Chemoprevention of dibenzo[a,l]pyrene transplacental carcinogenesis in mice born to mothers administered green tea: primary role of caffeine

Abstract

Our laboratory recently developed a mouse model of transplacental induction of lymphoma, lung and liver cancer by the polycyclic aromatic hydrocarbon, dibenzo[a,l]pyrene (DBP). Pregnant B6129SF1 females, bred to 129S1/SvIm males, were treated on day 17 of gestation with an oral dose of 15 mg/kg DBP. Beginning on day 0 of gestation, dams were given (ad lib) buffered water, 0.5% green tea, 0.5% decaffeinated green tea, caffeine or epigallocatechin-3-gallate (EGCG) (both at equivalent concentrations found in tea). The concentration of the teas (and corresponding caffeine and EGCG) was increased to 1.0% upon entering the second trimester, 1.5% at onset of the third trimester and continued at 1.5% until pups were weaned at 21 days of age. Offspring were raised with normal drinking water and AIN93G diet. Beginning at 2 months of age, offspring experienced significant mortalities due to an aggressive T-cell lymphoma as seen in our previous studies. Ingestion of caffeinated, but not decaffeinated, green tea provided modest but significant protection (P = 0.03) against mortality. Caffeine provided a more robust (P = 0.006) protection, but EGCG was without effect. Offspring also developed DBP-dependent lung adenomas. All treatments significantly reduced lung tumor multiplicity relative to controls (P < 0.02). EGCG was most effective at decreasing tumor burden (P = 0.005) by on average over 40% compared with controls. Induction of Cytochrome P450 (Cyp)1b1 in maternal liver may reduce bioavailability of DBP to the fetus as a mechanism of chemoprevention. This is the first demonstration that maternal ingestion of green tea, during pregnancy and nursing, provides protection against transplacental carcinogenesis.

Introduction

In animal models, a number of carcinogens are capable of inducing cancer in offspring of mothers exposed during pregnancy (reviewed in ref. 1). In humans, ionizing radiation and diethylstilbestrol are the two agents for which the evidence for such transplacental carcinogenesis is convincing (1), although a number of studies exist providing indirect evidence that environmental exposures of pregnant women to chemical carcinogens are associated with increased risks of cancer in children (2–6).

Previous studies by Drs Anderson and Miller et al. demonstrated that exposure of the dam during the third trimester to the synthetic polycyclic aromatic hydrocarbon (PAH), 3-methylcholanthrene, in crosses of B6D2F1 and D2 mice, produced lung and liver cancers in offspring at 1 year of age (7). We employed a different strain of mice and a different PAH. Treatment of B6129F1 mice, bred to 129 mice, resulted in high mortalities beginning at 2–3 months of age from an aggressive T-cell lymphoma (8). If the mice survived to 10 months of age without developing lymphoma, 100% exhibited lung neoplasia and most males had liver tumors (8). The significance of this model is enhanced by the fact that leukemias and lymphomas are the major childhood cancers and cancer is the major cause of death in children other than accidents (1).

In our first chemoprevention study employing this model, offspring born to mothers fed indole-3-carbinol, a major phytochemical found in cruciferous vegetables (and available as a dietary supplement), exhibited significant protection from the dibenzo[a,l]pyrene (DBP)-dependent lymphoma mortality (9). Furthermore, there was a significant reduction in lung tumor multiplicity at 10 months of age (9). Therefore, the addition of a chemoprotective phytochemical to the maternal diet could provide significant reduction in cancer, not just in young adults (3–6 months in mice), but out to middle age, as well.

In the present study, we utilized this model to examine the potential for green tea to act as a chemoprotective agent against transplacental DBP carcinogenesis. Tea is the second most consumed beverage worldwide (10). Studies with a number of animal models have demonstrated the efficacy of teas (black, green, white) in chemoprevention of cancers (11–18). Epidemiology studies in humans have yielded equivocal results with respect to the cancer chemoprotective properties of tea (reviewed in ref. 15).

Epigallocatechin-3-gallate (EGCG) is a major polyphenol constituent, especially abundant in green and white (the least processed) teas. EGCG is thought to be one of the most effective constituents in modulation of disease, including cancers, by tea (19–24). Another major constituent of green tea is caffeine and some cancer chemoprevention studies suggest that caffeine plays a major role in the beneficial (17,25,26) and, perhaps, harmful (27,28) effects observed. Caffeine is of special interest in our model as concerns about its potential harmful effects on the fetus have led agencies such as the Food and Drug Administration and the March of Dimes to recommend pregnant women refrain from its consumption during pregnancy. Caffeine is a teratogen in rodents, but only at high doses (29). The risk to the fetus from maternal ingestion of caffeine remains a concern, but the overall evidence would suggest that caffeine has not been responsible for a significant number of malformations in babies born to mothers consuming moderate amounts (30–32).

We now report that caffeinated, but not decaffeinated, green tea given to the pregnant and nursing mouse significantly (P = 0.03) reduced the mortality in her offspring from transplacental DBP induction of lymphoma. EGCG administered at the same concentration had no effect. Caffeine (again, administered at the concentration found in green tea), exhibited the greatest chemoprotection (P = 0.006) against lymphoma mortality. All treatment groups reduced lung tumor multiplicity and size and these effects were most marked with EGCG.

These results confirm that this model can be very useful for the study of transplacental (and, perhaps, lactational) chemoprotection against cancer by modulation of the maternal diet. Offspring born to mothers consuming chemoprotective phytochemicals exhibit significant protection against PAH-dependent cancers in young adults and out to at least middle age, even though they never consume the phytochemical following weaning.

Materials and methods

Chemicals and diets

DBP was provided by the Carcinogen Repository, supported by the National Cancer Institute, at Midwest Research Institute (Kansas City, MO) and was at least 98% pure as determined by high-performance liquid chromatography. The semi-purified diets, AIN93G and AIN93M, were purchased from Dyets (Bethlehem, PA). Green tea and decaffeinated green tea were obtained from Stash Teas (Portland, OR). EGCG was a gift from DSM Nutritional Products (Basel, Switzerland). Caffeine was purchased from Sigma Chemical Co. (St Louis, MO). 2,3,4,5-Tetramethoxystilbene was purchased from Cayman Chemical Co. (Ann Arbor, MI). The Cytochrome P450 (Cyp)1b1 assay utilized the P450-Glo system purchased from Promega (Madison, WI).

Preparation of teas and treatment of mice

The teas used in this study were analyzed for the composition of various polyphenols and caffeine by high-performance liquid chromatography as described previously (33). The teas were prepared by adding 0.5 g tea leaves per 100 ml buffered water (0.5% wt/vol) with a brew time of 2 min. Citric acid (0.5% wt/vol) was used to buffer the solution to pH 7.6 in order to stabilize polyphenols. Solutions of EGCG and caffeine were prepared by addition of each to the citric acid-buffered solution (pH 7.6) at the concentration present in the whole tea (0.13 and 0.09 mg/ml for EGCG and caffeine, respectively, for a 0.5% tea, etc.). The administration of the teas, EGCG and caffeine began at day 0 (detected by appearance of the vaginal plug) of pregnancy and were given at 0.5% initially. We have observed that mice begun immediately at 1.5% tend to exhibit aversion. We administered the test solutions at 0.5% for the first trimester of pregnancy and then increased the concentration to 1.0% for the second trimester and to 1.5% for the third trimester. The 1.5% solutions were continued throughout nursing (21 days post-partum). We observed no overt adverse effects of this treatment regiment on either the dams or the pups (e.g. no difference in birth weight, litter size, etc.).

Eight-week-old B6129SF1 female and 129S1/SvImJ male mice (Jackson Laboratories, Bar Harbor, ME) were housed at the Laboratory Animal Resource Center at Oregon State University at 20 ± 1°C and 50 ± 10% humidity, with a light–dark cycle of 12 h each in micro-isolator cages (Life Products, Seaford, DE) with CareFRESH bedding. After acclimation for 1 week, mice were bred and appearance of the vaginal plug was determined to be day 0. The mice were given 0.5% citric acid-buffered water, teas, EGCG or caffeine to drink and AIN93G diet ad libitum. After birth, offspring were nursed for 21 days and dams continued on 1.5% tea and EGCG or caffeine equivalents. After weaning, offspring of each sex from the same litter were housed together (up to five per cage) and administered water and AIN93G diet ad libitum. At 3 months of age, the diet was changed to AIN93M. The number of dams and offspring in each experimental group is shown in Table I. We also analyzed five additional groups that were identical with the exception that dams were administered vehicle alone on day 17 of gestation. No lymphomas were observed in offspring born to mothers receiving vehicle alone (data not shown).

Table I.

Numbers of dams and offspring in each experimental group

Group	Dams	Offspring	Survived to 10 months
Controls	16	108	8
Green tea	14	100	15
Decaffeinated tea	16	116	15
Caffeine alone	14	105	24
EGCG alone	17	110	8

Day 0 of gestation was set as the first day a vaginal plug was observed. Dams were immediately started on 0.5% brewed (2 min) tea (in 0.5% citric acid, pH 7.6) as their sole drinking source. The concentrations were increased to 1% for the second trimester and to 1.5% at the start of the third trimester and this concentration was continued until weaning. All dams received DBP (15 mg/kg) as a single dose by gavage (corn oil 5 ml/kg) on day 17 of gestation. There was no difference between groups with respect to litter size or birth weight.

The health of the colony was monitored by sentinels tested for viral or bacterial pathogens and parasites by the University of Missouri Research Animal Diagnostic Laboratory (Columbia, MO). All tests were negative throughout the course of the study. If any signs of distress or pain (or any sign of morbidity) were observed, mice were immediately euthanized with an overdose of CO₂ and a necropsy performed. Of those that did not survive to the end of the study, 13.9% died between monitorings and were not useful for necropsy. The proportion not useful was similar across all treatments (P > 0.7, chi-square test). At 10 months of age, the study was terminated and any surviving mice were euthanized and necropsied. All procedures used in the handling, treatment and husbandry of mice were approved by the Oregon State University Institutional Animal Care and Use Committee.

Histopathology

At necropsy, the heart, thymus, lung, spleen, liver, kidney, abnormal lymph nodes, testes or ovaries, colon and skin were collected, fixed in 10% formalin, stained with hematoxylin and eosin and analyzed by light microscopy. In our previous studies with this model (8,9), we established that the lymphoma producing the mortalities in mice at 3–6 months of age was of T-cell origin (CD3+, B220−) with various phenotypes (CD4−, CD8+; CD4+, CD8+). These were classified as T-cell lymphoblastic lymphomas. These lymphomas were very aggressive, resulting in invasion of numerous organs by the transformed lymphocytes (8,9).

In our previous studies all offspring born to mothers treated with DBP developed lung tumors by 10 months and most males had liver neoplasia as well (8,9). In this study, we carefully examined the lung for each mouse necropsied and determined the lung multiplicity as a function of age, out to the 10-month termination point.

Statistical analysis

Overall strategy for analysis of offspring responses included: (i) accounting for cluster (litter) effects when there is any evidence of them being present; (ii) assessment of treatment effects both overall and then with pairwise comparisons with the control and (iii) assessment of gender and aryl hydrocarbon receptor effects, both main effects and interactions with other factors. The statistical software was SAS for Windows 9.1.3 and S-plus 7.0 Windows.

For analysis of the offspring survival curves, treatment groups were compared using Cox (proportional hazards) regression with robust score tests to account for the fact that pups were clustered into litters [S-Plus coxph function with cluster term, (34)]. There was evidence of litter effects in three of the five treatments. Pairs of survival curves reasonably satisfied the proportional hazards assumption based on graphical and analytical checks (35) with the Lifetest and Phreg procedures in SAS. There was no evidence of gender or aryl hydrocarbon receptor effects on survival.

For analysis of number of tumors per offspring, the data set modeled included two subsets: (i) survivors and (ii) animals surviving long enough to have the opportunity to develop multiple tumors. For the second subset, a criterion of requiring every treatment group to have at least 50% of the animals with multiple tumors resulted in an inclusion cut-off for survival of at least 150 days. Over-dispersed (quasi-likelihood) Poisson regression with a log link and age at death as a simple linear covariate was chosen as the final model due to a reasonable pattern for the deviance residuals, homogeneity of variance between subsets and easily interpretable (multiplicative) effects. Over the range of ages being modeled, there was no evidence of curvilinearity or differences in slope between treatments (no age-by-treatment interaction). Models were fit with the SAS Genmod and Glimmix procedures and P values reported are for approximate F-tests and t-tests. There was no evidence of litter effects for this response (zero estimate for litter variance component).

For analysis of maternal Cyp1b1 activity, data were log transformed prior to analysis due to increasing standard deviation with mean of response and right skew for repeated measurements. After averaging across three measurements within each of three trials to achieve one mean response per dam, treatments were compared by one-way analysis of variance followed by pairwise contrasts (t-tests) with the control treatment using the SAS Mixed procedure.

Results

Maternal consumption of green tea and green tea components protects against mortality in offspring from transplacental DBP-induced lymphoma

As reported previously by our laboratory (8,9), offspring of these crosses, born to mothers administered DBP, exhibited a high rate of mortality between 3–6 months of age from an aggressive T-cell lymphoma. Using robust tests to account for any litter effects, there was clear evidence of treatment effects on offspring survival (P = 0.018, overall four degrees of freedom test). The survival curves for offspring born to mothers given drinking water, caffeinated tea and decaffeinated tea are shown in Figure 1 (top panel). A modest, but significant (P = 0.03) protective effect was observed for offspring born to mothers drinking regular green tea throughout pregnancy and nursing. If the mothers were given decaffeinated green tea, the protection was not significant (P = 0.25).

Fig. 1.

Survival curve of offspring born to mothers given DBP. DBP was administered at a dose of 15 mg/kg by gavage (corn oil) on day 17 of gestation. The pregnant and nursing dams were administered ad libitum citric acid-buffered water (open square), green tea (filled diamond, top panel), decaffeinated green tea (filled inverted triangle, top panel), caffeine (filled diamond, bottom panel) and EGCG (filled triangle, bottom panel). The concentrations and periods of administration are given in Materials and Methods.

Figure 1 (bottom panel) depicts the survival of offspring born to mothers drinking EGCG or caffeine solutions (at the equivalent tea concentration). There was no protective effect for the offspring born to mothers consuming EGCG, but the protective effect with caffeine was striking and highly significant (P = 0.006).

Maternal consumption of green tea and green tea components protects against DBP-dependent transplacental lung cancer throughout the lifetime

As we reported previously, DBP in this model is also a transplacental lung carcinogen. In our study of the chemoprotective effect of indole-3-carbinol in this model, we reported a 35% reduction in lung tumor multiplicity (9) at 10 months of age in mice not succumbing to lymphoma. In this study, we monitored lung tumor multiplicity from 4–10 months of age (Figure 2). Both survivors and those surviving >150 days (∼21 weeks, see Materials and Methods) exhibited reduced tumors per animal in all treatments compared with the control. After adjusting for age and gender as significant covariates (P < 0.0001, quasi-likelihood F-test), there was clear evidence of treatment effects on tumors per animal (P = 0.01, overall F-test). In contrast to the impact on lymphoma survival, EGCG was the most effective at reducing the tumor multiplicity by an average of over 40% compared with controls. Tumor burden was also reduced by 27%, 32% and 36% in the caffeine, caffeinated green tea and decaffeinated green tea-drinking regimens (P < 0.02). These results imply that the mechanisms of cancer chemoprotection by tea and components of tea for lymphoma and lung in this model are distinct. In order to rigorously test the transplacental chemoprevention potential of tea and tea components with respect to lung cancer, it would be necessary to employ a model, such as the 3-methylcholanthrene model in B6D2F1 × D2 crosses (7) or the A/J mouse model (36). This experimental design would produce lung tumors without the confounding lymphoma mortality.

Fig. 2.

Lung tumor multiplicity. Offspring were euthanized due to lymphoma-dependent morbidity (4–9 months) or at the conclusion of the study (10 months) and lung lesions (predominantly adenomas) quantified by histopathology. Offspring born to mothers given citric acid-buffered drinking water (open squares), green tea (filled triangle), decaffeinated green tea (filled inverted triangle), caffeine (filled diamond) or EGCG (filled circle). Numbers euthanized in the first group (<21 weeks) were 72, 55, 70, 53 and 80. For the second group (>21 and <44 weeks), 11, 14, 14, 12 and 4 were euthanized. Numbers surviving to 10 months are given in Table I. The concentrations and timing of administration to the dams are given in Materials and Methods.

Influence of drinking regimens on maternal Cyp1b1 activity

Measurement of Cyp1b1 activity in maternal liver microsomes (Figure 3A) via a P450-Glo assay revealed strong evidence of differences between treatments (P < 0.0001, overall analysis of variance). Dams administered caffeinated green tea and caffeine alone showed significant (over 2-fold) increases in Cyp1b1 activity compared with controls (P = 0.01 and P = 0.024, respectively). Hepatic microsomes from dams administered EGCG exhibited some reduction in Cyp1b1 activity (P = 0.066). However, analysis of messenger RNA for Cyp1A1 and Cyp1b1 revealed no significant changes in transcript levels among the treatments (data not shown). In order to confirm the contribution of Cyp1b1, 2,3′,4,5′-tetramethoxystilbene was co-incubated under the same assay conditions. The presence of 2,3′,4,5′-tetramethoxystilbene effectively diminished Cyp1b1 activity to near background levels in a clear dose–response manner (Figure 3B). These results suggest that at least one mechanism for the transplacental chemoprevention observed in this study is a reduction of DBP bioavailability to the fetus due to induction of DBP metabolism in maternal liver. Cyp1b1 exhibits the greatest activity of all the Cyps examined with respect to DBP metabolism (37–44). In addition, we employed Cyp1b1 (breeding heterozygous mice) knockout mice to demonstrate a clear gene dose effect; null mice were resistant to DBP compared with wild-type siblings, and hets exhibited intermediate sensitivity (45).

Fig. 3.

Effects of drink regimens on maternal hepatic Cyp1b1 activity. At gestation day 17, a subset of dams were killed and livers were extracted for preparation of microsomal fractions. Activity of Cyp1b1 was assayed fluorometrically via reaction with a luminogenic Cyp1b1 substrate (luciferin-6′ chloroethyl ether, P450-Glo assay). Values are expressed in total luminescence (relative luminescence units, log scale) for each drink regimen. (A) Bars indicate the mean of log-transformed Cyp1b1 activity for the respective dams. Circles represent individual dams. (B) Reactions were performed in the presence or absence of the Cyp1b1 inhibitor, 2,3,4,5-tetramethoxystilbene.

Discussion

Tea is the second most widely consumed beverage in the world and green tea (Camellia sinesis), the more popular tea in Asia, has been demonstrated in a number of animal models to have chemoprotective effects (10–18,46). In these models, green tea has been shown to be effective against cancers of the breast, colon, skin, oral cavity, esophagus, forestomach, small intestine, pancreas and lung (reviewed in ref. 15). The lung has received much of the attention in mouse models of green tea chemoprotection. The A/J mouse is commonly used for these studies as they are prone to lung cancer development and studies can be conducted in 4–6 months, rather than a year or more required in resistant strains. In the A/J mouse model of lung cancer, green tea is chemoprotective against numerous carcinogens requiring bioactivation including the tobacco-specific nitrosamine, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone, diethylnitrosamine, the PAH (benzo[a]pyrene), as well as direct-acting carcinogens such as nitrosomethylurea and cisplatin (15,17).

Initial studies on the mechanism of cancer chemoprevention by green tea highlighted the antioxidant properties of polyphenol constituents such as EGCG. Indeed, green tea and EGCG are effective in vitro and in vivo in the reduction of biomarkers of carcinogen-induced oxidative damage, such as the level of 8-oxo-deoxyguanosine (14). Other potential mechanisms of action (reviewed in ref. 15) are induction of apoptosis (19,20,24), inhibition of angiogenesis (16), inhibition of cell proliferation or cell cycle arrest (10,19,20), induction of phase II enzymes through the Nrf-2 pathway (10) and alteration in cell signaling through nuclear factor-κB (21). Fewer studies with green tea have focused on leukemias and lymphomas. In vitro, green tea (EGCG) induces apoptotic cell death in transformed human B-cells (24) and in a mouse lymphoma cell line (19).

In an earlier study, we demonstrated a marked chemoprotective effect if offspring were born to mothers consuming indole-3-carbinol in diet during pregnancy and nursing (9). We now report that green tea, given to the mother throughout pregnancy and nursing, also provides significant protection to offspring with respect to lymphoma-dependent mortality. Removal of caffeine from this green tea resulted in no significant protection. Consistent with the whole tea results, offspring born to mothers consuming caffeine alone exhibited a striking protection against development of lymphoma, whereas the major polyphenol (present in both caffeinated and decaffeinated green tea), EGCG, was without effect. It would appear that the major chemoprotective effect of green tea in this transplacental model of DBP lymphoma is due to caffeine.

These results are consistent with other animal models comparing whole tea with caffeine, i.e. caffeine appears to be the major chemoprotective component (17). Although the exact mechanism of the inhibition by caffeine is not entirely known, we postulate that in this model induction of maternal Cyp1b1 results in the decreased bioavailability of DBP to the fetal target organs. This mechanism has also been shown to underlie the inhibition of lung tumorigenesis initiated by 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone, in rats treated with various levels of caffeine (46). It is possible, however, that other mechanisms may also be involved and can be attributed to caffeine's broad range of biochemical and physiological activities (47). In addition, caffeine has been demonstrated to be a teratogen in animal models, including the mouse (29). The Food and Drug Administration and March of Dimes recommend that pregnant women limit intake of caffeine. Based on the study reported here, we certainly do not advocate any change in this recommendation. Our results highlight the need, in developing candidates for human transplacental chemoprotection, for a thorough understanding of mechanism of action and dose–response for beneficial and adverse effects in order to do proper risk assessment.

As we reported previously, administration of DBP in this model not only resulted in T-cell lymphoma, but all mice not succumbing to the lymphoma developed lung cancer. In this study, we also saw a 100% incidence of lung cancer in offspring born to DBP-treated mothers. The time course of lung tumor multiplicity demonstrated that offspring born to mothers in any of the four treatment groups exhibited a marked and significant delay in the appearance of lung tumors. The greatest effect and the longest delay were observed in offspring born to mothers given EGCG alone. Mortality from lung cancer is greater than any other cancer in the USA for both sexes. A significant delay in the appearance of lung tumors could save billions in health care dollars. However, the results here should not be over-interpreted. In this model, the offspring succumb to lymphoma (3–6 months) prior to the optimal time point to assess lung tumor response (10 months–1 year). The lymphoma mortality impacts the lung tumor results so that we cannot be sure that our statistical analysis of lung tumor multiplicity represents the true population. In order to definitively address the hypothesis that maternal dietary green tea components inhibit lung tumor development in offspring, we would need to use a model (e.g. A/J mouse) where lung tumors develop rapidly and are the primary tumor site.

In addition to PAHs, arsenic has been demonstrated to be a transplacental lung carcinogen in mice (48). Evidence suggests that transplacental exposure to air pollutants (including PAHs) are transplacental carcinogens in humans as well (49). In humans, the chemopreventive or therapeutic properties, with respect to lung cancer, of green tea are equivocal (15,50–53). Very few studies on transplacental cancer chemoprevention have been done. In addition to our model (9), the addition of oriental food seasoning spices (garam masala) or mustard seed oil to the maternal diet of mice provides chemoprotection against 7,12-dimethylbenz[a]anthracene-dependent transplacental carcinogenesis (54,55). Cyp1b1 expression is high in late gestation in mouse fetal thymus and lung; more importantly, the situation is similar in humans and, in fact, the thymus is the fetal organ with the highest Cyp1b1 expression in late gestation (56,57).

In summary, administration of green tea, decaffeinated green tea or the major bioactive components of green tea, EGCG and caffeine, to pregnant and nursing mice exposed to the potent PAH carcinogen, DBP, provided significant cancer chemoprotection for her offspring. In the case of lymphoma, caffeinated green tea or caffeine alone were protective, whereas decaffeinated green tea of EGCG alone were without effect. In the case of lung, all the treatments decreased lung tumor multiplicity and size and EGCG was the most effective. It is remarkable that offspring can be protected out to at least middle age following in utero or breast milk exposure to a phytochemical chemoprotective agent. These results suggest possible ‘imprinting’ or epigenetic mechanisms and may represent a paradigm shift in strategies of cancer chemoprevention by diet.

Funding

The National Institutes of Health (Public Health Service grants CA90890, ES07060 and ES00210); the Linus Pauling Institute at Oregon State University.

Glossary

Abbreviations

Cyp

Cytochrome P450

DBP

dibenzo[a,l]pyrene

EGCG

epigallocatechin-3-gallate

PAH

polycyclic aromatic hydrocarbon