Pharmacogenetics in Neuroblastoma: What Can Already Be Clinically Implemented and What Is Coming Next?

Pharmacogenetics is one of the cornerstones of Personalized Precision Medicine that needs to be implemented in the routine of our patients’ clinical management in order to tailor their therapies as much as possible, with the aim of maximizing efficacy and minimizing toxicity. This is of great importance, especially in pediatric cancer and even more in complex malignancies such as neuroblastoma, where the rates of therapeutic success are still below those of many other types of tumors. The studies are mainly focused on germline genetic variants and in the present review, state of the art is presented: which are the variants that have a level of evidence high enough to be implemented in the clinic, and how to distinguish them from the ones that still need validation to confirm their utility. Further aspects as relevant characteristics regarding ontogeny and future directions in the research will also be discussed.


Introduction
Neuroblastoma (NB), the most common solid extracranial malignancy during childhood, has its origin in the adrenal medulla or paraspinal ganglia (sympathetic nervous system) during the period of development [1] and shows great phenotypic heterogeneity: some infants have spontaneous regression of the tumor while others present disease progression event after intensive multimodal treatment [2].
Although translational and clinical research has evolved considerably in recent years, in the cases requiring treatment, the prognosis of children with High-Risk Neuroblastoma (HR-NB) remains very poor. Despite the fact that it only represents around 8% of all pediatric cancers, it causes 15% of all deaths due to cancer in children. Barely 40% of the children diagnosed survive longer than 5 years [3,4]. Indeed, when metastatic relapse occurs, there is no curative treatment, and the overall survival rate after relapse is around 12 months. Thus, the lack of effective treatment continues to be a major concern for pediatric oncologists. Another relevant aspect to take into account is that >50% of survivor children

Pharmacogenetic Variants for Clinical Implementation
Relevant institutions and scientific societies worldwide agree on the use of three main pillars as the references to identify those genetic variants with the highest level of evidence to be implemented in the clinic. These three cornerstones are the indications of drug regulatory agencies, PharmGKB (Pharmacogenomics Knowledge Base) [8], and relevant international consortia of experts developing clinical guidelines for PGx implementation.

Drug Regulatory Agencies
Regulatory agencies as the US Food and Drug Administration (FDA), the European Medicines Agency (EMA), the Pharmaceuticals and Medical Devices Agency from Japan (PMDA), and Health Canada Santé Canada (HCSC), amongst others [9][10][11][12], recommend in the Drug Labels a genetic test prior to the use of many drugs they approve. The Drug Label indicates if the test is required, recommendable, actionable, or simply informative [8]. At this point, it must be remarked that any requirement of a specific action in the drug label must be considered a legal requirement, with consequences in the case of disregard.

PharmGKB Clinical Annotations
PharmGKB, is a free access database created, curated, and managed by the University of Stanford and funded by US National Institutes of Health (NIH) and National Institute of General Medicine Science (NIGMS) [8]. It compiles most of the existing PGx information, from many different databases, including PubMed, under a Creative Commons license. It counts with a group of experts working on the dissemination of knowledge about the impact of human genetic variation on drug responses and on the translation of PGx into clinical practice. In fact, the website (www.pharmgkb.org, accessed on 25 June 2021) includes not only the information of its own curation from published data, the 'Clinical Annotations', but also that of the Drug Labels and the reports by Experts Consortia.
Clinical Annotations are the curated results obtained by PharmGKB's experts after publications review for assigning different levels of evidence in the association of a genetic variant with efficacy, toxicity, metabolism/pharmacokinetics, and dosage of a drug. This 'Level of Evidence' ranks from 1 to 4, being 1 the one meeting the highest criteria. Level 2 is tagged as "moderate", while 3 and 4 are "low" and "unsupported," respectively. Very recently, these criteria have incorporated a new scoring system.

Clinical Implementation Guidelines
These are drug adjustment guidelines published by experts' consortia to provide recommendations about what actions the prescriber should consider according to patient genotype. The main consortia are the Clinical Pharmacogenetics Implementation Consortium (CPIC) [13], the Royal Dutch Association for the Advancement of Pharmacy-Pharmacogenetics Working Group (DPWG) [14], but also other professional and scientific societies [15]. PharmGKB also compiles this information on the website and provides links to the complete articles under the section 'Prescribing Info-Clinical Guideline Annotations'.
According to these three pillars, Table 1 shows the revised drugs that meet at least one criterion of the three pillars: PGx information in the drug label according to Drug Regulatory Agencies, PharmGKB Clinical Annotations with Level of Evidence 1 or 2, Clinical Implementation Guideline elaborated by an international expert Consortium. It must be underlined that, as previously stated, PharmGKB has just performed relevant modifications, beginning approximately on 25 March 2021. For caution, in this table, we have considered the immediate previous information regarding Clinical Annotations, but the current modifications have been marked.
Three drugs have genetic recommendations in Drug Labels according to different regulatory agencies, but these recommendations are not strictly PGx. They are mainly related to the use for which the drug is intended, its clinical indication, and the genes to be analyzed are not in the constitutive DNA of the patient but in the tumor. These drugs are Busulfan (FDA: ABL1 and BCR genes), 13-cis-retinoic acid (not isotretinoin, but tretinoin has required testing for PML and RARA genes in the FDA Drug Label, the same in Health Canada Santé Canada Label, and also, but at an informative level, in the Japanese regulatory agency). The same situation applies to vincristine but, apart from Drug Label, it has a Clinical Annotation.
Dinutuximab contains information in the FDA label regarding MYCN in the clinical trials performed.
In this sense, only the indications for TPMT regarding Cisplatin are really related to germline variants and a toxic event that could be prevented: ototoxicity.
Regarding Clinical Implementation Guidelines, we have this kind of source for two of the revised drugs: cisplatin and doxorubicin [16,17]. The recommendations of the CPNDS for the PGx-guided use of these drugs are summarized in Table 2.

Pharmacogenetic Variants under Investigation Regarding NB Therapy
Regarding research, there is still a lot of work to conduct. There are multitudes of SNPs whose influence on drug response was proposed but is not yet sufficiently tested, and, therefore, validation is needed. As shown in Table 1, only a few drugs from the group included in this review have one pillar (at least) supporting their implementation in the clinical setting. However, for the rest of the drugs, many of the genes coding the transporters, metabolizing enzymes, and targets of these drugs are already known, and research should be addressed to identify the implications of the SNPs in those genes on the safety and efficacy of those drugs. Table 3 shows a compilation of the most promising SNPs that are under research, according to the literature search. In most of the manuscripts, the studies have been performed with the combinations of drugs that are currently used in chemotherapy regimens. For clarity, we have searched for literature with results attributable to single drugs and not to combinations of them. Another criterion for the selection, in most cases, has been the known relationship between the gene and the drug in terms of transport, metabolism, and/or mechanism of action.
Our knowledge about the genes involved in the mechanism of action, transport, and metabolism of the drugs is in many cases much more limited than expected. This happens not only with the newest drugs, but also with the classic chemotherapy that has been employed for decades. Therefore, focusing our attention on the SNPs contained in the genes responsible for the fate of the drug in our organism is not always easy. In addition, interactions exist, and these can happen immediately or with mid-long term effects, difficult to predict. All these together lead to "difficult to explain findings," except for a much-reduced group of genes directly related to specific events. For example, SNPs in genes such as MTHFR, TP53, or VDR have been correlated with overall survival and event-free survival in PGx studies of NB patients, whereas these genes are not directly involved in the body routes of the chemotherapeutic drugs employed. In our group's experience, rs1801133 in MTHFR (p = 0.02) and rs1544410 in VDR (p = 0.006) added an important predictive value for overall survival to the MYCN status, with a more accurate patients sub stratification than using MYCN alone [58][59][60].
If we check the bibliography, the most robust results should be reported from clinical trials, but these are obtained in predesigned and very much controlled situations different from the real clinical setting. For this reason, studies should be performed respecting the clinical reality, including concomitant treatments, especially considering that in pediatric oncology, many of the children participate in clinical trials evaluating their treatment and impeding them to participate in another specific for pharmacogenetics [61,62]. Table 3. Candidate SNPs and genes to evaluate regarding the efficacy and toxicity of the drugs in Neuroblastoma (based in PharmGKB and the included references).

Drug. Gene SNP Hypothetic Effect References
Busulfan CTH rs1021737 Pediatric patients with the TT genotype (receiving hematopoietic stem cell transplantation) may have an increased risk for sinusoidal obstruction syndrome (SOS) when treated with cyclophosphamide and busulfan as compared to patients with the GG or GT genotypes.   Patients with the AA or AG genotypes (and non-Hodgkin lymphoma) who are treated with doxorubicin may have an increased risk of cardiotoxicity as compared to patients with the GG genotype.
rs17222723 Patients with the AA or AT genotypes (and non-Hodgkin lymphoma) who are treated with doxorubicin may have an increased risk of cardiotoxicity as compared to patients with the TT genotype.

The Role of Ontogeny in NB Pharmacogenetics
The statement that children are not small adults is valid, particularly in pediatric clinical PGx. In order to offer children the optimal treatment, it is important not only to know the characteristics of the particular disease but also to integrate the changes of normal growth and development with their impact on the ontogeny of pharmacokinetic and pharmacogenomic factors. There are several age-related anatomic and physiological changes that have been found to influence drug ADME (Absorption, Distribution, Metabolism and Excretion) processes, such as differences in fat proportion in the body, gastric pH evolution, renal development, etc., but some of them are directly linked to the expression and role of relevant pharmacogenes. For instance, the activity of ABCB1 (P glycoprotein) and ABCG2, two of the most extensively studied ATP-binding cassette (ABC) transporters are decreased in the neonatal period [63,64]. Regarding metabolism, CYP3A7 shows the highest activity in the liver during embryonic, fetal, and newborn stages. After that, its activity declines, and other CYPs take the main roles. CYP3A4 appears during the first week after birth, reaching approximately 30-40% of adult activity by the first month and full adult activity by the 6th month of life. Its activity increases so much that it reaches 120% of the adults between 1 and 4 years of age, decreasing to normal adult levels after puberty [65][66][67]. In the case of CYP1A2, its expression is delayed until 3 months after birth. Regarding Phase II enzymes, as UGT1A family, it is remarkable that UGT1A1 starts to increase at birth and does not reach adult levels since 3-6 months later; and that UGT1A6 and UGT1A9 activity levels are smaller in people younger than 10 years old in comparison with adults [67].

What Else Do We Need to Take into Account? Research Integrating Epigenomics and Metabolomics
The next goal is complementing PGx with two other relevant technologies that must undoubtedly be integrated: Epigenomics and Metabolomics. Explained in a very simple way, Epigenetics will deal with the study of the methylation status of the promoters of certain pharmacogenes as a mechanism of higher regulation that, above the nucleotide sequence of DNA, will determine whether a gene is really being expressed or not [68,69]. Metabolomics, also in a simplistic manner for this context, could characterize the plasma presence of metabolites corresponding to the drugs administered to the patients to check if the drugs have been effectively metabolized or not [70]. Complementing pharmacogenetics with the other two Omics techniques must help understand relevant questions: phenomena occurring in the patients that represent an upper layer of complexity for understanding the patients' real response to drugs. For instance, metabolic pathways that are switched on when the main metabolic route is "off" or hyper/hypomethylation phenomena that regulate the expression of genes, masking potential effects of the nucleotide sequence. In order to understand the real characteristics of each patient, integration of all this information by means of very advanced Biostatistics, Systems Biology, Pharmacology, and Artificial Intelligence will be required.

Conclusions
PGx knowledge needs to be implemented in the clinical routine of NB patients to support a more personalized approach regarding chemotherapy. The field is divided, with drug-genetic variants with a high level of scientific and clinical evidence and many more drug-SNP pairs needing further research in real patient contexts in order to validate their effects. The available literature regarding those associations is in many cases scarce and old, and the results have not been confirmed or updated. Thus, we need to put our efforts into this type of research. Meanwhile, those associations with high levels of evidence should be assessed in all our patients with the aim of providing the clinician with an additional tool to modify the treatment, if possible, and/or to be alert to increased risks of immediate or late toxicities. The current scenario provides data on relationships between SNP-drugs, but the reality is different most of the time due to the use of a combination of drugs. Thus, we need to validate the proposed "one-to-one" relationships in the real clinical context because interactions do exist, and the expected effects of concrete SNPs could not be the same in the context of polytherapy.

Data Availability Statement:
The data reported is contained in the referenced papers as well as in PharmGKB website (www.pharmgkb.org).