Biological Responses to TGF-b in the Mammary EpitheliumShow a Complex Dependency on Smad3 Gene Dosage withImportant Implications for Tumor Progression
Ethan A. Kohn1, Yu-an Yang1, Zhijun Du1, Yoshiko Nagano1, Catherine M.H. Van Schyndle1,Michelle A. Herrmann2, Madeleine Heldman2, Jin-Qiu Chen2, Christina H. Stuelten3,Kathleen C. Flanders1, and Lalage M. Wakefield1
AbstractTGF-b plays a dual role in epithelial carcinogenesis with the potential to either suppress or promote tumor
progression. We found that levels of Smad3 mRNA, a critical mediator of TGF-b signaling, are reduced byapproximately 60% in human breast cancer. We therefore used conditionally immortalized mammary epithelialcells (IMEC) of differing Smad3 genotypes to quantitatively address the Smad3 requirement for different biologicresponses to TGF-b. We found that a two-fold reduction in Smad3 gene dosage led to complex effects on TGF-bresponses; the growth-inhibitory response was retained, the pro-apoptotic response was lost, themigratory responsewas reduced, and the invasion response was enhanced. Loss of the pro-apoptotic response in the Smad3þ/� IMECscorrelated with loss of Smad3 binding to the Bcl-2 locus, whereas retention of the growth-inhibitory response inSmad3 IMECs correlated with retention of Smad3 binding to the c-Myc locus. Addressing the integrated outcomeof these changes in vivo, we showed that reduced Smad3 levels enhanced metastasis in two independent models ofmetastatic breast cancer.Our results suggest that different biologic responses to TGF-b in themammary epitheliumare differentially affected by Smad3 dosage and that a mere two-fold reduction in Smad3 is sufficient to promotemetastasis. Mol Cancer Res; 10(10); 1389–99. �2012 AACR.
IntroductionTGF-bs are widely expressed, pleiotropic growth factors
that modulate a variety of key biologic responses. In the adultanimal, the TGF-bs are important in maintaining tissuehomeostasis and in orchestrating response to stress or injury,and perturbations in the TGF-b signaling pathway have beenimplicated in a wide variety of pathologic processes. Incarcinogenesis, TGF-bs play a complex dual role in manyepithelial tissues including the mammary gland, as evidencedby numerous preclinical studies and correlative clinical data(reviewed in refs. 1–4). Early in tumorigenesis, tumor-sup-pressive effects of TGF-bs generally dominate, involvingbiologic responses such as inhibition of proliferation, induc-
tion of apoptosis or replicative senescence, maintenance ofgenomic stability and regulation of differentiation. However,in advanceddisease, resistance to the tumor-suppressive effectsof TGF-b frequently develops, due to altered expression and/or activity of components of the TGF-b signaling pathway, orto progression-related changes that specifically affect theeffector arm of the tumor suppression program. At the sametime, tumor cell–autonomous responses to TGF-b that pro-mote progression, such as invasion and migration, areenhanced or unmasked by the accumulating genetic andepigenetic changes in the tumor. Furthermore, increasedTGF-b in the tumor bed leads to tumor-promoting effectsof TGF-b on the microenvironment, such as suppression ofimmunosurveillance and enhanced angiogenesis.The canonical TGF-b signaling pathway involves TGF-
b–induced phosphorylation of two central mediators,Smad2 and Smad3, which translocate to the nucleus andregulate gene expression (5, 6). Smad3 appears to be themore important mediator in the adult organism, as Smad2 iseither dispensable or actually opposes Smad3 in manybiologic responses (7–9). However, unlike Smad2, Smad3is rarely mutated or deleted in human cancer (10). Here, wehave shown that human breast tumors have Smad3 mRNAlevels that are reduced approximately 2-fold compared withthe adjacent normal breast tissue. We have used a geneticapproach to ask what effect this reduction in Smad3 hason TGF-b responses and tumorigenesis in the breast
Authors' Affiliations: 1Laboratory of Cancer Biology and Genetics,2Collaborative Protein Technology Resource, Laboratory of Cell Biology,and 3 Laboratory of Cellular and Molecular Biology, Center for CancerResearch, National Cancer Institute, Bethesda, Maryland
Note: Supplementary data for this article are available at Molecular CancerResearch Online (http://mcr.aacrjournals.org/).
E.A. Kohn and Y.-a. Yang contributed equally to the work.
Corresponding Author: Lalage M. Wakefield, National Cancer Institute,Bldg. 37, Rm. 4032A, 37 Convent Drive, MSC 4255, Bethesda, MD 20892.Phone: 301-496-8351; Fax: 301-480-2772; E-mail: [email protected]
epithelium. Unexpectedly, we found that a mere 2-foldreduction in Smad3 significantly affected multiple TGF-bresponses in vitro in a variety of different ways. In vivo, theintegrated outcome of these effects was an enhancement ofmetastasis in two different breast cancer models.
Materials and MethodsCell lines and genetic modificationConditionally immortalized mouse mammary epithelial
cells (IMEC) with 0, 1, or 2 alleles of Smad3 were generatedby intercrossing the Smad3DelEx8 knockout mouse (11)and the H-2Kb-tsA58 transgenic "Immortomouse," whichbroadly expresses a temperature-sensitive SV-40 large Tantigen (12). The IMECs were generated and propagatedas previously described (7). Briefly, IMECS were expandedand propagated under permissive conditions: growth at33�C (5%CO2) in Ham's F12 medium supplemented with10% FBS, 10 ng/mL EGF, 5 mg/mL insulin, 1 mg/mLhydrocortisone, 5 ng/mL cholera toxin, 50 mg/mL genta-mycin ("complete" medium) and 30 U/mL IFN-g . Forassays, cells were switched to nonpermissive conditions bygrowing at 37�C (5% CO2) in the absence of IFN-g for 2 to4 days before the assay, to provide sufficient time for Tantigen expression to fully decay. For the Smad3þ/� geno-type that was the focus of this study, 3 independent IMECisolates were generated: EAK4, IMEC49, and IMEC 52.Mvt-1, a metastaticmousemammary cell line derived from amammary tumor that arose in an MMTV-c-Myc/VEGFbitransgenic mouse (13), was a gift of the late Dr. RobertDickson (Georgetown University, Washington, D.C.) andwas cultured at 37�C (5% CO2) in Dulbecco's ModifiedEagle's Media with 10% FBS. For genetic modification ofthe Mvt-1 line to knockdown Smad3, short hairpin RNA(shRNA) against Smad3 (50-GGCCATCACCACGCA-GAAC-30) or GFP (50-AAGACCCGCGCCGAGGT-GAAG-30) were cloned into the pLKO.1 lentiviral vector.To add back Smad3 to the Smad3þ/� IMECs, humanSmad3 was cloned into a modified pFUGW lentiviralbackbone byGateway recombinational cloning, with expres-sion driven by the CMV13 promoter. A stuffer sequence wasused in place of Smad3 to generate the control construct. Inboth cases, transduced cells were selected with puromycin for5 days, and in vivo experiments were carried out within 3 to 5passages. For knockdown and add-back experiments, Smad3protein levels were quantitated by SimpleWestern (seebelow).
Conventional Western blot, Simple Western analyses,and zymographyConventional Western blot analyses, and gelatin zymo-
graphy to assess matrix metalloproteinase (MMP) activity,were conducted essentially as previously described (7). ForWestern blot analyses of phospho-Smads, serum-starved cellswere treated with 2 ng/mL TGF-b for 1 hour. For Westernblot analyses of other TGF-b target proteins, serum-starvedcells were treated for 24 hours. Antibodies used were asfollows: Smad2, Zymed # 51–1300 or Invitrogen #511300
(1:1,000); phospho-Smad2, Cell Signaling #3108S(1:1,000); Smad3, Abcam #ab28379 (1:1,000); phospho-Smad3, Epitomics #1880–1 (1:1,000); b-actin, Sigma#A1978 (1:5,000); c-myc, Active Motif #39012 (1:500);Bcl-2, Cell Signaling (1:1,000); cleaved caspase-3, CellSignaling #9661 (1:1,000); p15INK4B, Santa Cruz Bio-technology #sc-613 (1:1,000); and p21Cip1, Santa CruzBiotechnology #sc-6246 (1:1,000). Membranes wereblocked with 5% bovine serum albumin (Smad2) or 5%nonfat dry milk (all others).In select cases, Smad3 protein levels were quantitated
using the automated capillary-based Simon SimpleWesternSystem (ProteinSimple). With this technique, automationand the elimination of the blotting step allows more accurateand reproducible assessment of protein levels. 60 ng of celllysate protein in reducing buffer with fluorescent molecularweight standards was loaded into each capillary, and proteinswere separated by molecular weight through stacking andseparation matrices for 45 minutes at 250 V. Proteins wereimmobilized to capillary walls using proprietary, photoacti-vated capture chemistry. Capillaries were then incubatedwith a blocking reagent and target proteins were immuno-probed with rabbit anti-Smad3 (Epitomics #1735–1) andanti-a-tubulin (Cell Signaling Technology #2125) primaryantibodies and horseradish peroxidase–conjugated anti-rab-bit secondary antibodies (Jackson ImmunoResearch). Amixture of luminol and peroxide was added, the resultingchemiluminescent signal was captured by a CCD camera,and the signal intensities were quantified and analyzed usingCompass Software (ProteinSimple). During analysis, theSmad3 signal was normalized to the a-tubulin loadingcontrol in the same samples.
Quantitative real-time PCRReal-time PCR (RT-PCR) reactions were run according
to the manufacturer's instructions (Brilliant SYBR GreenMaster Mix, Stratagene) using a BioRad CFX96 Real TimeSystem. Expression was normalized to either PPIA or PPIB.For quantitative RT PCR (QRT-PCR) of TGF-b targetsrelated to the epithelial-to-mesenchymal (EMT) transition,IMECs were treated with TGF-b for 48 hours. NMuMGcells, which are very sensitive toTGF-b–induced EMT,weretreated with TGF-b for 24 hours and served as a positivecontrol. Primer pairs for the various targets are given inSupplementary Table S1.
Smad3 expression in human breast tissuesTotal RNA from tumor and normal adjacent breast tissue
from 10 female patients with breast cancer was purchasedfrom Oncomatrix Inc., and Smad3 mRNA was assessed byQRT-PCR and normalized to PPIA. Patient characteristicsare given in Supplementary Table S2. A meta-analysis toinvestigate the relationship of Smad3 mRNA expression toclinical parameters across 8 independent breast cancercohorts was conducted using the online GOBO tool(http://co.bmc.lu.se/gobo; ref. 14). For immunohistochem-istry to assess which cell types in the tumor express Smad3,unstained sections of an estrogen receptor (ER)-negative,
Kohn et al.
Mol Cancer Res; 10(10) October 2012 Molecular Cancer Research1390
lymph node–positive, grade 2 breast cancer, and matchedadjacent normal tissue from the same patient were purchasedfrom Capital Biosciences. Immunolocalization of Smad3using rabbit anti-Smad3 from Abcam (cat # ab28379) at0.04 mg/mL was conducted as previously described (15).Before immunostaining, antigen retrieval in 1 mmol/LEDTA (pH 8) was carried out at 95oC for 10 minutesfollowing deparaffinization of sections. A mouse embryosection was stained as a positive control, and omission ofprimary antibody served as the negative control.
Smad3 chromatin immunoprecipitationSerum-starved cells were treated with 2 ng/mL TGF-b or
vehicle for 1 hour, before dual cross-linking with 2 mmol/Ldi-(N-succinimidyl)glutarate (Thermo Scientific) and 1%formaldehyde, followed by sonication and immunoprecip-itation of sonicated DNA with anti-Smad3 antibody(#28379 Chip Grade, Abcam). After immunoprecipitation,chromatin immunoprecipitation (ChIP)ed DNA was quan-titated by Q-PCR. Primers (50-ACAGGACTTCTGCAA-ATGCT-30 and 50-AACCAGAGATCTCAAGAGCA-30)were used to amplify a 98-bp fragment encompassing theGC-rich repeat of the mouse Bcl-2 P2 promoter (16).Primers (50-CGACTCGCCTCACTCCAGCTC-30 and50-GTCCGCTCACTCCCTCTGTC-30) were designed toamplify a 175-bp fragment corresponding to the region inthe mouse c-Myc gene that is homologous to the repressiveSmad-binding element (RSBE) in the human c-MYC pro-moter (17). Results are expressed as the fold enrichment ofChIPed DNA compared with input DNA for 3 biologicreplicates, with the Smad3-null cells serving as an internalnegative control for nonspecific pull-down.
Growth inhibition, apoptosis, cell migration, andinvasion assaysRelative rates of proliferation of the IMEC lines were
assessed by real-time imaging of cell culture confluency usingan IncuCyte imaging system (Essen Instruments). Phasecontrast images were taken every 3 hours over a period of30 hours, and the culture confluency at each time point wascalculated and plotted. Inhibition of cell proliferation byTGF-b was assessed by [3H]-thymidine incorporation, andapoptosis was quantitated using the Cell Death DetectionELISA kit (Roche). Cell migration and invasion assays werecarried out using the Transwell system without or withMatrigel, respectively (8 mm, BD Biosciences). Conditionsfor all assays were as previously described (7).
Tumorigenesis in transgenic and syngeneic transplantmodelsAll animal studies were conducted under institutionally
approved animal study protocols. Smad3þ/�mice (11) wereintercrossed with MMTV-PyMT transgenic mice (18) togenerate oncogenically initiated cohorts that were wild-typeor heterozygous for Smad3. The Smad3þ/�mice were on aninbred, mixed C57Bl/6/BlackSwiss/129Sv backgroundwhereas the MMTV-PyMT mice were pure FVB/NJ.Tumorigenesis and metastasis in this transgenic model were
assessed in the F1 offspring of the intercross to control forbackground strain effects. It should be noted that the overallincidence of lungmetastases was lower on this mixed geneticbackground (�40%) than is seen with the pure FVB/Nstrain (�90%; ref. 18). This feature of the mixed back-ground made it possible to see a stimulatory effect of theSmad3þ/� genotype on metastasis incidence. Mice wereeuthanized 40 days after initial detection of a palpablemammary tumor and tissues were harvested for molecularand pathologic analysis. Lung metastases in histologic sec-tions were enumerated by a pathologist. For the syngeneictransplant studies, 100,000 Mvt-1 cells (derived from ametastatic tumor in an MMTV-c-Myc/VEGF bitransgenicmouse and syngeneic to the FVB/N strain; ref. 13) weresurgically implanted in the #4 mammary gland of 6- to 8-week-old Smad3þ/þ or Smad3 þ/� mice that had beenbackcrossed for 10 generations onto the FVB/NJ back-ground. Tumor-bearing mice were euthanized after 7 weeks,and lungs were harvested for assessment ofmetastatic burdenas above.
Results and DiscussionAlthough Smad3 is rarely, if ever, deleted in human breast
cancer (10), we wished to assess whether more subtlealterations in Smad3 expression levels might occur. Wefound that Smad3 mRNA expression was reduced in 9 of10 cases of human breast cancer when compared withmatched adjacent normal samples, for an average reductionin Smad3 of approximately 60% in the tumor (Fig. 1A andB). In a meta-analysis of 1,412 human breast cancer caseswith Affymetrix U133A gene expression data in the GOBOdatabase (14), Smad3 mRNA levels were significantly lowerin the highest grade (grade 2) tumors than in lower gradetumors (Fig. 1C). Furthermore, Smad3 mRNA was alsosignificantly lower in the poorer prognosis ER-negativebreast cancers than in ER-positive tumors in the same dataset (Fig. 1D).We have previously shown that the majority ofSmad3 mRNA in the mouse mammary gland is associatedwith the epithelial compartment (19). To confirm thepresence of Smad protein in the epithelial compartment inhuman breast cancer, we immunostained a matched breastcancer and adjacent normal breast tissue from the samepatient. The results show that Smad3 is highly expressed inthe normal ductal epithelium and that lower expression isobserved in the parenchymal cells of a breast cancer from thesame patient (Fig. 1E). Together, these observations allsuggest that a reduction in Smad3 levels in the tumorparenchyma might contribute to breast cancer progression.Our mRNA data are also in agreement with a previousimmunohistochemical study which showed significant cor-relations between a decrease in Smad3 nuclear abundanceand high tumor grade and hormone receptor negativity inbreast cancer (20).To ask what effect a 2-fold reduction in Smad3 levels
might have on TGF-b responses in the mammary epithe-lium, we generated conditionally IMECs of the 3 Smad3genotypes, Smad3 þ/þ, Smad3þ/�, and Smad3�/� (7). This
Smad3 Dosage Effects in Breast Cancer
www.aacrjournals.org Mol Cancer Res; 10(10) October 2012 1391
genetic approach allowed us to precisely achieve a reductionin Smad3 levels (in the Smad3þ/� IMECs) that was com-parable with that seen in the human tumors. At the proteinlevel, Smad3 expression and TGF-b-induced Smad3 phos-phorylation were reduced in the Smad3þ/� cells andcompletely absent in the Smad3�/� cells, whereas Smad2expression and activation were essentially unchanged (Fig.2A). Quantitation of Smad3 protein in the Smad3þ/�
IMECs using the SimpleWestern technology showedthat Smad3 protein was reduced to 47% � 7% (n ¼ 3
independent determinations) of wild-type levels (Fig. 2B).Morphologically, Smad3þ/þ and Smad3þ/� cells showedsimilar, loosely organized colonies in culture (Fig. 2C),whereas Smad3�/� cells formed colonies with clearly definedboundaries, as seen previously (7). Proliferation rates werebroadly similar between the 3 cell lines (Fig. 2D).We next used these cells to examine the effect of Smad3
dosage on two tumor-suppressive effects of TGF-b (cell pro-liferation and apoptosis), and two pro-progression responses(migration and invasion).While all responses toTGF-bwere
●●●●●● ●●●●●●
●●●●
●●●
−2
−1
0
1
2
395 1225
ER- ER+
●●●
●●●●●
●
●●●●●●●
●
−2
−1
0
1
2
239 677 495
1 2 3
Adjacent
normal
Tumor0
10
20
30
40
50
25.5
9.7
Sm
ad
3/P
PIA
(A
U)
Adjacent
normal
Tumor0
10
20
30
40
50
Sm
ad
3/P
PIA
(A
U)
A B
C DP (ANOVA) <0.00001P (ANOVA) <0.00001
Log
2 S
mad3 e
xpre
ssio
n
Log
2 S
mad3 e
xpre
ssio
n
Tumor grade ER status
P = 0.04 P = 0.02
Stage
II/III
Stage
I
EAdjacent normal Tumor Mouse embryo lung
anti-S
mad3
anti-S
mad3
H&
E
No p
rim
ary
Ab
50μm
50μm
50μm
50μm
50μm
50μm
Figure 1. Smad3 expression inhuman breast cancer. A, Smad3mRNA levels in paired humanbreast cancers and adjacentnormal tissue, normalized tocyclophilin A (PPIA). Clinical stageis indicated. Statistics: paired ttest. B, box plot showing Smad3mRNA levels for grouped samplesinA. Statistics: Fisher exact test. C,box plots showing meta-analysisof Smad3 gene expression as afunction of tumor grade in 1,412human breast cancers usingGOBO database and analysistools. D, meta-analysis of Smad3gene expression as a function ofER status in 1,620 human breastcancers, as in C. E, Smad3immunostaining of an ER-negative, stage II human breastcancer and adjacent normal tissue.Mouse embryonic lung was thepositive control. Negative control,no primary antibody. Scale bar,50 mm. AU, arbitrary units; H&E,hematoxylin and eosin.
Kohn et al.
Mol Cancer Res; 10(10) October 2012 Molecular Cancer Research1392
essentially lost in Smad3 null cells, the intermediateSmad3þ/� genotype showed a complex pattern of responses.Smad3þ/� IMECs retained their growth-inhibitoryresponse to TGF-b (Fig. 3A) but completely lost theirpro-apoptotic response (Fig. 3B). The loss of the pro-apoptotic response in the Smad3þ/� cells was seen for 3independent cell isolates (Fig. 3C). To confirm that this losswas Smad3-dependent and not a result of compensatorychanges in the Smad3þ/� cells, we used lentiviral transduc-tion to add back Smad3 to wild-type levels (Fig. 3D). Addback of Smad3 restored the apoptotic response to TGF-b tothe same level as in the Smad3þ/þ cells (Fig. 3E).In contrast to this pattern of either full retention (growth
inhibition) or complete loss (apoptosis) of response to TGF-b in the Smad3þ/� cells, a different pattern was observed forthe pro-progression responses. Smad3þ/� cells showed areducedmigratory response to TGF-b (Fig. 3F), whereas theinvasion response in Smad3þ/� cells was at least as strong,and often stronger, than in Smad3þ/þ cells (Fig. 3G). Theunexpected enhancement of invasion in Smad3þ/� cells was
seen in 3 independent experiments at one or more of thetwo TGF-b concentrations tested (data not shown). Thus, a2-fold reduction in Smad3 dosage can alter the outcome ofTGF-b signaling in ways that differ dramatically dependingon the response considered. The pattern of change inbiologic responses to TGF-b at the different Smad3 levelsis summarized in Table 1.To address why the antiproliferative and pro-apoptotic
effects of TGF-b were differentially sensitive to loss of oneSmad3 allele, we examined the known Smad3 targets c-Myc(17) and Bcl-2 (16, 21). In agreement with the biologicresponse outcome data, TGF-b effectively suppressed c-Mycprotein levels in both Smad3þ/þ and Smad3þ/� cells(Fig. 4A). Expression of two other major cell-cycle–relatedtargets of TGF-b, the cyclin-dependent kinase inhibitorsp21(Waf1/Cip1) and p15INK4B, was unaffected (data notshown). In contrast to the expression pattern seen with c-Myc, Bcl-2 protein was downregulated by TGF-b only inSmad3þ/þ and not Smad3þ/� cells (Fig. 4B). As expected,downregulation of Bcl-2 in Smad3þ/þ cells was associated
A
p-Smad3
Smad3
p-Smad2
Smad2
β-Actin
TGF−β- + - + - +
S3+/+ S3+/- S3-/-
IMEC GENOTYPE
40 66 90 116
Smad3
α-Tubulin600
400
200
0
Smad3+/+
Smad3+/-
Smad3-/-
Molecular weight (KDa)
Sig
nal
B
C
Smad3+/+ Smad3+/- Smad3-/-
D
116
90
66
40
KDaSm
ad3+
/+
Smad
3+/-
Smad
3-/-
Smad3
α-Tubulin
1.0 0.4 0
Normalized
Smad3/tubulin
Time (h)
% C
onfluence
Figure2. Molecular andbiologic characterizationof IMECswithdifferentSmad3 levels. A,Westernblot analysis showing relative levels of total andC-terminallyphosphorylated Smad2 and Smad3 in IMECs of different Smad3 (S3) genotypes. B, electrophoretogram (left) and corresponding "virtual blot" image (right) ofSimpleWestern analyses of IMEC lysates simultaneously probed for Smad3 and a-tubulin as an internal loading control. Smad3/tubulin ratios normalized tothe Smad3 wild-type condition are shown below the virtual blot image. C, phase contrast images of IMEC morphology in monolayer culture. D, growthcharacteristics of the IMEC cultures, assessed by quantitative imaging of culture confluence in real-time using an IncuCyte imaging system. Results aremean� SEM (n ¼ 3).
Smad3 Dosage Effects in Breast Cancer
www.aacrjournals.org Mol Cancer Res; 10(10) October 2012 1393
with an increase in cleaved caspase-3, a marker of apoptosis(Fig. 4B). Thus, TGF-b–induced repression of c-Myc andBcl-2 shows patterns of Smad3 dependency that correlatewell with the observed biologic outcomes.We then asked whether TGF-b–induced binding of
Smad3 to these target gene loci was affected by Smad3 genedosage. TGF-b represses transcription of the human c-MYClocus via binding of Smad3 to an RSBE in the first exon (17).By ChIP analysis, we showed strong TGF-b–inducedSmad3 occupancy of the homologous region in the mousec-Myc enhancer in both Smad3þ/þ and Smad3þ/� IMEC
cells, consistent with the protein and biologic response data(Fig. 4C). Transcriptional repression of the Bcl-2 locusinvolves Smad3 binding to a repressive GC-rich region inthe first exon (16). In contrast to results at the c-Myc locus,we found that TGF-b treatment induced Smad3 binding tothe GC-rich region of the Bcl-2 enhancer only in Smad3þ/þ
but not in Smad3þ/� IMECs (Fig. 4D). In Smad3þ/þ cells,binding of Smad3 to the GC-rich region of the Bcl-2 locuswas approximately 4 times weaker than binding of Smad3 tothe RSBE in the c-Myc locus (compare fold enrichments),suggesting that Smad3 binds with lower affinity to the Bcl-2
Figure 3. Biologic responses to TGF-b in IMECs with different Smad3 levels. A and B, effect of Smad3 genotype on growth-inhibitory (A) or apoptotic (B)responses of IMECs to TGF-b. Results are mean � SEM for 3 determinations and are normalized to the untreated control for each genotype. �, P < 0.05,2-wayANOVA. C, the lack of an apoptotic response in Smad3þ/� cells was confirmed in 3 independent isolates. Results aremean�SEM for 3 determinationsand are presented without normalization to show similarity in basal apoptosis levels between genotype groups. D, conventional Western (left) andSimpleWestern "virtual blot" (right) showing quantitative restoration of Smad3 to wild-type levels in Smad3þ/� cells following lentiviral-mediated add back ofSmad3. CON is a control lentiviruswith stuffer sequence in place of Smad3. E, restoration of apoptotic response to TGF-bon addback of Smad3 toSmad3þ/�
IMECs. Results are expressed as in C. F andG, effect of Smad3 genotype onmigration (F) and invasion (G) responses to TGF-b. Results aremean�SEM for arepresentative experiment (n ¼ 10–20 fields per condition) normalized to the untreated Smad3þ/þ condition. �, P < 0.05, 2-way ANOVA with Bonferronicorrection. AU, arbitrary units.
Kohn et al.
Mol Cancer Res; 10(10) October 2012 Molecular Cancer Research1394
enhancer than the c-Myc enhancer. Thus, the reduction inSmad3 protein in the Smad3þ/� state leads to selective loss ofSmad3 from the low affinity repressive site in the Bcl-2enhancer but retention at the higher affinity repressive sitein the c-Myc enhancer. The net effect is loss of the pro-apoptotic response but retention of the growth-inhibitoryresponse in Smad3þ/� cells (schematized in Fig. 4E).The migration and invasion responses are molecularly
more complex than the growth-inhibitory and pro-apoptoticresponses, involving many different molecular steps andprocesses. To address possible mechanisms underlying theunexpectedly enhanced invasion response to TGF-b inSmad3þ/� cells, we first assessed MMP activity by gelatinzymography. While most MMPs were unaffected or unde-tectable, TGF-b induced MMP9 activity in Smad3 wild-type but not Smad3þ/� cells, a pattern that cannot explainthe enhanced invasion response (Supplementary Fig. S1). Insome cell types, TGF-b can induce a strong EMT, which isthought to be important for invasion responses (22). Like themajority of breast cancer cell lines (23), the IMECs do notshow a strong morphologic EMT in response to TGF-b (7).However, we tested whether TGF-bmight induce any of themolecular changes characteristic of a partial EMT andwhether the patterns of regulated gene expression in cellsof the three different Smad3 genotypes correlated with theobserved pattern of the invasion response. In a preliminaryscreen for effects of TGF-b on mRNA levels for Cdh1,Cdh2, Cdh3, Snail1, Snail2, Twist1, Zeb1 and Vimentin,only Cdh2 and Snail2 emerged as possible candidates withan appropriate pattern of response (Supplementary Fig.S2A). Snail2 did not validate in subsequent independentexperiments, but Cdh2 showed an expression pattern likethat of Bcl-2, in that it was suppressed in Smad3þ/þ but notSmad3þ/� cells (Supplementary Fig. S2B). Cdh2 has beenpreviously shown to promote invasion (24), so we hypoth-esize that the enhanced TGF-b–induced invasion seen inSmad3þ/� cells likely involves a combination of selective lossof TGF-b–mediated repression of invasion-promotingcomponents such as Cdh2, and/or selective loss of TGF-b–mediated induction of inhibitors of invasion. Identifica-tion of additional contributing components will best beaddressed through unbiased genome-wide approaches infuture studies.Our in vitro data suggested that reduced Smad3 levels
might tip the balance of TGF-b responses in favor of tumorprogression, with the apoptotic response lost and the inva-
sion response enhanced. To address the integrated effect ofthe reduction in Smad3 levels on the tumorigenic processin vivo, we crossed the MMTV-PyMT transgenic mousemodel of metastatic breast cancer (18) with Smad3þ/�mice(11). As expected, Smad3 was decreased by 50% to 70% atthe mRNA (Fig. 5A) and protein (Fig. 5B) levels in primarytumors fromMMTV-PyMT� Smad3þ/�mice. There wasno difference in primary tumorigenesis between oncogeni-cally initiated Smad3þ/þ and Smad3þ/� mice (Fig. 5C).However Smad3þ/� mice had a significantly higher inci-dence of metastases than wild-type mice (45% vs. 20%, P¼0.02; Fisher exact test; Fig. 5D). Thus, a mere 2-foldreduction in Smad3 levels can significantly enhance meta-static efficiency in vivo.The Smad3þ/�mouse used in the above studies was from
a germline knockout model, and thus had reduced Smad3levels in all tissues. To address whether the metastasis-promoting effect of reduced Smad3was due to the reductionof Smad3 in the tumor parenchyma or microenvironment,we used the Mvt-1 transplantable model of metastaticmurine breast cancer (13). To assess the microenvironmen-tal contribution, Mvt-1 were orthotopically transplantedinto syngeneic, immunocompetent Smad3þ/þ andSmad3þ/� host mice. Mvt-1 cells gave significantlyincreased numbers of metastases in the lungs of theSmad3þ/� compared with Smad3þ/þ mice (Fig. 5E). Toassess the tumor parenchymal contribution, we knockeddown Smad3 levels in Mvt-1 tumor cells with shRNA(Fig. 5F). Quantitation by Simon SimpleWestern showedthat Mvt-1-shSmad3 cells had 50% less Smad3 protein thanMvt-1-shGFP control cells, so knockdown was comparablewith that seen in the genetic model. We saw significantlyincreased metastasis in Mvt-1-shSmad3 cells compared withcontrol Mvt-1 cells following orthotopic transplantationinto Smad3þ/þ hosts (Fig. 5G). Together these resultsconfirm that a reduction in Smad3 can increase metastasisin an independent breast cancer model and suggest thatreduced Smad3 in both the tumor parenchyma and thetumor microenvironment can contribute independently toenhanced metastatic efficiency.Overall, our data suggest that a mere 2-fold reduction in
Smad3 levels can significantly impact on tumor progres-sion. In this study, we looked at just four of the manybiologic responses to TGF-b that are potentially relevantto tumor development. However, our results illustrate theimportant general principle that relatively small reductions
Table 1. Summary of effects of Smad3 gene dosage on biologic responses to TGF-b
Biologic activity Class of response Response to TGF-b in IMECS of different genotypes
in Smad3 levels can cause a spectrum of different effectson the pattern of TGF-b responses, with individual out-comes ranging from total loss of response (apoptosis), orreduction in response (migration), through no effect(growth inhibition), to enhancement of response (inva-sion). For the growth-inhibitory and apoptotic responses,we were able to show that the impact of Smad3 reductionon the biologic readout correlated with differences in
TGF-b–induced Smad3 occcupancy at enhancer elementsof key target genes for the responses. The data areconsistent with a mechanism in which different targetenhancers have different affinities for Smad3, such that asmall reduction in Smad3 levels causes a selective loss ofSmad3 from some enhancers (e.g., Bcl-2) but not fromothers (e.g., c-Myc). In a similar manner, activin/nodalsignaling through Smad2 was recently shown to cause
Figure 4. TGF-b–induced Smad3occupancy of c-Myc and Bcl-2enhancer elements at differentSmad3 levels. A and B, Westernblot analyses showing TGF-bregulation of c-Myc protein (A) andBcl-2 protein (B) in IMECs ofdifferent Smad3 genotypes.Cleaved (Cl.) caspase-3 reflectsapoptosis. b-Actin is the loadingcontrol. C and D, ChIP for Smad3binding at the indicated sites in thec-Myc locus (C) andBcl-2 locus (D)in response to TGF-b treatment ofIMECs of the 3 Smad3 genotypes.TGF-b–induced Smad3 bindingwas assessed by ChIP-qPCR, andresults are expressed as foldenrichment of ChIPed DNAcompared with input DNA.The Smad3�/� cells serve as anegative control. In the locusdiagrams, white indicatesuntranslated regions and grayindicates coding regions within thespecified exons. Arrows indicateapproximate positions of primerpairs used. Results are mean �SEM (n ¼ 3). �, P < 0.05, t test.E, schematic summary of Smad3occupancy of the enhancerelements and the transcriptionaland biologic response outcomesat different levels of Smad3.
Kohn et al.
Mol Cancer Res; 10(10) October 2012 Molecular Cancer Research1396
Figure 5. Effect of reduced Smad3 levels on tumorigenesis and metastasis. A and B, Smad3 mRNA (A) and protein (B) in tumors from MMTV-PyMTmice of different Smad3 genotypes. For mRNA, results are mean � SEM for 4 tumors per group. Protein levels were assessed by conventional Westernblotting, with b-actin as the loading control. C and D, primary tumorigenesis (C) and cumulative lung metastasis incidence (D) in MMTV-PyMT miceof Smad3 þ/þ (n ¼ 47) and Smad3þ/� (n ¼ 39) genotypes. Metastasis incidence was statistically different between the genotype groups at the endpoint(P¼0.02, Fisher exact test). E,metastasis ofSmad3þ/þMvt-1 tumor cells followingorthotopic implantation intoSmad3þ/þ (n¼19) orSmad3þ/� (n¼22) hosts.Results aremedian� interquartile range,Mann–WhitneyU test. F, conventionalWestern blot (left) and quantitative SimpleWestern virtual blot (right) of Smad3levels following shRNA knockdown in Mvt-1 cells. G, metastasis of Mvt-1 cells transduced with shSmad3 (test) or shGFP (control) lentiviruses and thenorthotopically implanted into Smad3þ/þ hosts (15 mice per group). Statistics are as in E. AU, arbitrary units.
Smad3 Dosage Effects in Breast Cancer
www.aacrjournals.org Mol Cancer Res; 10(10) October 2012 1397
transcriptional activation of different sets of target genesdepending on the quantitative level of Smad2 activation,thereby eliciting different biologic fates in the embryonicstem cell (21).The integrated result of the various, sometimes opposing,
effects of reduced Smad3 dosage on TGF-b responses in vivois likely to be highly dependent on the specific constellationof TGF-b responses that are engaged and dominant in agiven context. Here, we have shown that the net outcome ofSmad3 reduction in themammary epithelium is an enhance-ment of metastatic progression, at least in the contextof oncogenic activation of the Src (PyMT model) or Myc(Mvt-1 model) pathways. However, it remains to be seenhow general this phenomenon is. In the 7,12-dimethylbenz(a)anthracene (DMBA)/12-O-tetradecanoylphorbol-l3-ace-tate (TPA)-initiated skin carcinogenesis model, global dele-tion of one allele of Smad3 resulted in a significant decreasein primary tumor incidence, although effects on metastasiswere not reported (25, 26). Further studywill be necessary todetermine whether Smad3 agonists might be useful for theprevention or treatment of metastasis in specific cancersubpopulations.
Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.
Authors' ContributionsConception and design: E.A. Kohn, Z. Du, L.M. WakefieldDevelopment of methodology: E.A. Kohn, Z. Du, J.-Q. ChenAcquisition of data (provided animals, acquired and managed patients, providedfacilities, etc.): E.A. Kohn, Y.-a. Yang, Z. Du, Y. Nagano, C.M.H. Van Schyndle, M.A. Herrmann, M. Heldman, J.-Q. Chen, C.H. Stuelten, K.C. FlandersAnalysis and interpretation of data (e.g., statistical analysis, biostatistics, compu-tational analysis): E.A. Kohn, Y.-a. Yang, Z. Du, M.A. Herrmann, M. Heldman,J.-Q. Chen, L.M. WakefieldWriting, review, and/or revision of the manuscript: E.A. Kohn, Z. Du, C.H.Stuelten, L.M. WakefieldAdministrative, technical, or material support (i.e., reporting or organizing data,constructing databases): Y.-a. YangStudy supervision: L.M. Wakefield
AcknowledgmentsThe authors thank the expert assistance of Dr. Mario Anzano, Anthony Vieira, and
HaoDu in theAnimal Core; themembers of theCancer Biology ofTGF-b Section andDr. Kent Hunter for helpful discussions and critical reading of the manuscript; Drs.Akira Ooshima and Miriam Anver for pathology expertise; and Abigail Collett fortechnical assistance.
Grant SupportThis research was supported by the Intramural Research Program of the NIH,
National Cancer Institute, Center for Cancer Research Z01 BC 005785 to L.M.Wakefield.
The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be herebymarked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.
Received March 7, 2012; revised July 5, 2012; accepted July 26, 2012;published OnlineFirst August 9, 2012.
References1. Massague J. TGFb in cancer. Cell 2008;134:215–30.2. Barcellos-HoffMH,AkhurstRJ. Transforminggrowth factor-b in breast
cancer: too much, too late. Breast Cancer Res 2009;11:202.3. Tan AR, Alexe G, Reiss M. Transforming growth factor-b signaling:
emerging stem cell target in metastatic breast cancer? Breast CancerRes Treat 2009;115:453–95.
4. Pardali K, Moustakas A. Actions of TGF-b as tumor suppressor andpro-metastatic factor in human cancer. Biochim Biophys Acta2007;1775:21–62.
5. Schmierer B, Hill CS. TGFb-SMAD signal transduction: molecularspecificity and functional flexibility. Nat Rev Mol Cell Biol 2007;8:970–82.
6. Brown KA, Pietenpol JA, Moses HL. A tale of two proteins: differentialroles and regulation of Smad2 and Smad3 in TGF-b signaling. J CellBiochem 2007;101:9–33.
7. Kohn EA, DuZ, SatoM, VanSchyndleCM,WelshMA, YangYA, et al. Anovel approach for the generation of genetically modified mammaryepithelial cell cultures yields new insights into TGFb signaling in themammary gland. Breast Cancer Res 2010;12:R83.
8. Kretschmer A, Moepert K, Dames S, Sternberger M, Kaufmann J,Klippel A. Differential regulation of TGF-b signaling through Smad2,Smad3 and Smad4. Oncogene 2003;22:6748–63.
9. DunnNR,KoonceCH, AndersonDC, IslamA, Bikoff EK, Robertson EJ.Mice exclusively expressing the short isoform of Smad2 developnormally and are viable and fertile. Genes Dev 2005;19:152–63.
10. SjoblomT, Jones S,Wood LD, Parsons DW, Lin J, Barber TD, et al. Theconsensuscoding sequencesof humanbreast andcolorectal cancers.Science 2006;314:268–74.
11. Yang X, Letterio JJ, Lechleider RJ, Chen L, Hayman R, Gu H, et al.Targeted disruption of SMAD3 results in impaired mucosal immunityand diminished T cell responsiveness to TGF-b. EMBO J 1991;81:1280–91.
12. Jat PS, NobleMD, Ataliotis P, Tanaka Y, Yannoutsos N, Larsen L, et al.Direct derivation of conditionally immortal cell lines from an H-2Kb-tsA58 transgenicmouse. ProcNatl Acad Sci U SA 1991;88:5096–100.
13. Pei XF, Noble MS, Davoli MA, Rosfjord E, Tilli MT, Furth PA, et al.Explant-cell culture of primary mammary tumors from MMTV-c-Myctransgenic mice. In Vitro Cell Dev Biol Anim 2004;40:14–21.
14. Ringner M, Fredlund E, Hakkinen J, Borg A, Staaf J. GOBO: geneexpression-based outcome for breast cancer online. PLoS One2011;6:e17911.
15. Figueroa JD, Flanders KC, Garcia-Closas M, Anderson WF, Yang XR,Matsuno RK, et al. Expression of TGF-b signaling factors in invasivebreast cancers: relationships with age at diagnosis and tumor char-acteristics. Breast Cancer Res Treat 2010;121:727–35.
16. Yang YA, Zhang GM, Feigenbaum L, Zhang YE. Smad3 reducessusceptibility to hepatocarcinoma by sensitizing hepatocytes toapoptosis through downregulation of Bcl-2. Cancer Cell 2006;9:445–57.
17. Frederick JP, Liberati NT, Waddell DS, Shi Y, Wang XF. Transforminggrowth factor b-mediated transcriptional repression of c-myc isdependent on direct binding of Smad3 to a novel repressive Smadbinding element. Mol Cell Biol 2004;24:2546–59.
18. Guy CT, Cardiff RD, Muller WJ. Induction of mammary tumors byexpression of polyomavirus middle T oncogene: a transgenic mousemodel for metastatic disease. Mol Cell Biol 1992;12:954–61.
19. Yang YA, Tang B, Robinson G, Hennighausen L, Brodie SG, Deng CX,et al. Smad3 in themammary epitheliumhas a nonredundant role in theinduction of apoptosis, but not in the regulation of proliferation ordifferentiation by transforming growth factor-b. Cell Growth Differ2002;13;123–30.
20. Jeruss JS, SturgisCD,Rademaker AW,Woodruff TK.Down-regulationof activin, activin receptors, and Smads in high-grade breast cancer.Cancer Res 2003;63:3783–90.
21. LeeKL, LimSK, Orlov YL, Yit le Y, YangH, Ang LT, et al. GradedNodal/Activin signaling titrates conversion of quantitative phospho-Smad2levels into qualitative embryonic stem cell fate decisions. PLoS Genet2011;7:e1002130.
22. Xu J, Lamouille S, Derynck R. TGF-b-induced epithelial to mesenchy-mal transition. Cell Res 2009;19:156–72.
Kohn et al.
Mol Cancer Res; 10(10) October 2012 Molecular Cancer Research1398
23. Brown KA, Aakre ME, Gorska AE, Price JO, Eltom SE, Pietenpol JA,et al. Induction by transforming growth factor-b1 of epithelial tomesenchymal transition is a rare event in vitro. Breast Cancer Res2004;6:R215–31.
24. Cavallaro U. N-cadherin as an invasion promoter: a novel target forantitumor therapy? Curr Opin Investig Drugs 2004;5:1274–8.
25. Tannehill-Gregg SH, Kusewitt DF, Rosol TJ, Weinstein M. The roles ofSmad2 and Smad3 in the development of chemically induced skintumors in mice. Vet Pathol 2004;41:278–82.
26. Li AG, Lu SL, Zhang MX, Deng C, Wang XJ. Smad3 knockout miceexhibit a resistance to skin chemical carcinogenesis. Cancer Res2004;64:7836–45.
Smad3 Dosage Effects in Breast Cancer
www.aacrjournals.org Mol Cancer Res; 10(10) October 2012 1399
2012;10:1389-1399. Published OnlineFirst August 9, 2012.Mol Cancer Res Ethan A. Kohn, Yu-an Yang, Zhijun Du, et al. Implications for Tumor ProgressionComplex Dependency on Smad3 Gene Dosage with Important
in the Mammary Epithelium Show aβBiological Responses to TGF-
Updated version
10.1158/1541-7786.MCR-12-0136-Tdoi:
Access the most recent version of this article at:
