYP3A4 and CYP3A5 polymorphisms and mRNA expression after adjusting for MYCN amplification. HRs and 95% confidence intervals (CIs) were calculated. Statistical analyses were performed using the Statistical Package for Social Sciences, version 18.0.0 (SPSS, Inc., Chicago, IL, USA).
Discussion
In the present study, the associations between survival time and the expression levels or genetic polymorphisms of CYP3A4 and CYP3A5, as well as MYCN amplification, were analyzed in Lebanese children diagnosed with NB. CYP3A4*1B and CYP3A5*3 polymorphisms were specifically targeted as being the most common among multiple populations, with some insight into their effect on enzymatic activity and expression (27). In concordance with previous studies, MYCN amplification was observed to be the strongest predictor of an unfavorable clinical outcomes in children with NB in the present study (10). In addition, after adjusting for MYCN amplification, the results of the present study demonstrated that individuals with the CYP3A5*3/*3 mutant genotype and lower tumor CYP3A5 expression levels demonstrated a higher mortality risk, while those with the CYP3A4*1B mutant allele and lower tumor CYP3A4 expression levels demonstrated a lower mortality risk; however, these associations did not achieve statistical significance, likely owing to the small sample size.
CYP3A4 and CYP3A5 are considered to code for important CYP isoenzymes pertaining to drug metabolism, due to the diversity of their anti-cancer substrates and their relative abundance in humans (32). While no previous studies have examined the association between these enzymes and clinical outcome in NB, both enzymes have been shown to be important prognostic modifiers in several common types of malignancy. High protein expression levels of the CYP3A4/5 genes have previously been reported to predict metastasis and poor prognosis in osteosarcoma (33). On the other hand, patients with breast cancer carrying the CYP3A4*1B and CYP3A5*1 alleles were found to have a significantly lower mean survival time compared with that of carriers of the wild-type alleles (17). Furthermore, CYP3A5*1 was observed to be associated with an improved prognosis in a group of Caucasian patients with T-cell acute lymphocytic leukemia (34) and CYP3A5*3 was found to modulate chemotherapy-induced toxicity in a group of Swedish patients with ovarian cancer (35).
CYP3A5*3 consists of an A to G transition in intron 3, creating a cryptic splice site (27). Inappropriate splicing leads to a premature termination codon, resulting in a truncated protein with 101 amino acids (AA) compared with 502 AA in the normal protein (28). mRNAs with premature stop codons are more unstable and rapidly degraded, which explains the lower levels of CYP3A5 mRNA in CYP3A5*3 homozygote carriers compared with those in wild-type carriers (36). CYP3A5*3/*3 carriers were determined to have >50% reduction in catalytic activity for the metabolism of midazolam compared with that of carriers of the CYP3A5*1 wild-type allele (37). The results of the present study regarding CYP3A5 are in agreement with these findings. The CYP3A5*3/*3 genotype and lower expression levels of CYP3A5 mRNA were consistent in predicting poor clinical outcomes in NB patients. As such, it can be hypothesized that homozygote carriers of the CYP3A5*3 allele may have much lower levels of enzymatic activity than carriers of other CYP3A5 variants, and that the unfavorable outcome observed in the patients with NB in the present study may be partially due to slower bioactivation of used chemotherapeutic drugs, particularly etoposide and the nitrogen mustard drugs, based on the literature (16,19,20,38). This may be of importance clinically, particularly due to the high prevalence of the defective allele in the studied population (87%). The frequency of the allele reported in the Lebanese patients is almost identical to that reported in Caucasian patients (91%), and higher than that reported for populations of African descent (50%) (24,39). The hypothesis proposed in the present study, that the CYP3A5 genotype and associated phenotype are part of the prognosis profile of patients with NB, was based on the reported hazard ratios and supporting molecular biological evidence from the literature. However, this hypothesis could not be accepted with confidence as the results were not statistically significant, possibly due to the small sample size of the study.
CYP3A4*1B is an A to G transition within the 5′-flanking region of the gene. This polymorphism is considered to be associated with enhanced expression due to reduced binding of a repressor that may affect the transcriptional rate (29,30). However, no associations have been found between the CYP3A4*1B allele genotype and pharmacokinetics of its substrates (40,41). Regulation of CYP3A4 expression is known to be essentially pre-translational and its mRNA levels allow a good estimate of its enzymatic activity (42). The CYP3A4*1B polymorphism, however, cannot explain the observed mRNA expression levels (43). In the current study, the possible association between a higher mortality risk and higher CYP3A4 expression levels may be more significant than the reported protective effect for the CYP3A4*1B polymorphism, particularly since it has a low frequency in the studied group (8.7%). This is similar to frequencies reported in Caucasian individuals (3.6–9.6%) and much lower than those reported in African individuals (53–67%) (44, 45). A potential association between higher CYP3A4 expression levels and clinical outcome in NB may be hypothesized, particularly since similar associations have been previously reported in breast cancer (46), however, the results of the present study are inconclusive. Phenotypes may vary with tissue and may be dependent upon other endogenous and environmental factors. Further studies are required to elucidate the function of CYP3A4 in the clinical outcome of NB.
There were a number of limitations in the present study that should be considered when interpreting the results. The sample size of the study was smaller than desired, due to the low incidence of NB in the Lebanese population and missing tumor samples for a number of identified patients. Another limitation of this study is the possibility that other DMEs may be modifying the response to drugs received by patients. Additional CYP3A4/5 SNPs, GSTs, N-acetyltransferases, and other CYPs not investigated by the current study may influence the metabolic outcome of anticancer drugs most commonly used in the NB setting (Table I). The effects of these factors were not adjusted for and may explain why the obtained results were not significant. In addition, due to the retrospective nature of the study, serum samples from patients were not available to conduct direct measurements of drug levels. Furthermore, investigation of the mechanisms by which CYP3A4/5 polymorphisms and expression levels may act to influence patient outcome following high-dose therapy was not possible and further studies should be conducted to address this point. Additionally, DNA from blood samples was not available to check genotype concordance between the blood and the tumor.
However, this study showed some strength. Despite the small sample size, the prognostic value of MYCN amplification was found to be statistically significant, in concordance with the literature. Furthermore, CYP3A5*3/*3 genotype and lower mRNA expression levels predicted similar clinical outcome, which reflects the internal validity of the reported results.
The hypothesis that CYP3A5*3 polymorphism and low expression levels may be associated with an inferior clinical outcome in children diagnosed with NB is the first of its kind. Although the observed effects did not achieve statistical significance, the suggested associations may apply for a substantial proportion of cases given the high prevalence of this particular polymorphism in numerous populations including Lebanese. The Lebanese population appears to have a high CYP3A5*3 mutant genotype frequency, similar to that of the Caucasian population. The results of the current study may have an important clinical value. Since CYP3A5 represents at least half the total hepatic CYP3A content, it may be an important genetic contributor to inter-individual and interracial differences in clearance and response to NB anticancer drugs, particularly in CYP3A5 homozygote mutant patients (33). Characterization of associations between CYP3A5 polymorphisms and clinical outcome could be used to genetically define subgroups of patients with NB who may be more predisposed to therapeutic failures or adverse drug reactions. However, the prognostic value of functional polymorphisms of CYP3A5 and CYP3A4 remains to be fully determined. We recommend building upon the observations of the current study by conducting cohort studies which incorporate larger sample sizes. Confirming these hypotheses and generating supportive mechanistic data will ultimately allow for an individualized and optimized NB therapy that may improve survival.