PharmGKB summary: ifosfamide pathways, pharmacokinetics and pharmacodynamics

Background

Ifosfamide (IFO) is a widely used antitumor prodrug. It is in the oxazaphosphorine class of DNA alkylating agents and is effective against solid tumors, such as sarcomas and hematologic malignancies [1]. Despite its proven clinical benefit, its use is associated with major clinical toxicities that include urotoxicity, nephrotoxicity, encephalopathy, cardiotoxicity, and neurotoxicity, which occur in ~20% of patients [2]. IFO is used in combination chemotherapy. Thus, drug–drug interactions may increase IFO toxicity. In addition, polymorphisms in IFO disposition and target genes may also play roles [3]. IFO’s mechanism of cross-linking DNA plays a major role in preventing cancer cells from proliferating [4]. IFO has lower myelotoxicity relative to its structural analog cyclophosphamide but higher rates (45% compared with 10%) of the nephrotoxic metabolite chloroacetaldehyde (CAA) [5]. Glomerular and tubular dysfunctions represent serious side effects, especially in children who are cotreated with other nephrotoxic drugs. In addition, renal failure may lead to urotoxicity, and thus children are often supplemented with an equimolar dose of the organosulfur compound Mesna that prevents the buildup of IFO’s toxic metabolites acrolein and CAA [3, 6]. This summary reviews the pharmacokinetics, pharmacodynamics, toxicity, and pharmacogenomics of IFO, highlighting the genes important in the variation in response to the drug. A thorough understanding of this pathway may provide insight into personalization of this drug.

Pharmacokinetics

IFO is a prodrug that requires bioactivation by the cytochrome P450 (CYP) system to exert its pharmacological and toxicological effects. The initial hepatic activation step of IFO to its pharmacologically active molecule is 4-hydroxylation to 4-hydroxyifosfamide. Activation of IFO to 4-hydroxyifosfamide is catalyzed by the hepatic CYP isoforms CYP3A4 and CYP2B6, with minor contributions from CYP2A6, CYP2C8, CYP2C9, and CYP2C19 [1, 7]. Although IFO requires CYP3A4 and CYP2B6 for bioactivation and metabolism, CYP3A5 is mainly responsible for the transformation of IFO into active product through autoinduction. IFO autoinduces its own biotransformation by activating the xenobiotic receptor PXR coded by NR1I2, which mediates auto-induction by transcriptional upregulation of CYP3A4 [8]. IFO’s ability to autoinduce suggests that its rate of metabolism may increase over time. In contrast to cyclophosphamide and other drugs, IFO’s metabolism by CYP3A4 leads to more toxic metabolites [1]. Once formed, 4-hydroxyifosfamide is unstable and rapidly interconverts with its tautomer, aldoifosfamide, or is oxidized by alcohol dehydrogenase into 4-keto-4-hydroxyifosfamide. It is likely that both 4-hydroxyifosfamide and its tautomer passively diffuse out of hepatic cells, circulate, and then passively enter other cells [9]. Aldoifosfamide partitions between aldehyde dehydrogenase (ALDH1A1)-mediated detoxification and the inactive metabolite carboxyifosfamide, catalyzed conversion to aldoifosfamide, and more importantly a spontaneous (nonenzymatic) elimination reaction to ultimately yield the therapeutically active metabolite cytotoxic nitrogen mustards [ifosforamide mustard or isophosphoramide mustard (IPM)] and an equimolar amount of byproduct acrolein, which is highly electrophilic and is responsible for the urotoxic effect of IFO [7]. IPM contains a highly reactive alkyl group and thus covalently alkylates specific nucleophilic groups of DNA molecules, resulting in DNA cross-link formation and apoptosis of tumor cells. IPM, the DNA cross-linking agent of clinical significance, is a circulating metabolite, but the anionic IPM does not enter cells as readily as its metabolic precursors [10] (Fig. 1).

Fig. 1.

Graphic representation of the candidate genes and toxic metabolites involved in the pharmacokinetics of ifosfamide. A fully interactive version of this pathway is available online at PharmGKB at http://www.pharmgkb.org/pathway/PA2037.

The anticancer effect of IFO resides in the phosphamide mustard generated from sequential metabolism/degradation of IFO. The cytotoxic phosphamide mustard is translocated into the nucleus by passive or transport processes, where it is converted into the chemically reactive carbonium ion through imonium ion in alkaline or neutral pH. It then reacts with the N7 of the guanine residue in DNA to form a covalent linkage. The second arm in phosphoramide mustard can react with a second guanine moiety in an opposite DNA strand or in the same strand to form cross-link, and eventually cell apoptosis and/or necrosis. This bifunctional alkylation of DNA is believed to be the main mechanism of IFO mediated anticancer effect [7].

Competing with, and split about evenly with, 4-hydroxylation is a major oxidative detoxification pathway that results in both the detoxified 2- or 3-dechloroethylifosfamide (DCE) and the formation of CAA (toxic), both primarily mediated by CYP2B6 and CYP3A4 in the liver [1, 11].

Like cyclophosphamide, IFO contains a chiral phosphorous atom. IFO is administered as a racemic mixture of equal proportions of S-IFO and R-IFO. Unlike cyclophosphamide, enantioselective metabolism of IFO has been shown. In-vitro microsomal studies indicate enantioselective IFO 4-hydroxylation (R>S), with CYP3A4/5 preferentially metabolizing R-IFO to 4-hydroxyifosfamide and CYP2B6 activity toward S-IFO. IFO metabolism through the N-dechloroethylation reaction is also enantioselective, with S-IFO metabolism showing higher intrinsic metabolic clearance than R-IFO. Both CYP3A and CYP2B6 catalyze this reaction, but striking differences are observed: CYP3A4/5 preferentially produced (R)-N-2-DCl-IFO and (R)-N-3-DCl-IFO from R-IFO and S-IFO, respectively, whereas CYP2B6 preferentially formed (S)-N-3-DCl-IFO and (S)-N-2-DCl-IFO. On the basis of these in-vitro data, R-IFO exhibits more rapid 4-hydroxylation and less efficient N-dechloroethylation to toxic metabolites than S-IFO, suggesting that R-IFO may have a distinct clinical advantage over racemic IFO [12]. Although in-vivo studies on the metabolism and disposition of the R-enantiomer and S-enantiomer of IFO have not been fully evaluated, the limited clinical data available are consistent with the in-vitro evidence. Therefore, understanding IFO stereoselective metabolism appears particularly relevant with respect to the balance of 4-hydroxylation (beneficial pathway) versus N-dechloroethylation products and the impact of this on the central nervous system toxicity associated with IFO therapy. This may also help to better predict genetic and nongenetic mechanisms of interpatient differences in IFO disposition and effect [13].

Pharmacodynamics

The main mechanism in IFO is believed to be the direct interaction of IPM with DNA in the nucleus. Other mechanisms by IPM that reduce cancer proliferation, including activation of proapoptotic pathways and reduction of inflammatory and antiapoptotic pathways, have also been suggested [12, 13] (Fig. 2).

Fig. 2.

Graphic representation of the candidate genes and toxic metabolites involved in the pharmacodynamics of ifosfamide. A fully interactive version of this pathway is available online at PharmGKB at http://www.pharmgkb.org/pathway/PA2038.

IPM is translocated into the nucleus probably by passive diffusion. In the nucleus, IPM reacts with DNA by covalently bonding its highly reactive alkyl groups with nucleophilic groups on DNA, forming intrastrand and interstrand cross-links. These DNA strand breaks result in an inability to synthesize DNA and lead to cell apoptosis mediated by the caspase cascade.

IFO employs the caspase cascade to induce tumor cell death. Early on following IFO treatment, there is an activation of caspases 3, 8, and 9. IFO increases the gene expression levels of caspases 3 and 9, whereas it decreases the expression of BCL2, a caspase inhibitor. Not only does IFO decrease the possibility of BCL2 blocking the release of cytochrome c from the mitochondria but IFO can activate BAX and BAK, which are procytochrome c releasers [12, 14].

In studies involving MDCK cells, it has been identified that IFO modulates a series of cell cycle and immune response regulators. IPM interacts with proliferation and apoptotic pathways by both modifying MAP kinase signaling by downregulating genes responsible for proliferation and apoptosis control and downregulating heat shock protein activity. IPM also decreases expression of the genes TP53 and CIP1, which modulate the cell cycle regulator p53 activity. The 72Arg variant of TP53 has been shown to induce apoptosis at a higher frequency than the 72Pro form, thus leading to higher response rates and survival in chemotherapy and radiation therapy patients [7, 15, 16].

In addition, IPM downregulates TXNRD1 in the NRF-2 pathway, responsible for oxidative stress response. With the inhibition of thioredoxin reductase activity by IFO, the cell is unable to either respond to oxidative stress or transcribe proteins involved in the apoptotic and proliferation pathways. Finally, IPM can interfere with both the innate and adaptive immune response by decreasing production of the transcription factor NF-κB [10, 13].

IER3, a radiation inducible early-response gene that regulates apoptosis, is typically upregulated in cancer cells. In vitro, when the IER3 gene is silenced, cells gain sensitivity to IFO, suggesting that IPM downregulates this gene as a mechanism of antitumor therapy. This also indicates that antiapoptotic mechanisms may be involved in a growing resistance to IFO [17].

Not only can IPM’s ability to cross-link DNA inhibit proliferation of tumor cells, but it also alters the human immune response. IPM causes a depletion of intracellular glutathione (GSH), a major antioxidant. Multiple IFO metabolites can react with GSH, resulting in the formation of various conjugates at different sites along the pathway [4]. This decreased level of GSH is evident in T cells and natural killer cells that reduce the functionality of the immune response. IFO influences differential dendritic cell-mediated effector activities, specifically natural killer cells. However, this suppression of dendritic cell-mediated natural killer cell proliferation is indirect and is accomplished through T cells. The low levels of GSH reduce the ability for dendritic cells to stimulate T cell interleukin-2 production [18]. In addition, GSH conjugates with toxic metabolites, and therefore a decrease in the levels of GSH results in increased toxicity [19].

Toxicity

High rates of toxicity in kidney cells are evident following IFO treatment. The kidney is capable of metabolizing IFO through its abundance of xenobiotic metabolizing enzymes [20]. Metabolism in the kidney and liver may differ on the basis of the relative abundance of CYP isoforms, specifically, CYP2B6, CYP3A4, and CYP3A5. Variance of these CYPs in the liver can be influenced by sex, age, and genetic variation. Overall, CYP3A4 in the liver can vary up to 100-fold, with two-fold higher levels of CYP3A4 protein in females than in males. There are fairly undetectable levels of the CYP2B6 protein in pediatric livers. The allele CYP2B6*6 results in 50–75% lower liver protein expression. In the kidney, there are low levels of CYP2B6 mRNA transcripts but CYP3A5 protein expression is relatively high. Renal biotransformation in the kidney can be quicker than in liver cells and this is based on the variation and abundance of the CYP2B6, CYP3A4, and CYP3A5 enzymes. It is estimated that the renal variability of these CYPs ranges from 40-fold to 400-fold [21, 22]. CYP3A4 and CYP3A5 are present in the kidney before birth; therefore, this pathway route is likely active early on and able to generate toxic metabolites. Children 3 years of age and younger suffer serious renal damage, and therefore in pediatric cases patients are coadministered Mesna, a uroprotective agent, which can prevent the toxic metabolites acrolein and CAA that typically lead to hemorrhagic cystitis [3, 6, 19].

Acrolein, a toxic byproduct of the IPM activation pathway, is an unsaturated and highly reactive aldehyde. Upon activation, acrolein can enter uroepithelial cells and activate intracellular reactive oxidative genes, leading to peroxynitrite production, which can damage proteins, DNA, and lipids. Accumulation of acrolein in the bladder results in hemorrhagic cystitis [19].

It is estimated that 25–60% of IFO is metabolized into CAA. CAA suppresses the activation of complex I in the mitochondrial respiratory chain, resulting in decreased levels of GSH and ATP, and can induce cell death. CAA can cross the blood–brain barrier and this may cause encephalopathy. Local accumulation of CAA in the kidney is also a cause of nephrotoxicity. At toxic concentrations, CAA can deplete the levels of GSH, which is typically an antagonist of the toxic metabolite [2, 19, 23, 24].

Pharmacogenomics

CYPs and pharmacokinetics

Large interpatient differences, which may be up to 20-fold, in the pharmacokinetics and biotransformation of IFO have been reported [11]. CYP2B6 allele CYP2B6*6 heterozygous and homozygous carriers have been linked with lower catalytic activity and protein expression in the liver, higher concentrations of IFO in plasma, and higher rates of CAA-associated toxicity. In contrast, the CYP2B6*1 allele has been shown to have higher catalytic activity, higher protein expression, and a higher rate of clearance of IFO [11]. In addition, like CYP2B6*6, 14–25% of White female CYP2B6*5 carriers display lower protein expression and catalytic activity in the liver, resulting in lower rates of clearance and higher amounts of CAA [11]. Carriers of CYP3A5*1 have shown to catalyze the detoxification pathway to DCE at a faster rate, leading to a greater amount of CAA and higher risk of nephrotoxicity due to CAA’s ability to rapidly degrade in human blood [25].

ALDH and resistance

Aldehyde dehydrogenase enzymes can detoxify acrolein and CAA through the oxidation of less toxic carboxylic acids, acrylic acid, and chloroacetic acid. Therefore, inhibition of aldehyde dehydrogenase contributes to increased toxicity. Overexpression of ALDH1A1 and ALDH3A1 is thought to contribute to IFO resistance. It is believed that the overexpression directs the metabolism of 4-hydroxyifosfamide (and its tautomer) into the inactive metabolite instead of the active IPM. It has been demonstrated in cyclophosphamide that ALDH3A1 contributes to increased detoxification and tumor resistance. A variant ALDH3A1*2 allele has been noted, but the effects of this polymorphism on cyclophosphamide or IFO metabolism is unknown. In addition, it has been documented that increased ALDH1A1 and ALDH3A1 expression mediates resistance to chemotherapy treatment in breast cancer cells [9, 26].

Another possible contributor to IFO resistance is the MGMTenzyme, responsible for repairing precarcinogenic and pretoxic DNA damage. MGMT has the ability to repair large adducts that have been formed because of methylating or chloroethylating agents, such as acrolein. Thus, MGMT may both recognize and repair acrolein-induced DNA adducts. Conversely, studies have indicated that when there are high concentrations of MGMT in tumor cells, those cells are less sensitive to the growth inhibitory mechanisms of IFO [27].

Conclusion

IFO is a widely used drug with major clinical benefits, yet there are relatively few published studies on its metabolism and pharmacogenetics. As with all cancer drugs, it is often difficult to reduce toxicity side effects because the metabolism of the drug in the liver and kidney is necessary for release of the active component, and toxic metabolites can be required for therapeutic efficacy. Knowledge of the relative production of toxic metabolites at different sites and how this is influenced by genotype could be useful for personalizing treatment. IFO interferes with both tumor cell regulators and the human immune response to combat tumor proliferation. Although much is known about the cellular pharmacodynamics of IFO, the pharmacogenomics have not been well characterized. Further insight regarding CYP haplotype variation and the resulting phenotypes would be beneficial and would contribute to a more comprehensive understanding of the kinetics pathway. Future studies should integrate the pharmacokinetics, pharmacodynamics, and pharmacogenomics to look for connections between the increased toxicity, the immune response, and cell cycle modulation.