Drug Design Structural Alert

Abstract

In the process of drug design, it is important to consider potential structural alerts that may lead to toxicosis. This work illustrates how using trifluoroethane as a part of a novel chemical entity led to cytochrome P450 – mediated N-dealkylation and the formation of trifluoroacetaldehyde, a known testicular toxicant, in exploratory safety studies in rats. Testicular toxicosis was noted microscopically in a dose-dependent manner as measured by testicular spermatocytic degeneration and necrosis and excessive intratubular cellular debris in the epididymis. This apparent toxic effect correlated well with the dose-dependent formation of trifluoroacetaldehyde, identified from in vitro rat liver microsome metabolism studies. A similar safety study performed with an N-tetrazole substitution in place of the N-trifluoroethane showed no evidence of testicular injury, implicating further the role of trifluoroacetaldehyde in the testicular lesion observed. These results highlight the relevance of early metabolic and safety testing in assessing potential structural alerts in drug design.

Introduction

The avoidance of structural alerts in drug design reduces the probability that a new drug will lead to toxicosis through a bioactivation mechanism.^1,2 At the same time, it is appreciated that simply having a structural alert within a new chemical entity does not by default lead to injury as other factors such as species differences, dose, duration of use, quantity, and detoxification of reactive metabolites, and multiple metabolic pathways play an important role in its ultimate adverse reaction profile. The purpose of this work was to highlight the need to consider the formation of a structural alert due to metabolism in the process of drug design. In particular, the use of trifluoroethane conjugated to a nitrogen, oxygen, or sulfur in a molecule can lead to the formation of trifluoracetaldehyde (TFALD) through cytochrome P450 (CYP450) N-, O-, or S-dealkylation. Trifluoracetaldehyde has been reported in the literature to cause testicular injury in the rat.^3–5 These effects were characterized by reduced testis weight and morphological changes including damage to late pachytene and dividing spermatocytes with effects increasing in severity with dose.^3–5 Similar morphological effects have been reported in the literature for other toxicants (eg, Ethylene Glycol Monomethyl Ether [EGME]) that resulted in a reduction in fertility in rats. ⁶ Being an alkyl aldehyde, TFALD is capable of covalently binding to biological macromolecules, which is generally recognized as an initial event in the etiology of many chemical-induced toxicoses. The parent compound (1) and its N-dealkylation pathway that forms TFALD is shown in Figure 1. Also shown in Figure 1 is the N-tetrazole of the same compound (2) which was tested to assess the toxic potential of the pharmacophore and the N-dealkylated fragment.

Figure 1.

 Formation of trifluoroacetaldehyde (TFALD) through cytochrome P450 – mediated N-dealkylation of parent compound 1. Also shown is the N-tetrazole analog compound 2.

Experimental Procedures

Materials

General materials used were of analytical grade or better unless indicated. Compounds 1, 2 and their common N-dealkylated fragment were synthesized by Pfizer Inc, Department of Chemistry (Chesterfield, Missouri).

In Vitro Metabolism Studies

Compound 1 at 1 μmol/L (metabolic stability) or 10 μmol/L (structure elucidation) was incubated with 1.0 mg/mL rat liver microsomes (BD Sciences, www.bdbiosciences.com) or 0.8 mg/mL human liver microsomes (BD Sciences, www.bdbiosciences.com) or recombinant CYP450 microsomes (1A2, 2C9, 2C19, 2D6, and 3A4; BD Gentest, www.bdbiosciences.com) with and without (negative control) 2 mmol/L nicotinamide adenine dinucleotide phosphate (NADPH) (Sigma, www.sigmaaldrich.com) in 1 mmol/L MgCl₂ and 100 mmol/L potassium phosphate buffer, pH 7.4, at 37°C. The reaction was quenched with 1× volume of acetonitrile. Samples were centrifuged (15 minutes at 6000 rpm) and supernatant was analyzed using liquid-chromatography -tandem mass spectrometry (LC/MS/MS). Structure elucidation samples were analyzed on a ThermoFinnigan TSQ Quantum triple quadrupole mass spectrometer with associated Agilent 1100 HPLC system (Agilent Technologies, www.chem.agilent.com). The mass spectrometer was set to collect full scan data with a range of 100 to 600 m/z with data-dependent acquisition of MS/MS data (product ion spectra). The instrument was set to acquire MS/MS spectra from parent ions selected in a 2-step process. The first priority ions were those corresponding to the most likely products of phase 1 metabolism. If none were found, the MS/MS spectra were acquired for the most abundant ion in each scan. The data set was analyzed using MetWorks software (Version 1.0, Thermo Electron Corp, San Jose, California).

In Vivo Safety Testing

The Pfizer Institutional Animal Care and Use Committee reviewed and approved the animal use in these studies. The Association for Assessment and Accreditation of Laboratory Animal Care International fully accredits the Pfizer animal care and use program.

The safety and exposure of compound 1 (and compound 2) was determined in 200 to 250 g male SD/IGS rats (Charles River, www.criver.com) following an acclimation period of 7 days. Animals were single housed in suspended stainless steel wire mesh cages at room conditions of 70°F ± 5°F, 30% to 70% humidity, and an approximate 12-hour dark/light cycle. Water and Purina 5002 Certified Rodent Diet was available ad libitum. The test article was suspended in 0.5% methylcellulose (weight/volume [w/v]) and 0.1% Tween-80 (v/v) in deionized water. Compound 1 doses of 0 (vehicle control), 10, 30, 70, 100, 200, and 300 mg/kg and compound 2 doses of 0, 100, 500, and 2000 mg/kg were administered to 5 animals/dose group by oral gavage (10 mL/kg), once daily, for 7 consecutive days. Animals were observed daily for clinical signs and weighed at pretest, days 1, 3, 5, and 7, and at necropsy. Blood samples for toxicokinetics were taken at various time points on day 7 and for hematology and blood chemistry were taken at necropsy. In addition, gross pathology, organ weights, and terminal body weights were evaluated. Samples for parent compound and metabolite exposure assessment were analyzed by LC-MS-MS (Applied Biosystems/MDS Sciex API 4000, www.appliedbiosystems.com).

At necropsy, the lungs, liver, kidney, mesentery, and mesenteric lymph nodes, heart, and spleen were immersed in 10% neutral-buffered formalin (NBF). The testes and epididymis from both sides were fixed in modified Davidson’s solution. All tissues were processed whole on a Sakura VIP 5 series by dehydrating through a series of graded ethanol solutions, cleared with xylene, and impregnated with paraffin. Sections of 4 μm thickness were cut to expose the microstructures of each organ. Sectioned tissues were heated in a 60°C oven for a minimum of 1 hour, stained via automated linear stainer with hematoxylin-eosin, and coverslipped. The processed glass slides were evaluated under an Olympus light microscope and the images were captured by a SPOT Insight Firewire Camera and analyzed by SPOT software Advanced (Diagnostic Instruments, Inc, Sterling Heights, Michigan).

Results

The in vitro intrinsic liver metabolic clearance of compound 1 was determined to be 80 mL/min per kg in rat microsomes and 3 mL/min per kg in human microsomes. The rat intrinsic clearance scaled to an in vivo rat clearance of 7 mL/min per kg with the well-stirred liver model, ⁷ which was consistent with the observed total in vivo rat clearance of 9 mL/min per kg. Human CYP450 profiling suggested that 3A4 was the predominant pathway of biotransformation. In vitro rat and human metabolite profiling of compound 1 qualitatively identified the formation of the N-dealkylated metabolite shown in Figure 1, as well as some other ring hydroxylations.

From in vivo safety studies with compound 1, the most significant microscopic finding was acute testicular germ cell degeneration and necrosis (Figure 2). There were multiple multinucleated giant cells and cellular debris in the affected seminiferous tubules of both testes. Many seminiferous tubules were either half collapsed with degenerated spermatogenic cells or lined by a layer of sertoli cells without any germ cell layers. Multifocally, the spermatogonia were pyknotic, and sertoli cells had cytoplasmic vacuoles. The interstitial cells and the connective tissue remained intact. The severity and incidence of the testicular lesions were clearly dose-dependent (Table 1). At the 70 mg/kg dose level, the testicular lesions were minimal and the incidence was low in only 2 of the 5 rats tested. Conversely, at the 300 mg/kg dose level, all 5 rats had minimal-to-moderate testicular lesions.

Figure 2.

 Representative histology of the rat testes in vehicle control (A) and compound 1-treated animals (B; 200 mg/kg dose group), indicating acute germ cell degeneration and necrosis (arrows). Magnification, ×200.

Table 1

 Systemic Exposure Data and Testicular Scoring in the Rat Following 7 Days of Dosing of Compound 1 and Compound 2

PO Dose Group	Mean C max (±SD) (μmol/L)	Mean AUC (±SD) (μmol/L h)	Testicular Spermatocytic Degeneration and Necrosis	Epididymis, Excessive Intratubular Cell Debris
Compound 1
Vehicle control, 0 mg/kg	ND	ND	0/5	0/5
10 mg/kg	5.2 (±0.9)	32.1 (±6.9)	0/5	0/5
30 mg/kg	7.9 (±2.7)	57.6 (±11.1)	0/5	0/5
70 mg/kg	31.9 (±8.6)	290 (±84)	2/5 Minimal	1/5 Minimal
100 mg/kg	44.7 (±15.3)	346 (±111)	2/5 Minimal to mild	2/5 Minimal
200 mg/kg	50.8 (±15.4)	419 (±84)	5/5 Minimal to mild	4/5 Minimal to mild
300 mg/kg	56.3 (±17.1)	576 (±149)	5/5 Minimal to moderate	5/5 Minimal to mild
Compound 2
Vehicle control, 0 mg/kg	ND	ND	0/5	0/5
100 mg/kg	27.0 (±3.5)	380 (±56)	0/5	0/5
500 mg/kg	28.3 (±3.8)	409 (±87)	0/5	0/5
2000 mg/kg	34.3 (±2.4)	540 (±47)	0/5	0/5

Abbreviations: C _max, maximum plasma concentration; AUC, area under the curve; SD, standard deviation; ND, compound not detectable; n = 5 animals/dose group.

In most of the rats with testicular lesions, the ductuli efferentes of the epididymis contained excessive multinucleated giant cells, degenerated cellular debris, and sperm debris (Figure 3). The level of cellular debris was consistent with the severity of the testicular lesion. In some cases, the body of the ductus epididymis had cellular debris but rarely did the tail of the epididymis have degenerative cells and cellular debris. All of the lining epithelial cells and the interstitium remained intact. The degenerated germ cells and cellular debris in the epididymis occurred in the rats dosed at both 200 and 300 mg/kg levels, which correlated to the severity of their testicular lesions.

Figure 3.

 Representative histology of the rat epididymis of vehicle control (A; normal) and compound 1 treatment (B; 200 mg/kg dose group). Arrows indicate intratubular spermatocytic debris and degenerated giant cells. Magnification, ×200.

There were no significant differences in body weights, hematology, blood chemistry, organ weights, and terminal body weights among the dosed and control groups. The clinical observations and gross pathology had no abnormal findings. Microscopic findings were limited to the testes and epididymis. All other organs evaluated were within normal limits.

Plasma exposures, in terms of maximum concentrations and area under the curve of the parent compound 1, are shown in Figure 4. The plasma exposures of the N-dealkylated fragment are shown in Figure 5. As TFALD is a difficult molecule to assess due to its high volatility and instability, the relative concentrations of TFALD were assessed based on the level of the N-dealkylated fragment that was formed, knowing that the molar ratio of formation is one to one. The actual testis concentration of TFALD is not known as the compound is expected to undergo further oxidation to trifluoroacetic acid. ⁵ Nevertheless, relative tissue concentrations of TFALD between dose groups should be reflected by its plasma level profile.

Figure 5.

 Relative exposure levels in the plasma from individual animals (C _max indicates maximal concentration; AUC, area under the curve 0-24 hours) of trifluoroacetaldehyde based on the assessment of the N-dealkylated fragment of the parent compound 1. Center bars represent the mean value; T bars are 1 standard deviation.

Figure 4.

 Parent compound 1 exposure levels in the plasma (C _max indicates maximal concentration; AUC, area under the curve 0-24 hours) from individual animals in the in vivo safety studies. Center bars represent the mean value; T bars are 1 standard deviation.

Discussion

Correlation of microscopic findings to parent compound 1 and metabolite exposures suggests a significant role for N-dealkylated products in testicular degeneration in the rat. In particular, the increase in testicular severity scoring tracked with the dose-proportional exposures to TFALD. Conversely, exposures of parent compound overlap significantly at doses between 100 and 300 mg/kg, in spite of the continued increase in the severity of testicular findings with dose.

The toxicokinetic/toxicodynamic correlation of compound 1 was better assessed by setting a composite severity score for the microscopic findings from Table 1 for both the testicular and the epididymis findings in the following manner: (i) animal incidence score as: 0/5 = 1, 1/5 = 2, 2/5 = 3,…, 5/5 = 6; (ii) severity score as: none = 1, minimal = 2, minimal to mild = 3, and minimal to moderate = 4; and (iii) multiplication of both testicular and epididymis animal incidence and severity score numbers together to obtain a composite severity score, for example the 70 mg/kg dose group yields a composite severity score of 3 × 2 × 2 × 2 = 24. Plots of the composite score versus area under the curve of the parent and fragment are shown in Figure 6. Linear regression of the data indicated the best correlation coefficient using the fragment area under the curve (r ² = .98) versus that of the parent (r ² = .74).

Figure 6.

 Toxicokinetic/toxicodynamic correlations of the composite histological severity score and area under the curve (AUC) of compound 1 and trifluoroacetaldehyde (TFALD), respectively.

Considering the literature precedence for the connection between TFALD and testicular lesions, the initial assessment of the toxicology findings of compound 1 suggested the role of TFALD. However, potential toxicity of either the pharmacophore or the N-dealkylated fragment could not be ruled out at that time. To test this concept, further repeat dose toxicology studies in rats were performed with compound 2, the N-tetrazole analog (Figure 1), which also undergoes N-dealkylation in vivo. The results indicated no testicular lesions, even at similar plasma exposures (Table 1), suggesting that the original testicular lesions were not related to the pharmacophore or its N-dealkylated fragment.

It is apparent that the there are (previously) marketed drugs which contain the N-, O-, or S-trifluorethane fragment as a part of their chemical structure such as halazepam, quazepam, lansoprazole, epitizid, and flecainide. However, only 1 of these compounds has been reported to result in testicular effects in rodents (lansoprazole). Presumably, these findings emphasize that the overall safety profile of a molecule is determined not only by the presence of a structural alert but also with consideration of species differences, dose, duration of use, detoxification of reactive metabolites, and the presence of multiple metabolic pathways.

In summary, this work emphasizes the importance of understanding overall risks as early as is feasible in drug discovery and development, in particular as it relates to the use of trifluoroethane linked to N, O, or S within novel chemical entities. Early metabolic and safety testing can be useful in this regard, as was illustrated by this work. While it is appreciated that a number of considerations go into an overall risk assessment, the authors recommend that the use of this fragment be carefully considered in a risk-benefit paradigm and/or avoided when located in a region of a molecule that may be subject to dealkylation.