**3. Genotoxic impurities (GIs)**

#### **3.1 Sulfonates**

Sulfonate salts (Figure 1) are the most frequently used compounds in pharmaceutical developments. Salt formation is a useful technique for optimizing the physicochemical processing (formulation), biopharmaceutical or therapeutic properties of active pharmaceutical ingredients (APIs), and sulfonate salts are widely used for this purpose (Elder and Snodin, 2009). In addition to the advantages of processing, sulfonate salts possess some advantages over other salts such as producing higher melting point of the sulfonated API. This helps to enhance the stability and provide good solubility and may have certain *in vivo* advantages as well. For instance, in contrast to other salts of strong acids, mesylates do not have a tendency to form hydrates, which makes them an attractive

Based on the importance of the mechanism of action and the dose-response relationship in the assessment of genotoxic compounds, the EMEA guideline presents two classes of

1. Genotoxic compounds with sufficient (experimental) evidence for a threshold-related

2. Genotoxic compounds without sufficient (experimental) evidence for a threshold-

Those genotoxic compounds with sufficient evidence would be regulated according to the procedure as outlined for class 2 solvents in the "Q3C Note for Guidance on Impurities: Residual Solvents". For genotoxic compounds without sufficient evidence for a thresholdrelated mechanism, the guideline proposes a policy of controlling levels to "as low as

On the other hand, this guideline provides no advice on acceptable TTCs for drugs during development, especially for trials of short duration (Jacobson-Kram and McGovern, 2007). The pharmaceutical research and manufacturing association (PhRMA) has established a procedure for the testing, classification, qualification, toxicological risk assessment, and control of impurities processing genotoxic potential in pharmaceutical products. As most medicines are given for a limited period of time, this procedure proposes a staged TTC to adjust the limits for shorter exposure time during clinical trials (Table 3). Thus, the staged TTC can be used for genotoxic compounds having genotoxicity data that are normally not

> > 1-3 month

Table 3. PhRMA genotoxic impurity task force proposal – allowable daily intake (µg/day) for genotoxic impurities during clinical development using the staged TTC approach

Sulfonate salts (Figure 1) are the most frequently used compounds in pharmaceutical developments. Salt formation is a useful technique for optimizing the physicochemical processing (formulation), biopharmaceutical or therapeutic properties of active pharmaceutical ingredients (APIs), and sulfonate salts are widely used for this purpose (Elder and Snodin, 2009). In addition to the advantages of processing, sulfonate salts possess some advantages over other salts such as producing higher melting point of the sulfonated API. This helps to enhance the stability and provide good solubility and may have certain *in vivo* advantages as well. For instance, in contrast to other salts of strong acids, mesylates do not have a tendency to form hydrates, which makes them an attractive

Duration of clinical trial exposure

> 3-6 month

120 60 20 10 1.5

0.5% 0.5% 0.5% 0.5% 0.5%

>6-12 month

>12 month

reasonably practicable" (ALARP) principle, where avoiding is not possible.

≤ 1 month

suitable for a quantitative risk assessment (Muller *et al.*, 2006).

genotoxic compounds:

mechanism,

related mechanism.

Allowable Daily Intake (µg/day) for all phases of

Alternative maximum level of allowable impurity based on percentage of impurity in API

**3. Genotoxic impurities (GIs)** 

development

**3.1 Sulfonates** 

salt form for secondary processing, especially wet granulation. Another benefit of these salts is their high melting point because APIs with low melting points often exhibit plastic deformation during processing which can cause both caking and aggregation. Typically, an increase in the melting point has an adverse effect on aqueous solubility owing to an increase in the crystal lattice energies. Sulfonic acid salts tend to be an exception to this rule, since they exhibit both high melting points as well as good solubility. In addition, as mentioned in the literature, the high solubility and high surface area of haloperidol mesylate result in enhanced dissolution rates (<2 min in pH 2 simulated gastric media), which are more rapid than the competing common ion formation (Elder and Snodin, 2009; Elder *et al.*, 2010a).

On the other hand, sulfonic acids can react with low molecular weight alcohols such as methanol, ethanol, or isopropanol to form the corresponding sulfonate esters. In general, sulfonic acid esters are considered as potential alkylating agents that may exert genotoxic effects in bacterial and mammalian cell systems and possibly carcinogenic effects *in vivo*; thus, these compounds have raised safety concerns in recent times (Snodin, 2006; Teasdale *et al.*, 2009).

Fig. 1. Structures of common sulfonate salts

#### **3.1.1 Genotoxicity profile**

Sulfonate impurities comprise the most investigated group of genotoxic impurities (GIs). Initially in 2007, sulfonate impurities raised major concern when over a period of three months (March to May 2007), several thousand HIV patients in Europe were exposed to ViraceptR (nelfinavir mesylate) tablets containing the contaminant ethyl methane sulfonate (EMS). However, the available *in vitro* and animal data indicated that the levels at which HIV patients were exposed to EMS (maximal dose of 0.055 mg/kg/d) did not induce any risk; nevertheless, any further level was of significant concern to their safety (Elder and Snodin, 2009). Since 2007 other drugs have been reported for contamination by sulfonate impurities, such as alkyl benzene sulfonates in amlodipine besylate (Raman *et al.*, 2008), dimethyl sulfate (DMS) in pazopanib hydrochloride (Liu *et al.*, 2009), EMS and methyl methane sulfonate (MMS) in imatinib mesylate (Ramakrishna *et al.*, 2008), EMS in zugrastat (Schülé *et al.*, 2010), alkyl sulfonates in flouroaryl-amine (Cimarosti *et al.*, 2010), and ethyl besylate in UK-369,003-26, a novel PDE5 inhibitor (Hajikarimian *et al.*, 2010).

EMS is a well-established genotoxic agent in this group which reacts with DNA producing alkylated (specifically ethylated) nucleotides. MMS, an analog of EMS, is a genotoxic compound both *in vitro* and *in vivo*. The international agency for research on cancer (IARC) has classified EMS and MMS in group 2B and 2A, respectively (Snodin, 2006; Gocke *et al.*, 2009a).

Genotoxic Impurities in Pharmaceuticals 393

As indicated by the *in vivo* test in rodent bioassay, these compounds are either noncarcinogens (1- chlorobutane, bromomethane) or low-potency carcinogens (chloroethane, bromoethane). According to *in vivo* tests, chloroethane and alkyl bromides seem to be nongenotoxic carcinogens rather than genotoxic carcinogens. Based on the available data, the United States environmental protection agency (USEPA), considers tert-butyl chloride to be a group D compound or ''not classifiable as to human carcinogenicity'' (Bercu *et al.*, 2009;

Hydrazine is used as a medicine or as a starting compound for synthesizing some medicines. Hydrazine and some of its *N*-alkyl, *N*-aryl, and *N*-acyl analogues have been subjected to extensive toxicological evaluations. Hydrazines, hydrazides, and hydrazones have structural alerts for genotoxic potential and the metabolism increases their effects. Hydrazines adduct with DNA and the mechanism of adduction could include the formation of methyldiazanium ions or methyl free radicals. In addition, it seems that hydrazine reacts with endogenous formaldehyde to produce formaldehyde hydrazone. Subsequent to some other reactions, alkylating compounds like diazomethane as the genotoxic moiety are

*In vitro* studies have shown genotoxic effects for three hydrazine derivatives (hydrazines, hydrazides, and hydrazones). These compounds induce gene mutations in human teratoma cells, mouse lymphoma cells, and in several strains of bacteria. Hydralazine (1 hydrazinylphthalazine) and its hydrochloride salt are Ames-positive. In another study, 20 hydrazine-derivatives were found to induce a direct DNA damage in *Escherichia coli* and 16 of them (80%) were Ames positive as well (Flora *et al.*, 1984; Agency for Toxic Substance and

Although it was seen that hydrazine did not induce unscheduled DNA synthesis in mouse sperm cells, *in vivo* studies on the genotoxicity of hydrazines have largely produced positive results. In addition, it was observed that 1, 2-dimethylhydrazine failed to induce micronuclei in rat bone marrow cells, while this effect had been observed in mouse bone

The non-carcinogenic effects of hydrazine were also evaluated; however, it was found that hydrazine, methyl hydrazine, 1,1- and 1,2-dimethylhydrazine, and other analogues are carcinogenic in rodents and possibly in human. In addition, it was seen that hydrazine derivatives like hydralazine and its hydrochloride salt were tumorigenic in rodents. It should be mentioned that the clinical use of hydralazine hydrochloride for several years has shown no evidence for carcinogenicity (Flora *et al.*, 1984; Bercu *et al.*, 2009; Snodin, 2010).

Epoxides are considered as electrophilic compounds owing to the strained epoxide ring. These alkylating agents directly react with DNA. Alkene oxides are more reactive than arene oxides and symmetrically substituted epoxides are less reactive than asymmetrically substituted compounds. Some examples for APIs with epoxide impurities are betamethasone acetate, atenolol, and some herbal remedies. Carbamazepine, cyproheptadine, and protriptyline have stable epoxide metabolites. In addition, phenytoin,

marrow cells (Agency for Toxic Substance and Disease Registry, 1997).

Snodin, 2010).

**3.3 Hydrazines** 

produced (Bercu *et al.*, 2009; Snodin, 2010).

Disease Registry, 1997; Snodin, 2010).

**3.3.1 Genotoxicity profile** 

**3.4 Epoxides** 

Gocke *et al.* (2009a) reviewed both *in vivo* and *in vitro* genotoxicity, carcinogenicity, general toxicity, and the effects on reproductive and embryo fetal development of EMS. They reported that the genotoxic effects induced by EMS were observed in viruses/phages, bacteria, fungi, plant, insect, and mammalian cells. In another study, the induction of gene mutations at the hprt locus and the induction of chromosomal damage were examined as evidenced by the formation of micronuclei in human lymphoblastoid cells. It was found that the lowest dose inducing a positive response was 1.40 g/ml, and a no observed effect level (NOEL) could be defined at 1.2 g/ml. Also, no toxicity was observed at doses up to 2.5 g/plate. This observation is in strong contrast to the largely linear dose–response observed in the previous studies. As a result of *in vivo* assays for the induction of DNA damage, EMS is distributed rather uniformly over the body and induces similar levels of DNA damage in the various organs. Also, EMS is clastogenic in all test systems. The minimal dose of EMS applied in these studies was either 50 mg/kg or 100 mg/kg. In the majority of studies the dose–response relationships appeared sub linear and a threshold below 50 mg/kg appeared possible. Gocke *et al.* (2009a) demonstrated that EMS in various gene mutation tests such as induction of hprt, lacZ, and dlb-1 mutations in mice was mutagenic. The carcinogenicity of EMS was confirmed in several animal models. In another study, three methanesulfonates and three benzenesulfonates were tested by micronucleus and Yeast deletion recombination (DEL) assays. It was observed that all six substances produced positive responses in the tests (Sobol *et al.*, 2007).

#### **3.2 Alkyl halides and esters**

Owing to their electrophilic nature, alkylating agents can introduce lesions at nucleophilic centers of DNA. Drug salt formation includes strong acid/base interactions in the presence of alcohols, and can form impurities such as alkyl halides. As salt formation is a common method in drug formulation processes, alkyl halides exist as impurities in several drugs (Sobol *et al.*, 2007; Elder *et al.*, 2008a).

#### **3.2.1 Genotoxicity profile**

The nucleophilic attack mechanisms of alkylating compounds determine their reactivity against DNA. The SN1 mechanism leads to *O*-alkylation (*O-6*-methylguanine) which is mutagenic but not clastogenic, whereas the SN2 mechanism leads to N-methylation which is clastogenic and not mutagenic. In this group, it seems that bromo compounds are more reactive as compared to chloro compounds (Sobol *et al.*, 2007; Snodin, 2010).

Various tests have been performed to study DNA damage and mutation in alkyl halides. In the Ames test, it was found that most alkyl halides, especially bromides, are Ames positive except 1-chloropropane, 1-chlorobutane, and neopentyl bromide. As chloro- and bromobenzene are not alkylating agents, these compounds are Ames-negative. In Yeast deletion recombination (DEL) and micronucleus assays, alkyl chlorides such as *n*-propyl chloride are found to be negative (Sobol *et al.*, 2007; Snodin, 2010).

It was observed that alkyl chlorides in the NBP [4-(*p* nitrobenzyl) pyridine] alkylation assay are not reactive and that allyl chloride has minimal activity. Although benzyl chloride is more active than other chloro compounds, ethyl, propyl, or butyl bromides have at least 1/40 MMS activity; however, allyl bromide appears to be more active (around one-eighth of the activity of MMS) (Sobol *et al.*, 2007).

As indicated by the *in vivo* test in rodent bioassay, these compounds are either noncarcinogens (1- chlorobutane, bromomethane) or low-potency carcinogens (chloroethane, bromoethane). According to *in vivo* tests, chloroethane and alkyl bromides seem to be nongenotoxic carcinogens rather than genotoxic carcinogens. Based on the available data, the United States environmental protection agency (USEPA), considers tert-butyl chloride to be a group D compound or ''not classifiable as to human carcinogenicity'' (Bercu *et al.*, 2009; Snodin, 2010).

## **3.3 Hydrazines**

392 Toxicity and Drug Testing

Gocke *et al.* (2009a) reviewed both *in vivo* and *in vitro* genotoxicity, carcinogenicity, general toxicity, and the effects on reproductive and embryo fetal development of EMS. They reported that the genotoxic effects induced by EMS were observed in viruses/phages, bacteria, fungi, plant, insect, and mammalian cells. In another study, the induction of gene mutations at the hprt locus and the induction of chromosomal damage were examined as evidenced by the formation of micronuclei in human lymphoblastoid cells. It was found that the lowest dose inducing a positive response was 1.40 g/ml, and a no observed effect level (NOEL) could be defined at 1.2 g/ml. Also, no toxicity was observed at doses up to 2.5 g/plate. This observation is in strong contrast to the largely linear dose–response observed in the previous studies. As a result of *in vivo* assays for the induction of DNA damage, EMS is distributed rather uniformly over the body and induces similar levels of DNA damage in the various organs. Also, EMS is clastogenic in all test systems. The minimal dose of EMS applied in these studies was either 50 mg/kg or 100 mg/kg. In the majority of studies the dose–response relationships appeared sub linear and a threshold below 50 mg/kg appeared possible. Gocke *et al.* (2009a) demonstrated that EMS in various gene mutation tests such as induction of hprt, lacZ, and dlb-1 mutations in mice was mutagenic. The carcinogenicity of EMS was confirmed in several animal models. In another study, three methanesulfonates and three benzenesulfonates were tested by micronucleus and Yeast deletion recombination (DEL) assays. It was observed that all six substances produced positive responses in the tests

Owing to their electrophilic nature, alkylating agents can introduce lesions at nucleophilic centers of DNA. Drug salt formation includes strong acid/base interactions in the presence of alcohols, and can form impurities such as alkyl halides. As salt formation is a common method in drug formulation processes, alkyl halides exist as impurities in several drugs

The nucleophilic attack mechanisms of alkylating compounds determine their reactivity against DNA. The SN1 mechanism leads to *O*-alkylation (*O-6*-methylguanine) which is mutagenic but not clastogenic, whereas the SN2 mechanism leads to N-methylation which is clastogenic and not mutagenic. In this group, it seems that bromo compounds are more

Various tests have been performed to study DNA damage and mutation in alkyl halides. In the Ames test, it was found that most alkyl halides, especially bromides, are Ames positive except 1-chloropropane, 1-chlorobutane, and neopentyl bromide. As chloro- and bromobenzene are not alkylating agents, these compounds are Ames-negative. In Yeast deletion recombination (DEL) and micronucleus assays, alkyl chlorides such as *n*-propyl

It was observed that alkyl chlorides in the NBP [4-(*p* nitrobenzyl) pyridine] alkylation assay are not reactive and that allyl chloride has minimal activity. Although benzyl chloride is more active than other chloro compounds, ethyl, propyl, or butyl bromides have at least 1/40 MMS activity; however, allyl bromide appears to be more active (around one-eighth of

reactive as compared to chloro compounds (Sobol *et al.*, 2007; Snodin, 2010).

chloride are found to be negative (Sobol *et al.*, 2007; Snodin, 2010).

(Sobol *et al.*, 2007).

**3.2 Alkyl halides and esters** 

**3.2.1 Genotoxicity profile** 

(Sobol *et al.*, 2007; Elder *et al.*, 2008a).

the activity of MMS) (Sobol *et al.*, 2007).

Hydrazine is used as a medicine or as a starting compound for synthesizing some medicines. Hydrazine and some of its *N*-alkyl, *N*-aryl, and *N*-acyl analogues have been subjected to extensive toxicological evaluations. Hydrazines, hydrazides, and hydrazones have structural alerts for genotoxic potential and the metabolism increases their effects. Hydrazines adduct with DNA and the mechanism of adduction could include the formation of methyldiazanium ions or methyl free radicals. In addition, it seems that hydrazine reacts with endogenous formaldehyde to produce formaldehyde hydrazone. Subsequent to some other reactions, alkylating compounds like diazomethane as the genotoxic moiety are produced (Bercu *et al.*, 2009; Snodin, 2010).

#### **3.3.1 Genotoxicity profile**

*In vitro* studies have shown genotoxic effects for three hydrazine derivatives (hydrazines, hydrazides, and hydrazones). These compounds induce gene mutations in human teratoma cells, mouse lymphoma cells, and in several strains of bacteria. Hydralazine (1 hydrazinylphthalazine) and its hydrochloride salt are Ames-positive. In another study, 20 hydrazine-derivatives were found to induce a direct DNA damage in *Escherichia coli* and 16 of them (80%) were Ames positive as well (Flora *et al.*, 1984; Agency for Toxic Substance and Disease Registry, 1997; Snodin, 2010).

Although it was seen that hydrazine did not induce unscheduled DNA synthesis in mouse sperm cells, *in vivo* studies on the genotoxicity of hydrazines have largely produced positive results. In addition, it was observed that 1, 2-dimethylhydrazine failed to induce micronuclei in rat bone marrow cells, while this effect had been observed in mouse bone marrow cells (Agency for Toxic Substance and Disease Registry, 1997).

The non-carcinogenic effects of hydrazine were also evaluated; however, it was found that hydrazine, methyl hydrazine, 1,1- and 1,2-dimethylhydrazine, and other analogues are carcinogenic in rodents and possibly in human. In addition, it was seen that hydrazine derivatives like hydralazine and its hydrochloride salt were tumorigenic in rodents. It should be mentioned that the clinical use of hydralazine hydrochloride for several years has shown no evidence for carcinogenicity (Flora *et al.*, 1984; Bercu *et al.*, 2009; Snodin, 2010).

#### **3.4 Epoxides**

Epoxides are considered as electrophilic compounds owing to the strained epoxide ring. These alkylating agents directly react with DNA. Alkene oxides are more reactive than arene oxides and symmetrically substituted epoxides are less reactive than asymmetrically substituted compounds. Some examples for APIs with epoxide impurities are betamethasone acetate, atenolol, and some herbal remedies. Carbamazepine, cyproheptadine, and protriptyline have stable epoxide metabolites. In addition, phenytoin,

Genotoxic Impurities in Pharmaceuticals 395

epoxide have the potential to initiate cellular damage if not adequately detoxified via

It was observed that owing to the role of metabolism, epoxides that are formed *in vivo*, such as those generated by epoxidation of alkenes and arenes, have a greater potential to cause adverse effects than preformed epoxides. This is because they are often produced at close proximity to their site of action and can thus reach their target quite readily. Therefore, this mechanism can explain the limited evidence of animal carcinogenicity tests for some

Aromatic compounds involve various impurities; some impurities, such as fentanyl impurities, tremogenic impurities, p-nitrophenol (PNP) that have aromatic structure and

Primary and secondary aromatic amines (generally after metabolism) generate an electrophilic species and thus produce a positive result in the Ames test when S9 mixture exists. 2, 4-Diaminotoluene, 2, 4-diaminoethylbenzene and a few amines containing a nitrogroup are direct mutagens. According to the *in vivo* carcinogenicity test, Ames positive compounds produce positive results, although *p*-anisidine and *p*-chloroaniline are

This synthetic chemical possesses fungicidal activity and is used as a starting material for the synthesis of some drugs. PNP and other substituted nitro benzenes after reduction produce arylhydroxylamines or hydroxamic esters which contain electrophilic nitrogen atoms. Thus, the electrophilic atoms might show genotoxic property for these compounds

It should be mentioned that negative results were obtained for Ames tests with the various strains of *Salmonella typhimurium* in the absence and presence of metabolic activation with rat liver S9. Another *in vitro* test, the hprt mutation test in Chinese hamster ovary (CHO) cells presented the same result as the Ames test for PNP. However, it was seen that PNP could induce chromosomal aberrations in mammalian cells, particularly in the presence of metabolic activation. Also, PNP was negative in the bone marrow micronucleus assay in mice at doses ranging from little toxicity to the maximum tolerated dose. In addition, PNP was cytotoxic to the bone marrow of male mice at tested doses (Eichenbaum *et al.*, 2009).

The forced degradation of fentanyl produced seven aromatic degradants. Among these, propionanilide (PRP), N-phenyl-1-(2-phenylethyl)-piperidin-4-amine (PPA), 1-phenethyl-1H-pyridin-2-one (1-PPO), fentanyl N-oxide, and 1-styryl-1H-pyridin-2-one (1-SPO) possibly indicate safety concerns. PPA was suggested as a potential genotoxic compound and the DNA damage in unscheduled DNA synthesis (UDS); the results were positive for PRP when *in vitro* rat hepatocytes were checked. In the ACD/Tox suite, 1-PPO and 1-SPO were identified as Ames hazards. These compounds were also predicted to have higher

conjugation with glutathione (Snodin, 2010).

epoxide compounds (Flora *et al.*, 1984).

aromatic amines will be discussed in this section.

noncarcinogenic in rodent bioassays (Snodin, 2010).

probabilities of being Ames positive (Garg *et al.*, 2010).

**3.5 Aromatic compounds** 

**3.5.1 Aromatic amines** 

**3.5.2 p-Nitrophenol** 

(Eichenbaum *et al.*, 2009).

**3.5.3 Fentanyl impurities** 

lamotrigine, amitryptiline, and diclofenac tend to form reactive arene oxide metabolic intermediates (Flora *et al.*, 1984; Elder *et al.*, 2010b; Snodin, 2010).

The metabolism of epoxides mainly involves epoxide hydrolase (EH) and glutathione *S*transferase (GST), which leads to either detoxification or production of epoxides. These pathways play a key role in the genotoxic action of epoxides (Snodin, 2010).

#### **3.4.1 Genotoxicity profile**

As indicated in *in vitro* studies, epoxides are genotoxic in bacterial reverse mutation assays; however, other studies have shown different results. Hude *et al.* (1990) reported that 12/51 epoxides were nongenotoxic in the Ames *Salmonella* assay. In this study, 51 epoxides were assessed with the SOS-Chromo test using *Escherichia coli* PQ37 followed by a comparison with the results of the Ames test. All compounds were tested with and without S9 mixture up to cytotoxicity. In tests without S9 mixture the SOS-repair induction of each experiment was controlled by the response to 4-nitroquinoline-N-oxide, and in tests with S9 mixture, it was controlled with benzo[a]pyrene. In the Ames test, 20 epoxides were tested for mutagenic activity with the *Salmonella typhimurium* strains TA100, TA1535, TA98, and TA1537. By comparing the results of the Ames test and the SOS-Chromo test, it was found that among 51 epoxide-bearing chemicals 39 induced base-pair mutations in at least one Salmonella strain.

Wade *et al.* (1978) studied the mutagenicity of 17 aliphatic epoxides using the specially constructed mutants of *Salmonella typhimurium* that were developed by Ames. It was found that all the compounds in the study, with the exception of 2-methyl-3,3,3-trichloropropylene oxide, *cis-stilbene* oxide, and cyclohexene oxide that were mutagenic in strain TA100 were also mutagenic, but-with reduced sensitivity, in the second strain TA1535. However, none of the epoxides in this study were found to be mutagenic in strains TA1537 and TA98 which detect frame-shift mutagens. The results indicate that the monosubstituted epoxides are the most potent mutagens and that the addition of a single methyl group to the oxirane ring could reduce or eliminate mutagenicity.

Glatt *et al.* (1983) investigated 35 epoxides for mutagenicity, using reversion of his-*Salmonella typhimurium* TA98 and TA100 as the biological end-point. The results obtained were negative with the antibiotics oleandomycin, anticapsin and asperlin, the cardiotonic drug resibufogenin, the widely used parasympatholytic drugs butylscopolamine and scopolamine, the sedatives valtratum, didovaltratum and acevaltratum, the tranquilizer oxanamide as well as the drug metabolites carbamazepine 10,11-oxide and diethylstilbestrol α and β oxide. It was found that among the drugs and drug metabolites, only the cytostatic ethoglucide was markedly mutagenic. Three barbiturate epoxides showed very weak mutagenicity only at extremely high concentrations such that the effects were probably of low practical relevance.

Later, the role of metabolism was also examined. For example, *in vitro* studies in rat-liver S9 fractions which contain both microsomal and cytosolic detoxifying enzymes, such as EH and GST showed a decrease of bacterial genotoxicity (Flora *et al.*, 1984).

*In vivo* rodent bioassays on epoxides are not always positive and several epoxides are carcinogenic only at the point of administration. For example, it was found that when given by oral gavage, both ethylene oxide and propylene oxide caused late-onset tumors only in the rat fore-stomach. Again, when administered by inhalation, propylene oxide is a nasal carcinogen. On the other hand, *in vivo* studies in rat have shown that carbamazepine-10, 11-

lamotrigine, amitryptiline, and diclofenac tend to form reactive arene oxide metabolic

The metabolism of epoxides mainly involves epoxide hydrolase (EH) and glutathione *S*transferase (GST), which leads to either detoxification or production of epoxides. These

As indicated in *in vitro* studies, epoxides are genotoxic in bacterial reverse mutation assays; however, other studies have shown different results. Hude *et al.* (1990) reported that 12/51 epoxides were nongenotoxic in the Ames *Salmonella* assay. In this study, 51 epoxides were assessed with the SOS-Chromo test using *Escherichia coli* PQ37 followed by a comparison with the results of the Ames test. All compounds were tested with and without S9 mixture up to cytotoxicity. In tests without S9 mixture the SOS-repair induction of each experiment was controlled by the response to 4-nitroquinoline-N-oxide, and in tests with S9 mixture, it was controlled with benzo[a]pyrene. In the Ames test, 20 epoxides were tested for mutagenic activity with the *Salmonella typhimurium* strains TA100, TA1535, TA98, and TA1537. By comparing the results of the Ames test and the SOS-Chromo test, it was found that among 51 epoxide-bearing chemicals 39 induced base-pair mutations in at least one

Wade *et al.* (1978) studied the mutagenicity of 17 aliphatic epoxides using the specially constructed mutants of *Salmonella typhimurium* that were developed by Ames. It was found that all the compounds in the study, with the exception of 2-methyl-3,3,3-trichloropropylene oxide, *cis-stilbene* oxide, and cyclohexene oxide that were mutagenic in strain TA100 were also mutagenic, but-with reduced sensitivity, in the second strain TA1535. However, none of the epoxides in this study were found to be mutagenic in strains TA1537 and TA98 which detect frame-shift mutagens. The results indicate that the monosubstituted epoxides are the most potent mutagens and that the addition of a single methyl group to the oxirane ring

Glatt *et al.* (1983) investigated 35 epoxides for mutagenicity, using reversion of his-*Salmonella typhimurium* TA98 and TA100 as the biological end-point. The results obtained were negative with the antibiotics oleandomycin, anticapsin and asperlin, the cardiotonic drug resibufogenin, the widely used parasympatholytic drugs butylscopolamine and scopolamine, the sedatives valtratum, didovaltratum and acevaltratum, the tranquilizer oxanamide as well as the drug metabolites carbamazepine 10,11-oxide and diethylstilbestrol α and β oxide. It was found that among the drugs and drug metabolites, only the cytostatic ethoglucide was markedly mutagenic. Three barbiturate epoxides showed very weak mutagenicity only at extremely high concentrations such that the effects were probably of

Later, the role of metabolism was also examined. For example, *in vitro* studies in rat-liver S9 fractions which contain both microsomal and cytosolic detoxifying enzymes, such as EH

*In vivo* rodent bioassays on epoxides are not always positive and several epoxides are carcinogenic only at the point of administration. For example, it was found that when given by oral gavage, both ethylene oxide and propylene oxide caused late-onset tumors only in the rat fore-stomach. Again, when administered by inhalation, propylene oxide is a nasal carcinogen. On the other hand, *in vivo* studies in rat have shown that carbamazepine-10, 11-

and GST showed a decrease of bacterial genotoxicity (Flora *et al.*, 1984).

intermediates (Flora *et al.*, 1984; Elder *et al.*, 2010b; Snodin, 2010).

**3.4.1 Genotoxicity profile** 

Salmonella strain.

could reduce or eliminate mutagenicity.

low practical relevance.

pathways play a key role in the genotoxic action of epoxides (Snodin, 2010).

epoxide have the potential to initiate cellular damage if not adequately detoxified via conjugation with glutathione (Snodin, 2010).

It was observed that owing to the role of metabolism, epoxides that are formed *in vivo*, such as those generated by epoxidation of alkenes and arenes, have a greater potential to cause adverse effects than preformed epoxides. This is because they are often produced at close proximity to their site of action and can thus reach their target quite readily. Therefore, this mechanism can explain the limited evidence of animal carcinogenicity tests for some epoxide compounds (Flora *et al.*, 1984).

#### **3.5 Aromatic compounds**

Aromatic compounds involve various impurities; some impurities, such as fentanyl impurities, tremogenic impurities, p-nitrophenol (PNP) that have aromatic structure and aromatic amines will be discussed in this section.

#### **3.5.1 Aromatic amines**

Primary and secondary aromatic amines (generally after metabolism) generate an electrophilic species and thus produce a positive result in the Ames test when S9 mixture exists. 2, 4-Diaminotoluene, 2, 4-diaminoethylbenzene and a few amines containing a nitrogroup are direct mutagens. According to the *in vivo* carcinogenicity test, Ames positive compounds produce positive results, although *p*-anisidine and *p*-chloroaniline are noncarcinogenic in rodent bioassays (Snodin, 2010).

#### **3.5.2 p-Nitrophenol**

This synthetic chemical possesses fungicidal activity and is used as a starting material for the synthesis of some drugs. PNP and other substituted nitro benzenes after reduction produce arylhydroxylamines or hydroxamic esters which contain electrophilic nitrogen atoms. Thus, the electrophilic atoms might show genotoxic property for these compounds (Eichenbaum *et al.*, 2009).

It should be mentioned that negative results were obtained for Ames tests with the various strains of *Salmonella typhimurium* in the absence and presence of metabolic activation with rat liver S9. Another *in vitro* test, the hprt mutation test in Chinese hamster ovary (CHO) cells presented the same result as the Ames test for PNP. However, it was seen that PNP could induce chromosomal aberrations in mammalian cells, particularly in the presence of metabolic activation. Also, PNP was negative in the bone marrow micronucleus assay in mice at doses ranging from little toxicity to the maximum tolerated dose. In addition, PNP was cytotoxic to the bone marrow of male mice at tested doses (Eichenbaum *et al.*, 2009).

#### **3.5.3 Fentanyl impurities**

The forced degradation of fentanyl produced seven aromatic degradants. Among these, propionanilide (PRP), N-phenyl-1-(2-phenylethyl)-piperidin-4-amine (PPA), 1-phenethyl-1H-pyridin-2-one (1-PPO), fentanyl N-oxide, and 1-styryl-1H-pyridin-2-one (1-SPO) possibly indicate safety concerns. PPA was suggested as a potential genotoxic compound and the DNA damage in unscheduled DNA synthesis (UDS); the results were positive for PRP when *in vitro* rat hepatocytes were checked. In the ACD/Tox suite, 1-PPO and 1-SPO were identified as Ames hazards. These compounds were also predicted to have higher probabilities of being Ames positive (Garg *et al.*, 2010).

Genotoxic Impurities in Pharmaceuticals 397

N C C

As discussed above, GIs possess unwanted effects and their contamination levels should be controlled. To achieve this, pharmaceutical R&D should employ robust and sensitive analytical methods for supporting drug development and monitoring the levels of GIs. In addition, analytical methods that are capable of measuring trace GIs must be employed to monitor the outcome of GIs during chemical synthesis. In recent years, manufacturers have developed sensitive methods for analyzing various GIs. In this context, conventional HPLC/UV methods are the first option for GIs analysis; however, these methods are often inadequate for the accurate determination of analytes at trace levels, depending on the properties of the analytes and sample matrices. Some of the challenges in the analytical determination of GIs in pharmaceuticals at trace levels include the diverse structural types of GIs, the unstable or chemically reactive nature of GIs, and an extremely high level of API

In general, non-volatile GIs are analyzed by HPLC separation techniques, among which reversed phase HPLC (RPLC) is the most widely used separation mode (Elder *et al.*, 2008a; Liu *et al.*, 2010). A simple isocratic RPLC method has been employed for the determination of four genotoxic alkyl benzenesulfonates (ABSs) viz. methyl, ethyl, *n*-propyl, and *iso*propyl benzenesulfonates (MBS, EBS, NPBS, and IPBS) in amlodipine besylate (ADB). The RPLC is also applicable for sulfonate impurities with phenyl moiety such as methyl (MTs), ethyl (ETs) and *iso*propyl tosylates (ITs), methyl (MBs), ethyl (EBs), butyl (BBs) and isopropyl

Epoxides/hydroperoxides were analyzed using HPLC, and simple RPLC methods employing direct analysis (no sample preparation) were used for some of them. Yasueda *et al.* (2004) described an HPLC method for the determination of loteprednol impurities including a minor photolytic epoxide degradation product. Lacroix *et al.* (1992) reported an HPLC method for the determination of related substances, including the epoxide impurity of nadalol. A rapid resolution HPLC method was used for separating and quantifying the related impurities of atorvastatin, including two epoxide impurities atorvastatin epoxy

H

O

H N

O

CH3

CH3

N

S

<sup>H</sup> COOH

S

N

CH3

H

O

O NH

CH3

N N C

O O

CH3 CH2

Fig. 3. Structure of piperacillin impurity-A

as contaminant (Bai *et al.*, 2010; Liu *et al.*, 2010).

**4. Analytical approaches** 

**4.1 HPLC methods** 

besylates (IBs) (Raman *et al.*, 2008).

O

H
