Carcinogenesis vol.33 no.1 pp.226–232, 2012doi:10.1093/carcin/bgr247Advance Access publication November 9, 2011

Cell signaling pathways associated with a reduction in mammary cancer burden bydietary common bean (Phaseolus vulgaris L.)

Matthew D.Thompson, Meghan M.Mensack1,Weiqin Jiang2, Zongjian Zhu2, Matthew R.Lewis3,John N.McGinley2, Mark A.Brick4 andHenry J.Thompson2,�

Department of Physiology, Medical College of Wisconsin, Milwaukee, WI53226, USA and 1Department of Chemistry, 2Cancer Prevention Laboratory,Department of Horticulture and Landscape Architecture, 3Proteomics andMetabolomics Facility–Office of the Vice President for Research and4Department of Soil and Crop Sciences, Colorado State University, FortCollins, CO 80523, USA

�To whom correspondence should be addressed. Department of Horticultureand Landscape Architecture, Cancer Prevention Laboratory, Colorado StateUniversity, 1173 Campus Delivery, Fort Collins, CO 80523, USA.Tel: þ1 970 491 7748; Fax: þ1 970 491 3542;Email: henry.thompson@colostate.edu

Emerging evidence indicates that common bean (Phaseolusvulgaris L.) is associated with reduced cancer risk in human pop-ulations and rodent carcinogenesis models. This study sought toidentify cancer-associated molecular targets that mediate the ef-fects of bean on cancer burden in a chemically induced rat modelfor breast cancer. Initial experiments were conducted using a highdietary concentration of bean (60% wt/wt) where carcinoma bur-den in bean-fed rats was reduced 62.2% (P < 0.001) and histolog-ical and western blot analyses revealed that the dominant cellularprocess associated with reduced burden was induction of apopto-sis. Further analysis of mammary carcinomas revealed changes inthe phosphorylation states of mammalian target of rapamycin(mTOR) substrates (4E-binding protein 1 and p70S6 kinase)and mTOR regulators adenosine monophosphate-activated pro-tein kinase and protein kinase B (Akt) (P < 0.001). Effects onmTOR signaling in carcinomas were also found at lower dietaryconcentrations of bean (7.5–30% wt/wt). Liquid chromatogra-phy–time of flight–mass spectrometry analysis of plasma providedevidence of altered lipid metabolism consistent with reducedmTOR network activity in the liver (P < 0.001). Plasma concen-trations of insulin and insulin-like growth factor-1 were reducedby 36.3 and 38.9%, respectively, (P < 0.001), identifying a link toAkt regulation. Plasma C-reactive protein, a prognostic markerfor long-term survival in breast cancer patients, was reduced by23% (P < 0.001) in bean-fed rats. Identification of a role for themTOR signaling network in the reduction of cancer burden bydietary bean is highly relevant given that this pathway is deregu-lated in the majority of human breast cancers.

Introduction

Common bean (Phaseolus vulgaris L.), also referred to as dry bean, isknown as a staple food because of its contribution to daily caloricintake in many populations. It is widely available and affordable (1).In many areas of the world, average consumption of common beancan reach 150–200 g (dry weight) per day (2), but typical USA intake(7.5–10 g/ day, dry weight) (3) is well below recommended levels, inpart the result of the use of animal products rather than legumes asa source of dietary protein (4). Given evidence that bean consumption

is inversely associated with cancer risk, current consumption patternsin the USA suggest that bean is an under-utilized food for cancerprevention and control.

Epidemiological studies, such as the Nurses’ Health Study II, havefound intake of common beans and lentils to be associated withreduced breast cancer risk (relative risk 5 0.76, P , 0.03) (5). Inthe Four-Corners Breast Cancer Study, a relationship between beanconsumption and reduced breast cancer risk was reported in whichbreast cancer incidence in Hispanic women who consumed a nativeMexican diet (characterized by higher pulse consumption, such ascommon bean) was two-thirds that of the non-Hispanic white popu-lation whose diet was characterized as high in red meat, sugar andprocessed foods (6). Additional epidemiological and preclinical stud-ies evaluating colon cancer (7–10) and prostate cancer (11,12) havelent further support for an inverse relationship between bean con-sumption and the development of cancer. Our laboratory has shownsignificant inhibition of the post-initiation stage of chemically in-duced mammary carcinogenesis in the rat by common bean (13).However, little is known about the cell signaling pathways by whichbean exerts its effect.

In cancer, host systemic factors, such as insulin and insulin-likegrowth factor-1 (IGF-1), are known to contribute to the survival andproliferation of cancer cells through the activation of intracellularsignaling pathways that can also act autonomously of growth factorsduring the progression of cancer (14). Therefore, our investigations ofdietary bean consumption have centered: (i) on systemic factors (e.g.glucose-dependent growth factor signaling, inflammatory pathways),(ii) on cell autonomous mechanisms (e.g. cellular energy and nutrient-sensing networks) such as the mammalian target of rapamycin(mTOR) network and (iii) on signaling pathways through which sys-temic factors regulate cell proliferation and apoptosis. Given theemerging evidence that the mTOR network is deregulated in cardio-vascular disease, type-2 diabetes, and in cancer, including breast can-cer (15–18), and that little is currently known about the deregulationof mTOR components during early stages of carcinogenesis, the focusof our analysis was on mechanisms operative in mammary carcino-mas. Information gleaned from the analysis of carcinomas has strongimplications for cancer control, i.e. for individuals undergoing cancertreatment and for cancer survivors.

Materials and methods

Chemicals

Primary antibodies used in this study were anti-cyclin D1, anti-E2F-1 and anti-p27Kip1 from Thermo Fisher Scientific (Fremont, CA); anti-retinoblastoma (Rb),anti-B-cell lymphoma 2 (Bcl-2), anti-X-linked inhibitor of apoptosis protein andanti-Bcl-2 associated X protein (Bax) from BD Biosciences (San Diego, CA);anti-apoptosis protease-activating factor-1 from Millipore (Billerica, MA); anti-pAMPK/adenosine monophosphate-activated protein kinase (AMPK), anti-phospho-acetyl-CoA carboxylase (pACC)/ACC, anti-pAkt/Akt, anti-pp70S6K/p70S6K, anti-p4E-PB1/4E-PB1, anti-pRaptor/Raptor, anti-pPRAS40/PRAS40,anti-rabbit immunoglobulin-horseradish peroxidase-conjugated secondary anti-body and LumiGLO reagent with peroxide were purchased from Cell SignalingTechnology (Beverly, MA); anti-p21Cip1 and anti-mouse immunoglobulin-horseradish peroxidase-conjugated secondary antibody were from Santa CruzBiotechnology (Santa Cruz, CA); mouse anti-b-actin primary antibody wasobtained from Sigma–Aldrich (St Louis, MO) rabbit anti-Ki-67, clone SP6,was obtained from Labvision (Fremont, CA). Biotinylated donkey anti-rabbitand normal donkey serum were obtained from Jackson ImmunoResearch (WestGrove, PA); horseradish peroxidase-conjugated streptavidin was obtained fromDako (Carpinteria, CA) and Stable DAB was obtained from Invitrogen (Carls-bad, CA). The following chemicals for metabolite extraction and liquid chro-matography–mass spectrometry were used as received: ethanol (HPLC grade;Fisher, Pittsburgh, PA), methanol (Optima LC-MS grade; Fisher) and formicacid (LC-MS grade; Fluka, St Louis, MO).

Abbreviations: ACC, acetyl-CoA carboxylase; Akt, protein kinase B; AMPK,adenosine monophosphate-activated protein kinase; Bcl-2, B-cell lymphoma 2;Bax, Bcl-2 associated X protein; CRP, C-reactive protein; IGF-1, insulin-likegrowth factor-1; IL-6, interleukin-6; mTOR, mammalian target of rapamycin;m/z, mass-to-charge ratio of ion; Rb, retinoblastoma; VIP, variable importancein the projection.

� The Author 2011. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com 226

by Moldovan N

icolae-Andrei on M

ay 10, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Animals and experimental design

The plasma and tissue evaluated in this study were obtained from two pre-viously reported experiments (13,19). Briefly, female Sprague–Dawley ratswere obtained from Taconic Farms, Germantown, NY at 20 days of age. An-imal rooms were maintained at 22 ± 2�C with 50% relative humidity and a 12 hlight/12 h dark cycle. During the experiment, rats were weighed three times perweek. At 21 days of age, rats were injected with 1-methyl-1-nitrosourea(50 mg/kg body wt, intraperitoneally), as described previously (20). For thefirst week of the study, rats were housed three per cage in solid-bottomedpolycarbonate cages equipped with a food cup; they were given free accessto AIN-93G control diet. Seven days following carcinogen injection, all ratswere randomized to diet groups based on body weight. Bean was incorporatedat 60% wt/wt or for the dose response, 7.5, 15, 30 and 60% wt/wt. Rats werefed their assigned diets ad libitum until the end of the study at 46 days post-carcinogen. The post-initiation design of this experiment simulates the pro-motion and progression events of the disease process, which are highly relevantto women at increased risk for breast cancer and to breast cancer survivors. Thework followed guidelines approved by the Colorado State University AnimalCare and Use Committee.

Composition of diets

Beans were kindly provided by Archer Daniels Midland Company (Decatur,IL), a commercial bean processor, as seed. Seed was grown at multiplelocations and mixed; therefore, it was representative of the variation in envi-ronmental and genetic differences within beans typically consumed in theUSA. Bean seed was sent to Bush Brothers and Company (Chestnut Hill,TN) for canning and all material was processed according to industry standardmethods. Cooked beans were packed in standard brine without the incorpora-tion of any additives. Beans were then sent to Van Drunen Farms (Momence,IL) where the beans were removed from the cans, drained and then immedi-ately freeze-dried. The freeze-dried product was milled into a homogenouspowder and sent to Colorado State University where bean powders were storedat �20�C until incorporated into diets. Diets were formulated using specificguidelines and adjusted using data from proximate analysis (Warren Analyti-cal, Greeley, CO). The diets were formulated to match macronutrient levels(i.e. protein, carbohydrate and crude fiber) across the diet groups. The differ-ences in macronutrient composition were balanced with purified diet compo-nents. Diet formulations are reported in (13,19). Diets were stored at �20�Cuntil fed to animals.

Necropsy

Following an overnight fast, rats were euthanized over a 3 h time interval viainhalation of gaseous carbon dioxide. The sequence in which rats were eutha-nized was stratified across groups so as to minimize the likelihood that ordereffects would masquerade as treatment effects. After the rats lost conscious-ness, blood was directly obtained from the retro-orbital sinus and gravity fedthrough heparinized capillary tubes (Fisher Scientific) into ethylenediaminete-traacetic acid-coated tubes (Becton Dickinson, Franklin Lakes, NJ) to obtainplasma. The bleeding procedure took �1 min/rat. Plasma was isolated bycentrifugation at 1000g for 10 min at room temperature (22 ± 2�C). Followingblood collection and cervical dislocation, rats were then skinned and the skin towhich the mammary gland chains were attached was examined undertranslucent light for detectable mammary pathologies. All grossly detectablemammary gland pathologies were excised, weighed and a section was fixed inneutral buffered formalin; the remainder of each lesion was snap frozen inliquid nitrogen. Mammary pathologies were histopathologically classified fol-lowing routine hematoxylin and eosin staining as previously reported (21).Cancer incidence, multiplicity and tumor burden were based on histologicallyconfirmed mammary adenocarcinomas.

Measurement of plasma glucose, insulin, IGF-1, interleukin-6 and C-reactiveprotein levels

Glucose was determined using a kit obtained from Thermo Fisher Scientific(Waltham, MA). Insulin was determined by commercial enzyme-linkedimmunosorbent assay kit from Millipore. IGF-1 was determined using a com-mercial rat enzyme immunoassay kit from Diagnostic Systems Laboratories(Webster, TX). Interleukin-6 (IL-6) and C-reactive protein (CRP) were deter-mined using enzyme-linked immunosorbent assay kits from BD Biosciences.All analyses were performed according to manufacturer’s instructions.

Cell proliferation and apoptosis

Ki-67 immunohistochemical staining was used as an index of tumor growthfraction and was determined as described previously (22). Ki-67-stained sectionswere analyzed using a CAS-200 image analysis system (Bacus Labs, Lombard,IL). Apoptosis was quantified using the criteria developed by Kerr for its de-tection (23,24); images of corresponding hematoxylin- and eosin-stained serial

sections were acquired using a Zeiss Axioskop II (Carl Zeiss, Thornwood, NY)at a magnification of �400. Apoptotic and normal cells were marked andcounted using the manual tag tools in Image Pro Plus 4.5 (Media Cybernetics,Bethesda, MD).

Western blotting

Liver and mammary carcinomas (seven to twenty per group) were homoge-nized in lysis buffer [40 mM Tris–HCl (pH 7.5), 1% Triton X-100, 0.25 Msucrose, 3 mM ethyleneglycol-bis(aminoethylether)-tetraacetic acid, 3 mMethylenediaminetetraacetic acid, 50 lM b-mercaptoethanol, 1 mM phenylme-thylsulfonyl fluoride and complete protease inhibitor cocktail (Calbiochem,San Diego, CA)]. The lysates were centrifuged at 7500g for 10 min at 4�Cand supernatant fractions collected and stored at �80�C. Supernatant proteinconcentrations were determined by the Bio-Rad protein assay (Bio-Rad, Her-cules, CA). Western blotting was performed as described previously (25).Briefly, 40 lg of protein lysate per sample was subjected to 8–16% sodiumdodecyl sulfate–polyacrylamide gradient gel electrophoresis after being dena-tured by boiling with sodium dodecyl sulfate sample buffer [63 mM Tris–HCl(pH 6.8), 2% sodium dodecyl sulfate, 10% glycerol, 50 mM dithiothreitol and0.01% bromophenol blue] for 5 min. After electrophoresis, proteins were trans-ferred to a nitrocellulose membrane. The levels of cyclin D1, E2F-1, Rb,p21Cip1, p27Kip1, ppRb/pRb, Bcl-2, X-linked inhibitor of apoptosis protein,Bax, apoptosis protease-activating factor-1, VEGF, pAMPK, AMPK, pACC,ACC, pAkt, Akt, pp70S6K, p70S6K, p4E-binding protein 1, 4E-binding pro-tein, pRaptor, Raptor, pPRAS40, PRAS40 and b-actin were determined usingspecific primary antibodies, followed by treatment with the appropriate per-oxidase-conjugated secondary antibodies and visualized by LumiGLO reagentwestern blotting detection system. The chemiluminescence signal was cap-tured using a ChemiDoc densitometer (Bio-Rad) that was equipped witha CCD camera having a resolution of 1300 � 1030. Quantity One software(Bio-Rad) was used in the analysis. The actin-normalized scanning densitydata were used for analysis.

Extraction and analysis of plasma metabolites using liquid chromatography–mass spectrometry

Plasma samples from rats (100 ll) were extracted in 65% ethanol (750 ll),centrifuged (1000g, 5 min), supernatant was transferred to a clean 1.5 mlmicrocentrifuge tube and dried using a rotary evaporator without heat. Drysamples were resuspended in eluent prior to liquid chromatography–massspectrometry. An Acquity UPLC controlled with MassLynx software, version4.1 (Waters, Millford, MA) equipped with a 1.0 � 100 mm Waters AcquityUPLC BEH C18 column with 1.7 lm particle size was used for sampleseparation. Column temperature was held constant at 40�C. Separation wasperformed by reverse phase chromatography at a flow rate of 0.14 ml/min. Inorder to prevent evaporation, samples were held at 10�C in a sample managerduring the analysis. The complete sample set was randomized and profiled intriplicate. One microliter sample injections were made from 100 ll samplevolumes. The eluent consisted of water and methanol (Optima LC-MS grade;Fisher) and formic acid (LC-MS grade; Fluka) in the following proportions:Solvent A 5 95:5 water:methanol þ 0.1% formic acid; Solvent B 5 100%methanol þ 0.1% formic acid. The separation method is described as follows(58 min total): 3 min hold at 100% A, 30 min linear gradient to 100% B, 12 minhold at 100% B, 3 min linear gradient to 100% A and 10 min hold at 100%for equilibration. A Q-TOF Micro quadrupole orthogonal acceleration time-of-flight mass spectrometer (Waters/MicroMass) using positive mode electro-spray ionization (ESIþ) was used to collect mass spectral data at a rate of twoscans per second between 50 and 1000 mass-to-charge ratio of ion (m/z). Thevoltage and temperature parameters were tuned for general profiling as fol-lows: capillary 5 3000 V; sample cone 5 30 V; extraction cone 5 2.0 V;desolvation temperature 5 250�C; source temperature 5 130�C. Desolvationand cone nitrogen flows were set to 400 and 50 l/h, respectively. Leucineenkephalin was infused via a separate orthogonal ESI spray and baffle system(LockMass) which allowed ions to be detected for a single half-second scanevery 10 s in an independent data collection channel. The standard mass wasaveraged across 10 scans providing a continuous reference for mass correctionof analyte data. Mass spectral scans were centered in real time producingcentroid data using MassLynx software (Waters).

Statistical analyses

Cancer burden, rates of apoptosis and Ki-67 staining and actin-normalizedwestern blot data were initially evaluated by the Kruskal–Wallis Test withDwass–Steel–Chritchlow–Fligner Test for pairwise comparisons. Differencesamong groups in plasma molecules were analyzed by analysis of variance. Testfor linear trends was done by regression analysis on dietary concentration ofbean. Data were evaluated within SAS v 9.1.3 (Cary, NC), STATA (Stata

Mammary cancer burden and common bean

227

by Moldovan N

icolae-Andrei on M

ay 10, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Corporation, College Station, TX) or Systat statistical analysis software,Version 12.02. All P values are two sided and statistical significance was seta priori at P ,0.05.

Liquid chromatography–time of flight–mass spectrometry data

Centroid and integrated data were detected, extracted and aligned usingMarkerLynx software (Waters). Chromatographic peaks were extracted from0 to 35 min with a retention time error window of 0.1 min and mass spectralpeaks detected from 50 to 1000 m/z with a mass error window of 0.07 m/zgenerating a data matrix consisting of retention time, m/z and peak intensity forall features as determined by peak area. Data were mean centered and normal-ized with Pareto scaling for principal components analysis, variable impor-tance in the projection (VIP) and orthogonal partial least squares to latentstructures discriminant analysis using SIMCA-Pþ v12 (Umea, Sweden). Listsof potential markers for both differences in plasma obtained from bean-fedversus control-fed animals were generated using VIP with a threshold of 5 andp-(corr)[1] , �0.5 or p-(corr)[1] . 0.5 obtained from orthogonal partial leastsquares to latent structures discriminant analysis. Corresponding S-plots andprincipal components analysis scores plots were generated using SIMCA-Pþ.

Metabolic pathway analysis

MassTRIX: Mass Translator into Pathways (www.masstrix.org) was used toidentify preliminary ions and Kyoto Encyclopedia of Genes and Genomespathways, using the Rattus norvegicus database, which distinguish plasmafrom control rats and bean-fed rats. Proton and sodium corrections were ap-plied in MassTRIX using a mass error of 7 p.p.m.

Results

Tumor burden and mechanisms of its reduction

For the initial mechanistic experiments reported in this paper, analysesfocused on the 60% wt/wt concentration of dietary bean in order toenhance sensitivity to detect treatment effects and because of otherpreviously reported rodent carcinogenesis experiments using similarlyhigh concentrations (7,9). The mass of mammary carcinoma per rat wasmarkedly reduced (62.2%) in bean-fed rats (P , 0.001, Table I). Alsoshown in the table are the effects of dietary bean on the Ki-67 labelingindex, a measurement of cell growth fraction, i.e. the number of cells inthe cell cycle, and the apoptotic index, measured histologically using thecriteria initially developed by Kerr (23). For Ki-67, a small numericalreduction was observed in carcinomas from bean-fed rats, but the effectwas not statistically significant. On the other hand, the apoptotic indexwas increased 3-fold (P , 0.001) in the carcinomas from bean-fed rats.

Molecular markers of cell proliferation. To extend the immunohisto-chemial findings on cell growth fraction determined by Ki-67 staining

to the molecular machinery underlying cell proliferation, western blot-ting of mammary carcinomas was carried out for the cellcycle regulatory proteins cyclin D1, E2F-1, p21, p27 and pRB(Table I). Levels of cyclin D1 (P , 0.001) and E2F-1 (P , 0.001)were significantly lower in carcinomas from bean-fed animals, consis-tent with the finding that the ratio of hyper-phosphorylated Rb tohypo-phosphorylated Rb (ppRB/pRB) was decreased (P , 0.001). Cy-clin-dependent kinase inhibitors p21 (P , 0.001) and p27 (P , 0.001)were increased in bean-fed animal carcinomas. Representative westernblots for proteins involved in cell proliferation are shown (Figure 1A).

Molecular markers of apoptosis. In order to assess candidate path-ways of apoptosis induction, mammary carcinomas were westernblotted to determine levels of anti-apoptotic and pro-apoptotic factors.Anti-apoptotic factors Bcl-2 (P , 0.001) and X-linked inhibitor ofapoptosis protein (P , 0.001) were decreased in mammary carcino-mas of bean-fed rats; whereas pro-apoptotic factors Bax (P , 0.001)and apoptosis protease-activating factor-1 (P , 0.001) increased withbean consumption. The ratio of Bax to Bcl-2, an overall indicator ofthe apoptotic potential of the intracellular environment, was signifi-cantly higher (2.9-fold, P , 0.001) in carcinomas from bean-fed rats,a finding consistent with an elevated rate of apoptosis and reducedcarcinoma burden. Representative western blots for proteins involvedin apoptosis are shown (Figure 1A).

Systemic factors. The effects of feeding bean on plasma concentra-tions of glucose, insulin, IGF-1, CRP and IL-6 are shown (Table II).The concentrations of each analyte were reduced in bean-fed rats versuscontrol: glucose (27.0% reduction, P , 0.001), insulin (36.3% reduc-tion, P , 0.001), IGF-1 (38.8% reduction, P , 0.001), CRP (23.1%reduction, P , 0.001) and IL-6 (24.4%, P , 0.001).

mTOR and mammary carcinomas. A selected number of protein tar-gets upstream or downstream of mTOR were investigated in mammarycarcinomas to test the hypothesis that activation of AMPK, downregu-lation of protein kinase B (Akt) and/or suppression of mTOR signalingare candidate pathways for carcinoma burden reduction by dietary bean(60% wt/wt). The activity of upstream mTOR regulator AMPK wassignificantly increased (P , 0.001) while Akt was significantly de-creased (P , 0.001) in bean-fed rats compared with control (TableII). The activity of ACC, a direct target of activated AMPK that regu-lates lipid biosynthesis, was also evaluated. Phosphorylation of ACCwas increased in carcinomas of bean-fed rats (P , 0.001). Lastly, theactivity of mTOR in carcinomas from bean-fed rats was evaluated usingp70S6K and 4EBP-1, two downstream targets of mTOR. Decreasedphosphorylation of both proteins in the carcinoma was observed inbean-fed rats, consistent with reduced mTOR activity and cancer bur-den. Representative western blots are shown (Figure 1B).

Liver mTOR. The analysis of mTOR activity was extended tothe liver. Hepatic mTOR response to dietary treatment is shown(Table II), with upstream mTOR regulators AMPK and Akt beingup (30.7% increase, P 5 0.011) and downregulated (29.2% decrease,P 5 0.001), respectively. Direct regulator of mTOR activity, Raptor,a target of activated AMPK, and the regulator PRAS40, a target ofactivated Akt, were also analyzed in liver. The AMPK phosphoryla-tion site on Raptor was modestly increased (14.3%, P 5 0.037) bybean feeding; whereas a decrease in phosphorylation of PRAS40,predicted by the lower levels of activated Akt in bean-fed rats,was more apparent (23.3% reduction, P , 0.001). p70S6K and4EBP-1, downstream targets of activated mTOR, were also assessed.The phosphorylation of both proteins was lower in livers of bean-fedrats (p70S6K 5 0.037 and 4EBP-1 5 0.001). Overall, effects of beanfeeding are consistent with a reduction in mTOR activity in the liver.Representative western blots are shown (Figure 1C).

Plasma metabolome. Plasma samples from control and bean-fedrats were subjected to alcohol extraction and analysis by liquid

Table I. Effect of feeding bean-containing diets on cancer burden andcellular processes regulating tumor size, cell proliferation and apoptosis

Process Units Control Bean-fed P

Cancer burden g/rat 1.85 ± 0.74 (0.48) 0.70 ± 0.22 (0.11) ,0.001Cell proliferation

Ki-67 (index) % 26.1 ± 2.5 23.1 ± 2.2 NSCyclin D1 auod � 105 3.56 ± 0.18 2.94 ± 0.13 ,0.001E2F-1 auod � 105 2.36 ± 0.14 1.78 ± 0.06 ,0.001p21 auod � 104 9.1 ± 0.6 16.9 ± 1.2 ,0.001p27 auod � 104 6.41 ± 0.58 7.51 ± 0.38 ,0.01ppRb/pRb 1.00 ± 0.06 0.77 ± 0.03 ,0.001

ApoptosisApoptoticindex

% 1.65 ± 0.26 4.92 ± 0.59 ,0.001

Bcl-2 auod � 104 8.31 ± 0.46 5.66 ± 0.17 ,0.001XIAP auod � 105 2.52 ± 0.16 1.64 ± 0.10 ,0.001Bax auod � 104 4.59 ± 0.37 9.02 ± 0.48 ,0.001Apaf-1 auod � 104 5.06 ± 0.38 6.87 ± 0.31 ,0.001Bax/Bcl-2 0.55 ± 0.03 1.60 ± 0.08 ,0.001

Values are means ± SEM, values in parentheses are medians (control, n 5 9;bean, n 5 20). Data were analyzed by the Kruskal–Wallis test. Apaf-1, apoptosisprotease-activating factor-1; auod, arbitrary units of optical density; NS, notstatistically different from control; XIAP, X-linked inhibitor of apoptosis protein.

M.D.Thompson et al.

228

by Moldovan N

icolae-Andrei on M

ay 10, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

chromatography–time of flight–mass spectrometry to determine ifbean consumption affected the plasma metabolome. Principal com-ponents analysis was performed on the resulting dataset and revealedseparation of treatment groups with 45.59% of the total varianceexplained by the first three principal components (PC 1 5 22.7%,PC 2 5 14.99%, PC 35 7.9%) (Figure 1D). The plasma metaboliteprofiles were further analyzed using VIP and orthogonal partial leastsquares to latent structures discriminant analysis as described in theMaterials and Methods. Features indicated by circles on the S-plot(Supplementary Figure 1 is available at Carcinogenesis Online) werethose which passed the VIP threshold of 5 and p(corr)[1] , �0.5 orp(corr)[1] . 0.5 when comparing control to bean-fed animals. Aninclusion list of ions contributing to the differences between bean-fed and control animals is shown (Supplementary Table 1 is availableat Carcinogenesis Online). All features were evaluated using Mass-TRIX to determine possible ion identities and candidate pathwaysdifferentially regulated in bean-fed animals versus control. Bioinfor-matic pathway analysis indicated fatty acid biosynthesis, steroid me-tabolism and liver bile acid metabolism were affected by feedingbean, consistent with the effect of dietary bean on the mTOR network.

Dose-dependent effects of dietary bean on cancer burden, induction ofapoptosis, systemic factors and mTOR signaling

Analysis of the mTOR network was extended to mammary carcinomasfrom rats fed 7.5, 15 or 30% (wt/wt) common bean, dietary concentra-tions that correspond to amounts of bean consumed by various popula-tions around the world. Effects on regulators of mTOR were dosedependent (Table III and Figure 2). While AMPK activation was mark-edly induced only at the highest dietary concentration, the phosphoryla-tion of Akt was reduced across the dose response (P5 0.008). Decreasedphosphorylation of p70S6K (P 5 0.037) and 4E-binding protein 1 (P 50.001) indicated a dietary dose-dependent downregulation of mTORactivity resulting from bean feeding.

Discussion

Tumor burden and mechanisms of its reduction

Evidence presented in this paper and previously (13,19) has demon-strated that rats fed common bean have reduced mammary carcinomaburden, implicating processes that regulate the proliferation and

Fig. 1. Effects of bean feeding on cell cycle and apoptosis regulators, mTOR signaling and the plasma metabolome. The images shown are those directly acquiredfrom the ChemiDoc work station that is equipped with a CCD camera having a resolution of 1300 � 1030. (A) A composite image of representative western blotsof lysates of carcinomas from control (CTRL) and bean-fed (bean) rats. Images are for cell cycle regulators, cyclin D1, E2F-1, p21Cip1, p27Kip1 and Rb (ppRb,hyper-phosphorylated Rb; pRb, hypo-phosphorylated Rb) and apoptosis regulators, Bcl-2, X-linked inhibitor of apoptosis protein (XIAP), Bax and Apaf-1. (B)A composite image of representative western blots of lysates of carcinomas from control (CTRL) and bean-fed (bean) rats. Images are for components of theAMPK-Akt-mTOR signaling network, phosphorylated and total: AMPK, ACC, Akt, p70S6 kinase (p70S6K) and 4E-binding protein 1 (4E-BP1). (C) A compositeimage of representative western blots of lysates of liver from control (CTRL) and bean-fed (bean) rats. Images are for phosphorylated and total: AMPK, Raptor,Akt, PRAS40, p70S6K and 4E-BP1. (D) Scores scatter plot from principal component analysis (PCA) of plasma demonstrating the separation between rats fedcontrol diet (circles) and bean diet (60% wt/wt) (squares). (control, n 5 7; bean fed, n 5 11).

Mammary cancer burden and common bean

229

by Moldovan N

icolae-Andrei on M

ay 10, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

survival of cancer cells. Summarizing this evidence, mammarycancer burden as well as the average number of cancers per rat, isreduced dose dependently over a range of dietary bean concentrations(Figure 3A). Data also indicate that induction of apoptosis is dosedependent and associated with reduced cancer burden and that deathinduction is primarily through the intrinsic mitochondrial pathwaysindicated by the ratio of Bax to Bcl-2 (Figure 3B). Unexpectedly,there is little evidence that the fraction of cells in the proliferativepool (Ki-67 staining index, Figure 3B) is affected by bean consump-tion despite effects on cell cycle machinery (Table I). This is possiblybecause these effects are secondary to induction of a pro-apoptoticenvironment within carcinomas. Apoptosis being a dominant mecha-nism is consistent with the argument that dietary bean may be bene-ficial in the cancer control context where agents that induce apoptosishave been reported to have greater clinical efficacy (26), an argumentthat also is supported by the dose-dependent reduction in CRP withincreasing dietary bean (Figure 3C). CRP, the synthesis of whichoccurs in the liver and responds to circulating levels of IL-6, hasrecently been shown to have prognostic value for long-term survivalfollowing treatment for breast cancer (27).

mTOR and mammary carcinomas

mTOR is an evolutionarily conserved serine/threonine kinase thatintegrates external cellular stimuli with intracellular energy and nu-trient-sensing pathways to regulate cellular metabolism and growth(28). Two upstream regulators were investigated. AMPK was chosenbecause it is the target of the widely used diabetes management drug,metformin, which is a nitrogen containing phytochemical derivative,and common bean is a rich source of small nitrogen containing com-pounds (29). Moreover, AMPK is a key regulator of lipid metabolismdirectly via effects on ACC, one of its downstream targets (30) anddietary bean affects altered lipid metabolism in cardiovascular disease(31). At the highest dietary concentration of bean, AMPK was acti-vated and ACC was downregulated. This finding is noteworthy be-cause AMPK-mediated suppression of ACC activity limits the abilityof the cell to carry out robust lipid synthesis required for tumor growth(32–34). It is not clear whether activation of AMPK was due toa specific perturbation of energy metabolism related to signalingthrough the IGF-1 receptor, to the effects on plasma cytokines, suchas adiponectin which is known to regulate the activity of AMPK (35),or to other factors. However, since consistent activation of AMPK was

Table II. Effect of feeding bean-containing diet on systemic factors andmTOR network components in mammary carcinomas and liver

Tissue Units Control Bean-fed P

CarcinomasAMPK Ratio 2.43 ± 0.28 3.18 ± 0.20 ,0.001ACC Ratio 3.23 ± 0.26 4.15 ± 0.14 ,0.001Akt Ratio 0.65 ± 0.06 0.46 ± 0.03 ,0.001p70S6K Ratio 0.33 ± 0.04 0.23 ± 0.02 ,0.0014E-BP1 Ratio 4.31 ± 0.42 2.71 ± 0.10 ,0.001

PlasmaGlucose mg/dl 125.3 ± 5.4 91.5 ± 2.6 ,0.001Insulin ng/ml 2.73 ± 0.1 1.74 ± 0.05 ,0.001IGF-1 ng/ml 1586 ± 54 969 ± 31 ,0.001C-reactive protein lg/ml 334 ± 9 257 ± 5 ,0.001IL-6 pg/ml 48.8 ± 1.2 36.9 ± 1.3 ,0.001

LiverAMPK Ratio 3.15 ± 0.10 3.65 ± 0.10 0.011Raptor Ratio 1.40 ± 0.10 1.60 ± 0.10 0.037Akt Ratio 0.34 ± 0.02 0.26 ± 0.01 ,0.001PRAS-40 Ratio 1.20 ± 0.04 0.92 ± 0.02 ,0.001p70S6K Ratio 0.12 ± 0.01 0.09 ± 0.01 0.0374E-BP1 Ratio 3.23 ± 0.29 3.02 ± 0.22 ,0.001

Values are means ± SEM, (control, n 5 9; bean, n 5 20). Actin-normalizedwestern blot data were analyzed by the Kruskal–Wallis test. Plasma analytedata were analyzed by analysis of variance. Ratio, the ratio of phospho-protein (arbitrary units of optical density) to total protein (arbitrary units ofoptical density); 4E-BP1, 4E-binding protein 1.

Table III. Effect of feeding bean-containing diets on the mTOR signaling network in mammary carcinomas

Diet Ratio of phosphorylated protein to total protein

AMPK ACC Akt p70S6K 4E-BP1

Control 2.43 ± 0.28a 3.23 ± 0.26a 0.65 ± 0.06a 0.33 ± 0.04a 4.31 ± 0.42a

7.5% Beans 2.47 ± 0.12a 3.84 ± 0.17a 0.50 ± 0.07b 0.28 ± 0.05a,b 4.36 ± 0.47a,b

15% Beans 2.65 ± 0.23a 3.97 ± 0.23a 0.51 ± 0.05b 0.27 ± 0.05a,b 4.18 ± 0.18a

30% Beans 2.37 ± 0.28a 3.99 ± 0.44a 0.52 ± 0.05b 0.26 ± 0.03b 3.27 ± 0.40b

60% Beans 3.18 ± 0.20b 4.15 ± 0.14b 0.46 ± 0.03c 0.23 ± 0.02c 2.71 ± 0.10c

P-linear trend 0.011 0.006 0.008 0.037 0.001

Values are means ± SEM (control, n 5 9; 7.5% bean, n 5 8; 15% beans, n 5 9; 30% beans, n 5 7; 60% bean, n 5 20). Data were analyzed by the Kruskal–Wallistest with Dwass–Steel–Chritchlow–Fligner test for pairwise comparisons as implemented in Systat Statistical Analysis Software. P-value for linear trend wasdetermined by regression analysis. Values within a column with different superscripts (a, b, c) are statistically different from each other (P , 0.05). 4E-BP1, 4E-binding protein 1.

Fig. 2. Dose-dependent effects of bean feeding on mTOR signalingcomponents. The images shown are those directly acquired from theChemiDoc work station that is equipped with a CCD camera havinga resolution of 1300 � 1030. A composite image of representative westernblots of lysates of carcinomas from control (CTRL) rats and from rats fed 7.5,15, 30 or 60% (wt/wt) bean in the diet. Images are for components of theAMPK-Akt-mTOR signaling network, phosphorylated and total: AMPK,ACC, Akt, p70S6 kinase (p70S6K) and 4E-binding protein 1 (4E-BP1).

M.D.Thompson et al.

230

by Moldovan N

icolae-Andrei on M

ay 10, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

not observed at lower dietary concentrations, despite evidence of in-creased phosphorylation of ACC, other mechanisms are likely to beinvolved in mTOR regulation in mammary carcinomas. The secondupstream regulator of the mTOR network that was investigated wasprotein kinase B (Akt). It was selected for analysis because it isa downstream effector of the insulin and IGF-1 receptors and bothplasma insulin and IGF-1 were reduced dose dependently with in-creasing dietary concentration of bean (Figure 3C). Reduced activa-tion of Akt was observed at lower concentrations of bean, a findingconsistent with the dose-dependent effects of dietary bean on insulinand IGF-1 and the induction of apoptosis (36).

Two downstream targets of mTOR are generally evaluated to assessmTOR activity. They are 4E-binding protein 1 and p70S6 kinase(37,38). In carcinomas, the phosphorylation of both proteins was dosedependently reduced, consistent with downregulation of mTOR(Table III and Figure 3D). Through mTOR-mediated phosphorylation,4E-binding protein 1 is de-repressed which enhances ribosomal bio-genesis, and phosphorylation of p70S6 kinase by mTOR stimulatescap-dependent protein synthesis. The decreased level of phosphory-lation of both proteins is consistent with reduced cancer burden andinduction of apoptosis via the mitochondrial pathway (36).

In addition to mammary carcinomas, the liver was investigatedbecause of its central role in sensing the availability to the organismof energy and nutrients and in turn controlling metabolism and thehomeostatic regulation of size in peripheral tissues like the breast.

mTOR activity was downregulated in the liver providing evidencethat feeding bean in the diet was exerting a systemic effect on thehost and as noted above, evidence of these effects was observed notonly in the dose-dependent suppression of plasma growth factors,hormones and inflammatory factors but also in lower concentrationsof fasting blood glucose (Table II). Moreover, the effects of bean onvarious nodes in the mTOR signaling network is consistent with thealtered patterns of plasma lipids detected by metabolomics analysis(Figure 1D; Supplementary Table 1 is available at CarcinogenesisOnline). It is noteworthy that mTOR itself has direct effects on lipidbiosynthesis that are mediated via sterol response element-bindingprotein (38). Therefore, alterations in hepatic mTOR activity are con-sistent with the evidence that dietary bean impacts lipid profiles ina number of chronic disease states (31). In cancer, the possible alter-ation of lipid biosynthesis within carcinomas would suggest an addi-tional mechanism of suppressed tumor burden given the reliance oftumor growth on de novo lipid biosynthesis (32–34).

Limitations

These findings provide an important foundation for new experiments tofurther investigate the role of dietary bean in cancer control as well asprevention. Nonetheless, the design of the current experiments neces-sitates that the reported data be interpreted with appropriate caution. Interms of effects on molecular pathways that regulate tissue size

Fig. 3. Summary of dose-responsive effects of dietary bean. Concentrations of dietary bean investigated were 7.5, 15, 30 and 60% wt/wt. All data are expressed asa percent of the response observed in rats fed control diet. (A) Data from Table I or previously reported in ref. 19. Multiplicity refers to the average number ofcancers per rat. Tumor burden refers to the average mass of carcinoma per rat originally reported as grams per rat. (B) Data from Table I or previously reported inref. 19. Apoptosis refers to the percent of cells in a carcinoma undergoing apoptosis, Bax/Bcl-2 refers to the ratio of these pro- and anti-apoptotic factorsdetermined in carcinoma by western blot analysis and Ki-67 refers to the percent of cells within a carcinoma in the growth fraction as determined byimmunohistochemical analysis. (C) Data from Table II or previously reported in ref. 19. Analytes measured in plasma using the methodology described in theMaterials and Methods. (D) Western blot data from Table III showing dose-dependent effects of bean feeding on components of the mTOR signaling network inmammary carcinoma.

Mammary cancer burden and common bean

231

by Moldovan N

icolae-Andrei on M

ay 10, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

homeostasis, carcinomas from untreated rats may respond differently todietary bean than carcinomas that have emerged in bean-fed rats.

Concluding comments

Collectively, common bean appears to reduce mammary cancer burdenprimarily through inducing apoptosis and modifying key metabolicsignaling networks linked to cell growth and survival. Part of this effecton cellular pathways is attributed to alterations in levels of systemicfactors linked to mTOR-associated intracellular signaling. Understand-ing how dietary bean influences the interaction of systemic factors andcellular pathways, particularly its dose dependence, will be especiallybeneficial in guiding public health decisions regarding bean consump-tion for cancer control in the USA where consumption of common beanis markedly below recommended levels.

Supplementary material

Supplementary Figure 1 and Table 1 can be found at http://carcin.oxfordjournals.org/

Funding

United States Agency for International Development (REE-A-00-03-00094-00); American Institute for Cancer Research (08A032); Colo-rado State University Crops For Health� research program.

Acknowledgements

The authors would like to thank Leslie Brick, Vanessa Fitzgerald, ElizabethNeil, Denise Rush and Jennifer Sells for their excellent technical assistance.Thanks are extended to Archer Daniels Midland Co. and Bush Brothers, forproviding and processing the beans for this study.

Conflict of Interest Statement: None declared.

References

1.Geil,P.B. et al. (1994) Nutrition and health implications of dry beans: a re-view. J. Am. Coll. Nutr., 13, 549–558.

2.Broughton,W.J. et al. (2003) Beans (Phaseolus spp.)—model food legumes.Plant Soil, 252, 55–128.

3.Economic Research Service. (2007) Food Availability. USDA, Washington,DC.

4.U.S. Department of Agriculture. (2010) Dietary Guidelines for Americans.http://www.cnpp.usda.gov/DGAs2010-PolicyDocument.htm. (July 2011,date last accessed).

5.Adebamowo,C.A. et al. (2005) Dietary flavonols and flavonol-rich foodsintake and the risk of breast cancer. Int. J. Cancer, 114, 628–633.

6.Murtaugh,M.A. et al. (2008) Diet patterns and breast cancer risk in His-panic and non-Hispanic white women: the Four-Corners Breast CancerStudy. Am. J. Clin. Nutr., 87, 978–984.

7.Bobe,G. et al. (2008) Dietary cooked navy beans and their fractions atten-uate colon carcinogenesis in azoxymethane-induced ob/ob mice. Nutr. Can-cer, 60, 373–381.

8.Mentor-Marcel,R.A. et al. (2009) Inflammation-associated serum and co-lon markers as indicators of dietary attenuation of colon carcinogenesis inob/ob mice. Cancer Prev. Res. (Phila Pa), 2, 60–69.

9.Hangen,L. et al. (2002) Consumption of black beans and navy beans (Pha-seolus vulgaris) reduced azoxymethane-induced colon cancer in rats. Nutr.Cancer, 44, 60–65.

10.Lanza,E. et al. (2006) High dry bean intake and reduced risk of advancedcolorectal adenoma recurrence among participants in the polyp preventiontrial. J. Nutr., 136, 1896–1903.

11.Mills,P.K. et al. (1989) Cohort study of diet, lifestyle, and prostate cancer inAdventist men. Cancer, 64, 598–604.

12.Kolonel,L.N. et al. (2000) Vegetables, fruits, legumes and prostate cancer:a multiethnic case-control study. Cancer Epidemiol. Biomarkers Prev., 9,795–804.

13.Thompson,M.D. et al. (2009) Chemical composition and mammary cancerinhibitory activity of dry bean. Crop Sci., 49, 179–186.

14.Hanahan,D. et al. (2000) The hallmarks of cancer. Cell, 100, 57–70.15.Marshall,S. (2006) Role of insulin, adipocyte hormones, and nutrient-

sensing pathways in regulating fuel metabolism and energy homeostasis:a nutritional perspective of diabetes, obesity, and cancer. Sci. STKE, 2006,re7.

16.Um,S.H. et al. (2006) Nutrient overload, insulin resistance, and ribosomalprotein S6 kinase 1, S6K1. Cell Metab., 3, 393–402.

17.Yang,Q. et al. (2007) Expanding mTOR signaling. Cell Res., 17, 666–681.18.Hynes,N.E. et al. (2006) The mTOR pathway in breast cancer. J. Mammary

Gland Biol. Neoplasia., 11, 53–61.19.Thompson,M.D. et al. (2008) Mechanisms associated with dose-dependent

inhibition of rat mammary carcinogenesis by dry bean (Phaseolus vulgaris,L.). J. Nutr., 138, 2091–2097.

20.Thompson,H.J. et al. (1995) Rapid induction of mammary intraductal pro-liferations, ductal carcinoma in situ and carcinomas by the injection ofsexually immature female rats with 1-methyl-1-nitrosourea. Carcinogene-sis, 16, 2407–2411.

21.Thompson,H.J. et al. (2000) Classification of premalignant and malignantlesions developing in the rat mammary gland after injection of sexuallyimmature rats with 1-methyl-1-nitrosourea. J. Mammary Gland Biol.Neo-plasia., 5, 201–210.

22.McGinley,J.N. et al. (2000) Effect of fixation and epitope retrieval onBrdU indices in mammary carcinomas. J. Histochem. Cytochem., 48,355–362.

23.Kerr,J.F. et al. (1972) Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. Br. J. Cancer, 26, 239–257.

24.Walker,N.I. et al. (1989) Cell death by apoptosis during involution of thelactating breast in mice and rats. Am. J. Anat., 185, 19–32.

25. Jiang,W. et al. (2003) Effect of energy restriction on cell cycle machinery in1-methyl-1-nitrosourea-induced mammary carcinomas in rats. CancerRes., 63, 1228–1234.

26.Christov,K. et al. (2007) Short-term modulation of cell proliferation andapoptosis and preventive/therapeutic efficacy of various agents in a mam-mary cancer model. Clin. Cancer Res., 13, 5488–5496.

27.Pierce,B.L. et al. (2009) Elevated biomarkers of inflammation are associ-ated with reduced survival among breast cancer patients. J. Clin. Oncol., 27,3437–3444.

28.Howell,J.J. et al. (2011) mTOR couples cellular nutrient sensing toorganismal metabolic homeostasis. Trends Endocrinol. Metab., 22, 94–102.

29.Wink,M. (2003) Evolution of secondary metabolites from an ecologicaland molecular phylogenetic perspective. Phytochemistry, 64, 3–19.

30.Hardie,D.G. (2004) AMP-activated protein kinase: a master switch in glu-cose and lipid metabolism. Rev. Endocr. Metab. Disord., 5, 119–125.

31.Bennink,M.R. et al. (2008) Beans and Health: A Comprehensive Review.The Bean Institute, Frazee, MN, pp. 1–31.

32.Shackelford,D.B. et al. (2009) The LKB1-AMPK pathway: metabolismand growth control in tumour suppression. Nat. Rev. Cancer, 9, 563–575.

33.Shaw,R.J. (2009) LKB1 and AMP-activated protein kinase control ofmTOR signalling and growth. Acta Physiol. (Oxf.), 196, 65–80.

34.Hardie,D.G. (2008) AMPK: a key regulator of energy balance in thesingle cell and the whole organism. Int. J. Obes. (Lond.), 32 (suppl. 4),S7–S12.

35.Hardie,D.G. (2008) AMPK and Raptor: matching cell growth to energysupply. Mol. Cell, 30, 263–265.

36.Stiles,B.L. (2009) PI-3-K and AKT: onto the mitochondria. Adv. DrugDeliv. Rev., 61, 1276–1282.

37.Efeyan,A. et al. (2010) mTOR and cancer: many loops in one pathway.Curr. Opin. Cell Biol., 22, 169–176.

38.Laplante,M. et al. (2009) An emerging role of mTOR in lipid biosynthesis.Curr. Biol., 19, R1046–R1052.

Received July 11, 2011; revised October 27, 2011; accepted October 29, 2011

M.D.Thompson et al.

232

by Moldovan N

icolae-Andrei on M

ay 10, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from