Can the Pharmacogenetics of GST Gene Polymorphisms Predict the Dose of Busulfan in Pediatric Hematopoietic Stem Cell Transplantation?

In the last half of the 20th Century remarkable progress has been made in pediatric oncology; pediatric cancer survival increased from 10% to 70% in less than 50 years. Still, too many children relapse or suffer from treatment-related toxicities. Moreover, the increase in cancer survivor population will also increase the incidence long-term effects of chemotherapeutics.

Interindividual differences in drug responses are an important cause of resistance to treatment and adverse drug reactions. Identifying pharmacogenomic determinants of a given drug may allow for prospective identification of patients with suboptimal drug responses, allowing for complementation of traditional treatment protocols by genotype-based drug dose adjustment.

Pharmacogenetics has an impact on both health and economy. Each year there are approximately 2.1 million hospitalized patients who suffer from adverse drug response (ADR). and more than 105,000 deaths are caused by ADRs alone [1]. It has been shown that efficacy is increased and toxicity reduced when a genetically-guided dose adjustment strategy is used [2,3]. Besides being advantageous for the patients, it is also highly cost effective and therefore desired from a socioeconomic perspective [4].

Busulfan (BU) is an alkylating agent used in a wide variety of myeloablative conditioning regimens for hematopoietic stem cell transplantation. BU has a narrow therapeutic index. High-drug exposure leads to increased risk of hepatic veno-occlusive disease (HVOD), while low-drug exposure has been associated with a higher risk of disease recurrence and graft failure [5–8]. Accordingly, steady state BU plasma concentration and area under the concentration–time curve (AUC) correlate with the incidence of these clinical events.

Intravenous (IV) administration has gained popularity, especially in children, as IV pharmacokinetics can be more predictive than pharmacokinetics following oral administration, reducing the variability of BU exposure among patients [9–12]. Dose targeting based on therapeutic drug monitoring seems to improve event-free survival and survival rates, and in some studies HVOD-free survival as well as incidence of graft versus host disease (GVHD) compared with conventional BU treatment [13–15]. In spite of these improvements, interpatient variability still exists and the percentage of patients reaching the target AUC after the first dose remains low [13,16,17]. Several factors have been identified to account for differences such as sex, weight, age or body surface area [8,18,19]. Nevertheless, important interindividual variability still persists and several additional factors contribute to such observations [16].

Glutathione S-transferases (GSTs) are involved in the metabolism/detoxification of a variety of drugs used in cancer treatment, such as alkylating agents, anthracyclines, topoisomerase II inhibitors and corticosteroids. Genotypes with no or lower GST activity may increase the effectiveness of the drug, given the reduced detoxification capacity of the enzymes, but may also render patients more prone to drug side effects. Metabolism by GST is the main route of BU biotransformation. Several GST subfamilies exist, such as GSTA1, GSTM1, GSTP1 and GSTT1. GSTA1 is the predominant GST isoform catalyzing BU conjugation with glutathione, whereas GSTM1 and GSTP1 have 46 and 18% of the activity of GSTA1, respectively [20]. Interindividual variability in BU metabolism may be explained, at least in part, by GST polymorphisms. Two GSTP1 variations leading to Ile105Val and Ala114Val amino acid replacements result in different catalytic activities [21]. A significant number of individuals (~50%) lack GSTM1 activity due to the homozygous deletion of the GSTM1 gene [22]. Several polymorphisms, shown in some studies to be functional, have been identified in the GSTA1 gene [23–25]. A recent study analyzed the impact of different polymorphisms in GSTA1, GSTM1 and GSTP1 genes, and found a significant correlation between the pharmacokinetics of IV BU and GSTM1 gene deletion [26]. GSTM1-null individuals had a significantly higher plasma drug concentration and lower clearance compared with non-null individuals, as would be expected for subjects lacking GSTM1 activity. GSTM1-null patients also received lower cumulative BU doses.

Srivastava et al. reported a higher incidence of HVOD in GSTM1-null thalassemia patients that did not have BU dose adjustments to achieve the predetermined target level [27]. This finding is in concordance with other studies that suggest a relationship between HVOD and higher BU exposure exists [6–8]. However, they reported lower drug levels and higher clearance in GSTM1-null individuals, suggesting that HVOD may be mediated by glutathione depletion, possibly through higher GSTA1 activity. The study of Bredschneider et al. contrasts with this explanation by demonstrating that neither GSTA1 protein expression nor conjugation activity was affected by GSTM1 status in human liver tissue [28]. Other studies that analyzed the GSTM1 gene did not find any significant correlation between first dose pharmacokinetic variability and the GSTM1-null genotype [18,29]. The difference in the patient population, sample size and treatment protocol or drug administration routes and schedule may account for discrepancies seen across studies. It remains to be seen in larger studies whether this genotype may predict BU-related side effects. It seems that many factors are equally important for HVOD development including underlying disease, pre-existing risk for HVOD and particularly other drugs administered concomitantly with BU [6,13,29].

The C69T promoter polymorphism of GSTA1 defines so called haplotypes GSTA1*A and *B. In some studies this polymorphism was shown to affect the gene expression and protein levels in hepatic cells [23–25], whereas others did not find an association with GSTA1 protein levels or the rate of busulfan conjugation [28], suggesting that a functional role of this polymorphism remains to be clarified.

Ansari et al. did not find any association of haplotype GSTA1*B with pharmacokinetic variability, which might be due to the small number of patients analyzed [26]. The absence of such an association in pediatric patients was also reported by Zwaweling et al.[18], whereas others, in contrast, reported that GSTA1*B can have an impact on BU pharmacokinetics or treatment outcome [29,30]. Johnson et al. found that children who were carriers of GSTA1*B had almost threefold higher BU concentration compared with noncarriers [29]. Using population pharmacokinetic modeling they also demonstrated that carriers of GSTA*1B have reduced BU clearance. Similar association was reported in Japanese adult patients by Kusama et al.[30]. GSTA1*A homozygosity would thus be expected to be associated with lower BU levels. Indeed Kim et al. found lower incidence of GVHD in adult Korean patients with hematological malignancy [31]. However, no association with HVOD was found and no pharmacokinetic data were available for the patients analyzed. There was no association reported between GSTP1 variants and BU pharmacokinetics [18,26].

The studies published so far demonstrated that GSTA1*B variant and GSTM1-null genotype may contribute to pharmacokinetic variability following the first BU dose. Bartelink et al. recently demonstrated a link between the pharmacokinetics and BU toxicity as well as event-free survival for patients receiving BU for hematopoietic stem cell transplantation; therefore, it is possible that correlation between the GST polymorphisms and clinical outcomes of BU treatment may also exist [13]. Nevertheless, the results are still preliminary, as the studies were mostly conducted in a small number of patients. Further and larger studies are needed to replicate and confirm the findings obtained for GSTA1 and GSTM1 gene variations. Such studies will also be needed to understand how association between genetic factors and drug concentration or therapeutic end points depends on the protocol and stratification parameters (i.e., age, sex, population groups, disease characteristics, risk classes and so on), thus reducing the presently observed heterogeneity across the studies. It is also possible that several variants of several genes influence drug effects. A haplotype rather than genotype-based approach and a gene–gene interaction rather than a single gene may provide in some instances a more complete and useful assessment of genetic differences between individuals [32,33].

In conclusion, pharmacogenomics is evolving rapidly due to many recent discoveries such as the Human Genome project, the expansion of genomics and proteomics, emerging technologies, the knowledge of molecular basis of disease, and of the drug pathways. Pharmacogenomics is a novel and promising field of research and holds new possibilities for tailoring medical therapy in individual patients. Genotyping patients for particular polymorphisms, even prior to drug administration, could individualize drug treatment including BU. The understanding of how GST and other gene polymorphisms affect the response to busulfan will be essential to improving the efficacy and safety of this drug, while allowing for reduction in acute and long-term drug side events. Importantly, individualization of treatment will also reduce the cost associated with suboptimal response and drug side effects. Different population-based modeling in combination with the identification of covariates that account for the pharmacokinetic and pharmacodynamic variability of BU may allow better estimation of first drug dose and efficacy/safety profiles in pediatric patients. Identified genetic variants should be subsequently tested in a prospective manner, ultimately allowing their future incorporation in a protocol, which adjusts treatment according to patients‘ genotypes.

Acknowledgements

We are thankful to all patients and their parents who consented to participate in the genetics study.

Financial & competing interests disclosure

This work was supported by the Foundation Télémaque and the Foundation du centre de cancérologie Charles-Bruneau. Marc Ansari is a scholar of the Foundation Télémaque. Maja Krajinovic is a scholar of the Fonds de la Recherche en Santé du Québec. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending or royalties.

No writing assistance was utilized in the production of this manuscript.

Additional information

Funding

This work was supported by the Foundation Télémaque and the Foundation du centre de cancérologie Charles-Bruneau. Marc Ansari is a scholar of the Foundation Télémaque. Maja Krajinovic is a scholar of the Fonds de la Recherche en Santé du Québec. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending or royalties. No writing assistance was utilized in the production of this manuscript.