**A Comparison Between Lignans from Creosote Bush and Flaxseed and Their Potential to Inhibit Cytochrome P450 Enzyme Activity**

Jennifer Billinsky, Katherine Maloney, Ed Krol and Jane Alcorn *Drug Design and Discovery Research Group, College of Pharmacy and Nutrition, University of Saskatchewan Canada* 

## **1. Introduction**

144 Drug Discovery Research in Pharmacognosy

Woelkar, K.; Xu, W.; Pei, Y.; Makriyannis, A.; Picone, R.P. & Bauer, R. (2005). The

Wu, L.C.; Fan, N.C.; Lin, M.H.; Chu, I.R.; Huang, S.J.; Hu, C.Y. & Han, S.Y. (2008). Anti-

Wu, T.-S..; Chang, F.-C. & Wu, P.-L. (1995). Flavonoids, amidosulfoxides and an alkaloid

Yue, J.M.; Xu, J.; Zhao, Y.; Sun, H.D. & Lin, Z.W. (1997). Chemical components from

Zacares, L.; Lopez-Gresa, M.P.; Fayos, J.; Primo, J.; Belles, J.M. & Conejero, V. (2007).

Zhao, J.; Muhammad, I.; Dunbar, D.Ch.; Mustafa, J. & Khan, I.A. (2005). New Alkamides

doi:10.1016/S0305-1978(00)00030-2

pp. 56–59, doi:10.1248/cpb.53.56

pp. 1453-1457, doi:10.1016/0031-9422(95)00171-3

No.3, (January 2005), pp. 690−693, doi:10.1021/jf048529t

1997), pp. 1031–1033, doi:10.1021/np97004

doi:10.1021/jf073057e

1439

*Systematics and Ecology*, Vol.29, No.2, (February 2001), pp. 209–211,

endocannabinoid system as a target for alkamides from *Echinaceae angustifolia* roots. *Planta Medica*, Vol.71, No.8, (August 2005), pp. 701-705, doi:10.1055/s-2005-871290 Wu, P.L.; Wu, T.S.; He, C.X.; Su, C.H. & Lee, K.H. (2005). Constituents from the stems of

*Hibiscus taiwanensis*. *Chemical & Pharmaceutical Bulletin*, Vol.53, No.1, (January 2005),

inflammatory effect of spilanthol from *Spilanthes acmella* on murine macrophage by down-regulating LPS-induced inflammatory mediators. *Journal of the Agricultural and Food Chemistry*, Vol.56, No.7, (March 2008), pp. 2341–2349,

from the leaves of *Glycosmis citrifolia*. *Phytochemistry*, Vol.39, No.6, (August 1995),

*Ceratostigma willmottianum*. *Journal of Natural Products*, Vol.60, No.10, (October

Induction of *p*-coumaroyldopamine and feruloyldopamine, two novel metabolites, in tomato by the bacterial pathogen *Pseudomonas syringae*. *Mol. Plant-Microbe Interact.*, Vol.20, No.11, (November 2007), pp. 1439–1448, doi:10.1094/MPMI-20-11-

from Maca (*Lepidium meyenii*). *Journal of the Agricultural and Food Chem*istry, Vol.53,

The popularity in natural product use we witness today arose from a growing public skepticism about taking "pharmaceutical chemicals" to treat illness. Such skepticism was supplanted by a public perception that "medications" from natural sources are safer to use and have similar efficacies as their pharmaceutical equivalents. The diverse assortment of natural products on the shelves of pharmacies, health food stores, and grocery stores attest to this enhanced public demand, but has compelled regulatory agencies to question the adequacy of the safety and efficacy data associated with the use of these products (Natural Health Products Directorate 2007). In the current regulatory environment, full realization of the wellness and therapeutic value of these natural products can only come about with more rigorous assessments of their safety and efficacy.

This is particularly true of natural products that contain lignans as the principal bioactive component. Interest in lignans continues to grow due to an increased awareness of their putative health benefits. One such product, Chaparral, contains lignans extracted from creosote bush. Creosote bush had centuries of traditional use by aboriginal peoples of the Southwestern United States as an effective natural medicine and was marketed as an extract of the plant in capsule form based on this historical medicinal value (Clark & Reed 1992). While traditional creosote bush use appears to be quite safe, chronic use of Chaparral led to reports of toxicity (Clark & Reed 1992; Gordon et al., 1995; Batchelor et al., 1995; Grant et al., 1998). Nordihydroguaiaretic acid (NDGA) is the major lignan in creosote bush, which is believed to be responsible for both the efficacious and toxic properties of Chaparral (Grice et al., 1968; Moore 1989; Arteaga et al., 2005; Lambert et al., 2004). Wagner and Lewis 1980 previously reported that NDGA undergoes oxidation to "activated NDGA" (Wagner & Lewis 1980). Billinsky et al. 2008 suggest this "activated NDGA" is the result of an autoxidation process to a stable dibenzocyclooctadiene product of NDGA (Billinsky & Krol 2008). Whether this dibenzocyclooctadiene is present in traditional creosote bush products or is formed *in vivo* is not known.

The recent popularity of lignans from flaxseed, which are currently marketed as concentrated extracts of the principal plant lignan, secoisolariciresinol diglucoside (SDG), in

A Comparison Between Lignans from Creosote Bush and

(Fowler & Zhang 2008).

Flaxseed and Their Potential to Inhibit Cytochrome P450 Enzyme Activity 147

body and mediate the elimination of a wide variety of drugs, phytochemicals, and environmental toxicants. Consequently, interactions involving P450 enzymes are widely reported in the literature. Such P450 enzyme interactions often involve P450 inhibition, which may proceed either through mechanism-based, irreversible interaction with a P450 enzyme or by a reversible competitive interaction between two substrates for the same P450. Both inhibition mechanisms represent common underlying mechanisms for toxicity associated with pharmaceutical products, phytochemicals and environmental toxicants

Mechanism-based, irreversible inhibition of P450 usually follows from bioactivation by P450 to a reactive metabolite that, in turn, covalently binds to the enzyme (heme or apoprotein) and prevents further P450 activity (Kalgutkar et al., 2007). Enzyme function is only restored with synthesis of new enzyme. Alternatively, the reactive metabolite may bind to other macromolecules (i.e. protein, nucleic acid) to inhibit their function or act as a hapten leading to an immune response with toxicological outcomes (Kalgutkar et al., 2007). Hence, P450 mediated bioactivation to reactive metabolites can represent an important mechanism of toxicity associated with natural product use. Reversible competitive interactions at the same P450 can also result in toxicity when natural products are present simultaneously with other chemicals whose metabolic clearance is predominantly dependent upon a particular P450 enzyme (Fowler & Zhang 2008). Competitive interactions, then, may lead to the accumulation of a compound and eventual toxicity. Given the potential for significant adverse outcomes, evaluation of P450 inhibition is commonly investigated during standard

safety assessments for drugs, environmental contaminants, and phytochemicals.

of natural products containing lignan as the principal bioactive component.

**2. Cytochrome P450-mediated formation of reactive metabolites** 

Cytochrome P450 enzyme interactions are commonly touted as a mechanism of toxicity (Fowler & Zhang 2008; Stresser et al., 2000). In particular, bioactivation, or the metabolic activation of a compound to an electrophilic reactive intermediate, which subsequently undergoes covalent binding to critical cellular macromolecules and interferes with their

Our purpose was to determine whether mechanism-based, irreversible inhibition and/or reversible, competitive P450 inhibition might explain the differences in apparent toxicity between oral consumption of naturally occurring lignans from creosote bush, which have known toxicity, and flaxseed, which show relative safety. Additionally, such information is important to determine whether lignans present a significant concern with respect to their potential for P450-mediated interactions. As a proof-of-concept study, rat hepatic microsomal systems were employed to investigate the hypothesis that lignan from creosote bush undergoes P450-enzyme mediated bioactivation to form reactive metabolites while flaxseed lignans do not. Furthermore, we hypothesized that lignan from creosote bush inhibits P450 enzyme activity via mechanism-based and/or reversible inhibition mechanisms and that glutathione (GSH) can attenuate the inhibition. The use of rat also supports current investigations in our laboratories concerning lignan efficacy and pharmacokinetics in rat models of hyperlipidemia and hyperglycemia (Felmlee et al., 2009). Such investigations will make an invaluable contribution to our understanding of the safety

products such as Brevail™ and Beneflax™, arises from recent promising clinical trial evidence of their chemopreventive and therapeutic properties for a variety of chronic diseases (Pan et al., 2007; Hallund et al., 2008; Zhang et al., 2008). Following oral consumption, the glucose groups of SDG are cleaved to form the aglycone, secoisolariciresinol (SECO), which is further metabolized to the mammalian lignans, enterodiol (ED) and enterolactone (EL), by colonic bacteria (Rickard et al., 1996; Borriello et al., 1985a). Flaxseed also contains smaller amounts of other lignans such as mataresinol, and lariciresinol. The presence of anhydrosecoisolariciresinol (ASECO) was reported in flaxseed (Charlet et al., 2002). However, its actual presence is questioned as the acid hydrolysis method used to extract and quantify lignans from flaxseed can convert SECO into ASECO (Mazur et al., 1996). Which lignan form mediates the putative health effects is not known, but little evidence of toxicity exists with their use, and most clinical trial data identifies their relative safety. Interestingly, a comparison of the structure of NDGA from creosote bush with the flaxseed lignan, SDG and its aglycone form, SECO, identifies remarkable structurally similarity, yet with obvious differences in their safety profiles (Figure 1).

Fig. 1. Structures of lignans derived from Flaxseed1 and Creosote bush2.

Why lignan in chaparral is associated with toxicity while lignans from flaxseed have limited reported toxicity may relate to differences in their ability to inhibit cytochrome P450 (P450) enzymes. Cytochrome P450 enzymes are the principal detoxification mechanisms of the

products such as Brevail™ and Beneflax™, arises from recent promising clinical trial evidence of their chemopreventive and therapeutic properties for a variety of chronic diseases (Pan et al., 2007; Hallund et al., 2008; Zhang et al., 2008). Following oral consumption, the glucose groups of SDG are cleaved to form the aglycone, secoisolariciresinol (SECO), which is further metabolized to the mammalian lignans, enterodiol (ED) and enterolactone (EL), by colonic bacteria (Rickard et al., 1996; Borriello et al., 1985a). Flaxseed also contains smaller amounts of other lignans such as mataresinol, and lariciresinol. The presence of anhydrosecoisolariciresinol (ASECO) was reported in flaxseed (Charlet et al., 2002). However, its actual presence is questioned as the acid hydrolysis method used to extract and quantify lignans from flaxseed can convert SECO into ASECO (Mazur et al., 1996). Which lignan form mediates the putative health effects is not known, but little evidence of toxicity exists with their use, and most clinical trial data identifies their relative safety. Interestingly, a comparison of the structure of NDGA from creosote bush with the flaxseed lignan, SDG and its aglycone form, SECO, identifies remarkable

structurally similarity, yet with obvious differences in their safety profiles (Figure 1).

H3CO

OH OH

OH

OH

OH OH

OH OH

HO

HO

Why lignan in chaparral is associated with toxicity while lignans from flaxseed have limited reported toxicity may relate to differences in their ability to inhibit cytochrome P450 (P450) enzymes. Cytochrome P450 enzymes are the principal detoxification mechanisms of the

OH Enterodiol<sup>1</sup>

Fig. 1. Structures of lignans derived from Flaxseed1 and Creosote bush2.

O

OCH3

HO

Anhydrosecoisolariciresinol<sup>1</sup>

OH

Secoisolariciresinol Diglucoside<sup>1</sup>

OCH3

O O

<sup>O</sup> HO

<sup>O</sup> HO

OH

H3CO

HO

H3CO

HO

Enterolactone1

OH

HO

HO

HO

HO

Dibenzocyclooctadiene<sup>2</sup>

O O

HO

OH

OH Nordihydroguaiaretic acid<sup>2</sup>

OH Secoisolariciresinol<sup>1</sup>

OH

H3CO

HO

OH <sup>O</sup>

OH

**Lariciresinol1**

OCH3

OCH3

HO OH

body and mediate the elimination of a wide variety of drugs, phytochemicals, and environmental toxicants. Consequently, interactions involving P450 enzymes are widely reported in the literature. Such P450 enzyme interactions often involve P450 inhibition, which may proceed either through mechanism-based, irreversible interaction with a P450 enzyme or by a reversible competitive interaction between two substrates for the same P450. Both inhibition mechanisms represent common underlying mechanisms for toxicity associated with pharmaceutical products, phytochemicals and environmental toxicants (Fowler & Zhang 2008).

Mechanism-based, irreversible inhibition of P450 usually follows from bioactivation by P450 to a reactive metabolite that, in turn, covalently binds to the enzyme (heme or apoprotein) and prevents further P450 activity (Kalgutkar et al., 2007). Enzyme function is only restored with synthesis of new enzyme. Alternatively, the reactive metabolite may bind to other macromolecules (i.e. protein, nucleic acid) to inhibit their function or act as a hapten leading to an immune response with toxicological outcomes (Kalgutkar et al., 2007). Hence, P450 mediated bioactivation to reactive metabolites can represent an important mechanism of toxicity associated with natural product use. Reversible competitive interactions at the same P450 can also result in toxicity when natural products are present simultaneously with other chemicals whose metabolic clearance is predominantly dependent upon a particular P450 enzyme (Fowler & Zhang 2008). Competitive interactions, then, may lead to the accumulation of a compound and eventual toxicity. Given the potential for significant adverse outcomes, evaluation of P450 inhibition is commonly investigated during standard safety assessments for drugs, environmental contaminants, and phytochemicals.

Our purpose was to determine whether mechanism-based, irreversible inhibition and/or reversible, competitive P450 inhibition might explain the differences in apparent toxicity between oral consumption of naturally occurring lignans from creosote bush, which have known toxicity, and flaxseed, which show relative safety. Additionally, such information is important to determine whether lignans present a significant concern with respect to their potential for P450-mediated interactions. As a proof-of-concept study, rat hepatic microsomal systems were employed to investigate the hypothesis that lignan from creosote bush undergoes P450-enzyme mediated bioactivation to form reactive metabolites while flaxseed lignans do not. Furthermore, we hypothesized that lignan from creosote bush inhibits P450 enzyme activity via mechanism-based and/or reversible inhibition mechanisms and that glutathione (GSH) can attenuate the inhibition. The use of rat also supports current investigations in our laboratories concerning lignan efficacy and pharmacokinetics in rat models of hyperlipidemia and hyperglycemia (Felmlee et al., 2009). Such investigations will make an invaluable contribution to our understanding of the safety of natural products containing lignan as the principal bioactive component.

## **2. Cytochrome P450-mediated formation of reactive metabolites**

Cytochrome P450 enzyme interactions are commonly touted as a mechanism of toxicity (Fowler & Zhang 2008; Stresser et al., 2000). In particular, bioactivation, or the metabolic activation of a compound to an electrophilic reactive intermediate, which subsequently undergoes covalent binding to critical cellular macromolecules and interferes with their

A Comparison Between Lignans from Creosote Bush and

(Nesbitt et al., 1999; Axelson et al., 1982d).

was not identified in our studies.

incubated with ED for 20 min.

only 36.2 ± 7.26%, 65.7 ± 8.36% and 64.4 ± 6.36%, respectively.

lignan concentration.

Flaxseed and Their Potential to Inhibit Cytochrome P450 Enzyme Activity 149

We examined the ability of creosote bush and flaxseed lignans to inhibit the activity of CYP3A, CYP2C and CYP1A2 enzymes as these isoforms are principally involved in the metabolism of drugs and other natural products on the market today. To determine whether lignans result in mechanism-based inhibition of P450 enzymes initial experiments examined the extent of inhibition as a function of pre-incubation time and

SDG did not cause mechanism-based inhibition of any P450 isoform at any concentration tested. The data also suggests that SDG does not cause reversible inhibition of P450 isoforms, as metabolite formation remained the same at pre-incubation time 0 at all SDG concentrations tested. This result was not unexpected given the physicochemical characteristics of SDG. Hydrophilic molecules tend not to be substrates for P450 enzymes and are primarily eliminated from the body via renal elimination mechanisms. Furthermore, studies have failed to detect systemic levels of SDG (Kuijsten et al., 2005b), suggesting that it is not absorbed from the gastrointestinal tract or undergoes extensive first pass metabolism. Studies further suggest that this first-pass metabolism occurs at the gastrointestinal level, where the glucosidic groups of SDG are cleaved to produce the aglycone, SECO (Thompson et al., 1991; Borriello et al., 1985c). SECO is likely the flaxseed lignan form available for absorption following the oral consumption of flaxseed products

SECO did not cause significant inhibition of CYP1A2. SECO inhibited CYP3A in a concentration-dependent but not time-dependent manner suggesting that SECO is a reversible inhibitor of the three isoforms. Reversible inhibition is consistent with the lack of GSH adduct formation following microsomal incubation of SECO in the presence of GSH. However, at 2000 μM SECO inhibited 6β-, 16α- and 2α-hydroxytestosterone formation by

ASECO caused limited inhibition of CYP1A2, 2B, and 2C11, with moderate inhibition of CYP3A at the highest concentration (100 µM) resulting in 49.2 ± 1.8% inhibition after 20 minutes preincubation. As with SDG and SECO, ASECO showed concentration-dependent but not time-dependant inhibition which suggests P450 inhibition by reversible mechanisms. These results are interesting, as in the presence of GSH, we observed an ASECO adduct that was consistent with *para-*quinone methide formation (Bolton et al., 1994; Awad et al., 2002). Despite this observation, mechanism-based inhibition of P450 activity

Enterodiol (ED) failed to inhibit P450 activity and rather caused activation of CYP1A2, 3A, 2B and 2C11 activity particularly at shorter preincubation times (Figure 2). Although an apparent inhibition of CYP3A activity was observed after a 20 min pre-incubation (Figure 2b), control samples without ED showed a large %CV of 91.8% relative to the samples pre-

EL caused concentration-dependent, but not time-dependent, inhibition of CYP1A2 and only at high concentrations with 500 µM EL inhibiting CYP1A2 to 14.3 ± 23% that of control (data not shown). EL also caused concentration-dependent inhibition of CYP3A.

**3. Mechanism-based inhibition of cytochrome P450 enzymes** 

function, has long-standing recognition as a biochemical mechanism of organ toxicity (Miller & Miller 1947; Mitchell et al., 1973; Masubuchi et al., 2007). Several examples exist of natural products undergoing P450-mediated bioactivation and irreversible P450 enzyme inhibition and hepatotoxicity (Johnson et al., 2003; Zhou et al., 2004; Surh & Lee 1995; He et al., 1998; Kent et al., 2002). Furthermore, natural products are known to result in significant interactions with coadministered therapeutic agents resulting in adverse effects or therapeutic failures (Dietz & Bolton 2007; Yuan et al., 2004; Kupiec & Raj 2005). With the growing interest in lignans for their health and wellness or therapeutic values, assessments of their safety are necessary to support their use.

Structural features of the lignans suggest their potential for oxidation to quinone derivatives, a class of compounds known to be electrophilic reactive intermediates (Bolton et al., 1994; Iverson et al., 1995). Since SDG is unlikely absorbed due to the polar nature of the glycosidic groups, it was not investigated for their potential to undergo hepatic P450 mediated bioactivation to a reactive intermediate. Previous studies in our laboratory identified NDGA's (the major lignan of creosote bush) ability to undergo autoxidation to a reactive quinone species (Billinsky et al., 2007; Billinsky & Krol 2008). Billinsky et al., 2007 had previously reported that NDGA underwent autoxidation and conversion via an *ortho*quinone reactive intermediate to three glutathione (GSH) adducts (Billinsky et al., 2007). Furthermore, in the absence of GSH NDGA autoxidation resulted in the formation of a novel dibenzocyclooctadiene (Billinsky & Krol 2008). The autoxidation of NDGA to a reactive intermediate in rat hepatic microsomes (Billinsky et al., 2007) did not result in the inactivation of P450 enzyme activity. These findings are not unexpected as the literature suggests *ortho*-quinones do not usually undergo adduct formation with cellular macromolecules, instead their damage is usually mediated through an increase in oxidative stress (Chichirau et al., 2005; O'Brien 1991; Monks et al., 1992). Although *ortho*-quinones may isomerize to *para-*quinone methides *in vivo*, a quinone derivative known to form covalent adducts with cellular macromolecules (Powis 1987), this process is unlikely to occur as NDGA is substituted at the benzylic carbon, which severely hinders the isomerization process (Iverson et al., 1995).

Current studies indicate that SECO (the aglycone form of the major flaxseed lignan, SDG) underwent conversion to lariciresinol, both in the presence and absence of GSH, without formation of a stable reactive quinone intermediate. No GSH adducts were observed following microsomal incubations. ASECO was also examined as its structure is intermediate of NDGA and SECO. For ASECO, all control experiments showed no product formation; however, in the presence of GSH a product peak in MS-ESI(-) with a m/z 648.19 [M-H]- was observed. This mass is consistent with the formation of an ASECO-GSH adduct, which is likely derived from an ASECO *para-*quinone methide.

These studies, in conjunction with the growing interest in their use, compelled an investigation into their potential for mechanism-based inhibition and/or reversible inhibition of P450 isoforms. The use of *in vitro* microsomal systems for assessments of cytochrome P450 inhibition potential by drug candidates and natural products is a commonly employed technique to determine the likelihood for pharmacokinetic interactions resulting from P450 enzyme inhibition (Stresser et al., 2000).

function, has long-standing recognition as a biochemical mechanism of organ toxicity (Miller & Miller 1947; Mitchell et al., 1973; Masubuchi et al., 2007). Several examples exist of natural products undergoing P450-mediated bioactivation and irreversible P450 enzyme inhibition and hepatotoxicity (Johnson et al., 2003; Zhou et al., 2004; Surh & Lee 1995; He et al., 1998; Kent et al., 2002). Furthermore, natural products are known to result in significant interactions with coadministered therapeutic agents resulting in adverse effects or therapeutic failures (Dietz & Bolton 2007; Yuan et al., 2004; Kupiec & Raj 2005). With the growing interest in lignans for their health and wellness or therapeutic values, assessments

Structural features of the lignans suggest their potential for oxidation to quinone derivatives, a class of compounds known to be electrophilic reactive intermediates (Bolton et al., 1994; Iverson et al., 1995). Since SDG is unlikely absorbed due to the polar nature of the glycosidic groups, it was not investigated for their potential to undergo hepatic P450 mediated bioactivation to a reactive intermediate. Previous studies in our laboratory identified NDGA's (the major lignan of creosote bush) ability to undergo autoxidation to a reactive quinone species (Billinsky et al., 2007; Billinsky & Krol 2008). Billinsky et al., 2007 had previously reported that NDGA underwent autoxidation and conversion via an *ortho*quinone reactive intermediate to three glutathione (GSH) adducts (Billinsky et al., 2007). Furthermore, in the absence of GSH NDGA autoxidation resulted in the formation of a novel dibenzocyclooctadiene (Billinsky & Krol 2008). The autoxidation of NDGA to a reactive intermediate in rat hepatic microsomes (Billinsky et al., 2007) did not result in the inactivation of P450 enzyme activity. These findings are not unexpected as the literature suggests *ortho*-quinones do not usually undergo adduct formation with cellular macromolecules, instead their damage is usually mediated through an increase in oxidative stress (Chichirau et al., 2005; O'Brien 1991; Monks et al., 1992). Although *ortho*-quinones may isomerize to *para-*quinone methides *in vivo*, a quinone derivative known to form covalent adducts with cellular macromolecules (Powis 1987), this process is unlikely to occur as NDGA is substituted at the benzylic carbon, which severely hinders the isomerization

Current studies indicate that SECO (the aglycone form of the major flaxseed lignan, SDG) underwent conversion to lariciresinol, both in the presence and absence of GSH, without formation of a stable reactive quinone intermediate. No GSH adducts were observed following microsomal incubations. ASECO was also examined as its structure is intermediate of NDGA and SECO. For ASECO, all control experiments showed no product formation; however, in the presence of GSH a product peak in MS-ESI(-) with a m/z 648.19 [M-H]- was observed. This mass is consistent with the formation of an ASECO-GSH adduct,

These studies, in conjunction with the growing interest in their use, compelled an investigation into their potential for mechanism-based inhibition and/or reversible inhibition of P450 isoforms. The use of *in vitro* microsomal systems for assessments of cytochrome P450 inhibition potential by drug candidates and natural products is a commonly employed technique to determine the likelihood for pharmacokinetic interactions

which is likely derived from an ASECO *para-*quinone methide.

resulting from P450 enzyme inhibition (Stresser et al., 2000).

of their safety are necessary to support their use.

process (Iverson et al., 1995).

## **3. Mechanism-based inhibition of cytochrome P450 enzymes**

We examined the ability of creosote bush and flaxseed lignans to inhibit the activity of CYP3A, CYP2C and CYP1A2 enzymes as these isoforms are principally involved in the metabolism of drugs and other natural products on the market today. To determine whether lignans result in mechanism-based inhibition of P450 enzymes initial experiments examined the extent of inhibition as a function of pre-incubation time and lignan concentration.

SDG did not cause mechanism-based inhibition of any P450 isoform at any concentration tested. The data also suggests that SDG does not cause reversible inhibition of P450 isoforms, as metabolite formation remained the same at pre-incubation time 0 at all SDG concentrations tested. This result was not unexpected given the physicochemical characteristics of SDG. Hydrophilic molecules tend not to be substrates for P450 enzymes and are primarily eliminated from the body via renal elimination mechanisms. Furthermore, studies have failed to detect systemic levels of SDG (Kuijsten et al., 2005b), suggesting that it is not absorbed from the gastrointestinal tract or undergoes extensive first pass metabolism. Studies further suggest that this first-pass metabolism occurs at the gastrointestinal level, where the glucosidic groups of SDG are cleaved to produce the aglycone, SECO (Thompson et al., 1991; Borriello et al., 1985c). SECO is likely the flaxseed lignan form available for absorption following the oral consumption of flaxseed products (Nesbitt et al., 1999; Axelson et al., 1982d).

SECO did not cause significant inhibition of CYP1A2. SECO inhibited CYP3A in a concentration-dependent but not time-dependent manner suggesting that SECO is a reversible inhibitor of the three isoforms. Reversible inhibition is consistent with the lack of GSH adduct formation following microsomal incubation of SECO in the presence of GSH. However, at 2000 μM SECO inhibited 6β-, 16α- and 2α-hydroxytestosterone formation by only 36.2 ± 7.26%, 65.7 ± 8.36% and 64.4 ± 6.36%, respectively.

ASECO caused limited inhibition of CYP1A2, 2B, and 2C11, with moderate inhibition of CYP3A at the highest concentration (100 µM) resulting in 49.2 ± 1.8% inhibition after 20 minutes preincubation. As with SDG and SECO, ASECO showed concentration-dependent but not time-dependant inhibition which suggests P450 inhibition by reversible mechanisms. These results are interesting, as in the presence of GSH, we observed an ASECO adduct that was consistent with *para-*quinone methide formation (Bolton et al., 1994; Awad et al., 2002). Despite this observation, mechanism-based inhibition of P450 activity was not identified in our studies.

Enterodiol (ED) failed to inhibit P450 activity and rather caused activation of CYP1A2, 3A, 2B and 2C11 activity particularly at shorter preincubation times (Figure 2). Although an apparent inhibition of CYP3A activity was observed after a 20 min pre-incubation (Figure 2b), control samples without ED showed a large %CV of 91.8% relative to the samples preincubated with ED for 20 min.

EL caused concentration-dependent, but not time-dependent, inhibition of CYP1A2 and only at high concentrations with 500 µM EL inhibiting CYP1A2 to 14.3 ± 23% that of control (data not shown). EL also caused concentration-dependent inhibition of CYP3A.

A Comparison Between Lignans from Creosote Bush and

and 13.9 ± 4.3% of control, respectively, at 100 µM.

**4. Reversible inhibition of cytochrome P450 enzymes** 

flaxseed, such inhibition required rather high lignan concentrations.

either the lignans of creosote bush or flaxseed.

substrate concentrations.

al., 1962; Atkins 2005).

testosterone is summarized in Table 1.

Flaxseed and Their Potential to Inhibit Cytochrome P450 Enzyme Activity 151

The NDGA dibenzocyclooctadiene did not cause significant inhibition of CYP2B/2C11 and did not cause time-dependant inhibition of any of the tested isoforms (data not shown). The NDGA dibenzocyclooctadiene significantly inhibited CYP3A and CYP1A2 to 42.9 ± 3.6%,

Our studies provided no evidence of mechanism-based inhibition of the P450 enzymes by

To determine whether lignans inhibit specific P450 enzymes through reversible mechanisms, we monitored the metabolite formation of P450-probe substrates by rat hepatic microsomes in the presence of different concentrations of lignan and probe substrates. Since time- and concentration-dependent experiments indicated that SDG and ED were neither reversible nor irreversible inhibitors of P450 enzyme activity, we excluded SDG and ED from our evaluations for reversible inhibition. We estimated the % inhibition at each lignan concentration and when possible estimated the IC50 values to provide a measure of potency of the inhibition. Lineweaver-Burke plots were constructed to determine the mechanism of inhibition. HPLC results from lignan incubations indicated substrate depletion at a testosterone concentration of 25 µM, representing ½ KM for the substrate. Consequently, this concentration was not included in any of the graphical representations and analysis. Reversible inhibition of P450 enzyme activity by the flaxseed lignans tended to show more specificity than the creosote bush lignans. Although our studies provide evidence of reversible inhibition of rat P450 enzyme activity by the lignans of creosote bush and

Our data suggest SECO is a more specific inhibitor of P450 enzymes and causes a marked reduction in CYP3A activity only at high SECO concentrations. Such concentrations are substantially greater than the levels anticipated under physiological conditions (25-100 μM) (Hu et al., 2007). Our data also show that SECO activates CYP2B activity at lower probe

SECO caused a concentration-dependent decrease in 6β-OH formation, which is mediated by CYP3A (Figure 3a). A plot of 6β-OH formation (at KM) as a function of the logarithmic concentration of SECO yielded an IC50 value of 373 µM (95% CI 266-523). A Lineweaver-Burke plot showed a pattern almost consistent with that of competitive inhibition, but the lines intersected in the upper right hand quadrant near the y-axis (not shown). This suggests SECO inhibits CYP3A activity consistent with atypical Michaelis-Menten kinetics (Folk et

The formation of 16α-OH testosterone, representing CYP2B/2C11 activity, was largely unaffected by SECO at testosterone concentrations of 100 and 250 µM (Figure 3b). At 50 µM testosterone, SECO concentrations ≥50 µM increased the formation of 16α-OH testosterone. Since CYP2C11 activity was generally unaffected by SECO (Figure 3c), SECO-mediated activation of CYP2B likely explains the increased formation of 16α-OH testosterone. The percent of control activity of the various CYP isoforms tested at KM, 2×KM and VMax

Interestingly, EL caused activation of 16α-hydroxytestosterone (16α-OH) formation, a metabolite pathway shared by CYP2B and CYP2C11. Given the ability of EL to inhibit 2α-OH formation, a pathway largely catalyzed by CYP2C11, increased 16α-OH formation may be due to activation of CYP2B activity by EL.

Fig. 2. Cytochrome P450 enzyme activity (as percent of control) as a function of Enterodiol concentration and pre-incubation time. a) CYP1A2, b) CYP3A, c) CYP2B/2C11 and d) CYP2C11. Enterodiol (closed diamond = 0 µM; closed square = 50 µM; closed triangle = 100 µM; symbol 'x' = 250 µM; open circle = 500 µM) was pre-incubated in pooled male rat hepatic microsomes (n=4) for different time periods. At the end of each pre-incubation period, testosterone (50 µM) and methoxyresorufin (0.5 µM) was added and metabolite formation was determined after a 15 min and 8 min reaction time, respectively. Each point is the mean of 3 replicates ± percent coefficient of variation.

NDGA generally caused concentration-dependent but not time-dependent inhibition of all CYP isoforms (data not shown). We anticipated that NDGA could be an irreversible inhibitor given its ability to form a reactive *ortho*-quinone intermediate. Since this is an autoxidation process and not a P450 catalyzed process, it appears that the *ortho*-quinone does not inhibit P450 *in vitro* (Billinsky et al., 2007; Billinsky & Krol 2008). Generally, marked inhibition was not observed until NDGA concentrations reached 100 μM or higher. At a preincubation time of 20 minutes and a concentration of 200 µM, NDGA inhibited CYP1A2, 2B/2C11 and 2C11 to 3.55 ± 12.0%, 21.2 ± 8.81% and 21.2 ± 9.37% of control, respectively. NDGA caused complete inhibition of formation of 6β-OH via CYP3A at pre-incubation time of 20 minutes and a concentration of 200 µM.

Interestingly, EL caused activation of 16α-hydroxytestosterone (16α-OH) formation, a metabolite pathway shared by CYP2B and CYP2C11. Given the ability of EL to inhibit 2α-OH formation, a pathway largely catalyzed by CYP2C11, increased 16α-OH formation may

**% of Control**

**% of Control**

Fig. 2. Cytochrome P450 enzyme activity (as percent of control) as a function of Enterodiol concentration and pre-incubation time. a) CYP1A2, b) CYP3A, c) CYP2B/2C11 and d) CYP2C11. Enterodiol (closed diamond = 0 µM; closed square = 50 µM; closed

triangle = 100 µM; symbol 'x' = 250 µM; open circle = 500 µM) was pre-incubated in pooled

pre-incubation period, testosterone (50 µM) and methoxyresorufin (0.5 µM) was added and metabolite formation was determined after a 15 min and 8 min reaction time, respectively.

NDGA generally caused concentration-dependent but not time-dependent inhibition of all CYP isoforms (data not shown). We anticipated that NDGA could be an irreversible inhibitor given its ability to form a reactive *ortho*-quinone intermediate. Since this is an autoxidation process and not a P450 catalyzed process, it appears that the *ortho*-quinone does not inhibit P450 *in vitro* (Billinsky et al., 2007; Billinsky & Krol 2008). Generally, marked inhibition was not observed until NDGA concentrations reached 100 μM or higher. At a preincubation time of 20 minutes and a concentration of 200 µM, NDGA inhibited CYP1A2, 2B/2C11 and 2C11 to 3.55 ± 12.0%, 21.2 ± 8.81% and 21.2 ± 9.37% of control, respectively. NDGA caused complete inhibition of formation of 6β-OH via CYP3A at pre-incubation time

male rat hepatic microsomes (n=4) for different time periods. At the end of each

Each point is the mean of 3 replicates ± percent coefficient of variation.

**d)**

**b)**

0 5 10 15 20 **Preincubation Time (min)**

0 5 10 15 20 **Preincubtion Time (min)**

be due to activation of CYP2B activity by EL.

0 5 10 15 20 **Preincubation Time (min)**

0 5 10 15 20 **Preincubation Time (min)**

of 20 minutes and a concentration of 200 µM.

**% of Control**

**c)**

**% of Control**

**a)**

The NDGA dibenzocyclooctadiene did not cause significant inhibition of CYP2B/2C11 and did not cause time-dependant inhibition of any of the tested isoforms (data not shown). The NDGA dibenzocyclooctadiene significantly inhibited CYP3A and CYP1A2 to 42.9 ± 3.6%, and 13.9 ± 4.3% of control, respectively, at 100 µM.

Our studies provided no evidence of mechanism-based inhibition of the P450 enzymes by either the lignans of creosote bush or flaxseed.

## **4. Reversible inhibition of cytochrome P450 enzymes**

To determine whether lignans inhibit specific P450 enzymes through reversible mechanisms, we monitored the metabolite formation of P450-probe substrates by rat hepatic microsomes in the presence of different concentrations of lignan and probe substrates. Since time- and concentration-dependent experiments indicated that SDG and ED were neither reversible nor irreversible inhibitors of P450 enzyme activity, we excluded SDG and ED from our evaluations for reversible inhibition. We estimated the % inhibition at each lignan concentration and when possible estimated the IC50 values to provide a measure of potency of the inhibition. Lineweaver-Burke plots were constructed to determine the mechanism of inhibition. HPLC results from lignan incubations indicated substrate depletion at a testosterone concentration of 25 µM, representing ½ KM for the substrate. Consequently, this concentration was not included in any of the graphical representations and analysis. Reversible inhibition of P450 enzyme activity by the flaxseed lignans tended to show more specificity than the creosote bush lignans. Although our studies provide evidence of reversible inhibition of rat P450 enzyme activity by the lignans of creosote bush and flaxseed, such inhibition required rather high lignan concentrations.

Our data suggest SECO is a more specific inhibitor of P450 enzymes and causes a marked reduction in CYP3A activity only at high SECO concentrations. Such concentrations are substantially greater than the levels anticipated under physiological conditions (25-100 μM) (Hu et al., 2007). Our data also show that SECO activates CYP2B activity at lower probe substrate concentrations.

SECO caused a concentration-dependent decrease in 6β-OH formation, which is mediated by CYP3A (Figure 3a). A plot of 6β-OH formation (at KM) as a function of the logarithmic concentration of SECO yielded an IC50 value of 373 µM (95% CI 266-523). A Lineweaver-Burke plot showed a pattern almost consistent with that of competitive inhibition, but the lines intersected in the upper right hand quadrant near the y-axis (not shown). This suggests SECO inhibits CYP3A activity consistent with atypical Michaelis-Menten kinetics (Folk et al., 1962; Atkins 2005).

The formation of 16α-OH testosterone, representing CYP2B/2C11 activity, was largely unaffected by SECO at testosterone concentrations of 100 and 250 µM (Figure 3b). At 50 µM testosterone, SECO concentrations ≥50 µM increased the formation of 16α-OH testosterone. Since CYP2C11 activity was generally unaffected by SECO (Figure 3c), SECO-mediated activation of CYP2B likely explains the increased formation of 16α-OH testosterone. The percent of control activity of the various CYP isoforms tested at KM, 2×KM and VMax testosterone is summarized in Table 1.

A Comparison Between Lignans from Creosote Bush and

0 25 50 100

**[ASECO] (µM)**

**% of Control**

**a)**

examination of ASECO's ability to inhibit P450 enzyme activity.

Flaxseed and Their Potential to Inhibit Cytochrome P450 Enzyme Activity 153

ASECO could be formed during the isolation process for lignan enriched flaxseed products such as Beneflax™, and with oral consumption of flaxseed lignan products small amounts of ASECO could form within the acidic environment of the stomach. Such factors warranted an

Pre-incubation time- and concentration- dependant experiments showed that ASECO was not an inhibitor of CYP2B/2C11. Due to overlapping and unresolvable HPLC peaks, the inhibition of CYP2C11, measured by 2α-OH testosterone formation could not be assessed. ASECO maximally inhibited CYP1A2 when methoxyresorufin concentration was at the KM of the enzyme, with the greatest extent of inhibition occurring at 100 µM ASECO (70.8 ± 3.9% of control) (Figure 4a). Further increases in the extent of inhibition at higher ASECO concentrations was not observed, but this may be due to an inability to completely solubilize ASECO. Therefore, data for 150 and 200µM are not shown. ASECO generally caused a concentration-dependent inhibition of CYP3A at all testosterone concentrations, although the pattern of inhibition became inconsistent at ASECO concentrations beyond 100 µM, which is likely due to solubility issues (Figure 4b). The percent of control activity of the formation of resorufin and 6β-OH tested at KM, 2×KM and VMax of methoxyresorufin and testosterone is summarized in Table 2. A plot of the 6β-OH formation (at KM) as a function of the logarithmic concentration of ASECO yielded an IC50 value of 36.4 µM (95% CI 21.9- 60.3). The IC50 value for CYP1A2 inhibition was greater than 200 µM. A Lineweaver-Burke plot gave parallel lines, a pattern consistent with uncompetitive inhibition (not shown).

**% of Control**

Fig. 4. Anhydrosecoisolariciresinol (ASECO) concentration dependent inhibition of a) CYP1A2 and b) CYP3A using methoxyresorufin (solid bar = 0.5 µM; open bar = 1 µM;

(8 min and 15 min, respectively) with pooled (n=4) male, rat liver microsomes. Each point

SDG (and SECO) is often referred to as the precursor for the enterolignans, ED and EL (Thompson et al., 1991; Borriello et al., 1985a; Kuijsten et al., 2005; Axelson et al., 1982c; Bambagiotti-Alberti et al., 1994). Much of the putative health benefits of flaxseed lignans are ascribed to the enterolignan forms, although definitive data for such assertions is lacking. Following oral consumption, SDG is converted to SECO, which further undergoes metabolism to ED and then to EL by colonic bacterial activity (Axelson et al., 1982b). The literature reports significant systemic and urinary levels of EL and its conjugated forms

stipled bar = 2.5 µM) and testosterone (solid bar = 50 µM; open bar = 100 µM; stipled bar = 250 µM) as the probe substrates, respectively, in incubation reactions

represents the mean of 3 replicates ± percent coefficient of variation.

**b)**

0 25 50 100

**[ASECO] (µM)**

Fig. 3. Secoisolariciresinol (SECO) concentration dependent inhibition of a) CYP3A, b) CYP2B/2C11 and c) CYP2C11 using testosterone (solid bar = 50 µM; open bar = 100 µM; stipled bar = 250 µM) as the probe substrate in incubation reactions (15 min) with pooled (n=4) male, rat liver microsomes. Each point represents the mean of 3 replicates ± percent coefficient of variation.


Table 1. The percent of control activity (mean ± % CV) for the formation of 6β-, 16α- and 2α-hydroxytestosterone (OHT) in pooled (n=4) rat liver microsomes by 50 and 1600 μM Secosiolariciresinol (SECO) at the KM, 2×KM and ~VMax of testosterone.

We further examined the ability of an additional flaxseed lignan, ASECO, to inhibit P450 enzyme activity. Whether ASECO is present in flaxseed remains controversial as the identification of ASECO in flaxseed may be an artifact of the analytical techniques employed by Charlet et al. (Charlet et al., 2002). Our lab has produced ASECO by acid hydrolysis of SECO using the same methods as Charlet et al., 2002 (Charlet et al., 2002). Furthermore,

**% of Control**

**b)**

0 10 25 50 100 200 400 800 1600 **[SECO] (µM)**

(CYP2B/2C11)

Percent of Control Activity (mean ± % CV)

> SECO 1600 μM

172 ± 8.1

87.4 ± 5.2

80.9 ± 5.2

SECO 50 μM

117.5 ± 5.1

91.0 ± 0.5

98.4 ± 2.8

Fig. 3. Secoisolariciresinol (SECO) concentration dependent inhibition of a) CYP3A,

b) CYP2B/2C11 and c) CYP2C11 using testosterone (solid bar = 50 µM; open bar = 100 µM; stipled bar = 250 µM) as the probe substrate in incubation reactions (15 min) with pooled (n=4) male, rat liver microsomes. Each point represents the mean of 3 replicates ± percent

0 10 25 50 100 200 400 800 1600 **[SECO] (µM)**

> 2α-OHT (CYP2C11)

Percent of Control Activity (mean ± % CV)

> SECO 1600 μM

86.0 ± 7.6

75.0 ± 4.0

71.9 ± 3.3

SECO 50 μM

106.1 ± 1.7

97.3 ± 0.5

97.1 ± 2.3

coefficient of variation.

Testosterone/ Methoxyresorufin

**% of Control**

**a)**

0 10 25 50 100 200 400 800 1600 **[SECO] (µM)**

6β-OHT (CYP3A) 16α-OHT

SECO 50 μM

5.0

1.1

3.6

Secosiolariciresinol (SECO) at the KM, 2×KM and ~VMax of testosterone.

KM 90.7 ±

2KM 86.4 ±

VMax 91.9 ±

Percent of Control Activity (mean ± % CV)

> SECO 1600 μM

29.0 ± 11

31.8 ± 11

42.4 ± 3.1

Table 1. The percent of control activity (mean ± % CV) for the formation of 6β-, 16α- and 2α-hydroxytestosterone (OHT) in pooled (n=4) rat liver microsomes by 50 and 1600 μM

We further examined the ability of an additional flaxseed lignan, ASECO, to inhibit P450 enzyme activity. Whether ASECO is present in flaxseed remains controversial as the identification of ASECO in flaxseed may be an artifact of the analytical techniques employed by Charlet et al. (Charlet et al., 2002). Our lab has produced ASECO by acid hydrolysis of SECO using the same methods as Charlet et al., 2002 (Charlet et al., 2002). Furthermore,

**% of Control**

**c)**

ASECO could be formed during the isolation process for lignan enriched flaxseed products such as Beneflax™, and with oral consumption of flaxseed lignan products small amounts of ASECO could form within the acidic environment of the stomach. Such factors warranted an examination of ASECO's ability to inhibit P450 enzyme activity.

Pre-incubation time- and concentration- dependant experiments showed that ASECO was not an inhibitor of CYP2B/2C11. Due to overlapping and unresolvable HPLC peaks, the inhibition of CYP2C11, measured by 2α-OH testosterone formation could not be assessed. ASECO maximally inhibited CYP1A2 when methoxyresorufin concentration was at the KM of the enzyme, with the greatest extent of inhibition occurring at 100 µM ASECO (70.8 ± 3.9% of control) (Figure 4a). Further increases in the extent of inhibition at higher ASECO concentrations was not observed, but this may be due to an inability to completely solubilize ASECO. Therefore, data for 150 and 200µM are not shown. ASECO generally caused a concentration-dependent inhibition of CYP3A at all testosterone concentrations, although the pattern of inhibition became inconsistent at ASECO concentrations beyond 100 µM, which is likely due to solubility issues (Figure 4b). The percent of control activity of the formation of resorufin and 6β-OH tested at KM, 2×KM and VMax of methoxyresorufin and testosterone is summarized in Table 2. A plot of the 6β-OH formation (at KM) as a function of the logarithmic concentration of ASECO yielded an IC50 value of 36.4 µM (95% CI 21.9- 60.3). The IC50 value for CYP1A2 inhibition was greater than 200 µM. A Lineweaver-Burke plot gave parallel lines, a pattern consistent with uncompetitive inhibition (not shown).

Fig. 4. Anhydrosecoisolariciresinol (ASECO) concentration dependent inhibition of a) CYP1A2 and b) CYP3A using methoxyresorufin (solid bar = 0.5 µM; open bar = 1 µM; stipled bar = 2.5 µM) and testosterone (solid bar = 50 µM; open bar = 100 µM; stipled bar = 250 µM) as the probe substrates, respectively, in incubation reactions (8 min and 15 min, respectively) with pooled (n=4) male, rat liver microsomes. Each point represents the mean of 3 replicates ± percent coefficient of variation.

SDG (and SECO) is often referred to as the precursor for the enterolignans, ED and EL (Thompson et al., 1991; Borriello et al., 1985a; Kuijsten et al., 2005; Axelson et al., 1982c; Bambagiotti-Alberti et al., 1994). Much of the putative health benefits of flaxseed lignans are ascribed to the enterolignan forms, although definitive data for such assertions is lacking. Following oral consumption, SDG is converted to SECO, which further undergoes metabolism to ED and then to EL by colonic bacterial activity (Axelson et al., 1982b). The literature reports significant systemic and urinary levels of EL and its conjugated forms

A Comparison Between Lignans from Creosote Bush and

**<sup>0</sup> <sup>100</sup> <sup>200</sup> <sup>300</sup> <sup>400</sup> <sup>500</sup> <sup>0</sup>**

**[Enterolactone] (µM)**

Resorufin (CYP1A2)

Percent of Control Activity (mean ± % CV)

> EL 500 μM

24.7 ± 23.7

> 42.6 ±1.0

64.7 ± 1.6

EL 100 μM

74.2 ± 13.8

97.4 ± 12.2

96.7 ± 1.1

**% of Control**

percent coefficient of variation.

Testosterone/ Methoxyresorufin

KM

2KM

VMax

testosterone.

**% of Control**

Flaxseed and Their Potential to Inhibit Cytochrome P450 Enzyme Activity 155

**% of Control**

**<sup>0</sup> <sup>100</sup> <sup>200</sup> <sup>300</sup> <sup>400</sup> <sup>500</sup> <sup>0</sup>**

Fig. 5. Enterolactone concentration dependent inhibition of a) CYP1A2, b) CYP3A and c) CYP2C11 using methoxyresorufin (solid bar = 0.5 µM; open bar = 1 µM; stipled bar = 2.5 µM) and testosterone (solid bar = 50 µM; open bar = 100 µM; stipled bar = 250 µM) as the probe substrates, respectively, in incubation reactions (8 min and 15 min, respectively) with pooled (n=4) male, rat liver microsomes. Each point represents the mean of 3 replicates ±

> 6β-OHT (CYP3A)

Percent of Control Activity (mean ± % CV)

> EL 500 μM

309.9 ± 26.0

55.0 ± 0.7

50.2 ± 2.3

EL 100 μM

95.1 ± 26.8

45.5 ± 4.2

52.0 ± 6.4

Table 3. The percent of control activity (mean ± % CV) for the formation of resorufin, and 6β-, 16α- and 2α-hydroxytestosterone (OHT) in pooled (n=4) rat liver microsomes by 100 and 500 μM Enterolactone (EL) at the KM, 2×KM and ~VMax of methoxyresorufin or

**[Enterolactone] (µM)**

**<sup>0</sup> <sup>100</sup> <sup>200</sup> <sup>300</sup> <sup>400</sup> <sup>500</sup> <sup>0</sup>**

16α-OHT (CYP2B/2C11)

Percent of Control Activity (mean ± % CV)

> EL 500 μM

302.5 ± 11.4

61.9 ± 2.1

60.1 ± 2.7

EL 100 μM

262.7 ± 17.1

86.9 ± 1.8

75.4 ± 5.4

2α-OHT (CYP2C11)

Percent of Control Activity (mean ± % CV)

> EL 500 μM

120.2 ± 48.1

42.4 ± 1.9

43.0 ± 1.7

EL 100 μM

176.2 ± 13.2

51.0 ± 2.9

53.9 ± 5.5

**[Enterolactone] (µM)**


(Jacobs et al., 1999; Axelson et al., 1982a) and much lower levels of ED suggesting that EL and ED are absorbed following their conversion within the gastrointestinal tract. This warranted an investigation of their potential to inhibit P450 enzyme activity.

Table 2. The percent of control activity (mean ± % CV) for the formation of 6β-hydroxytestosterone (OHT) and resorufin in pooled (n=4) rat liver microsomes by 25 and 100 μM Anhydrosecosiolariciresinol (ASECO) at the KM, 2×KM and ~VMax of testosterone or methoxyresorufin.

Enterodiol caused CYP2B/2C11 activation in time- and concentration-dependent experiments and thus was not examined for reversible inhibition of these enzymes. EL maximally inhibited CYP1A2 when methoxyresorufin concentration was at the KM of the enzyme (Figure 5a). EL also caused a concentration-dependent decrease in resorufin formation at 2× KM and VMax methoxyresorufin concentrations (Figure 5a). At the KM of testosterone, EL did not inhibit CYP3A activity and at 500 µM EL, a 3-fold increase in metabolite formation was observed (Figure 5b). However, at 2×KM and VMax testosterone concentrations, EL generally caused a concentration-dependent decrease in CYP3Amediated 6β-OH formation (Figure 5b). For CYP2C11, at the KM of testosterone, EL generally increased the rate of 2α-OH formation at all EL concentrations. However, EL caused pronounced inhibition at 2×KM and VMax testosterone concentrations for the enzyme (Figure 5c). Plots of metabolite formation (at KM) as a function of the logarithmic concentration of EL yielded IC50 values of 441 µM (95% CI 115-1695), 72.9 µM (95% CI 54.0- 98.2) (determined at the VMax due to activation at KM) and 104 µM (95% CI 85.7-127) (determined at the VMax due to activation at KM) for CYP1A2, CYP3A and CYP2C11, respectively. The percent of control activity of the various CYP isoforms tested at KM, 2× KM and VMax of methoxyresorufin or testosterone is summarized in Table 3.

Lineweaver-Burke plots were difficult to interpret for EL inhibition of CYP1A2, although the pattern was somewhat consistent with that of competitive inhibition, with the lines intersecting in the upper right hand quadrant (not shown). However, the pattern of the plots for CYP2C11 and CYP3A were consistent with noncompetitive inhibition by EL.

(Jacobs et al., 1999; Axelson et al., 1982a) and much lower levels of ED suggesting that EL and ED are absorbed following their conversion within the gastrointestinal tract. This

> Percent of Control Activity (mean ± % CV)

> > ASECO 100 μM

KM 89.6 ± 19.5 70.8 ± 3.9 74.0 ±3.5 40.9 ± 6.5

2KM 87.5 ± 14.1 86.4 ± 14.0 55.5 ± 7.8 15.6 ± 10.2

VMax 102.6 ± 3.8 94.7 ± 12.5 59.7 ± 12.2 65.2 ± 28.7

6β-hydroxytestosterone (OHT) and resorufin in pooled (n=4) rat liver microsomes by 25 and 100 μM Anhydrosecosiolariciresinol (ASECO) at the KM, 2×KM and ~VMax of testosterone or

Enterodiol caused CYP2B/2C11 activation in time- and concentration-dependent experiments and thus was not examined for reversible inhibition of these enzymes. EL maximally inhibited CYP1A2 when methoxyresorufin concentration was at the KM of the enzyme (Figure 5a). EL also caused a concentration-dependent decrease in resorufin formation at 2× KM and VMax methoxyresorufin concentrations (Figure 5a). At the KM of testosterone, EL did not inhibit CYP3A activity and at 500 µM EL, a 3-fold increase in metabolite formation was observed (Figure 5b). However, at 2×KM and VMax testosterone concentrations, EL generally caused a concentration-dependent decrease in CYP3Amediated 6β-OH formation (Figure 5b). For CYP2C11, at the KM of testosterone, EL generally increased the rate of 2α-OH formation at all EL concentrations. However, EL caused pronounced inhibition at 2×KM and VMax testosterone concentrations for the enzyme (Figure 5c). Plots of metabolite formation (at KM) as a function of the logarithmic concentration of EL yielded IC50 values of 441 µM (95% CI 115-1695), 72.9 µM (95% CI 54.0- 98.2) (determined at the VMax due to activation at KM) and 104 µM (95% CI 85.7-127) (determined at the VMax due to activation at KM) for CYP1A2, CYP3A and CYP2C11, respectively. The percent of control activity of the various CYP isoforms tested at KM, 2× KM

Lineweaver-Burke plots were difficult to interpret for EL inhibition of CYP1A2, although the pattern was somewhat consistent with that of competitive inhibition, with the lines intersecting in the upper right hand quadrant (not shown). However, the pattern of the plots

Resorufin (CYP1A2) 6β-OHT (CYP3A)

Percent of Control Activity (mean ± % CV)

> ASECO 100 μM

ASECO 25 μM

warranted an investigation of their potential to inhibit P450 enzyme activity.

ASECO 25 μM

Table 2. The percent of control activity (mean ± % CV) for the formation of

and VMax of methoxyresorufin or testosterone is summarized in Table 3.

for CYP2C11 and CYP3A were consistent with noncompetitive inhibition by EL.

Testosterone/ Methoxyresorufin

methoxyresorufin.

Fig. 5. Enterolactone concentration dependent inhibition of a) CYP1A2, b) CYP3A and c) CYP2C11 using methoxyresorufin (solid bar = 0.5 µM; open bar = 1 µM; stipled bar = 2.5 µM) and testosterone (solid bar = 50 µM; open bar = 100 µM; stipled bar = 250 µM) as the probe substrates, respectively, in incubation reactions (8 min and 15 min, respectively) with pooled (n=4) male, rat liver microsomes. Each point represents the mean of 3 replicates ± percent coefficient of variation.


Table 3. The percent of control activity (mean ± % CV) for the formation of resorufin, and 6β-, 16α- and 2α-hydroxytestosterone (OHT) in pooled (n=4) rat liver microsomes by 100 and 500 μM Enterolactone (EL) at the KM, 2×KM and ~VMax of methoxyresorufin or testosterone.

A Comparison Between Lignans from Creosote Bush and

0 25 50 100 150 200 **[NDGA] (µM)**

0 5 10 25 50 75 100 150 200 **[NDGA] (µM)**

noncompetitive inhibition (not shown).

represents the mean of 3 replicates ± percent coefficient of variation.

**% of Control**

**c)**

**% of Control**

**a)**

Flaxseed and Their Potential to Inhibit Cytochrome P450 Enzyme Activity 157

**% of Control**

**b)**

**% of Control**

Fig. 6. Nordihydroguaiaretic acid (NDGA) concentration dependent inhibition of a) CYP1A2, b) CYP3A c) CYP2B/2C11, and d) CYP2C11 using methoxyresorufin (solid bar = 0.5 µM; open bar = 1 µM; stipled bar = 2.5 µM) and testosterone (solid bar = 50 µM; open bar = 100 µM; stipled bar = 250 µM) as the probe substrates, respectively, in incubation reactions (8 min and 15 min, respectively) with pooled (n=4) male, rat liver microsomes. Each point

Although previous studies showed that in the presence of GSH NDGA is oxidized to NDGA-GSH adducts and prevents the formation of NDGA dibenzocyclooctadiene (Billinsky & Krol 2008a), inhibition of P450 activity was still observed. For NDGA dibenzocyclooctadiene, the activity of CYP2C11, as monitored by 2α-OH testosterone formation, could not be assessed, as the dibenzocyclooctadiene eluted at the same retention time as 2α-OH testosterone during the HPLC run. There was sufficient peak overlap which could not be separated so we could not accurately achieve reliable data. For CYP1A2, the dibenzocyclooctadiene caused more prominent inhibition at the KM of methoxyresorufin (Figure 7a), but for CYP3A and CYP2B/2C11 slightly greater inhibition of testosterone metabolite formation at the testosterone VMax was observed (Figure 7b, 7c). At the VMax of testosterone, dibenzocyclooctadiene concentrations of 50 to 150 µM increased the formation of resorufin, an index pathway for CYP1A2 (Figure 7a). Dibenzocyclooctadiene did not cause significant inhibition of CYP2B/2C11 as evidenced by no substantial decrease in 16α-OH formation (Figure 7c). Plots of metabolite formation as a function of the logarithmic concentration of NDGA dibenzocyclooctadiene yielded an IC50 value of 36.8 µM (95% CI 25.3-53.6) for CYP3A. For CYP1A2, the data did not yield an interpretable value due to extensive variation in the data. The percent of control activity of the various CYP isoforms tested at KM, 2×KM and VMax of methoxyresorufin or testosterone is summarized in Table 5. Lineweaver-Burke plots for dibenzocyclooctadiene inhibition of CYP1A2 showed a pattern consistent with competitive inhibition, but with the lines intersecting in the upper right hand quadrant near the y-axis and for CYP3A the pattern was consistent with

**d)**

0 5 10 25 50 75 100 150 200 **[NDGA] (µM)**

0 5 10 25 50 75 100 150 200 **[NDGA] (µM)**

In general, NDGA caused a concentration-dependent decrease in CYP1A2, CYP3A, CYP2B and CYP2C11 activity (Figure 6). For CYP1A2, NDGA caused more prominent inhibition at the KM of methoxyresorufin (Figure 6a), but for the CYP3A, CYP2B/2C11, and CYP2C11, NDGA caused most pronounced inhibition of testosterone metabolite formation at the VMax for testosterone (Figure 6b, 6c, 6d). Furthermore, at the KM of testosterone, NDGA increased the 16α- and 2α-hydroxylation of testosterone, index pathways for CYP2B/2C11 and CYP2C11, respectively. Activation of metabolism was more pronounced for 16α-OH formation. The IC50 values were calculated at the KM of the substrate for CYP1A2 and CYP3A and at VMax for testosterone for CYP2B/2C11 and CYP2C11, as activation was observed at the KM for CYP2B/2C11 and CYP2C11. Plots of metabolite formation as a function of the logarithmic concentration of NDGA yielded IC50 values of 63.5 µM (95% CI 11.8-341), 97.3 µM (95% CI 49.6-191), 68.7 µM (95% CI 46.4-102) and 96.6 µM (95% CI 55.3- 169) for CYP1A2, CYP3A, CYP2B/2C11 and CYP2C11, respectively. The percent of control activity of the various CYP isoforms tested at KM, 2×KM and VMax of methoxyresorufin or testosterone is summarized in Table 4.


Table 4. The percent of control activity (mean ± % CV) for the formation of resorufin, and 6β-, 16α- and 2α-hydroxytestosterone (OHT) in pooled (n=4) rat liver microsomes by 25 and 200 μM Nordihydroguaiaretic acid (NDGA) at the KM, 2×KM and ~VMax of methoxyresorufin or testosterone.

Lineweaver-Burke plots for NDGA inhibition of CYP1A2 showed a pattern consistent with competitive inhibition, but with the lines intersecting in the upper right hand quadrant near the y-axis (not shown). For CYP3A the pattern was consistent with noncompetitive inhibition. For the CYP2B/2C11 and CYP2C11 pathways, the Lineweaver-Burke plots were more difficult to interpret. At higher NDGA concentrations, parallel lines suggested uncompetitive inhibition. However, activation of these pathways at the KM of testosterone (Figure 6c and 6d) likely affected the overall pattern observed in these plots (not shown).

In general, NDGA caused a concentration-dependent decrease in CYP1A2, CYP3A, CYP2B and CYP2C11 activity (Figure 6). For CYP1A2, NDGA caused more prominent inhibition at the KM of methoxyresorufin (Figure 6a), but for the CYP3A, CYP2B/2C11, and CYP2C11, NDGA caused most pronounced inhibition of testosterone metabolite formation at the VMax for testosterone (Figure 6b, 6c, 6d). Furthermore, at the KM of testosterone, NDGA increased the 16α- and 2α-hydroxylation of testosterone, index pathways for CYP2B/2C11 and CYP2C11, respectively. Activation of metabolism was more pronounced for 16α-OH formation. The IC50 values were calculated at the KM of the substrate for CYP1A2 and CYP3A and at VMax for testosterone for CYP2B/2C11 and CYP2C11, as activation was observed at the KM for CYP2B/2C11 and CYP2C11. Plots of metabolite formation as a function of the logarithmic concentration of NDGA yielded IC50 values of 63.5 µM (95% CI 11.8-341), 97.3 µM (95% CI 49.6-191), 68.7 µM (95% CI 46.4-102) and 96.6 µM (95% CI 55.3- 169) for CYP1A2, CYP3A, CYP2B/2C11 and CYP2C11, respectively. The percent of control activity of the various CYP isoforms tested at KM, 2×KM and VMax of methoxyresorufin or

> 6β-OHT (CYP3A)

Percent of Control Activity (mean ± % CV)

> NDGA 200 μM

> > 17.6 ± 11.7

> > 12.4 ± 30.7

> > 11.1 ± 3.5

NDGA 25 μM

81.8 ± 6.4

66.6 ± 1.8

63.6 ± 3.9

Table 4. The percent of control activity (mean ± % CV) for the formation of resorufin, and 6β-, 16α- and 2α-hydroxytestosterone (OHT) in pooled (n=4) rat liver microsomes by 25 and 200 μM Nordihydroguaiaretic acid (NDGA) at the KM, 2×KM and ~VMax of methoxyresorufin

Lineweaver-Burke plots for NDGA inhibition of CYP1A2 showed a pattern consistent with competitive inhibition, but with the lines intersecting in the upper right hand quadrant near the y-axis (not shown). For CYP3A the pattern was consistent with noncompetitive inhibition. For the CYP2B/2C11 and CYP2C11 pathways, the Lineweaver-Burke plots were more difficult to interpret. At higher NDGA concentrations, parallel lines suggested uncompetitive inhibition. However, activation of these pathways at the KM of testosterone (Figure 6c and 6d) likely affected the overall pattern observed in these

16α-OHT (CYP2B/2C11)

Percent of Control Activity (mean ± % CV)

> NDGA 200 μM

123.2 ± 8.8

42.2 ± 26.0

27.6 ± 0.3

NDGA 25 μM

158.3 ± 6.2

88.7 ± 4.5

79.8 ± 3.4

2α-OHT (CYP2C11)

Percent of Control Activity (mean ± % CV)

> NDGA 200 μM

> > 57.3 ± 8.6

> > 37.5 ± 26.1

> > 29.7 ± 0.1

NDGA 25 μM

118.0 ± 3.9

93.0 ± 2.9

88.7 ± 3.8

testosterone is summarized in Table 4.

NDGA 25 μM

84.1 ± 16.2

100.6 ± 1.7

109.1 ± 1.1

Testosterone/ Methoxyresorufin

KM

2KM

VMax

or testosterone.

plots (not shown).

Resorufin (CYP1A2)

Percent of Control Activity (mean ± % CV)

> NDGA 200 μM

> > 6.7 ± 13.5

17.9 ± 4.2

40.2 ± 7.9

Fig. 6. Nordihydroguaiaretic acid (NDGA) concentration dependent inhibition of a) CYP1A2, b) CYP3A c) CYP2B/2C11, and d) CYP2C11 using methoxyresorufin (solid bar = 0.5 µM; open bar = 1 µM; stipled bar = 2.5 µM) and testosterone (solid bar = 50 µM; open bar = 100 µM; stipled bar = 250 µM) as the probe substrates, respectively, in incubation reactions (8 min and 15 min, respectively) with pooled (n=4) male, rat liver microsomes. Each point represents the mean of 3 replicates ± percent coefficient of variation.

Although previous studies showed that in the presence of GSH NDGA is oxidized to NDGA-GSH adducts and prevents the formation of NDGA dibenzocyclooctadiene (Billinsky & Krol 2008a), inhibition of P450 activity was still observed. For NDGA dibenzocyclooctadiene, the activity of CYP2C11, as monitored by 2α-OH testosterone formation, could not be assessed, as the dibenzocyclooctadiene eluted at the same retention time as 2α-OH testosterone during the HPLC run. There was sufficient peak overlap which could not be separated so we could not accurately achieve reliable data. For CYP1A2, the dibenzocyclooctadiene caused more prominent inhibition at the KM of methoxyresorufin (Figure 7a), but for CYP3A and CYP2B/2C11 slightly greater inhibition of testosterone metabolite formation at the testosterone VMax was observed (Figure 7b, 7c). At the VMax of testosterone, dibenzocyclooctadiene concentrations of 50 to 150 µM increased the formation of resorufin, an index pathway for CYP1A2 (Figure 7a). Dibenzocyclooctadiene did not cause significant inhibition of CYP2B/2C11 as evidenced by no substantial decrease in 16α-OH formation (Figure 7c). Plots of metabolite formation as a function of the logarithmic concentration of NDGA dibenzocyclooctadiene yielded an IC50 value of 36.8 µM (95% CI 25.3-53.6) for CYP3A. For CYP1A2, the data did not yield an interpretable value due to extensive variation in the data. The percent of control activity of the various CYP isoforms tested at KM, 2×KM and VMax of methoxyresorufin or testosterone is summarized in Table 5.

Lineweaver-Burke plots for dibenzocyclooctadiene inhibition of CYP1A2 showed a pattern consistent with competitive inhibition, but with the lines intersecting in the upper right hand quadrant near the y-axis and for CYP3A the pattern was consistent with noncompetitive inhibition (not shown).

A Comparison Between Lignans from Creosote Bush and

**glutathione** 

**lignans** 

metabolism.

isoform.

the creosote bush lignan, NDGA.

Flaxseed and Their Potential to Inhibit Cytochrome P450 Enzyme Activity 159

The effect of adding GSH to microsomal incubations before the addition of NDGA was studied to determine whether GSH could attenuate inhibition at various time and inhibitor concentrations. GSH had little impact on NDGA inhibition of CYP1A2 activity (data not shown). The calculated IC50 for NDGA in the presence and absence of GSH was 63.5 µM (95% CI 11.8-341) and 80.1 µM (95% CI 22.6-284), respectively, for CYP1A2. These observations suggest that GSH *in vitro* does not protect against reversible P450 inhibition by

Interestingly, at lower probe substrate concentrations (50 µM) for NDGA, GSH attenuated inhibition, and appeared to cause activation of CYP1A2 (not shown), while dibenzocyclooctadiene increased the metabolite formation of the index pathway for CYP2B/2C11. Although P450 enzyme activation has been observed for other compounds (Stresser et al., 2000c; Lasker et al., 1984), the relevance of this phenomenon on *in vivo*

metabolism is not known (Stresser et al., 2000b; Houston & Kenworthy 2000).

the CYP2B/2C11 index pathway, the relevance of which is not known.

**6. Summary of inhibition of rat CYP3A, CYP2B, CYP2C11 and CYP1A2 by** 

Lineweaver-Burke plots of the inhibition data suggest that lignans largely caused reversible inhibition via a combination of competitive, noncompetitive, and uncompetitive mechanisms. The lignans EL, NDGA and its dibenzocyclooctadiene autoxidation product, inhibited a combination of CYP3A, CYP2B/2C11 and CYP1A2 activity suggesting that these lignans act as general reversible inhibitors of P450 activity. Interestingly, EL also activated

The ability of lignans to activate P450 enzyme activity in the presence of low probe substrate concentrations for the P450 index pathways may account for the mixed type inhibition we observed (Stresser et al., 2000). Nonetheless, our *in vitro* results suggest that potential for pharmacokinetic interactions via inhibition of P450-mediated elimination by flaxseed lignans is extremely limited. Concentration at the P450 active site is a principal determinant of the potential for clinically relevant interactions (Bjornsson et al., 2003). Systemic levels reported in the literature suggest that lignan concentrations achieved at the P450 active site would be insufficient to cause significant inhibition. Even following oral consumption, when portal vein concentrations and hence, hepatic concentrations, of lignans are expected to be much greater than systemic levels, the competing phase II reactions (i.e. glucuronidation, sulfation) (Axelson & Setchell 1980; Dean et al., 2004) would diminish the availability of lignans at the P450 enzyme active sites. The clinical use of lignans should not be associated with adverse outcomes resulting from inhibition of P450-mediated

Table 6 summarizes the estimated IC50 values for lignan-mediated reversible inhibition different P450 isoforms. For lignans that failed to inhibit specific P450 isoforms the value reported represents the solubility limit for the respective lignan. Table 7 summarizes the type of reversible inhibition mechanism exhibited by each lignan for a particular P450

**5. Inhibition of CYP1A2 by nordihydroguaiaretic acid in the presence of** 

Fig. 7. Dibenzocyclooctadiene (cyclolignan) concentration dependent inhibition of a) CYP1A2, b) CYP3A, and c) CYP2B/2C11 using methoxyresorufin (solid bar = 0.5 µM; open bar = 1 µM; stipled bar = 2.5 µM) and testosterone (solid bar = 50 µM; open bar = 100 µM; stipled bar = 250 µM) as the probe substrates, respectively, in incubation reactions (8 min and 15 min, respectively) with pooled (n=4) male, rat liver microsomes. Each point represents the mean of 3 replicates ± percent coefficient of variation.


Table 5. The percent of control activity (mean ± % CV) for the formation of resorufin, and 6β-, 16α- and 2α-hydroxytestosterone (OHT) in pooled (n=4) rat liver microsomes by 12.5, 50, 100 and 200 μM Dibenzocyclooctadiene (DIB) at the KM, 2×KM and ~VMax of methoxyresorufin or testosterone.

**% of Control**

0 12.5 25 50 75 100 **[Cyclolignan] (µM)**

(CYP1A2) 6β-OHT (CYP3A) 16α-OHT

Percent of Control Activity (mean ± % CV)

> DIB 100 μM

> > 21.8 ±

DIB 12.5 μM

71.5 ±

74.0 ± 3.5

70.0 ±

Fig. 7. Dibenzocyclooctadiene (cyclolignan) concentration dependent inhibition of a) CYP1A2, b) CYP3A, and c) CYP2B/2C11 using methoxyresorufin (solid bar = 0.5 µM; open bar = 1 µM; stipled bar = 2.5 µM) and testosterone (solid bar = 50 µM; open bar = 100 µM; stipled bar = 250 µM) as the probe substrates, respectively, in incubation reactions (8 min and 15 min, respectively) with pooled (n=4) male, rat liver microsomes. Each point

represents the mean of 3 replicates ± percent coefficient of variation.

DIB 50 μM

107.4 ± 12.4

153.8 ± 12.7

KM 90.3 ± 2.5 45.3 ±

Resorufin

Percent of Control Activity (mean ± % CV)

> DIB 200 μM

> > 1.0

68.6 ± 12.1

103.4 ± 4.8

Table 5. The percent of control activity (mean ± % CV) for the formation of resorufin, and 6β-, 16α- and 2α-hydroxytestosterone (OHT) in pooled (n=4) rat liver microsomes by 12.5,

50, 100 and 200 μM Dibenzocyclooctadiene (DIB) at the KM, 2×KM and ~VMax of

**b)**

0 12.5 25 50 75 100 **[Cyclolignan] (µM)**

(CYP2B/2C11)

Percent of Control Activity (mean ± % CV)

> DIB 100 μM

DIB 12.5 μM

10.0 91.3 ± 4.1 60.4 ± 8.2

7.4 26.9 ± 3.5 96.8 ± 7.0 82.0 ± 2.5

2.2 26.4 ± 8.0 85.4 ± 2.9 59.6 ± 7.0

**% of Control**

**a)**

0 50 75 100 150 200 **[Cyclolignan] (µM)**

**% of Control**

Testosterone/ Methoxyresorufin

2KM

VMax

methoxyresorufin or testosterone.

**c)**

## **5. Inhibition of CYP1A2 by nordihydroguaiaretic acid in the presence of glutathione**

The effect of adding GSH to microsomal incubations before the addition of NDGA was studied to determine whether GSH could attenuate inhibition at various time and inhibitor concentrations. GSH had little impact on NDGA inhibition of CYP1A2 activity (data not shown). The calculated IC50 for NDGA in the presence and absence of GSH was 63.5 µM (95% CI 11.8-341) and 80.1 µM (95% CI 22.6-284), respectively, for CYP1A2. These observations suggest that GSH *in vitro* does not protect against reversible P450 inhibition by the creosote bush lignan, NDGA.

Interestingly, at lower probe substrate concentrations (50 µM) for NDGA, GSH attenuated inhibition, and appeared to cause activation of CYP1A2 (not shown), while dibenzocyclooctadiene increased the metabolite formation of the index pathway for CYP2B/2C11. Although P450 enzyme activation has been observed for other compounds (Stresser et al., 2000c; Lasker et al., 1984), the relevance of this phenomenon on *in vivo* metabolism is not known (Stresser et al., 2000b; Houston & Kenworthy 2000).

#### **6. Summary of inhibition of rat CYP3A, CYP2B, CYP2C11 and CYP1A2 by lignans**

Lineweaver-Burke plots of the inhibition data suggest that lignans largely caused reversible inhibition via a combination of competitive, noncompetitive, and uncompetitive mechanisms. The lignans EL, NDGA and its dibenzocyclooctadiene autoxidation product, inhibited a combination of CYP3A, CYP2B/2C11 and CYP1A2 activity suggesting that these lignans act as general reversible inhibitors of P450 activity. Interestingly, EL also activated the CYP2B/2C11 index pathway, the relevance of which is not known.

The ability of lignans to activate P450 enzyme activity in the presence of low probe substrate concentrations for the P450 index pathways may account for the mixed type inhibition we observed (Stresser et al., 2000). Nonetheless, our *in vitro* results suggest that potential for pharmacokinetic interactions via inhibition of P450-mediated elimination by flaxseed lignans is extremely limited. Concentration at the P450 active site is a principal determinant of the potential for clinically relevant interactions (Bjornsson et al., 2003). Systemic levels reported in the literature suggest that lignan concentrations achieved at the P450 active site would be insufficient to cause significant inhibition. Even following oral consumption, when portal vein concentrations and hence, hepatic concentrations, of lignans are expected to be much greater than systemic levels, the competing phase II reactions (i.e. glucuronidation, sulfation) (Axelson & Setchell 1980; Dean et al., 2004) would diminish the availability of lignans at the P450 enzyme active sites. The clinical use of lignans should not be associated with adverse outcomes resulting from inhibition of P450-mediated metabolism.

Table 6 summarizes the estimated IC50 values for lignan-mediated reversible inhibition different P450 isoforms. For lignans that failed to inhibit specific P450 isoforms the value reported represents the solubility limit for the respective lignan. Table 7 summarizes the type of reversible inhibition mechanism exhibited by each lignan for a particular P450 isoform.

A Comparison Between Lignans from Creosote Bush and

**8. Acknowledgements** 

experiments.

**9. References** 

343-351.

pp. 659-660.

pp. 595-598.

1615.

Vol.122, No.1, pp. 49-53.

Flaxseed and Their Potential to Inhibit Cytochrome P450 Enzyme Activity 161

flaxseed lignan administration where no adverse effects have been reported as of yet. Such promising clinical trial data associated with the flaxseed lignans warrants further

This study was performed with the assistance of Natural Sciences and Engineering Research Council of Canada (NSERC) and Canada Foundation for Innovation grants. J.B. was the recipient of an NSERC Postgraduate Scholarship. The authors would like to thank Alister Muir from Agriculture and Agrifood Canada, Saskatoon, SK for the kind gifts of SDG and SECO. In addition we thank Krista Thompson for running the UPLC-MS

Arteaga, S., Andrade-Cetto, A., and Cardenas, R. (2005). Larrea tridentata (Creosote bush),

Awad, H. M., Boersma, M. G., Boeren, S., van Bladeren, P. J., Vervoort, J., and Rietjens, I. M.

Axelson, M. and Setchell, K. D. (1980). Conjugation of lignans in human urine. *FEBS Lett.,*

Axelson, M., Sjovall, J., Gustafsson, B. E., and Setchell, K. D. (1982). Origin of lignans in

Bambagiotti-Alberti, M., Coran, S. A., Ghiara, C., Giannellini, V., and Raffaelli, A. (1994).

Batchelor, W. B., Heathcote, J., and Wanless, I. R. (1995). Chaparral-induced hepatic injury.

Billinsky, J. L. and Krol, E. S. (2008). Nordihydroguaiaretic acid autoxidation produces a

Billinsky, J. L., Marcoux, M. R., and Krol, E. S. (2007). Oxidation of the lignan nordihydroguaiaretic acid. *Chem.Res.Toxicol.,* Vol.20, No.9, pp. 1352-1358. Bjornsson, T. D., Callaghan, J. T., Einolf, H. J., Fischer, V., Gan, L., Grimm, S., Kao, J., King, S.

Bolton, J. L., Acay, N. M., and Vukomanovic, V. (1994). Evidence that 4-allyl-o-quinones

nordihydroguaiaretic acid. *J.Ethnopharmacol.,* Vol.98, No.3, pp. 231-239. Atkins, W. M. (2005). Non-Michaelis-Menten kinetics in cytochrome P450-catalyzed

reactions. *Annu.Rev.Pharmacol.Toxicol.,* Vol.45, pp. 291-310.

*Am.J.Gastroenterol.,* Vol.90, No.5, pp. 831-833.

*Drug Metab Dispos.,* Vol.31, No.7, pp. 815-832.

an abundant plant of Mexican and US-American deserts and its metabolite

(2002). The regioselectivity of glutathione adduct formation with flavonoid quinone/quinone methides is pH-dependent. *Chem.Res.Toxicol.,* Vol.15, No.3, pp.

mammals and identification of a precursor from plants. *Nature,* Vol.298, No.5875,

Revealing the mammalian lignan precursor secoisolariciresinol diglucoside in flax seed by ionspray mass spectrometry. *Rapid Commun.Mass Spectrom.,* Vol.8, No.8,

schisandrin-like dibenzocyclooctadiene lignan. *J.Nat.Prod.,* Vol.71, No.9, pp. 1612-

P., Miwa, G., Ni, L., Kumar, G., McLeod, J., Obach, R. S., Roberts, S., Roe, A., Shah, A., Snikeris, F., Sullivan, J. T., Tweedie, D., Vega, J. M., Walsh, J., and Wrighton, S. A. (2003). The conduct of *in vitro* and *in vivo* drug-drug interaction studies: a Pharmaceutical Research and Manufacturers of America (PhRMA) perspective.

spontaneously rearrange to their more electrophilic quinone methides: potential

evaluations of their safety and efficacy, which remains a focus in our laboratory.


ND = not determined

aNo inhibition was observed at any lignan concentration

bCould not assess due to overlapping peak on high performance liquid chromatography cIC50 assessed at VMax

Table 6. Summary of estimated lignan IC50 values (µM) (with 95% confidence intervals displayed in brackets) for CYP1A2, CYP 3A, CYP2B/2C11 and CYP2C11. IC50 values were determined at probe substrate KM, except where noted.


NO = No inhibition was observed

aActivation of P450 activity was observed.

Table 7. Summary of the type of P450 enzyme inhibition caused by lignans of creosote bush and flaxseed.

#### **7. Conclusions**

In conclusion, our data does not support the hypothesis that the differential toxicity between lignans of creosote bush and flaxseed may be due to differences in their capacity to undergo P450-mediated bioactivation to electrophilic reactive intermediates or reversible P450 enzyme inhibition. NDGA's autoxidation to a reactive quinone intermediate warrants further investigation as a possible mechanism associated with its known toxicity. Nonetheless, our *in vitro* data suggests the potential for inhibition of P450-mediated elimination of compounds by the lignans of creosote bush and flaxseed is limited. Their use for promotion of health and wellness or for therapeutic reasons should not be associated with adverse outcomes resulting from inhibition of P450-mediated metabolism. For the flaxseed lignans, our *in vitro* metabolism data is supported by the emerging clinical data on flaxseed lignan administration where no adverse effects have been reported as of yet. Such promising clinical trial data associated with the flaxseed lignans warrants further evaluations of their safety and efficacy, which remains a focus in our laboratory.

## **8. Acknowledgements**

160 Drug Discovery Research in Pharmacognosy

(115-1695)

63.5 (11.8-341)

bCould not assess due to overlapping peak on high performance liquid chromatography

CYP1A2 CYP3A CYP2B/

Table 6. Summary of estimated lignan IC50 values (µM) (with 95% confidence intervals displayed in brackets) for CYP1A2, CYP 3A, CYP2B/2C11 and CYP2C11. IC50 values were

Secoisolariciresinol NO Competitive NOa NO

secoisolariciresinol Uncompetitive Uncompetitive NO NO

Enterolactone Competitive Noncompetitivea NOa Noncompetitivea

Dibenzocyclooctadiene Competitive Noncompetitive NO Could not assess

Table 7. Summary of the type of P450 enzyme inhibition caused by lignans of creosote bush

In conclusion, our data does not support the hypothesis that the differential toxicity between lignans of creosote bush and flaxseed may be due to differences in their capacity to undergo P450-mediated bioactivation to electrophilic reactive intermediates or reversible P450 enzyme inhibition. NDGA's autoxidation to a reactive quinone intermediate warrants further investigation as a possible mechanism associated with its known toxicity. Nonetheless, our *in vitro* data suggests the potential for inhibition of P450-mediated elimination of compounds by the lignans of creosote bush and flaxseed is limited. Their use for promotion of health and wellness or for therapeutic reasons should not be associated with adverse outcomes resulting from inhibition of P450-mediated metabolism. For the flaxseed lignans, our *in vitro* metabolism data is supported by the emerging clinical data on

acid Competitive Noncompetitive Uncompetitivea Uncompetitivea

Secoisolariciresinol NDa <sup>373</sup>

Anhydrosecoisolariciresinol > 200 36.4

Dibenzocyclooctadiene NDa 36.8

Enterolactone <sup>441</sup>

aNo inhibition was observed at any lignan concentration

determined at probe substrate KM, except where noted.

Nordihydroguaiaretic acid

ND = not determined

cIC50 assessed at VMax

Anhydro-

Nordihydroguaiaretic

and flaxseed.

**7. Conclusions** 

NO = No inhibition was observed aActivation of P450 activity was observed.

CYP1A2 CYP3A CYP2B/

72.9c (54.0-98.2)

97.3 (49.6-191) CYP2C11 CYP2C11

> 500 104c

(85.7-127)

96.6c (55.3-169)

(266-523) > 1600 > 1600

(21.9-60.3) > 200 NDb

68.7c (46.4-102)

(25.3-53.6) NDa NDb

CYP2C11 CYP2C11

This study was performed with the assistance of Natural Sciences and Engineering Research Council of Canada (NSERC) and Canada Foundation for Innovation grants. J.B. was the recipient of an NSERC Postgraduate Scholarship. The authors would like to thank Alister Muir from Agriculture and Agrifood Canada, Saskatoon, SK for the kind gifts of SDG and SECO. In addition we thank Krista Thompson for running the UPLC-MS experiments.

## **9. References**


A Comparison Between Lignans from Creosote Bush and

Vol.45, No.11, pp. 2219-2227.

Vol.68, No.5-6, pp. 211-218.

Vol.42, No.2, pp. 149-158.

162-170.

No.2, pp. 169-180.

No.3, pp. 246-254.

pp. 1092-1104.

537-544.

Flaxseed and Their Potential to Inhibit Cytochrome P450 Enzyme Activity 163

Houston, J. B. and Kenworthy, K. E. (2000). *In vitro*-*in vivo* scaling of CYP kinetic data not

Hu, C., Yuan, Y. V., and Kitts, D. D. (2007). Antioxidant activities of the flaxseed lignan

Iba, M. M., Soyka, L. F., and Schulman, M. P. (1977). Characteristics of the liver microsomal

Iverson, S. L., Hu, L. Q., Vukomanovic, V., and Bolton, J. L. (1995). The influence of the p-

Jacobs, E., Kulling, S. E., and Metzler, M. (1999). Novel metabolites of the mammalian

Johnson, B. M., Qiu, S. X., Zhang, S., Zhang, F., Burdette, J. E., Yu, L., Bolton, J. L., and van

Kent, U. M., Aviram, M., Rosenblat, M., and Hollenberg, P. F. (2002). The licorice root

Kuijsten, A., Arts, I. C., Vree, T. B., and Hollman, P. C. (2005). Pharmacokinetics of

Kupiec, T. and Raj, V. (2005). Fatal seizures due to potential herb-drug interactions with

Lambert, J. D., Dorr, R. T., and Timmermann, B. N. (2004). Nordihydroguaiaretic Acid: A

Lasker, J. M., Huang, M. T., and Conney, A. H. (1984). *In vitro* and *in vivo* activation of

Lowry, O. H., ROSEBROUGH, N. J., FARR, A. L., and RANDALL, R. J. (1951). Protein measurement with the Folin phenol reagent. *J.Biol.Chem.,* Vol.193, No.1, pp. 265-275. Masubuchi, N., Makino, C., and Murayama, N. (2007). Prediction of *in vivo* potential for

Mazur, W., Fotsis, T., Wahala, K., Ojala, S., Salakka, A., and Adlercreutz, H. (1996). Isotope

methysticum Forst (Kava). *Chem.Res.Toxicol.,* Vol.16, No.6, pp. 733-740. Kalgutkar, A. S., Obach, R. S., and Maurer, T. S. (2007). Mechanism-based inactivation of

reactions. *Curr.Drug Metab,* Vol.8, No.5, pp. 407-447.

2B6, and 2C9. *Drug Metab Dispos.,* Vol.30, No.6, pp. 709-715.

Ginkgo biloba. *J.Anal.Toxicol.,* Vol.29, No.7, pp. 755-758.

secoisolariciresinol diglucoside. *J.Nutr.,* Vol.135, No.4, pp. 795-801.

consistent with the classical Michaelis-Menten model. *Drug Metab Dispos.,* Vol.28,

secoisolariciresinol diglucoside, its aglycone secoisolariciresinol and the mammalian lignans enterodiol and enterolactone *in vitro*. *Food Chem.Toxicol.,*

drug-metabolizing enzyme system of newborn rats. *Mol.Pharmacol.,* Vol.13, No.6,

alkyl substituent on the isomerization of o- quinones to p-quinone methides: potential bioactivation mechanism for catechols. *Chem.Res.Toxicol.,* Vol.8, No.4, pp.

lignans enterolactone and enterodiol in human urine. *J.Steroid Biochem.Mol.Biol.,*

Breemen, R. B. (2003). Identification of novel electrophilic metabolites of piper

cytochrome P450 enzymes: chemical mechanisms, structure-activity relationships and relationship to clinical drug-drug interactions and idiosyncratic adverse drug

derived isoflavan glabridin inhibits the activities of human cytochrome P450S 3A4,

enterolignans in healthy men and women consuming a single dose of

Review of Its Numerous and Varied Biological Activities. *Pharmaceutical Biology,*

oxidative drug metabolism by flavonoids. *J.Pharmacol.Exp.Ther.,* Vol.229, No.1, pp.

metabolic activation of drugs into chemically reactive intermediate: correlation of *in vitro* and *in vivo* generation of reactive intermediates and *in vitro* glutathione conjugate formation in rats and humans. *Chem.Res.Toxicol.,* Vol.20, No.3, pp. 455-464.

dilution gas chromatographic-mass spectrometric method for the determination of isoflavonoids, coumestrol, and lignans in food samples. *Anal.Biochem.,* Vol.233,

bioactivation mechanism for the hepatocarcinogen safrole. *Chem.Res.Toxicol.,* Vol.7, No.3, pp. 443-450.


Borriello, S. P., Setchell, K. D., Axelson, M., and Lawson, A. M. (1985a). Production and

Borriello, S. P., Setchell, K. D., Axelson, M., and Lawson, A. M. (1985b). Production and

Charlet, S., Bensaddek, L., Raynaud, S., Gillet, F., Mesnard, F., and Fliniaux, M. (2002). An

Chichirau, A., Flueraru, M., Chepelev, L. L., Wright, J. S., Willmore, W. G., Durst, T.,

Clark, F. and Reed, R. (1992). Chaparral-Induced Toxic Hepatitis- California and Texas,

Dean, B., Chang, S., Doss, G. A., King, C., and Thomas, P. E. (2004). Glucuronidation,

Dietz, B. and Bolton, J. L. (2007). Botanical dietary supplements gone bad. *Chem.Res.Toxicol.,*

Elbarbry, F. A., McNamara, P. J., and Alcorn, J. (2007). Ontogeny of hepatic CYP1A2 and CYP2E1 expression in rat. *J.Biochem.Mol.Toxicol.,* Vol.21, No.1, pp. 41-50. Felmlee, M. A., Woo, G., Simko, E., Krol, E. S., Muir, A. D., and Alcorn, J. (2009). Effects of

Folk, J. E., Wolff, E. C., Schirmer, E. W., and Ornfield, J. (1962). The kinetics of

Fowler, S. and Zhang, H. (2008). *In vitro* evaluation of reversible and irreversible cytochrome

Gordon, D. W., Rosenthal, G., Hart, J., Sirota, R., and Baker, A. L. (1995). Chaparral

Grant, K. L., Boyer, L. V., and Erdman, B. E. (1998). Case Report- Chaparral-Induced

Grice, H. C., Becking, G., and Goodman, T. (1968). Toxic properties of nordihydroguaiaretic

Hallund, J., Tetens, I., Bugel, S., Tholstrup, T., and Bruun, J. M. (2008). The effect of a lignan

1992. *JAMA,* Vol.268, No.23, pp. 3295-3298.

*Arch.Biochem.Biophys.,* Vol.429, No.2, pp. 244-251.

hepatic lipids in hyperlipidaemic rats. *Br.J.Nutr.,* pp. 1-9.

activities; kinetic models. *J.Biol.Chem.,* Vol.237, pp. 3105-3109.

drug-drug interactions. *AAPS.J.,* Vol.10, No.2, pp. 410-424.

Hepatotoxicity. *Integrative Medicine,* Vol.1, No.2, pp. 83-87.

acid. *Food Cosmet.Toxicol.,* Vol.6, No.2, pp. 155-161.

juice. *Chem.Res.Toxicol.,* Vol.11, No.4, pp. 252-259.

No.3, pp. 443-450.

37-43.

37-43.

pp. 225-229.

pp. 344-355.

Vol.20, No.4, pp. 586-590.

*JAMA,* Vol.273, No.6, pp. 489-490.

bioactivation mechanism for the hepatocarcinogen safrole. *Chem.Res.Toxicol.,* Vol.7,

metabolism of lignans by the human faecal flora. *J.Appl.Bacteriol.,* Vol.58, No.1, pp.

metabolism of lignans by the human faecal flora. *J.Appl.Bacteriol.,* Vol.58, No.1, pp.

HPLC procedure for the quantification of anhydrosecoisolariciresinol. Application to the evaluation of flax lignan content. *Plant Physiology and Biochemistry,* Vol.40,

Hussain, H. H., and Charron, M. (2005). Mechanism of cytotoxicity of catechols and a naphthalenediol in PC12-AC cells: the connection between extracellular autoxidation and molecular electronic structure. *Free Radic.Biol.Med.,* Vol.38, No.3,

oxidative metabolism, and bioactivation of enterolactone in rhesus monkeys.

the flaxseed lignans secoisolariciresinol diglucoside and its aglycone on serum and

carboxypeptidase B activity. III. Effects of alcohol on the peptidase and esterase

P450 inhibition: current status on methodologies and their utility for predicting

ingestion. The broadening spectrum of liver injury caused by herbal medications.

complex isolated from flaxseed on inflammation markers in healthy postmenopausal women. *Nutr.Metab Cardiovasc.Dis.,* Vol.18, No.7, pp. 497-502. He, K., Iyer, K. R., Hayes, R. N., Sinz, M. W., Woolf, T. F., and Hollenberg, P. F. (1998).

Inactivation of cytochrome P450 3A4 by bergamottin, a component of grapefruit


**8** 

*Taiwan* 

**Potential Applications of** 

Mei Fen Shih1 and Jong Yuh Cherng2\*

*Euphorbia hirta* **in Pharmacology** 

*1Department of Pharmacy, Chia-Nan University of Pharmacy & Science, Tainan,* 

*2Department of Chemistry & Biochemistry, National Chung Cheng University, Chia-Yi,* 

*Euphorbia* is a genus of plants belonging to the family Euphorbiaceae*.* Botanist and taxonomist Carl Linnaeus assigned the name *Euphorbia* to the entire genus in the physician's honor. *Euphorbia hirta* is a very popular herb amongst practitioners of traditional herb medicine, widely used as a decoction or infusion to treat various ailments including intestinal parasites, diarrhoea, peptic ulcers, heartburn, vomiting, amoebic dysentery, asthma, bronchitis, hay fever, laryngeal spasms, emphysema, coughs, colds, kidney stones, menstrual problems, sterility and venereal diseases. Moreover, the plant is also used to treat affections of the skin. In this chapter we explore those investigations related to their

Phytochemical analysis of *Euphorbia hirta* (*E. hirta*) revealed the presence of reducing sugar, alkaloids, flavonoids, sterols, tannins and triterpenoids in the whole plant. Some of them are

Epidemiological studies have revealed that polyphenols, including flavonoids, provide a significant protection against development of several chronic diseases such as cardiovascular diseases, cancer, diabetes, infections, aging, and asthma. Two flavonoids have been isolated from *E. hirta*, namely quercitrin and myricitrin (Johnson et al, 1999; Chen, 1991). In general, flavonoids have been reported to possess several proven medicinal properties including antioxidant (Kandaswami & Middleton, 1994), anti-allergic (Singh et al, 2006), antiinflammatory component of asthma (Miller, 2001) and antidiarrheal activity (Galvez et al, 1993; Mallavadhani et al, 2002). Many of the biological actions of flavonoids have been shown to attribute to their antioxidant properties, either through their reducing capacities or as a result of their possible influence on intracellular redox status (Williams et al, 2004). Flavonoids can also interact selectively within the mitogen-activated protein

**1. Introduction** 

**2.** *Euphorbia hirta*

**2.1.1 Flavonoids** 

pharmacological activities (see the section 2.2).

**2.1 Chemical Composition of** *Euphorbia hirta*

well known to possess biological activities (as shown in table 1).

