In vivo disposition of the natural anti-carcinogen indole-3-carbinol after PO administration to rainbow trout

Fd Chem. Toxic. Vol. 27, No. 6, pp. 385-392, 1989 Printed in Great Britain. All rights reserved

0278-6915/89 $3.00 + 0.00 Copyright © 1989 Pergamon Press plc

IN VIVO DISPOSITION OF THE N A T U R A L ANTI-CARCINOGEN INDOLE-3-CARBINOL* AFTER PO ADMINISTRATION TO RAINBOW TROUT R. H. DASHWOODt, L. UYETAKE,A. T. FONG, J. D. HENDRICKS and G. S. BAILEY Department of Food Science and Technology, Oregon State University, Corvallis, OR 97331, USA (Received 15 August 1988; revisions received 22 February 1989)

A~tract--Indole-3-carbinol (13C), a compound found naturally as a glucosinolate in cruciferous vegetables such as broccoli and cabbage, has been shown to modulate the carcinogenic process in a number of animal species. The lack of detailed information on the disposition of I3C in vivo provided the main impetus for the study reported here, in which the distribution and metabolic fate of I3C was assessed in selected tissues and excreta after po administration to rainbow trout (Salmo gairdneri). Animals were fasted for 3 days and given [5-3H]I3C either in the diet or by single oral gavage (40 mg/kg body weight; 15 pCi/kg body weight). Following administration, 75% of the initial 3H-dose was detected within the stomach between 0.5 and 12 hr, after which it was released to distal regions of the gut for subsequent uptake, distribution and elimination. At the end of the study (72 hr) 25% of the administered dose was recovered from the water which reflected excretion through the gills and urinary tract. Significant excretion also occurred in the bile, with approximately 5% of the initial 3H-dose recovered from the bile sacs at 72 hr. Further analyses of the radioactive components in the bile indicated that one or more derivatives of I3C, but not the parent compound itself, are excreted as glucuronide conjugates using this route. Radioactivity accumulated in the liver throughout most of the study, reaching levels of 1-1.5% between 48 and 72 hr of the administered dose. High-performance liquid chromatography analyses indicated the presence of four main radiolabelled species in these livers, one of which co-eluted with the parent compound, I3C. The major radiolabelled species recovered from the liver was tentatively identified as the dimer, 3,Y-diindolylmethane (I3Y), which comprised 40% of the total hepatic radiolabel. This dimer, I33", was also found to accumulate in the diet containing I3C, which reflected a time-dependent dimerization of the parent compound in vitro. These findings are discussed in view of recent postulates of a role for 13C condensation products such as I33' in the mechanism of I3C anti-carcinogenesis.

INTRODUCTION

B 1 (AFB1)-induced hepatocarcinogenesis in rats (Selivonchick et al., unpublished data, 1988) and in rainbow trout (Salmo gairdneri) (Nixon et al., 1984). Investigation of the mechanism responsible for the anti-carcinogenic effects of I3C in trout revealed substantial changes in the in vivo uptake, distribution and metabolism of AFBI such that c a r c i n o g e n - D N A binding in the target organ was significantly attenuated (Goeger et al., 1986). The inhibitory effect of I3C on in vivo D N A binding and tumorigenesis subsequently was studied in trout over a range of both AFB~ and I3C doses, providing evidence for a pure anti-initiating mechanism (solely through the inhibition of c a r c i n o g e n - D N A binding) at dietary levels below 2000 ppm (Dashwood et al., 1988 and 1989a). In addition, the inhibitory response appeared to be linearly related to low doses of I3C (Dashwood et aL, 1988), suggesting a possible protective effect of I3C at concentrations that are encountered in human diets. While considerable effort has been devoted to clarifying the effects of I3C on carcinogen disposition in vivo, only recently has attention been focused on the in vivo disposition of I3C itself. It has been suggested that the mechanism by which dietary indoles inhibit D N A binding and tumorigenesis may be related to their potency as inducers of cytochrome P-448 monooxygenases (Wattenberg, 1983). In recent studies performed by Bradfield and Bjeldanes (1987), I3C was treated with 0.05 M-hydrochloric acid

Cruciferous vegetables such as cabbage, cauliflower and broccoli contain a variety of naturally occurring compounds that modulate the carcinogenic process (Aspry and Bjeldanes, 1983; Boyd et al., 1982; Hendrich and Bjeldanes, 1983; McDanell et al., 1988; Srisangam et al., 1980; Stoewsand et al., 1978; Wattenberg, 1983). One modulator found in cruciferous vegetables is the c o m p o u n d indole-3-carbinol (I3C). This c o m p o u n d was first shown to inhibit tumorigenesis in rodents exposed to polycyclic aromatic hydrocarbons (Wattenberg and Loub, 1978). In further studies, I3C was found to inhibit aflatoxin *According to the IUPAC Nomenclature of Organic Chemistry (2nd Ed, 1969, Butterworths, London) the 'carbinol' nomenclature should be abandoned. Thus, using substitutive nomenclature the compound indole-3carbinol should be named indol-3-ylmethanol. However, for reasons of consistency with respect to the references the former compound name is used throughout the text. tTo whom correspondence should be addressed. Abbreviations: AFB t = aflatoxin B~; dpm = disintegrations/min; GI = gastro-intestinal; HPLC = high-performance liquid chromatography; I3Y=3,Ydiindolylmethane; I3C = indole-3-carbinol; KOAc buffer = 10% acetonitrile in 33/~M-potassium acetate buffer; Rt= retention time; TCDD = 2,3,7,8-tetrachlorodibenzo-pdioxin; TLC = thin-layer chromatography. 385

386

R.H. DASHWOODet al.

under simulated gastric conditions to give a reaction mixture that contained a number of I3C 'condensation products'. Some of the latter reportedly resemble 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in shape and molecular dimensions and, like TCDD, bind to the Ah receptor protein and induce hepatic monooxygenase activity (Bradfield and Bjeldanes, 1987). To date, these monooxygenase induction studies have been reported only for the rat, and have been limited to studies of reaction products generated in vitro. Little information is available on the metabolic fate of 13C in the rat (Shertzer et al., 1988), and 13C condensation products have yet to be identified in vit, o.

In this study we report the in vivo disposition of I3C in selected tissues and excreta of rainbow trout, following administration of [5-3H]I3C (40 mg/kg body weight; 15/~Ci/kg body weight) in the diet or by single oral gavage. A dose level of 2000 ppm I3C in the diet was administered to trout since it is known that at this level carcinogenesis is modified in this species (Dashwood et al., 1989a; Goeger et al., 1986; Nixon et al., 1984). Preliminary characterization of the radioactive components isolated from the liver and bile, and from the dosing media, was also undertaken. MATERIALS AND METHODS

Chemicals. Tritium-labelled I3C (sp. act. 21.14 Ci/ mmol) was synthesized using the route described by Dashwood et al. (1989b). Following elution with ethyl acetate-chloroform (1:1) on thin-layer chromatography (TLC) silica gel plates, a single radioactive spot confirmed the radiochemical purity as greater than 98% (Dashwood et al., 1989b). Ultraviolet (UV) spectra, high-performance liquid chromatography (HPLC) retention times, and TLC characteristics of [5-3H]I3C were identical with those of unlabelled I3C (Sigma Chemical Company, St Louis, MO, USA). The mass spectrum of [5-3H]I3C contained all the major fragments present in the unlabelled sample, but with m / z values greater by a value of 2 (indicating the replacement of the 5-position hydrogen atom with tritium). Samples of tissues, blood and bile were digested with NCS tissue solubilizer and counted in OCS or ACS scintillation fluids (Amersham Corp., Arlington Heights, IL, USA). Arylsulphatase, protease, RNAase and /Lglucuronidase were purchased from Sigma. All other chemicals and reagents were of the purest grade available and from sources described earlier (Dashwood et al., 1988). 3,3'-Diindolylmethane (I3Y) was the gift of Dr L. Bjeldanes of the University of California, Berkeley, CA, USA. Animals. Mount Shasta strain rainbow trout, reared at the Oregon State University Food Toxicology and Nutrition Laboratory, were used in all studies. Animals were housed as described previously (Sinnhuber et al., 1977) except in the gills-urinary excretion studies (see below). Fifty fish (250-320 g) were divided into two groups, one consisting of 21 fish and the other consisting of 29, and fasted for three days prior to dosing. The latter group received [5-3H]I3C incorporated into a dextrose and caseingelatin based semi-purified diet (Sinnhuber et al., 1977) at a concentration of 2000 ppm. Based on the

amount of diet consumed/total fish body weight, each animal received a single average dose of 40 mg/kg body weight; 15 ~Ci/kg body weight. This dose was also employed in a parallel study in which 21 animals were given [5-3H]I3C by single oral gavage in a salmon-oil carrier. At various times after dosing, groups of fish were killed by decapitation and the liver, bile sacs and selected tissues were removed and analysed for their radioactive content. In a second study, groups of three trout were housed individually in glass chambers in 1 litre of continuously aerated water in order to determine the amount of I3C-associated radioactivity excreted into the water. Water was replaced every 6 hr by syphoning. After acclimatization for 36 he, animals received a single oral exposure of [5-3H]I3C by gavage (40 mg/kg body weight; 15 p C i / k g body weight). Excretion of I3C-associated radioactivity into the water (gills plus urine output) was measured at 0.5, 1, 6, 12, 24, 30, 36, 48 and 72 hr after dosing. At each time period, the entire 1-1itre vol of water was replaced to minimize re-uptake of labelled I3C metabolites. Sample collection and processing. Immediately after death, an aliquot of blood (0.1 ml) was collected from each animal by cardiac puncture and the livers were perfused in situ by way of the hepatic vein with 0.9% (w/v) NaCI at room temperature to clear the tissue of blood. Prior to removal and weighing of the livers, the bile sacs were removed and the total volume of bile in each sac was recorded. Aliquots (0.1 ml) were removed for determination of radioactivity and the remainder of the bile was frozen in dry ice together with the livers. The entire length of the gastrointestinal (GI) tract was removed and divided into three sections--the stomach, pyloric caecum and lower GI tract. Each section was weighed before and after removal of its contents. Triplicate samples of washed liver tissue and GI tract contents (0.5 g) were digested in vials containing NCS tissue solubilizer (2 ml). Digestions were conducted at 4Y~C overnight, or until homogeneous. After neutralization with acetic acid, OCS scintillation fluid (18 ml) was added. Blood and bile samples (0.1 ml) were digested for 1 hr at 45°C in vials containing NCS tissue solubilizer (1 ml). Each sample was then decolourized by addition of 0.2 ml benzoyl peroxide solution (0.2 g/ml, in toluene), prior to addition of ACS scintillation fluid (18 ml). Aliquots (1 ml) from the 1-1itre vol of water collected at each time-point in the gills-urinary excretion studies were counted directly after addition of ACS scintillation fluid. All samples were counted using a Beckman LS7500 liquid scintillation counter (Palo Alto, CA, USA). Analysis o f biliary metabolites. Aliquots of bile (0.2ml) were diluted with 10ml 0.02M-sodium acetate buffer (pH 6) and applied to a C-18 reversedphase Sep-Pak cartridge (Waters Associates, Milford, MA, USA). The cartridge was washed with 10ml 10% MeOH in the same buffer and then with 10ml 50% MeOH in buffer. Approximately 85-90% of the radiolabel applied to the cartridge was eluted in the 50% methanolic buffer. After removal of the solvent in a rotary evaporator, the eluate was divided into three equal aliquots (two of which were treated with either arylsulphatase or /3-glucuronidase and the third was left untreated as described previously

Disposition of [3H]I3C in rainbow trout (Loveland et al., 1984). Each aliquot was then reapplied to fresh Sep-Pak columns and the resulting 50% methanolic eluates were analysed by TLC and HPLC. TLC was performed on commercially prepared Silica Gel 60 aluminium sheets (0.2 mm; E. Merck, Darmstadt, FRG) using the solvent systems ethyl acetate-chloroform (85: 15) and butanol-acetonitrile-acetic acid-water (15 : 10 : 3 : 12). Radiolabelled bands were detected using a TLC radiochromatoscan 4-pi scanner (Berthold Varian Aerograph Walnut Creek, CA, USA) or by liquid scintillation counting of 5-mm strips cut from the TLC plate. In HPLC studies, 50% methanolic eluates from the Sep-Pak cartridges were applied to a Nova-Pak C-18 reversedphase column (3.9 mm × 15 cm ID; Waters Associates, Milford, MA, USA). Using a constant flow rate of 1 ml/min, compounds were eluted from the HPLC column and detected with a Shimadzu SPD-6AV spectrophotometer detector (Kyoto, Japan) linked to an SCL-6A system controller and a Beckman 171 radioisotope detector/110B solvent delivery module (Palo Alto, CA, USA). The HPLC solvent programme was initiated with 10% MeOH in water for 5 min, a linear gradient was then established over i 5 min to increase the MeOH concentration to 100%. After eluting with 100% MeOH for 5 min, the MeOH concentration was allowed to return to 10% in water using a 15-min linear gradient. H P L C analysis of liver metabolites. Pooled liver samples were homogenized in phosphate buffered saline using a glass Teflon homogenizer. Homogenized samples were extracted twice with 10 vol ethyl acetate, and then the combined ethyl acetate fractions, which accounted for greater than 85% of the initial hepatic radioactivity, were taken to dryness under nitrogen. The residue was dissolved in 10% acetonitrite in 33 #M potassium acetate buffer (pH 5) ('KOAc buffer'). After centrifugation for 2 min at 15,000 g using a TOMY-Seiko MC-150 microcentrifuge (Tokyo, Japan) the supernatants were passed through a 45 # filter and injected onto a Nova-Pak C-18 reversed-phase HPLC column. During the initial phase the column was eluted with 10%-55% acetonitrile in KOAc buffer over 25 min using a linear gradient (flow rate 0.8 ml/min). After eluting with 55% acetonitrile in KOAc buffer for 10min, the acetonitrile concentration was further increased to 100% over 15 min using a linear gradient. The acetonitrile concentration was maintained at 100% for 5min before returning to the starting conditions using a 5-min linear gradient. Extraction of [3H]13C from diet and oral gavage media. Diet (2 g) containing [3H]I3C was blended with 5 vol ice-cold buffer (0.15 M-potassium chloride, 0.01 M-potassium phosphate; pH 7.4) and extracted with 10 vol ethyl acetate. After centrifugation at 5000 g for 15 min, the ethyl acetate layer was removed and the extraction process repeated. The organic fractions, containing typically greater than 88% of the initial radioactivity, were combined and taken to dryness in vaeuo. The residue was dissolved in 10% acetonitrile-KOAc buffer and analysed using HPLC under the conditions described above for liver samples. Information on the stability of I3C under the storage conditions used in this and previous studies FCT 27'6~C

387

(Dashwood et al., 1989b) was determined by extraction of the I3C-associated radioactivity at various times from the diet, which had been stored at 4°C. [3H]I3C was also incorporated into salmon oil used in the oral gavage studies, and was extracted (twice) with 10 vol ice-cold MeOH. The combined methanolic extracts accounted for greater than 90% of the initial radioactivity. After drying in vacuo, the methanolic extracts were dissolved in 10% acetonitrile-KOAc buffer and analysed using HPLC under the conditions described above for the liver samples. RESULTS

Passage of [3H]I3C through the GI tract Transport of radiolabel down the gut following oral administration of [5-3H]I3C is shown in Fig. 1. " Stomach Q Garage

TTI

O x u =E a. (3

["]Diet

[tiI

; ~1! 24~4St72 f

2.0

>

.~_

1.5

1.0

g=

0.5

>~

8

2.0

Lower GT Trncf

1.5

1.0

0.,5

'0.5 ~ 1 ' 6 ' 1 2 ' 2 4 ' 4 8 ' 7 2 I Time a f t e r administration of [ 5 - 3 H 3 1 3 C (hr)

Fig. l. Movement of [3H]13C-associated radioactivity through the G! tract of rainbow trout following administration (40mg/kg body weight; 15/zCi/kg body weight) either in the diet or by single oral gavage. Values are means + SD for groups of three animals except at time points 48 and 72 hr in the dietary study, for which groups of seven animals were employed. At each time point, the data from the oral gavage and dietary studies were compared using Student's t-test. *Values for animals of the oral gavage group differ significantly from those obtained for animals of the dietary group. P < 0.05.

R.H. DASHWOODet al.

388

Tritium levels recovered from the stomach over the initial 12-hr period were about 7 x 106dpm in animals given [5-3H]I3C in the diet or by oral gavage (approximately 75% of the administered dose) (Fig. 1). Levels of radioactivity declined thereafter (Fig. 1) such that by 72 hr virtually all of the radiolabel had passed from the stomach to the distal regions of the gut and elsewhere. Coincident with the loss of radioactivity from the stomach, levels of radioactivity increased in the pyloric caeca and shortly thereafter in the lower GI tract (Fig. 1). In the dietary study, peak values were approximately 0.8 x 10 6 dpm and 0.5 × l06 dpm in the pyloric caeca and lower GI tract, respectively. Following oral gavage administration, peak values both in the pyloric caeca and lower GI tract were approximately 1.5 x 106 dpm. These values represent 16% and 8.5% of the administered dose in the pyloric caeca, and 16% and 5.5% in the lower GI tract in trout given [5-3H]I3C by oral gavage or in the diet, respectively. Thus, oral gavage administration of [5-3H]I3C in salmon oil gave similar stomach retention values over a 12-hr time period, followed by a more rapid transfer to the pyloric caeca and lower GI tract (P < 0.05; Fig. 1) when compared with animals given [5-3H]I3C in the diet.

125

% x

lO0 ~ 75 o

~ 5o ~ 25

5 Time after administration of [ 5 - 3 H ] I 3 C ( h r )

Fig. 3. Hepatic radioactivity in rainbow trout at 0.5, 1, 6, 12, 24, 48 and 72 hr following administration of [5-3H]I3C in the diet or by oral gavage (40mg/kg body weight; 15 #Ci/kg body weight). For the number of animals in each group, see legend to Fig. I. *Values are mean + SD. Values for livers from animals of the oral gavage group differ significantly (Student's t-test) from those obtained for livers from animals in the dietary group at 72 hr: P < 0.05.

Absorption o f [3H]I3C As shown in Fig. 2, peak tritium levels in the blood were detected within the first hour after both oral gavage and dietary exposure. A rapid decline (approximately 50%) was observed in both groups over the next 24 hr leading to a steady-state level over the period 24-72 hr after dosing. Peak values in the blood were about 55 x l03 dpm/ml and 30 x 103 dpm/ml for animals given labelled I3C by oral gavage or in the

0 x

~o 6O

diet, respectively. Based on the mean blood vol of 4.09 ml/100 g rainbow trout (Gingerich et al., 1987), these values correspond to about 7% (diet) and 4% (oral gavage) of the administered dose. In the liver (Fig. 3) I3C-derived radioactivity was detected at all the time periods examined. M a x i m u m levels of radioactivity detected were 70 x l 0 3 dpm for animals in the dietary study (48 hr) and about 125 x 103 dpm for animals administered [5-3H]I3C by oral gavage (72 hr). These values represent approximately 0.79 (diet) and 1.5% (oral gavage) of the administered dose. Subsequent analyses indicated that 20% of this total hepatic radioactivity was associated with the parent molecule, I3C (see below).

50

Q 40 ~,,,,,,,,,~

~ ~°

Diet

T

500

o

T

Gavage

~

Diet

4OO

~0 ~ 300 ~

20

~

10

E

x

OQ .o

"s

0

i

I

I

0 6 12

24

418

712

Time after administration of [ 5 - 3 H ] I 3 C (hr)

Fig. 2. Levels of radioactivity detected in the blood at 0.5, 1, 6, 12, 24, 48 and 72 hr after administration of [5-3H]I3C (40 mg/kg body weight; 15 #Ci/kg body weight) in the diet or by oral gavage to rainbow trout. Values are means + SD. For the number of animals in each group see legend to Fig. 1. *Values for animals of the oral gavage group differ significantly (Student's t-test) from those obtained for animals of the corresponding dietary group: P < 0.05.

200 100

o

0.5 1 6 1 2 2 4 4 8 7 2

0.5 1 6 1 2 2 4 4 8 7 2

Time of'(er administration [ 5 - 3H] I3C (hr)

Fig. 4. Cumulative excretion of radioactivity in the bile of rainbow trout following oral gavage and dietary administration of [5-3H]I3C (40mg/kg body weight; 15 #Ci/kg body weight). Values are means + SD. The number of animals used in these studies is given in legend of Fig. 1. *Values for bile from animals of the oral gavage group differ significantly (Student's t-test) from those obtained for animals of the corresponding dietary exposed group at 48 and 72 hr following administration: P < 0.0f

Disposition of [3H]I3C in rainbow trout

Biliary and gills-urinary excretion Cumulative excretion of radioactivity in the bile following oral dosing with [5)H]I3C is shown in Fig. 4. Maximum levels of radioactivity recovered in the bile were 430 x 103 dpm for animals in the oral gavage group and 140 x 103dpm for animals in the dietary study. These values correspond to 4.6% (oral gavage) and 1.5% (diet) of the administered dose. Radioactivity excreted in the urine and through the gills is presented in Fig. 5 in terms of percent administered dose excreted (%)/hr (diamond-shaped symbols) and the cumulative dose (%) excreted (triangular symbols). After an initial, very rapid appearance of the radiolabel within the first halfhour of dosing (diamond symbols), the rates of excretion declined sharply over the next 6 hr and then fell at a slower rate for the remainder of the study. As a result, 3H levels accumulated in the water continuously over 72 hr (triangles), with a final cumulative excretion mean value of 25% of the administered dose.

Analysis of biliary metabolites Biliary metabolites of [3H]I3C were analysed using TLC and HPLC. These analyses indicated the absence of the parent compound and the presence of one product whose migration on the TLC plates or elution from the HPLC columns could be altered by incubation with/~-glucuronidase. In the solvent system ethyl acetate-chloroform (85:15), the biliary metabolites remained at the origin of the silica plates (Table I), whereas control bile, from trout treated with the salmon-oil carrier only, that contained pure tritiated I3C migrated with an R r of 0.29. Biliary metabolites of [3H]I3C incubated with /3-glucuronidase migrated in the solvent system butanol-acetic acid-acetonitrile-water (15 : 10: 3 : 12) with an Rf of

8

40

6

3O o

o ¢3

10

2

I

o

o

0

12

24

36

48

60

Time a f t e r administration of [ 5 - 3 H ]

g~

72 13C (hr)

Fig. 5. Excretion of [3H]I3C metabolites by the gills and urinary routes in rainbow trout following administration of [5-3H]I3C (40 mg/kg body weight; 15 pCi/kg body weight) by single oral gavage. The 3H dose excreted (%)/hr ( 0 ) and cumulative (%) a . Values are means + SD for three animals.

389

Table 1. TLC and HPLC analyses of the radioactive components in the bile of rainbow trout following administration of 15-3H]I3C * by single oral gavage TLCt

HPLC~

Sample

RF~

R~2

R t (min)

Untreated bile Bile + arylsulphatase Bile + #-glucuronidase Control bile + I3C 'spike'

0.00 0.00 0.00

0.62 0.64 0.81

13.9 14.0 18.0

0.29 (0.79)§

--

15.4

[5)H]I3C = [5-3H]indole-3-carbinol *Animals received a dose level of 40 mg/kg body weight; 15/~Ci/kg body weight. tRF~ and RF2 values were obtained using the solvent systems ethyl acetate-chloroform (85:15) and butanol acetic acid-aceto-nitrilewater (15: 3 : 10:12), respectively. ~Samples for HPLC were prepared by Sep-Pak purification (twice) and TLC, see Materials and Methods. §The value in parenthesis was obtained from a minor band at heavy loading.

0.81 compared with 0.62 for untreated bile. Treatment with arylsulphatase had no effect on the TLC characteristics of the biliary metabolites of [3H]I3C (Table 1). Analysis of the bile using HPLC revealed a single radiolabelled peak with a retention time (Rt) of 13.9 rain for the untreated bile sample. Incubation of the bile sample with/3-glucuronidase increased the R t of the biliary metabolite to 18.0min. This retention time is different from that of the I3C parent compound present in the spiked control sample (R t 15.4 min). Incubation of the bile sample with arylsulphatase did not alter the R t of the radioactive metabolite (Table 1). While these data were from trout administered [5-3H]I3C by oral gavage, similar results were observed on analysis of the bile from trout fed diet containing [5-3H]I3C (data not shown).

Hepatic metabolite studies HPLC retention times of radioactive components in the ethyl acetate extracts of liver homogenates from trout given [5-3H]I3C by oral gavage are given in Fig. 6. Four major radioactive peaks were observed, with elution times of 5.5, 28, 29 and 36 min. Two minor radioactive peaks with elution times of 21 and 25.5min also were detected (Fig. 6). In the corresponding UV absorbance spectrum (280nm), shown in Fig. 6, two main peaks were identified by co-elution with known standards as I3C and the dimer 3,3'-diindolylmethane (I33') (Rt of 5.15 and 29.4min, respectively). Studies are in progress to further characterize these liver metabolites.

Stability of [3H]13C after incorporation in the diet As shown in Table 2, both I3C and the dimer 133' were identified in extracts taken at various times after incorporation of [5-3H]I3C in the diet. For example, 4% of the radioactivity extracted from the diet after 2 hr co-eluted with I33', whereas the remaining radioactivity corresponded with the parent compound, I3C. At 72 hr, 77% of the radioactivity extracted from the diet corresponded with I3C, compared with 22% for 133'. Extraction of [5-3H]I3C from salmon oil

390

R . H . DASHWOOD et al. DISCUSSION

H

H

~,3'- Oiindolylmet bane

(T33']

[~H

A280

CH20N I

1000i DPM

5°0t 0 i 0

10

nB I

20 30 MINUTES

410

6O

Fig. 6. HPLC analysis of the radioactive components in ethyl acetate extracts of liver from rainbow trout 48 hr after administration of [5-3H]I3C by oral gavage. Four major and two minor radioactive peaks were observed. Two of the main peaks had R~ values of 5.5 and 29 min, respectively. Identification of these two main UV absorbing peaks as I3C and I33' (top) was based on co-elution with known standards. The two minor radioactive peaks shown in this figure (R, of 20.5 and 26 min) were not detected in all liver samples. (Values for radioactivity are corrected for background). gave similar results to those described above at 2 hr in that after incorporation approximately 88 and 12% of the total radioactivity corresponded with I3C and I33', respectively. After 72 hr, 20% of the radioactivity co-eluted with I3Y compared with 68% for I3C (data not shown). These data indicate a timedependent conversion of the parent molecule to its dimer. The conversion appeared to plateau after 24 hr at an approximate ratio of 80% I3C:20% I3Y (Table 2), which indicated a possible equilibrium between the monomeric and dimeric structures. Table 2. HPLC analysis of 13C-associated radioactivity extracted from diet at various times following incorporation of the radiolabel Total radioactivity eluted (%) Time (hr)

I3C

I33'

0 2 6 24 48 72

96 96 88 78 80 77

1 4 9 22 20 22

13C = indole-3-carbinol Values represent the percentage total radioactivity eluted under two main HPLC peaks, both of which corresponded with the I3C and 133' peaks shown in Fig. 6. Identification of the peaks as I3C and I33' was based on co-elution with known standards. The results presented here are from a single HPLC run, and are representative from at least three independent determinations at each time-point.

In vivo disposition of" 13C: uptake, distribution and excretion Studies on the movement of I3C-associated radioactivity down the GI tract indicated that the majority of the administered dose remained within the stomach for at least 12 hr (Fig. 1), after which it was released to distal regions of the gut for subsequent uptake and distribution. Measurements of radioactivity in the blood indicated rapid absorption of I3C-associated radioactivity from the gut, with the highest blood levels of radioactivity occurring within the earliest assay time points (less than 24 hr). However, there was also evidence for absorption across the gut into the blood after 24 hr (Fig. 2), and levels of radioactivity continued to increase in the liver, bile and water over the entire time-course of this study (Figs 3-5). Biliary excretion plays an important role in the elimination of certain I3C metabolites in the trout, with up to 5.3% of the initial 3H-dose recovered from the bile sacs 72hr after administration (430 × 10~ dpm; oral gavage group, Fig. 4). Further studies established that greater than 85% of the biliary radiolabel applied to a reversed-phase Sep-Pak column was eluted over a narrow range of solvent conditions (about 37% MeOH), which indicated the presence of one or a limited number of discrete I3C metabolite(s) of similar polarity. Subsequent analysis of the eluate with TLC and HPLC provided some indication as to the identity of this metabolite (Table 1). Thus, chromatography after digestion with /%glucuronidase provided evidence that one or more derivatives of I3C are excreted in bile as glucuronide conjugates. These data are consistent with previous findings in trout, which indicate that the metabolites of compounds such as AFB] (Loveland et al., 1984), tetrachlorobiphenyl and di(2-ethylhexyl) phthalate (Melancon and Lech, 1976a, b), l-naphthyI-Nmethylcarbamate (Statham et al., 1975) and 2',5dichloro-4'-nitrosalicylanilide (Statham and Lech, 1975) are excreted in the bile predominantly as glucuronides. The importance of the gills and urinary tract as routes of excretion is provided by the fact that 6% of the initial dose was detected in the water within 30 rain of exposure, and that about 25% of the label was recovered from the water after 72 hr. This indicates that one or both routes of excretion (gills and urine) constitute a primary means of eliminating I3C metabolites after po administration to trout. Limiting the amount of [3H]I3C-containing diet or administration of the radiolabel by single oral gavage to trout minimized faecal elimination, which could constitute an additional route of excretion during daily feeding or following repeated oral gavage administration. No attempts were made in this investigation to identify gills v. urinary excretion, or to further characterize the I3C biliary metabolites. Previous studies in trout have employed almost exclusively the dietary route of administration in tumour studies or when investigating the mechanism of I3C action. It was not clear whether or not gavage would provide an equivalent route for I3C exposure. Hence, the oral gavage route of administration also

Disposition of [3H]I3C in rainbow trout was employed in this study for comparative purposes and to extend the current I3C database in this species. Some quantitative differences were observed between the dietary and oral gavage routes of administration (Figs 1-4). These were most evident in the analysis of radioactivity in the blood (Fig. 2), where the values measured in the oral gavage study 1-72 hr after dosing were consistently lower than those observed for the corresponding animals given [5-3H]I3C in the diet. Despite these quantitative differences, similar overall 'trends' were observed with both routes of administration in the tissues and excreta studied here; and we have interpreted these findings as supporting the use of oral gavage administration as an alternative route of I3C exposure in the trout. Condensation products and the mechanism anti-carcinogenesis

o f 13C

Recent studies by Bradfield and Bjeldanes (1987) have highlighted the importance of route of administration in the mechanism of action of I3C in vivo. When administered ip to rats, thus bypassing the acidic conditions of the stomach, I3C had little effect on hepatic cytochrome P-448 monooxygenase activity. The same I3C dose when administered orally produced a 15-fold enzyme induction. In contrast, Bradfield and Bjeldanes (1987) found that the condensation products generated in vitro by the action of hydrochloric acid on I3C (see Introduction) were able to induce hepatic monooxygenase activities following both ip and oral routes of exposure. This was considered to be indicative of a role for the condensation products of I3C in the mechanism of action of the dietary modulator in vivo (Bradfield and Bjeldanes, 1987). However, the metabolic fate of I3C was not studied, and there was no direct evidence for the formation of I3C condensation products in vivo. One such condensation product, the dimer I3Y, was tentatively identified both in trout liver 48 hr after dosing (Fig. 6) and in the dosing media at various times after preparation (Table 2). Other than the parent compound, I3C, 133' was the only radiolabelled species present in extracts from the diet (Table 2) and salmon oil. Several additional, and as yet uncharacterized metabolites were detected in the 48-hr liver extracts and these accounted for 40% of the total radioactivity (Fig. 6). In comparison, 20% of the total radioactivity in liver corresponded with the parent molecule, I3C (Fig. 6). Collectively, these data highlight two important points. First, that I3C undergoes limited dimerization when incorporated into the diet or the salmon oil used in these studies. It is important to stress that [3H]I3C was incorporated into the diet in an identical fashion as described in previous studies (Dashwood et al., 1989b) that showed a potent anti-carcinogenic effect by this compound. The presence of I33' in the diet and oral gavage media is of interest because this dimer represents an intermediate in the formation of several other I3C condensation products, the potential anticarcinogenic activities of which remain to be evaluated. Secondly, the spectrum of I3C-derived species in liver at 48 hr differed both quantitatively and qualitatively from the HPLC profiles for the dosing media. It is presently unclear whether these radiolabelled species in liver represent additional I3C condensation

391

products formed in vivo, or whether they are enzymecatalysed products of I3C and I33'. Studies currently in progress should provide a detailed chemical characterization of the I3C species present in the liver and GI tract of trout at various times after oral administration. The results from a number of recent studies have raised questions about the proposed relationship between altered cytochrome P-448 activities and the mechanism of I3C anti-carcinogenesis. It has been reported that I3C protects against the DNA-damaging effects of benzo[a]pyrene and N-nitrosodimethylamine in mice without inducing monooxygenase activities (Shertzer, 1983 and 1984). Shertzer et al. (1988) have suggested that I3C and related indoles may be capable of interacting with reactive intermediates of carcinogens and toxins (free-radicals and electrophiles) and that this scavenging role, rather than monooxygenase induction, constitutes the central mechanism of I3C inhibition. Studies in the rainbow trout indicate clearly that 13C inhibition of AFB~ carcinogenesis proceeds without concomitant alterations, either induction or repression, in hepatic cytochrome P-448 monooxygenase activities (Dashwood et al., 1988; Eisele et al., 1983; Goeger et al., 1986; Nixon et al., 1984), but appears to involve direct reversible enzyme inhibition by one or more of the I3C condensation products discussed herein (Swanson et al., unpublished results, 1988). For example, Eisele et al. (1983) reported that I3C does not increase hepatic microsomal cytochrome P-450 or P-488 levels, and fails to induce the associated mixed-function oxidase activities, following administration of I3C to trout at a dietary concentration of 500 ppm. Furthermore, in trout fed diets containing 500 to 2000 ppm I3C, no detectable changes were found in the specific contents of hepatic cytochrome P-450 isoenzymes LM 2 and LM4b, or in the liver microsomal uridine diphosphate glucuronyl-transferase and cytosolic glutathione S-transferase activities (Swanson et al., unpublished results, 1989). However, enzyme kinetic studies have indicated that a I3C reaction mixture (see Introduction) inhibits the metabolic activation of AFB~ by rat or trout liver microsomes, such that the apparent K m for AFB1-DNA binding was increased without changing the associated V~ax (Swanson et al., unpublished results, 1989). These findings suggest that mechanism(s) other than induction of cytochrome P-448 monooxygenases by I3C condensation products may exist in vivo. thank J. Casteel, T. Will and S. Cleveland for the animal care, M. Dutchuk for assistance during oral gavage procedures, and Dr D. E. Williams for the use of HPLC equipment. This work was supported in part by grants ES03850 and ES00210 from the NIEHS and grant CA34732 from the National Cancer Institute. Technical Paper No. 8774, Oregon Agricultural Experiment Station. Acknowledgements--We

REFERENCES

Aspry K. E. and Bjeldanes L. F. (1983) Effects of dietary broccoli and butylated hydroxyanisole on liver-mediated metabolism of benzo[a]pyrene. Fd Chem. Toxic. 21, 133.

392

R. H. DASHWOODet al.

Boyd J. N., Babish J. G. and Stoewsand G. S. (1982) Modification by beet and cabbage diets of aflatoxin Bt-induced rat plasma ~-foetoprotein elevation, hepatic tumorigenesis and mutagenicity of urine. Fd Chem. Toxic. 20, 47. Bradfield C. A. and Bjeldanes L. F. (1987) Structure activity relationships of dietary indoles: a proposed mechanism of action as modifiers of xenobiotic metabolism. J. Toxicol. envir. Hlth 21, 311. Dashwood R. H., Arbogast D. N., Fong A. T., Hendricks J. D. and Bailey G. S. (1988) Mechanisms of anti-carcinogenesis by indole-3-carbinol: detailed in vivo DNA binding dose-response studies after dietary administration with aflatoxin B1 (AFBt). Carcinogenesis 9, 427. Dashwood R. H., Arbogast D. N., Fong A. T., Pereira C., Hendricks J. D. and Bailey G. S. (1989a) Quantitative interrelationships between aftatoxin B1 carcinogen dose, indole-3-carbinol anti-carcinogen dose, target organ DNA adduction, and final tumor response. Carcinogenesis 10, 175. Dashwood R. H., Uyetake L., Fong A. T., Hendricks J. D. and Bailey G. S. (1989b) The synthesis of [3H]-indole3-carbinol, a natural anticarcinogen from cruciferous vegetables. J. label. Cpds Radiopharm. In press. Eisele T. A., Bailey G. S. and Nixon J. E. (1983) The effect of indole-3-carbinol, an aflatoxin BI hepatocarcinoma inhibitor, and other indole analogs on the rainbow trout hepatic mixed function oxidase system. Toxicology Lett. 19, 133. Gingerich W. H., Pityer R. A. and Rach J. J. (1987) Estimates of plasma, packed cell and total blood volume in tissues of the rainbow trout (Salmo gairdneri). Comp. Biochem. Physiol. 87A, 251. Goeger D. E., Shelton D. W., Hendricks J. D. and Bailey G. S. (1986) Mechanisms of anti-carcinogenesis by indole-3-carbinol: effect on the distribution and metabolism of aflatoxin BI in rainbow trout. Carcinogenesis 7, 2025. Hendrich S. and Bjeldanes L. F. (1983) Effects of dietary cabbage, Brussels sprouts, lllicium verum, Schizandra chinensis and alfalfa on the benzo[a]pyrene metabolic system in mouse liver. Fd Chem. Toxic. 21, 479. Loveland P. M., Nixon J. E. and Bailey G. S. (1984) Glucuronides in bile of rainbow trout (Salmo gairdneri) injected with [3H]-aflatoxin BI and the effects of dietary b-naphthoflavone. Comp. Biochem. Physiol. 78, 13. McDanell R., McLean A. E. M., Hanley A. B., Heaney R. K. and Fenwick G. R. (1988) Chemical and biological

properties of indole glucosinolates (glucobrassicins): a review. Fd Chem. Toxic, 26, 59. Melancon M. J. and Lech J. J. (1976a) Isolation and identification of a polar metabolite of tetrachlorobiphenyl from bile of rainbow trout exposed to t4C-tetrachlorobiphenyl. Bull. envir. Contam. Toxicol. 15, 181. Melancon M. J. and Lech J. J. (1976b) Distribution and biliary excretion products of di-2-ethylhexyl phthalate in rainbow trout. Drug Metab. Dispos. 4, 112. Nixon J. E., Hendricks J. D., Pawlowski N. E., Pereira C. B., Sinnhuber R. O. and Bailey G. S. (1984) Inhibition of aflatoxin B 1 carcinogenesis in rainbow trout by flavone and indole compounds. Carcinogenesis 5, 615. Shertzer H. G. (1983) Protection by indole-3-carbinol against covalent binding of benzo[a]pyrene metabolites to mouse liver DNA and protein. Fd Chem. Toxic. 21, 3I. Shertzer H. G. (1984) Indole-3-carbinol protects against covalent binding of benzo(a)pyrene and N-nitrosodimethylamine metabolites to mouse liver macromolecules. Chemico-Biol. Interactions 48, 81. Shertzer H. G., Berger M. L. and Tabor M. W. (1988) Intervention in free radical mediated hepatotoxicity and lipid peroxidation by indole-3-carbinol. Biochem. Pharmac. 37, 333. Sinnhuber R. O., Hendricks J. D., Wales J. H. and Putnam G. B. (1977) Neoplasms in rainbow trout, a sensitive animal model for environmental carcinogenesis. Ann. N.Y. Acad. Sci. 298, 389. Srisangam C., Hendricks. D. G., Sharma R. P.. Salinkhe D. K. and Mahoney A. W. (1980) Effects of dietary cabbage on the tumorigenicity of 1,2 dimethylhydrazine in mice. J. Fd Saf. 4, 235. Statham C. N. and Lech J. J. (1975) Metabolism of 2',5-dichloro-4'-nitrosalicylanilide (Bayer 73) in rainbow trout (Salmo gairdneri). J. Fish. Res. Bd Can. 32, 515. Statham C. N., Pepple S. K. and Lech. J. J. (1975) Biliary excretion products of l-[1-t4C]-N-methylcarbamate in rainbow trout (Salmo gairdneri). Drug Metab. Dispos. 3, 400. Stoewsand G. S., Babish J. G. and Wimberley H. C. (1978) Inhibition of hepatic toxicities from polybrominated biphenyls and aflatoxin BI in rats fed cauliflower. J. Envir. Path. Toxicol. 2, 399. Wattenberg L. W. (1983) Inhibition of neoplasia by minor dietary constituents. Cancer Res. (Suppl.) 43, 2448s. Wattenberg L. W. and Loub W. D. (1978) Inhibition of polycyclic aromatic hydrocarbon-induced neoplasia by naturally occurring indoles. Cancer Res. 38, 1410.