Nutrition, Pharmacology, and Toxicology

Feeding of Potato, Tomato and Eggplant AlkaloidsAffects Food Consumption and Body and Liver Weightsin Mice1

MENDEL FRIEDMAN,2 P. R. HENIKA AND B. E. MACKEY

Food Safety and Health Research Unit, USDA-ARS Western Regional Research Center,800 Buchanan Street, Albany, California 94710

ABSTRACT Reduced liverweight was used to evaluatethe potential toxicity in mice of four naturally occurringsteroidal glycoalkaloids: a-chaconine and a-solanine, or-tomatine and solasonine. Increased liver weight wasused to evaluate the three corresponding steroidal agly-cones: solanidine, tomatidine, and solasodine and thenon-alkaloid adrenal steroid dehydroepiandrosterone(DHEA). Adult female Swiss-Webster mice were feddiets containing test compound concentrations of 0(control), 1.2, 2.4 or 4.8 mmol/kg diet for 7, 14 or 28d. Absolute liver weights (LW) and relative liver weights(liver weight/body weight x 100, %LW/BW)were determined at autopsy. The %LW/BWwas lower than thatof the controls in mice fed the potato glycoalkaloid a-chaconine (-10%, P -~0.05) for 7 d with the 2.4 mmol/kg diet dose. Under these same conditions, %LW/BWwas greater than that of controls in mice fed two agly-cones: solanidine (27%, P s 0.001 ) and solasodine (8%,P £0.01). Relative liver weight increases induced by theaglycones were determined under time and dose conditions in which differences in body weight and food consumption were not significant (2.4 mmol/kg diet for 28d). under these conditions, the observed %LW/BW increases relative to the controls were as follows: solanidine (32%, P -.. 0.001), solasodine (22%, P ^ 0.001)and DHEA (16%, P -- 0.001). Solanidine, solasodineand DHEA were equally potent and were more potentthan tomatidine. We also observed that the greater%LW/BW in mice fed 2.4 mmol/kg diet solasodine orsolanidine for 14 d declined to near control values if theywere fed control diets for another 14 d. The increase inrelative Ihrerweight induced by solanidine and solasodineis a reversible adaptive response. These findings and theapparent effects of structure on biological activity shouldserve as a guide for the removal of the most toxic compounds from plant foods. The implications of the resultsfor food safety and health are discussed. J. Nutr. 126:989-999, 1996.

INDEXING KEY WORDS:

•glycoalkaloids •food consumption•hepatomegaly •mice

food safety

Steroidal glycoalkaloids have been found in potatoes(Friedman and Dao 1992), green tomatoes (Friedmanand Levin 1995) and eggplants (Aubert et al. 1989).Symptoms of glycoalkaloid toxicity experienced by animals and humans include colic pain in the abdomenand stomach, gastroenteritis, diarrhea, vomiting, burning sensation about the lips and mouth, hot skin, fever,rapid pulse and headache (van Gelder 1990). The reported toxicity of these glycoalkaloids may be due tosuch adverse effects as 1) anticholinesterase effects onthe central nervous system (Roddick 1989), 2) induction of hepatic ornithine decarboxylase, a cell proliferation marker enzyme (Caldwell et al. 1991), and 3) disruption of cell membranes affecting the digestive system (Blankemeyer et al. 1992 and 1995, Roddick et al.1992). Toxicity does not seem to occur at the geneticlevel (Friedman and Henika 1992). One manifestationof these adverse effects may be alkaloid-induced terato-genicity (Keeler et al. 1991, Renwick et al. 1984).The estimated highest safe level of total glycoalka

loids for human consumption is ~1 mg/kg bodyweight, a level that may cause gastrointestinal irritation (Slanina 1990). The acute toxic dose is estimatedto be -1.75 mg/kg body weight (van Gelder 1990). Alethal dose may be as low as 3-6 mg/kg body weight(Morris and Lee 1984).Human consumption of potatoes, an excellent

source of carbohydrates and good quality protein, variesby country. For example, the average daily per capitaintake in the United States is -167 g (Friedman 1996);in the United Kingdom, it is 140 g (Hopkins 1995) and

1The costs of publication of this article were defrayed in part bythe payment of page charges. This article must therefore be herebymarked "advertisement" in accordance with 18 USC section 1734solely to indicate this fact.2To whom correspondence should be addressed.

0022-3166/96 $3.00 ©1996 American Institute of Nutrition.Manuscript received 15 May 1995. Initial review completed 21 June 1995. Revision accepted 12 December 1995.

989

by on Novem

ber 22, 2009 jn.nutrition.org

Dow

nloaded from

990 FRIEDMAN ET AL.

in Sweden, 300 g (Slanina 1990). The cited amount forthe United Kingdom is estimated to contain 14 mgglycoalkaloids. Although the glycoalkaloid concentration of most commercial potato varieties is usually below a suggested guideline of 200 mg/kg fresh potatoes(Maga 1994), the concentration can increase substantially on exposure to light and as a result of mechanicalinjury, including peeling and slicing (Dao and Friedman1994). Glycoalkaloids are largely unaffected by foodprocessing, including baking, cooking and frying (Friedman and Dao 1992, Friedman and Levin 1995). Becausethe major potato glycoalkaloids, a-chaconine and a-solanine, differ in biological potency, may act synergis-tically and their ratio may vary in different potato culti-vars (Friedman and Dao 1992, Friedman et al. 1991and 1992), care should be exercised in relating dose-response values of individual glycoalkaloids to totalglycoalkaloid concentration of potatoes. New potatovarieties may also contain glycoalkaloids of unknownstructure and function inherited from their progenitors(Maga 1994). These considerations suggest a need toreduce steroidal glycoalkaloid levels in the diet, possibly through suppression of enzymes responsible fortheir biosynthesis in plants (Stapleton et al. 1991).For the purposes of this study, we define the follow

ing terms: glycoalkaloids—naturally occurring, nitrogen-containing plant steroids with a carbohydrate sidechain attached to the 3-hydroxy position, e.g., a-chaconine and a-solanine from potatoes, a-tomatine fromtomatoes and solasonine from eggplants,- aglycones—the steroidal parts of the glycoalkaloid lacking the carbohydrate side chain, e.g., solanidine from a-chaconineand a-solanine, tomatidine from a-tomatine and sola-sodine from solasonine; alkaloids—glycoalkaloids andaglycones as above. Figure 1 illustrates the structures ofthe glycoalkaloids, aglycones, and the adrenal steroid,dehydroepiandrosterone (DHEA),3 evaluated in thisstudy.Previously, we have found that dietary consumption

of solasodine induces dose-related increases in absoluteliver weight (hepatomegaly) and relative liver weight(g liver weight (LW)/g body weight (BW) x 100, %LW/BW) in male weanling mice treated for 14 d with 0.16%(3.8 mmol/kg diet) solasodine (Friedman 1992). Thishepatomegaly was accompanied by increases in the activities of four liver enzymes: serum alkaline phospha-tase, glutamate-oxaloacetate transaminase, glutamate-pyruvate transaminase, and Af-demethylase. Histologi-cal examination of the tissue from these mice also revealed a distinctive liver cholangiohepatitis and gastricgland dilation/degeneration.The purpose of this study is to define possible effects

of four glycoalkaloids and three aglycones on food consumption, body weight, and liver weight changes in

•'Abbreviations used: BW, body weight; DHEA, dehydroepiandrosterone; LW, liver weight.

mice. Because the naturally occurring adrenal steroiddehydroepianderosterone is reported to induce hepatomegaly in rats (Beilei et al. 1992), it was also includedin the study. Our specific objectives were as follows:Õ)to use the dose-response and time-course data todetermine relative potencies and structure/function relationships with regard to liver weight gain in miceand 2) to investigate whether this hepatomegaly is reversible with minimal changes in food consumptionand body weight. We also speculate on structural requirements for activity and propose a possible mechanism for the observed hepatomegaly.

MATERIALS AND METHODS

Test compounds. Solasodine (mol wt 416), tomatidine (mol wt 417), a-tomatine (mol wt 1034), andDHEA (mol wt 288) were obtained from Sigma Chemical (St. Louis, MO). Solanidine (mol wt 397) was obtained from Atomergic Chemetals (Farmingdale, NY).a-Solanine (mol wt 868) and a-chaconine (mol wt 852)were isolated from potato sprouts (Friedman et al.1993). Solasonine (mol wt 884) was purchased fromBiosynth AG (Basel, Switzerland). The purity of allcompounds was established by HPLC (Friedman andLevin 1992 and 1995). We did not extract the test compounds from the diets and analyze them because theyare very stable even when subject to cooking and frying(Friedman and Dao 1992).Diets. Control mice were fed a pulverized Purina

Mouse Chowâ„¢ diet (Purina St. Louis, MO). The proportion of each test compound is expressed as mmol/kg diet. In the case of solasodine, for example, 0.2%(2.0 g per kg diet) corresponds to [2 g/416 (mol wt)] X1000 (mmol/mol) or 4.8 mmol/kg diet. To prepare themaster diet, all alkaloids were triturated into a finepowder, weighed into one third of the total diet andmixed for 10 min in a Twin Shell Dry Blender (Patterson-Kelly Co., East Stroudsberg, PA). The remainingthirds were added and mixed for 10 min. Other dietlevels were prepared by splitting the master diet intohalves or fourths, adding the difference and mixing for10 min.Feeding studies. The study protocol was approved

by our Animal Care and Use Committee, which follows the guidelines of the National Institutes of Healthon the humane use of laboratory animals. Adult 8-wk-old female Swiss-Webster mice (26-32 g) were obtainedfrom Simonsen Laboratories (Gilroy, CA). Mice werehoused individually in plastic shoebox cages with hardwood shavings for litter. An automatic timer regulatedthe 12-h light:dark cycle (lights on at 0600). Water wasfreely available to all groups. All mice were fed 60 g ofdiet per week in glass food cups with stainless steelretainers. Food consumption was estimated weekly bycomparing the remainder in the cup with the original

by on Novem

ber 22, 2009 jn.nutrition.org

Dow

nloaded from

LIVER EFFECTS OF STEROIDAL ALKALOIDS IN MICE 991

Solanidine R = H Solasodine R = H

glucosea-Solanine R = galactose< Solasonine R = galactose<

glucose

rhamnose rhamnose

rhamnosea-Chaconine R = glucose<

rhamnose

Dehydroepiandrosterone Tomatidine R = H

glucosea-Tomatine R = galactose-glucose<

xylose

FIGURE 1 Structures of steroidal glycosides, steroidal aglycones and dehydroepiandrosterone.

amount. Food consumption was considered to be anestimate because food could be kicked from the cup orcontaminated with feces, urine or litter.The mouse feeding studies reported here were based

on either dose in mmol/kg diet and/or time of exposurein days. Groups of eight mice were fed control or experimental diets with doses of 1.2, 2.4 or 4.8 mmol/kg dietof test compound for 7, 14 or 28 d. Doses and timeswere based on an earlier study with solasodine (Friedman 1992) and limited by toxicity and/or availability ofthe compounds. The glycoalkaloids, for example, wereused at 2.4 mmol/kg diet but only for 7 d due to significant body weight loss. The aglycones were studied overthe entire range of doses and times. DHEA was studiedover the entire range of doses for 28 d. In the reversibil

ity study, mice were fed as follows: Õ)control diets (14or 28 d); 2) positive control diets of 2.4 mmol/kg dietsolasodine or solanidine (14 or 28 d); and 3) combinations of either control diets (d 0-14) followed by 2.4mmol/kg diet solasodine or solanidine diets (d 15-28)or 2.4 mmol/kg diet solasodine or solanidine (d 0-14)followed by control diets (d 15-28).Variables assessed in all studies included estimated

daily food consumption (g/d) and [g/(kg mouse x d)],changes in body weight (final weight minus initial bodyweight in g), liver weight (g upon autopsy), and theliver weight to body weight ratio (%LW/BW). To bettervisualize the magnitude of the differences in the figures, all variables are expressed as the "percentage difference from controls."

by on Novem

ber 22, 2009 jn.nutrition.org

Dow

nloaded from

992 FRIEDMAN ET AL.

TABLE1Effects of dietary glycoalkaloids on body weight and liver size in adult female Swiss-Webster mice fed 2.4 mmol/kg diets for 7 d1

Group'InitialbodyweightFinalbodyweightBodyweightchangeFoodintakeFoodintakeLiverweightLiver weight/body

weight

g/7d g/d g/(kg x d) g/lOOg

Control27.7a-Chaconine27.5a-Solanine29.0a-Tomatine29.4Control

28.61.4

28.11.724.82.226.31.428.71.0

30.2Solasonine28.7 ±1.1 29.21.3

0.41.0-2.71.5-2.81.1-0.60.71.61.1

0.51.0

6.41.7*6.11.3*5.70.86.50.96.10.8b

5.60.9

229 ±31 1.65 ±0.125.870.8246 ±421.310.9216 ±351.460.9225 ±251.640.5204 ±161.620.8190 ±25 1.560.12C

5.290.17d5.560.17

5.690.085.380.09

5.340.300.39d0.460.530.260.21

1Values are means ±SD, n = 8. *P s 0.001, significantly different than the corresponding control, 2-tailed Dunnett Test, bp s 0.05,significantly different than the corresponding control, 2-tailed Dunnett Test, c p «0.001, significantly different than the corresponding control,1-tailed Dunnett Test, dp s 0.05, significantly different than the corresponding control, 1-tailed Dunnett Test.

Statistics. The figures are intended to give a visualoverview of the results, whereas tables contain themore detailed information, including results of thetests of significance. Specifically, data were checkedfor the need for transformation using the Box-Cox approach (Box et al. 1978) and, in some cases, transformedto stabilize the variance among treatments. Instancesfor which transformations were used are denoted byfootnotes in the tables. Excluding Table 1, the following transformations were used: Table 2, 7 d solasod-ine—reciprocals for LW and %LW/BW and 28 d sola-sodine—ranks for BW change, LW, and %LW/BW; Table 3, squared reciprocals for LW and %LW/BW.ANOVA (SAS 1987) and Dunnett's test (Dunnett 1955)were used for data in Tables 1, 2 and Table 4 to comparetreatment means with their corresponding controlmeans. A variety of tests are required to deal with thedifferent types of mean comparisons. One- and two-tailed tests were used to correspond with their associated null hypotheses. Two-tailed tests were used forfood consumption and body weight variables, whereasone-tailed tests were used for liver weight measurements. Liver weights were hypothesized to decreasewith diets containing glycoalkaloid and to increasewith diets containing aglycone (Tables 1-4). Plannedcontrasts were used for the reversibility results in Table3. The 14-d diets containing the test compounds werecompared with the corresponding 14-d controls. The28-d diets, including the switchover diets, were compared with the corresponding 28-d controls. The Bon-ferroni adjustment of P' = P/8 was used for these tests(Milliken and Johnson 1984) where P values are thesignificance levels given in Table 3, and the P' valuesare the significance levels given by Student's t test.Values in the text are means ±so.

RESULTS AMD DISCUSSIONFood consumption. The use of alkaloid-supple

mented diets requires estimation of food consumption

to approximate dose ranges and select doses (Dugan etal. 1989). In a previous study, Friedman (1992) usedestimated food consumption (in terms of daily foodintake and intake per kg mouse) to gauge the possibledose ranges for weanling male mice fed various levelsof solasodine. Daily food intakes (g/d) of mice exposedto the high dose (0.16% or 3.8 mmol/kg diet) for 7 dwas significantly below controls (3.2 vs. 2.8 g/d). However, food intake relative to body weight did not differ[214 vs. 216 g/(kg mouse x d)] because the experimentalmice did not gain as much weight as the controls. Thus,in that study, food consumption appeared to be inconsequential for estimation of dose range.In our current study with adult female mice, daily

food intake was significantly below controls in the solasodine group treated with 4.8 mmol/kg diet for 7 d (5.2±0.5 vs. 6.0 ±0.5 g/d, P s 0.001) and the solanidinegroup treated with 4.8 mmol/kg diet for 14 d (4.9 ±0.5vs. 6.2 ±0.7 g/d, P s 0.001). Food intake per kilogrambody weight was significantly below controls in the solasodine group treated with 4.8 mmol/kg diet for 7 d[190 ±15 vs. 216 ±17 g/(kg mouse x d), P ±s0.001] andin the solanidine groups treated with 2.4 mmol/kg dietfor 14 d [198 ±11 vs. 219 ±21 g/(kg mouse x d), P =s0.05] and with 4.8 mmol/kg diet for 14 d [198 ±15 vs.219 ±21 g/(kg mouse x d), P s 0.05]. Differences infood consumption between experimental groups andtheir controls were not observed in mice treated withthe 2.4 mmol/kg diet dose for 28 d (see Table 2 for dataon solasodine, tomatidine, and solanidine and Table 3for data on DHEA). Thus, we selected the 28-d exposureperiod for comparison of liver weight increases withoutsignificant differences in estimated food consumptionin studies involving the aglycones.Changes in body weight. The %LW/BW could vary

based on changes in either variable. It was thereforenecessary to define conditions in which differences in%LW/BW were due to liver weight gain only. We alsoused body weight loss as a means of assessing relativetoxicity of the alkaloids tested because loss of body

by on Novem

ber 22, 2009 jn.nutrition.org

Dow

nloaded from

LIVER EFFECTS OF STEROIDAL ALKALOIDS IN MICE 993

TABLE2Dose response and time courses effects of dietary solasodine, solaaidine and tomatidine on liver weight in female mice1

GroupDietmmol/kgGroupTime sized

nFoodintakeg/dFoodintakeg/(kg x d)Bodyweight

Liverweight/changeLiver weight bodyweightg

g/100g

Control0Solasodine1.2Solasodine2.4Solasodine4.8Control0Solasodine1.2Solasodine2.4Solasodine4.8Control0Solasodine

1.2Solasodine2.4Solasodine4.8Control0Tomatidine2.4Tomatidine4.8Control0Tomatidine

1.2Tomatidine2.4Tomatidine4.8Control0Tomatidine2.4Tomatidine4.8Control0Solanidine2.4Control0Solanidine1.2Solanidine2.4Solanidine4.8Control0Solanidine

2.4777714141414282828287771414141428282877141414142828161616168888888888888888881616888816136.0

±0.5 216 ±17 -1.0 0.9 1.39 ±0.102 5.03 ±0.2725.7±0.6 207 ±20 -1.0 1.0 1.40 ±0.122 5.10 ±0.4625.7±0.6 206 ±23 -0.6 1.1 1.52 ±0.13<U 5.45 ±0.29d,25.2±O.Sb 190 ±15b -0.6 1.6 2.00 ±0.26e,2 7.29 ±0.72e,26.2±0.4 207 ±20 0.6 0.8 1.60 ±0.15 5.36 ±0.436.2±0.2 215 ±13 0.9 0.8 1.63 ±0.16 5.64 ±0.426.3±0.3 213 ±11 -0.4 1.7 2.01 ±0.23e 6.80 ±0.56*5.9±0.4 212 ±23 -0.7 1.6 2.35 ±0.25e 8.40 ±0.74e6.0±0.9 204 ±31 0.4 1.53 .53 ±0.123 5.18 ±0.1936.3±0.3 212 ±13 0.1 0.73 .55 ±0.093 5.24 ±0.1835.8±0.6 194 ±23 1.4 1.13 .89 ±O.lle,3 6.32 ±0.37e,36.3±0.7 228 ±26 -1.0 2.93 2.27 ±0.38e,3 8.31 ±2.04e,36.3±0.7 223 ±21 -0.4 1.1 .48 ±0.08 5.27 ±0.335.7±0.7 203 ±23 0.1 0.5 .59 ±0.13«: 5.65 ±0.4lc5.5±0.4a 204 ±13 -0.6 1.1 .70 ±0.08e 6.28 ±0.22e5.9±0.5 211 ±24 0.3 0.9 .46 ±0.16 5.13 ±0.356.1±0.6 218 ±17 0.0 1.0 .46 ±0.09 5.23 ±0.335.8±0.6 214 ±31 -0.2 1.0 .48 ±0.19 5.37 ±0.446.2±0.8 222 ±35 -0.2 1.2 .72 ±0.21d 6.10 ±0.66e6.5±0.4 232 ±17 1.5 ±1.2 .45 ±0.14 5.12 ±0.356.4±0.4 221 ±17 1.0 ±1.3 .58 ±0.16 5.42 ±0.436.4±0.4 216 ±20 1.4 1.1 .72 ±0.15d 5.75 ±0.32d5.7±1.0 201 ±42 0.4 1.5 1.51 ±0.16 5.27 ±0.355.2±0.9 204 ±38 -2.7 l.Sb 1.73 ±0.1 7e 6.70 ±0.50e6.2±0.7 219 ±21 -0.1 1.1 1.53 ±0.10 5.40 ±0.345.7±0.4 208 ±14 -0.2 1.4 1.50 ±0.08 5.47 ±0.265.6±0.4 198 ±lia Q.3 1.2 1.82±0.1le 6.42 ±0.32e4.9±O.Sb 198 ±15a -4.3 l.lb 1.80 ±0.21e 7.25 ±0.53e5.8±1.0 200 ±37 1.0 1.3 1.47 ±0.12 5.12 ±0.286.0±0.8 204 ±20 1.2 1.3 1.99 ±0.27* 6.77 ±0.58e

1Values are means ±SD.aP s 0.05, significantly lower than control, 2-tailed Dunnett Test. ^ P ^ 0.001, significantly lower than control,2-tailed Dunnett Test. c P =s 0.05, significantly greater than control, 1-tailed Dunnett Test. dP s 0.01, significantly greater than control, 1-tailed Dunnett Test, e p s 0.001, significantly greater than control, 1-tailed Dunnett Test.

2 Reciprocal transformation.3 Rank transformation.

weight is perhaps one of the best indicators of toxicity(Friedman and Gumbmann 1984, 1988).In the study mentioned above (Friedman 1992), dif

ferences in body weight from controls were due tolower body weight gains in weanling mice fed solasodine diets for 14 d. Friedman found significantly lowerbody weight gains [P =s 0.01) in weanling mice fed0.08% (1.9 mmol/kg diet) and 0.16% (3.8 mmol/kg diet)solasodine diets when compared with controls. Thepresent study used 8-wk-old adult female mice. Thus,we were concerned with absolute decreases in overallbody weight. Significant decreases relative to controlswere observed for the solanidine groups treated with2.4 mmol/kg diet for 7 d (-2.7 ±1.5 vs. 0.4 ±1.5 g, P=s 0.001) and with 4.8 mmol/kg diet for 14 d (-4.3 ±1.1 vs.-O.l ±1.1 g, P =s0.001). Significant body weightlosses were also observed in 7-d studies with the gly-

coalkaloids a-chaconine and a-solanine. The relativemagnitudes of body weight losses could be assessedbased on percentage body weight differences relativeto an average initial body weight and on the followingP values for body weight differences relative to controlsfrom the two-tailed Dunnett's tests: a-chaconine(-12.0%, P =s 0.001), a-solanine (-6.5%, P ^ 0.001),solanidine (-10.0%, P ^ 0.001), solasonine (-3.3%, PA 0.05), a-tomatine (2.1%, NS), solasodine (-0.7%, NS)and tomatidine (-0.1%, NS) (values for %BW difference vs. average initial body weight are derived fromdata not shown and P values for body weight differences are from Tables 1 and 2). Body weight differencesvs. controls were not significant in groups treated withthe aglycones or DHEA for 28 d (Tables 2 and 4). Thus,differences in %LW/BW in these groups were due onlyto increased liver weight.

by on Novem

ber 22, 2009 jn.nutrition.org

Dow

nloaded from

994 FRIEDMAN ET AL.

TABLE 3

Effects on relative liver weight in mice fed a solasodine or solanidine diet followed by a control diet or a control dietfollowed by a solasodine or solanidine diet for 14 d each1-2

GroupGroupFoodFoodDiet

Time size intake intakeBodyweightchangeLiverweight3Liverweight/body

weight3

mmol/kg g/d g/(kg x d) g/100g

ControlSolasodineSolanidineControlSolasodineSolanidineSolasodine/ControlSolanidine/ControlControl/SolasodineControl/Solanidine02.42.402.42.42.4,02.4,00,

2.40,2.414141428282814,

1414,1414,1414,

141688168888885.4

1.1 186 ±41 0.4 ±1.3 .53 0.17 5.28 ±0.446.21.0 210 ±30 0.8 ± .4 .96 0.28d 6.59 ±0.83d5.71.2 200 ±46 -1.0 ± .1« .77 0.13c 6.14 ±0.56d5.20.6 180 ±21 0.1 ± .2 .45 0.10 5.04 ±0.235.41.0 182 ±32 1.3 ± .1 .96 0.32d 6.58 ±0.89d6.10.9 205 ±23 1.2 ± .1 2.09 0.2Qd 7.00 ±0.62d6.01.0 210 ±41 -0.4 ± .4 .55 0.20 5.35 ±0.445.60.6 179 ±19 2.4 ±0.6*> .48 0.13 4.73 ±0.315.50.9 186 ±31 0.5 ±0.9 .91 0.27d 6.46 ±0.88d5.81.1 202 ±43 -0.2 ±0.9 .80 0.18d 6.23 ±0.43d

1The Bonferroni adjustment of a/8 was used for statistical analysis. Values are means ±SD. a P2-tailed planned contrast, b p s 0.001, significantly greater than control, 2-tailed planned contrast, c P -1-tailed planned contrast, d p s 0.001, significantly greater than control, 1-tailed planned contrast.2 Two separate experiments with solasodine and solanidine; negative control values pooled.3 Squared reciprocal transformations.

0.05, significantly lower than control,0.01, significantly greater than control,

TABLE4Effects of dietary dehydroepiandrosterone (DHEA)on liver weight in adult, female mice treated for 28

GroupControlDHEADHEADHEADietmmol/kg01.22.44.8Foodintakelid5.7

±0.45.5±0.25.6±0.95.5±0.5Food

intakeg/(kg

xd)187

±26176±11178±29185±18Body

weightchangeg/28d1.9

±2.43.2±1.62.2±1.41.3±1.6Liver

weightS1.65

±0.191.78±0.121.95±0.16a1.99±0.17bLiver

weight/bodyweightg/100g5.37

±0.395.68±0.206.22±0.36b6.62±0.18b

1Values are means ±SD,n = 8. a P •.control, 1-tailed Dunnett Test.

0.01, significantly greater than control, 1-tailed Dunnett Test, b P s 0.001, significantly greater than

Glycoalkaloids. Mice were fed a-chaconine, a-solanine, a-tomatine and solasonine for 7 d at 2.4 mrnol/kg diet. The 7-d exposure with the 2.4 mmol/kg dietwas selected for experiments with the glycoalkaloidsbecause mice treated with a-chaconine and a-solanineresponded with decreased body and liver weights, andthe glycoalkaloids were not available in sufficientquantity for longer studies. Because both body and liverweights decreased relative to controls in these mice,we were interested only in the difference in %LW/BWwhich was significant (P Ä0.05) for a-chaconine butnot for the other glycoalkaloids (Table 1). This lower%LW/BW was in contrast to the %LW/BW observedwith the corresponding aglycone, solanidine, whichwas higher than controls under comparable dietaryconditions (see Fig. 2 for contrast). The %LW/BWwasnot significantly different than controls in mice treatedwith the eggplant glycoalkaloid, solasonine, whereas a

higher %LW/BWwas observed with its correspondingaglycone, solasodine. Significant differences in %LW/BWs were not observed for the tomato glycoalkaloid,a-tomatine, or its corresponding aglycone, tomatidine.However, an increased %LW/BW (P s 0.001) was observed in mice fed a higher dose for 7 d with tomatidine(4.8 mmol/kg diet, Table 2).The significantly lower liver weights of mice treated

with a-chaconine (P =s0.05) and a-solanine (P Ä0.05)and the significantly lower %LW/BW of mice treatedwith a-chaconine (P & 0.05) suggest that these effectsmay be due to hepatotoxicity. Previously, Dalvi (1985)and Dalvi and Jones (1986) found higher activities ofserum glutamate-oxaloacetate transaminase and serumglutamate-pyruvate transaminase as well as lower activities of serum cholinesterase and microsomal enzymes in rats injected intraperitoneally with 20 mg/kgof a-solanine when compared with controls. Differ-

by on Novem

ber 22, 2009 jn.nutrition.org

Dow

nloaded from

LIVER EFFECTS OF STEROIDAL ALKALOIDS IN MICE 995

potato

alkaloid

tomato vegetable

FIGURE 2 Liver weight (percentage difference from controls) in adult female Swiss-Webster mice treated for 7 d witheggplant, potato, and tomato glycoalkaloids (2.4 mmol/kgdiet) and their corresponding aglycones. The percentage difference from controls is calculated by subtraction of individual values from the group control value, dividing the resultby the group control value, and then multiplying by 100.Bars represent group mean values. Error bars represent 95%confidence intervals based on nonpooled variances. See Tables 1 and 2 for results of statistical tests.

enees in enzyme activities were not observed in ratsgavaged once with 250 mg/kg and sampled after 24 h.Our feeding studies with glycoalkaloid diets with miceinvolved a 7-d exposure time and higher estimateddoses [a-chaconine, 542 mg/(kg mouse x d); a-solam'ne,476 mg/(kg mouse X d); a-tomatine, 563 mg/(kg mouseX d); solasonine, 418 mg/(kg mouse x d)]. The measurement of these hepatic marker enzymes is an obviousnext step for future work with glycoalkaloid diets.Aglycones and DHEA. In a previous study, the ad

ministration of 0.16% (3.8 mmol/kg diet) solasodine inthe diet for 14 d led to a significant 64% increase in%LW/BW in male weanling mice (calculated from datain Friedman 1992). We chose adult female mice for thisstudy for the following reasons: Õ)we wanted to testfor differences in liver weight under conditions inwhich food consumption and body weight were notdifferent than controls to determine whether the testcompounds can increase liver weight without affectingbody weight; and 2) we wanted data that would helpus in dose and time selection for subsequent studies

with pregnant mice. The objectives of these follow-upstudies with dietary aglycones involved further observations on dose response, time course and possible reversibility in terms of differences in liver weights.The dietary treatment of female mice for only 7 d with

2.4 mmol/kg diet of the eggplant aglycone, solasodine,led to significant increases in liver weight (P & 0.01) and%LW/BW (P ^ 0.01) without significant differences inbody weight and food consumption (Table 2). In the highest dose (4.8 mmol/kg diet), we observed a 45% increasein %LW/BW (calculated from data in Table 2). However,estimated food consumption was significantly lower thanthat of controls in this group. The %LW/BW was on theorder of 8% higher than controls in the group treatedwith 2.4 mmol/kg diet for 7 d without significant changesin estimated food consumption or body weight. In animals treated for 14 d, %LW/BW gains on the order of 57and 28% were observed in the 4.8 and 2.4 mmol/kg dietgroups, respectively, without significant differences inestimated food consumption or body weight (calculatedfrom data in Table 2). These values were sustained formice treated for 28 d in both the 4.8 mmol/kg diet dose(60%, calculated from data in Table 2) and the 2.4 mmol/kg diet dose (22%, Fig. 3). When mice were treated witha solasodine-containing diet followed by a solasodine-freediet for 14 d each, liver weights and %LW/BW returnedto near control values (Table 3, Fig. 4). This indicates areversible effect on the mouse liver rather than permanent damage. Changing the sequence, i.e., feeding a control diet for 14 d followed by a solasodine-containing dietfor another 14 d, resulted in liver weight and %LW/BWgains comparable to treatments in the first 14-d exposureperiod (Table 3, Fig. 4).Although solasodine and tomatidine are structurally

similar, the dietary treatment of female mice with tomatidine led to %LW/BW increases which were ~onethird of those found with solasodine (e.g., gains in micefed the 2.4 mmol/kg diet for solasodine and tomatidinewere 22 vs. 6%, respectively, for the 28-d exposure period) (Fig. 3).The dietary treatment of female mice with solani-

dine for 7, 14 and 28 d led to LW/BW increases of 27%,19% and 32%, respectively, in the 2.4 mmol/kg dietgroups (7- and 14-d values calculated from data in Table2; 28-d value, see Fig. 3). The %LW/BW increases werecomparable to those observed in the solasodine grouptreated with 2.4 mmol/kg diet for 28 d in which bodyweight differences were not significantly different thancontrols i.e., solanidine, 32% (P < 0.001 )vs. solasodine,22% (P s 0.001) (Fig. 3). As with solasodine, the observed solanidine-induced hepatomegaly was reversible (Fig. 4). We also observed increased %LW/BW(16%, P < 0.001 ) in mice treated with the natural adrenal steroid, DHEA, at the 2.4 mmol/kg diet dose for 28d (calculated from the data in Table 4).Based on conditions in which there were no signifi

cant differences in body weight and food consumption,i.e., 2.4 mmol/kg doses for 28 d, we evaluated the relative

by on Novem

ber 22, 2009 jn.nutrition.org

Dow

nloaded from

996 FRIEDMAN ET AL.

-10

BW LW LW/BW variable

FIGURE 3 Relative potency in terms of absolute and relative liver weight differences of potato and tomato aglyconesin adult female Swiss-Webster mice treated for 28 d with 2.4mmol/kg diets. The percentage difference from controls iscalculated as in Figure 2. Bars represent group mean values.Error bars represent 95% confidence intervals based on non-pooled variances. See Tables 2 and 4 for results of statisticaltests.

potency of the aglycones and DHEA with regard to%LW/BWdifference vs. controls (Fig.3) and %LW/BWgains in mice based on P values from the one-tailedDunnett's test (Tables 2, 4).The results were as follows:solanidine (32%, P Ä0.001),solasodine (22%,P =s0.001),DHEA (16%, P s 0.001) and tomatidine (6%, NS).Structure/biological activity relationships. Solani

dine, solasodine and DHEA are structurally similar inthat they share identical A-D rings of the steroid (Fig.1). Solanidine and solasodine differ in the E- and F-rings. Solasodine's E-ring is an oxygen heterocycle andthe F-ring is a nitrogen heterocycle, whereas solanidinehas fused nitrogenous E- and F-rings. In spite of thesestructural differences, solanidine and solasodine appearto have similar effects on liver weight and %LW/BWin terms of potency and reversibility. DHEA, whichalso induces hepatomegaly (Table 4), lacks the E- andF-rings entirely. Thus, structural differences in the E-and F-rings apparently have little effect on the observedhepatomegaly.Solasodine and tomatidine are structurally similar.

The only differences between them are that Õ) tomati-

50

40

30

20

10

-10

-20

.1

I

14 days

§ .58 8

28 days

3g g aglycone

S S

trt/cntl

"Q "U

l ll 8cntl/trt diet

FIGURE 4 Reversibility of relative liver weight changesin adult female Swiss-Webster mice exposed to solasodine orsolanidine for 14 d at the 2.4 mmcl/kg diet dose. The percentage difference from controls is calculated as in Figure 2. Barsrepresent group mean values. Error bars represent 95% confidence intervals based on nonpooled variances. See Table 3for results of statistical tests. Abbreviations: trt/cntl =treated (aglycone-containing) diet followed by control diet;cntl/trt = control diet followed by treatment diet.

dine lacks the double bond in the 5,6 position of theB-ring and 2) the 25-methyl group in the nitrogenous F-ring is epimeric. Because solasodine induces significanthepatomegaly and tomatidine does not, and becausestructural differences in the F-ring appear to have littleeffect, we conclude that the 5,6 double bond is likelyresponsible for the observed differences in potency ininducing liver enlargement.Although we do not have specific evidence to ex

plain the underlying mechanism for the observed liverenlargement, Beilei et al. (1992) suggest that the hepatomegaly induced by DHEA in rat livers may be dueto a dramatic increase in the size and number of hepato-cyte peroxisomes. The activities of the peroxisomalmarker enzymes catalase and fatty acyl-CoA oxidasealso increase significantly. The study of these enzymesis an obvious next step for future work with solanidineand solasodine.As previously mentioned, the difference in biologi

cal effect of glycosylating the 3-OH group of the aglycones is profound. Instead of causing liver enlargement,

by on Novem

ber 22, 2009 jn.nutrition.org

Dow

nloaded from

LIVER EFFECTS OF STEROIDAL ALKALOIDS IN MICE 997

feeding mice the glycosides either has little effect ormay cause liver weight loss. The most likely explanation for this is that glycosylation radically alters solubility and polarity by adding a hydrophilic componentto an essentially hydrophobic molecule (Friedman andMcDonald 1995). Although the nitrogen atoms of agly-cones and glycoalkaloids can participate in acid-baseequilibria, the aglycones are more hydrophobic (lesswater-soluble and more fat-soluble) than the glycoalkaloids which have both hydrophobic and hydrophiliccharacter. These differences could influence interactions with other dietary constituents in vitro and invivo, metabolism by gut bacterial enzymes, absorptionfrom the digestive tract, transport to target organs,binding to receptor sites, disruption of cell membranes,induction of liver enzymes, residence time in the liver,or histopathology, as discussed in more detail elsewhere (Blankemeyer et al. 1992 and 1995, Caldwell etal. 1991, Dalvi and Jones 1986, Friedman 1992, Friedman and Henika 1992, Friedman et al. 1991 and 1992,Rayburn et al. 1994). Additional studies are required tobetter define these possibilities.Dietary consumption of solanaceous steroidal agly-

cones, but not the glycoalkaloids, results in increasesin mouse liver weights. Of the three aglycones studied,we were most interested in solanidine for the followingreasons: Õ) it may be formed in vivo via either acid orenzymatic hydrolysis from the potato glycoalkaloidsa-chaconine or a-solanine; 2} it may be absorbed and/or persist in tissues after hydrolysis (Alozie et al. 1979,Ciaringbold et al. 1982, Harvey et al. 1985a and 1985b,Hellenas et al. 1992b); and 3) it may accumulate intransgenic potatoes in which the enzymes for glycoal-kaloid synthesis are suppressed (Stapleton et al. 1991).We also found that the hepatomegaly induced by

solanidine and solasodine apparently is a reversiblecondition, in that liver weights return to near normalvalues when the mice are taken off the solanidine diets.Some unanswered questions with regard to solanidineinclude: 1) Does the solanidine-induced hepatomegalyinvolve hepatocyte peroxisomal proliferation?; 2) Doother species vary in susceptibility to this hepatomegaly?; 3) Is solanidine an antif ertility agent like solasodine (Dixit et al. 1989)?; and 4}Does the consumptionby mice fed 5-7 mg/d of solanidine have any relevanceto humans whose weight (about 60 kg) is ~2000 timesthat of a mouse?Some evidence suggests that solanidine may not ad

versely affect human health. This evidence includesthe following: 1) the high dose required for hepatomegaly in mice,- 2) our earlier observations that solanidinedoes not appear to be embryotoxic or teratogenic tofrog embryos (Friedman et al. 1991 and 1992); 3) thefinding that oral administration via gavage did not induce omithine decarboxylase in rats (Caldwell et al.1991); and 4} our observation that the hepatomegalyinduced by solanidine is reversible (Table 3, Fig. 4).

This suggests that the aglycone-induced liver enlargement may be an adaptive response.The two other aglycones, tomatidine and solasodine,

also appear to be relatively benign in terms of hepatomegaly. Solasodine was comparable in potency to solanidine, and tomatidine had only one third the potency of solanidine. However, there is some concernthat increased human exposure to solasodine may occur because of its proposed use as an antifertility drug.This study raises several questions which merit fur

ther study. These include the following: Õ)What isthe species-dependence and mechanism of the citedchanges in liver weights?; 2) Can some of the test compounds act synergistically? (Fewell et al. 1994, Rayburnet al. 1995b); 3) Can dietary constituents protect animals and humans against adverse effects of glycoalkaloids? (Rayburn et al. 1995a); 4) Can the pharmacoki-netics and metabolism of glycoalkaloids and aglyconesbe defined in humans (Claringbold et al. 1992, Harveyet al. 1985a and 1985b, Hellenas et al. 1992b)?; and 5)Can intravenous infusion studies (Hellenas et al.1992a) predict adverse effects of glycoalkaloids consumed orally?A critical assessment of some of these issues (Hop

kins 1995) suggests that the status of glycoalkaloids inthe human diet is poorly defined and requires furtherstudy. Because humans seem to be more sensitive thanrodents to the toxicological effects of glycoalkaloids(Morris and Lee 1984), it is difficult to estimate a human "margin of safety" based on our studies in micewithout answers to some of these questions.Finally, since DHEA is reported to exhibit anti-carci-

nogenesis (Schwartz and Pashko 1993), and solasonineis effective against malignant skin lesions (Cham et al.1991), possibly by inhibiting sodium ion active transport across and disruption of cell membranes (Blankemeyer et al. 1995), it may also be worthwhile to investigate in detail the anti-carcinogenic and anti-viral(Thome et al. 1985) potential of the steroidal aglyconesand glycoalkaloids.

ACKNOWLEDGMENTS

We thank Carol E. Levin and Gary M. McDonald forconstructive contributions.

LITERATURECITEDAlozie, S. O., Sharma, R. P. & Salunkhe, D. (1979) Excretion ofalpha-chaconine-3H, a steroidal glycoalkaloid from Solarium tu-beiosum L. and its metabolites in hamsters. Pharm. Res. Comm.11:483-490.

Aubert, S., Daunay, M. C. & Pochard, E. (1989) Saponins and steroidal alkaloids of eggplant (Solanum melongena}. I. Foodinterest,analytical methods, and localization in fruit. Agronomie 9: 641-651.

by on Novem

ber 22, 2009 jn.nutrition.org

Dow

nloaded from

998 FRIEDMAN ET AL.

Beilei, M., Rateili, D., Formen, C., Mor, G., Muscatello, U., Lardy,H. & Bobyleva, V. (1992) Changes in liver structure and function after short-term and long-term treatment of rats with dehy-droepiandrosterone. J. Nutr. 122: 967-976.

Blankemeyer, J. T., Atherton, R. & Friedman, M. (1995) Effect ofpotato glycoalkaloids a-chaconine and a-solanine on sodium active transport in frog skin. J. Agrie. Food Chem. 43: 636-639.

Blankemeyer, J. T., Stringer, B. K., Raybum, I. R., Bande, J. A. &Friedman, M. (1992) Effects of potato alkaloids a-chaconineand a-solanine on membrane potential of frog embryos. J. Agrie.Food Chem. 40: 2022-2025.

Box, G.E.P., Hunter, W. G. & Hunter, J. S. (1978) Statistics forExperimenters: An Introduction to Design, Data Analysis, andModel Building, pp. 232-237. Wiley, New York, NY.

Caldwell, K. A., Grosjean, O. K., Henika, P. R. & Friedman, M.11991) Hepatic omithine decarboxylase induction by potato glycoalkaloids in rats. Food Chem. Toxicol. 29: 531-535.

Cham, B. E., Daunter, B. & Evans, R. A. (1991) Topical treatmentof malignant and premalignant skin lesions by very low concentrations of a standard mixture (BEC)of solasodine glycosides. Cancer Lett. 59: 55-58.

Claringbold, W.D.B., Few, ). D. & Renwick, J. (1982) Kinetics ofretention of solanidine in man. Xenobiotica 12: 293-302.

Dalvi, R. R. (1985) Comparative assessment of the effect of sola-nine administered orally and intraperitoneally on hepatic dysfunction in male rats. Jpn. J. Vet. Sci. 47: 657-659.

Dalvi, R. R. & Jones, R. H. (1986) Role of microsomal cytochromeP-450 containing monooxygenases in the metabolism and toxic-ity of certain alkaloids. Curr. Sci. 55: 558-561.

Dao, L. & Friedman, M. (1994) Chlorophyll, chlorogenic acid, glycoalkaloids, and protease inhibitor content of fresh and greenpotatoes. I. Agrie. Food Chem. 42: 633-639.

Dixit, V. P., Gupta, R. S. & Gupta S. (1989) Antifertility plantproducts: testicular cell population dynamics following solasodine administration in rhesus monkey [Macca mulatta}. Andro-Iogia21: 542-546.

Dugan,G.M.,Gumbmann,M.R.&Friedman,M. (1989) Toxicolog-ical evaluation of Jimson weed (Datura stramonium) seed. FoodChem. Toxicol. 27: 501-510.

Dunnett, C. W. (1955) A multiple comparisons procedure forcomparing several treatments with a control. J. Am. Stat. Assoc.50: 1096-1121.

Fewell, A.M., Roddick, J.G. &Weissenberg, M. (1994) Interactionbetween the glycoalkaloids solasonine and solamargine in relation to inhibition of fungal growth. Phytochemistry 37: 1007-1011.

Friedman, M. (1996) Nutritional value of proteins from differentfood sources. A review. J. Agrie. Food Chem. 44: 6-29.

Friedman, M. (1992) Composition and safety evaluation of potatoberries, potato and tomato seeds, potatoes, and potato alkaloids.In: Food Safety Evaluation (Finley, J. W., Robinson, S. & Armstrong, A., eds.|, pp. 429-462. American Chemical Society, Washington, DC.

Friedman, M. Si Dao, L. (1992) Distribution of glycoalkaloids inpotato plants and commercial potato products. J. Agrie. FoodChem. 40: 409-413.

Friedman, M. & Gumbmann, M. R. (1984) The utilization andsafety of isomerie sulfur amino acids in mice. J. Nutr. 114: 2301 -2310.

Friedman, M. & Gumbmann, M. R. (1988) Nutritional value andsafety of methionine derivatives, isomerie dipeptides, and hy-droxy analogs in mice. J. Nutr. 118: 388-397.

Friedman, M. & Henika, P. R. (1992) Absence of genotoxicity ofpotato alkaloids a-chaconine, a-solanine and solanidine in theAmes Salmonella and adult and foetal erythrocyte micronucleusassays. Food Chem. Toxicol. 30: 689-694.

Friedman, M. &. Levin, C. E. (1992) Reversed-phase high-performance Chromatographie separation of potato glycoalkaloids and

hydrolysis products on acidic columns. J. Agrie. Food Chem. 40:2157-2163.

Friedman, M. &. Levin, C. E. (1995) a-Tomatine content of tomatoand tomato products determined by HPLC with pulsed ampero-metric detection. J. Agrie. Food Chem. 43: 1507-1511.

Friedman, M. & McDonald, G. M. (1995) Acid catalyzed partialhydrolysis of carbohydrate groups of the potato glycoalkaloid a-chaconine in alcoholic solutions. J. Agrie. Food Chem. 43: 1501-1506.

Friedman, M., McDonald, G. & Haddon, W. F. (1993) Kinetics ofacid catalyzed hydrolysis of potato glycoalkaloids a-chaconineand a-solanine. J. Agrie. Food Chem. 41: 1397-1406.

Friedman, M., Raybum, J. R. & Bantle, J. A. (1991 ) Developmentaltoxicology of potato alkaloids in the frog embryo teratogenesisassay Xenopus (FETAX). Food Chem. Toxicol. 29: 537-547.

Friedman, M., Rayburn, J. R. & Bantle, J. A. (1992) Structural relationships and developmental toxicity of Solanum alkaloids in thefrog embryo teratogenesis assay Xenopus (FETAX). J. Agrie. FoodChem. 40: 1617-1624.

Harvey, M. H., McMillan, M., Morgan, M.R.A. & Chan, H.W.S.(1985a) Solanidine is present in serum of healthy individualsand in amounts dependent on their dietary potato consumption.Hum. Toxicol. 4: 187-194.

Harvey, M. H., Morris, B. A., McMillan, M. & Marks, V. (1985b)Measurement of potato steroidal alkaloids in human serum andsaliva by radioimmunoassay. Human Toxicol. 4: 503-512.

Hellenas, K. E., Cekan, E., Slanina, P. & Bergman, K. (1992a) Studies of embryotoxicity and the incidence of external malformationsafter continuous intravenous infusion of a-chaconine in pregnantrats. Pharmacol. Toxicol. 70: 381-383.

Hellenas, K. E., Nyman, A., Slanina, P., Loeoef, L. & Gabrielsson, J.(1992b) Determination of potato glycoalkaloids and their agly-cone in blood serum by high-performance liquid chromatography.Application to pharmacokinetic studies in humans. J. Chro-matogr. 573: 69-78.

Hopkins, J. (1995) The glycoalkaloids: naturally of interest (but ahot potato?). Food Chem. Toxicol. 33: 323-329.

Keeler, R. F., Baker, D. C. & Gaffield, W. (1991 ) Teratogenic Solanum species and the responsible teratogens. In: Handbook of Natural Toxins (Keeler, R. F. &. Tu, A. T., eds.), vol. 6, pp. 83-97.Marcel Dekker, New York, NY.

Maga, J. A. (1994) Glycoalkaloids in Solanaceae. Food Rev. Int.10: 385-418.

Milliken, G. A. & Johnson, D. E. (1984) Analysis of Messy Data.Volume I: Designed Experiments, pp. 33-34. Lifetime Learning,Belmont, CA.

Morris, S. C. & Lee, T. H. (1984) The toxicity and teratogenicityof Solanaceae glycoalkaloids, particularly those of the potato {Solanum tuberosum]. Food Technol. Aust. 36: 118-124.

Raybum, J. R., Bantle, J. A. &.Friedman, M. (1994) The role of thecarbohydrate side chains of potato glycoalkaloids in developmental toxicity. J. Agrie. Food Chem. 42: 1511-1515.

Raybum, J. R., Bantle, J. A., Quails, C. W., Jr. & Friedman, M.(1995a) Glucose-6-phosphate and NADP protect frog embryosagainst developmental toxicity induced by a-chaconine. FoodChem. Toxicol. 33: 1021-1025.

Rayburn, J. R., Friedman, M. & Bantle, J. A. (1995b) Synergism ofpotato glycoalkaloids a-chaconine and a-solanine in the developmental toxicity of Xenopus embryos. Food Chem. Toxicol. 33:1013-1019.

Renwick, J. H., Claringbold, D. B., Earthy, M. E., Few, J. D. &McLean, A.C.S. (1984) Neural-tube defects produced in Syrianhamsters by potato glycoalkaloids. Teratology 30: 371-381.

Roddick, J. G. (1989) The acetylcholinesterase-inhibitory activityof steroidal glycoalkaloids and their aglycones. Phytochemistry28: 2631-2634.

Roddick, J. G., Rijenenberg, A. L. & Weissenberg, M. (1992) Alterations to the permeability of liposome membranes by the solasod-

by on Novem

ber 22, 2009 jn.nutrition.org

Dow

nloaded from

LIVER EFFECTS OF STEROEDAL ALKALOIDS IN MICE 999

ine-based glycoalkaloids solasonine and solamargine. Phytochem-istrySl: 151-154.

SAS Institute Inc. (1987) SAS Guide for Personal Computers, version6. SAS Institute, Cary, NC.

Schwartz, A. G. & Pashko, L. L. (1993) Cancer chemopreventionwith the adrenocortical steroid dehydroepiandrosterone andstructural analogs. J. Cell. Biochem. 17: 73-79.

Slanina, P. (1990) Assessment of health-risks related to glycoalkaloids ("solanine") in potatoes: a Nordic view. Var Foda 43, suppl.1: 1-14.

Stapleton, A., Allen, P. V., Friedman, M. & Belknap, B. L. (1991)Isolation and characterization of solanidine glucosyltransferasefrom potato sprouts. J. Agrie. Food Chem. 39: 1187-1193.

Thome, H. V., Clarke, G. F. & Skuce, R. (1985) The inactivationof herpes simplex virus by some Solanaceae glycoalkaloids. Antiviral Res. 5: 335-343.

van Gelder, W.M.J. (1990) Chemistry, toxicology, and occurrenceof steroidal glycoalkaloids: potential contaminants of the potato.In: Poisonous Plants Contaminating Edible Plants (Rizk, A. F.,ed.), pp. 117-156. CRC, Boca Raton, FL.

by on Novem

ber 22, 2009 jn.nutrition.org

Dow

nloaded from