A Novel Functional Polymorphism in the Transforming Growth Factor-β₂ Gene Promoter and Tumor Progression in Breast Cancer

Transforming growth factor-β (TGF-β), a multifunctional growth factor (1, 2), plays a dual role in breast cancer development by acting as a tumor suppressor in early tumor formation and a promoter of tumor progression and invasion in advanced breast cancer (3). During early stages of breast cancer development, TGF-β inhibits cell cycle progression and tumor growth, but in later stages TGF-β promotes malignant progression (4, 5). TGF-β potently inhibits the proliferation of mammary epithelial cells and breast cancer cells (6, 7). Overexpression of TGF-β₁ in transgenic mice inhibits mammary tumor formation (8), supporting the tumor-suppressive role of TGF-β. In later stages of tumor progression, growth-inhibitory effects of TGF-β seem to be selectively lost, whereas tumor-promoting effects of TGF-β gain in importance. Enhanced TGF-β expression in advanced breast cancer may further stimulate tumor progression. Several studies have shown that elevated TGF-β expression levels are associated with invasiveness of breast cancer and disease progression (9–12). Furthermore, increased TGF-β expression correlates with the presence of lymph node metastases, suggesting that TGF-β promotes metastasis (13, 14). It has been postulated that the loss of growth-inhibitory action of TGF-β is caused by reduced expression of TGF-β receptors. More recent studies revealed that inhibition of TGF-β signaling suppresses metastasis of breast tumors, suggesting that intact TGF-β signaling through its receptors rather promotes tumor progression (15–20). Previously, we have shown that expression of TGF-β receptor type II is a negative prognostic factor in estrogen receptor–negative breast cancer, emphasizing the tumor-promoting role of TGF-β signaling (21).

Polymorphisms affecting the transcription of TGF-β genes and subsequently TGF-β expression levels might play a role in the development and progression of breast cancer. Little is known about polymorphisms in the TGFB2 gene. Thus far, one polymorphism in the 5′-untranslated region and one polymorphism in exon 1 have been described lacking an association with disease (22, 23). In the present study, we identified a novel polymorphism −246ins in the TGFB2 gene promoter whose functionality has thus far not been evaluated. Therefore, we analyzed the effect of the −246ins polymorphism on transcriptional activity and expression of the TGFB2 gene. We investigated whether the potentially functional −246ins polymorphism is associated with breast cancer progression and metastasis and examined TGF-β₂ protein expression in breast cancer tissues in relationship to TGFB2 genotypes.

Cell culture. The human breast cancer cell lines, MCF-7, MDA-MB-231, and MDA-MB-435, were cultured in DMEM containing 4.5 g of glucose/L supplemented with 50 μg/mL gentamicin and 10% FCS.

Plasmids constructs. Fragments of the TGFB2 promoter were amplified by PCR using genomic DNA from MCF-7 cells as a template. Three different 5′-specific primers and a common 3′-specific primer were used for PCR reaction. The primers were described by Noma et al. (24). Amplified DNA fragments were sequenced and inserted into the promoterless luciferase vector pGL3-Basic (Promega, Madison, WI). All TGFB2 promoter reporter constructs contained the same 3′-end of +64 relative to the transcription initiation site. The open reading frame of human Sp1 was amplified by PCR from a cDNA clone (IMAGp958H102105Q2⁷

Transfections and luciferase assay. One day before transfection, MCF-7, MDA-MB-231, and MDA-MB-435 cells were seeded into 24-well culture plates at a density of 50,000 per well. Cells were transfected using FuGENE 6 reagent (Roche, Indianapolis, IN) according to the protocol of the manufacturer. TGFB2 promoter luciferase constructs (0.14 pmol of each construct) were cotransfected with phRL-TK plasmid (10 ng; Promega) as an internal control for transfection efficiency. Total transfected DNA was kept constant by addition of unspecific DNA. Forty-eight hours after transfection, cells were harvested in passive lysis buffer (Promega) and luciferase activity was measured using the Dual Luciferase Reporter Assay System (Promega). For cotransfection experiments, varying amounts of pUB6Sp1 plasmid were cotransfected with 0.10 pmol reporter plasmid and 10 ng phRL-TK plasmid as an internal control for transfection efficiency. Total amount of transfected DNA was kept constant by addition of unspecific DNA. Transfections were done in triplicates and repeated at least thrice in independent experiments.

Nuclear protein preparation. Nuclear extracts from MCF-7, MDA-MB-231, and MDA-MB-435 cells were prepared by minor modifications of the protocol by Neurath et al. (25). Cells were harvested by scraping, washed in cold PBS, and centrifuged at 750 × g. The pellet was rinsed once in buffer A [10 mmol/L HEPES (pH 7.9), 1.5 mmol/L MgCl₂, 10 mmol/L KCl, 0.5 mmol/L phenylmethylsulfonyl fluoride, 1 mmol/L DTT] and then resuspended in buffer A containing 0.4% NP40. Samples were chilled on ice for 10 minutes and centrifuged at 750 × g. The pellet was resuspended in 5 volumes of buffer B [20 mmol/L HEPES (pH 7.9), 1.5 mmol/L MgCl₂, 0.42 mol/L NaCl, 0.2 mmol/L EDTA, 0.5 mmol/L phenylmethylsulfonyl fluoride, and 1 mmol/L DTT] and samples were stirred for at least 30 minutes at 4°C. After centrifugation at 7,500 × g, the supernatants were frozen and stored at −80°C.

Electrophoretic mobility shift assays. Oligonucleotides were annealed at 95°C for 10 minutes by mixing 10 μL each of two complementary oligonucleotide stocks (100 μmol/L) with 10 μL of 10 × Klenow buffer [500 mmol/L NaCl, 500 mmol/L Tris-Cl (pH 7.5), and 100 mmol/L MgCl₂] and 170 μL H₂O. The oligonucleotides were designed so that the annealed probes had 5′-overhangs that permitted radioactive labeling by Klenow fill-in reaction. For radioactive labeling, 2 μL of the annealed oligonucleotides, 5 μL of 10 × Klenow buffer, 25 μCi of [α³²P]dCTP, 0.5 μL of Klenow fragment (2.5 units), and H₂O to a final volume of 20 μL were incubated at 37°C for 60 minutes and purified through Sephadex columns (MicroSpin G-25 columns, Amersham Pharmacia Biotech, Braunschweig, Germany). Oligonucleotides were as follows: wild-type oligonucleotide, 5′-GTGATTGTCCTTCAACAGACCTGTC-3′; insertion oligonucleotide, 5′-GTGATTGTCCTTCCTTCAACAGACCTGTC-3′; and Sp1 consensus oligonucelotide, 5′-GATTCGATCGGGGCGGGGCGAGC-3′. Each gel shift reaction contained 50,000 cpm of ³²P-labeled probe and 5 μg nuclear protein, 10 mmol/L HEPES (pH 7.8), 60 mmol/L KCl, 0.2% NP40, 6% glycerol, 2 mmol/L DTT, and 2 μg poly(deoxyinosinic-deoxycytidylic acid) in a final volume of 20 μL. Samples were incubated on ice for 30 minutes after proteins had been added. For competition experiments, unlabeled double-stranded oligonucleotides were added to binding reaction as specific competitor. For the supershift assay, 6 μg Sp1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA) was incubated with nuclear extract for 30 minutes at room temperature before incubation with the radiolabeled probe. The same reaction was done using mouse IgG as a control. DNA/protein complexes were separated on a pre-electrophoresed 5% polyacrylamide gel in 0.5× TBE buffer at 200 V and 4°C. Gels were dried at 80°C for 2 hours. DNA/protein complexes were visualized with the BAS-1800 II phosphor-storage scanner (Fuji, Kanagawa, Japan) and AIDA software (Raytest, Straubenhardt, Germany).

Study subjects, DNA, and tissue samples. TGFB2 −246ins genotypes were determined from 78 patients with a confirmed histologic diagnosis of breast cancer established between 1996 and 1998 at the University Hospital Hamburg, Germany. Anonymous DNA samples were available from all patients for the purpose of frequency testing of genetic polymorphisms. Baseline information on tumor size, nodal status, and distant metastasis was obtained by tumor-node-metastasis classification following WHO guidelines (26). To elucidate allele and genotype frequencies of the TGFB2 −246ins polymorphism in a healthy control group, 303 randomly selected blood donors (143 female and 160 male) at the University Hospital, Tuebingen, Germany, were genotyped. Only subjects of White origin were included to avoid confounding due to mixed ethnicity. Additionally, to correct for a possible bias due to age, we did Mantel-Haenszel test after subjects were arranged in eight age groups, each spanning 10 years. The study protocol was approved by the ethical committee of the University Hospital Tuebingen, Germany, and all blood donors gave written informed consent.

Finally, both TGFB2 genotype and TGF-β₂ expression in primary breast tumor tissues were investigated in an additional population of 78 breast cancer patients, of which anonymous formalin-fixed and paraffin-embedded tissues were available from the pathology archives of the Robert Bosch Hospital (Stuttgart, Germany). Following deparaffinization by a xylene extraction step, DNA was isolated from 10-μm tissue sections using QIAamp DNA Mini kit (Qiagen, Hilden, Germany) according to the instructions of the manufacturer (Tissue Protocol).

Immunohistochemical methods. Immunohistochemistry was essentially done as described earlier using the EnVision+ System kit (DakoCytomation, Glostrup, Denmark; ref. 27). A monoclonal antibody against TGF-β₂ (R&D Systems, Wiesbaden, Germany) was used as primary antibody in a dilution 1:25. The staining intensity was determined differentiating between no staining (score 0), weak staining (score 1), and strong staining (score 2). For statistical analysis, staining intensities of score 1 and 2 were defined as showing TGF-β₂ expression, whereas absent staining was defined as showing no TGF-β₂ expression. Stained tissue sections were evaluated by an expert pathologist and a scientist without knowledge of other sample characteristics.

Genotyping. Genotyping was done by fluorescent-labeled PCR product sizing and 5′-nuclease assay (TaqMan). For fragment analysis, the region containing the TGFB2 polymorphism was amplified with flanking oligonucleotides bearing fluorescent labels designed for GeneScan analysis. The following primers were used for amplification: 5′-GAAGCCTTCCCTTCTAGAGCA-3′ (forward) and 5′-CGCCCTGACAA-CAGTGATTTA-3′ (reverse). The forward primer was labeled with 6-carboxyfluorescein (6-FAM). PCR amplification was done in a final reaction volume of 50 μL containing 100 ng genomic DNA, 200 nmol/L forward and reverse primer, 200 μmol/L of each deoxynucleotide triphosphate (dATP, dCTP, dGTP, and dDTT), 1.5 mmol/L MgCl₂, and 25 units Taq polymerase (Invitrogen, Karlsruhe, Germany). Thermal cycling conditions were as follows: 95°C for 2 minutes followed by 35 cycles of 95°C for 30 seconds, 54°C for 1 minute, and 72°C for 1 minute with a final extension step of 72°C for 10 minutes. Amplification of the wild-type allele leads to a 146-bp fragment, whereas amplification of the mutated allele leads to a 150-bp fragment. After confirmation of successful amplification by agarose gel electrophoresis, PCR products were separated by fluorescence capillary electrophoresis (ABI PRISM 310 Genetic Analyzer). One to 2 μL of the amplified and purified PCR products were added to 12 μL Hi-Di Formamide and 0.5 μL GeneScan-500 ROX size standard (Applied Biosystems, Foster City, CA). Denatured PCR fragments were separated on the genetic analyzer. Allele fragment size was determined using the internal size standard GeneScan-500 ROX and collected data were analyzed with GeneScan Analysis software.

For analysis of a larger number of samples, a 5′-nuclease assay for allelic discrimination (TaqMan) was established. Primer and probe sets were designed using PrimerExpress program (Applied Biosystems). All oligonucleotide sequences were checked using the BLAST search program. Amplifications were done with forward 5′-GGAGGCTGTGACTGAGCTACACTA-3′ and reverse 5′-CGGTATTTACTTTCCAGTC-ATTTTGG-3′ primers. Specific probes for each allele were labeled with the fluorescence reporter dyes VIC and FAM at their 5′-end; wild-type probe: 5′-VIC-TGATTGTCC-TTCAACAGA-MGB-3′ and probe with insertion sequence: 5′-6-FAM-ATTGTCCTTCCTT-CAACAG-MGB-3′. Each PCR reaction mixture (25 μL) contained 50 ng of DNA, 900 nmol/L of each forward and reverse primer, 300 nmol/L of each allele specific probe, and 12.5 μL of Taqman Universal PCR Master Mix (Applied Biosystems). Amplification was carried out under the following conditions: 50°C for 2 minutes and 95°C for 10 minutes, followed by 40 cycles of 92°C for 15 seconds, and 60°C for 1 minute. Fluorescence in each sample well was measured after PCR using the ABI Prism 7000 Sequence Detection System (Applied Biosystems). Data were analyzed with SDS2.0 software (Applied Biosystems).

Statistical analysis. Fisher's exact test and Mantel-Haenszel test were used to evaluate differences in genotype frequencies between breast cancer patients and controls. The association between TGFB2 promoter polymorphism and clinical variables, such as tumor size, nodal status, metastases, and hormonal status, were examined using Fisher's exact test and, if reasonable, χ² test for trend. Because of the low frequency of TGFB2 −246ins homozygotes (Table 1), minor allele heterozygous and homozygous individuals were grouped together for statistical analyses. For multivariate analysis, multiple logistic regression was used. P < 0.05 (two-sided) was considered as statistically significant. Data were analyzed using SPSS (version 10; SPSS, Inc., Chicago, IL) software, statistical package R (version 2.2.0),⁸

A novel polymorphism in the TGFB2 promoter. Cloning and sequence analysis of the TGFB2 gene promoter region from genomic MCF-7 mammary carcinoma cell DNA revealed a new sequence variation of the originally published nucleotide sequence (24). A 4-bp insertion of CTTC at position −246 relative to the transcription start site (−246ins) was identified. Further analysis showed that the −246ins was also present in MDA-MB-435, MDA-MB-468, and SK-BR-3 breast cancer cell lines, indicating that the −246ins could represent a common variation. Therefore, we decided to investigate the functional role of the insertion.

Effects of the −246ins polymorphism on the TGFB2 promoter activity in human breast cancer cell lines. To examine the potential effects of the −246ins polymorphism on TGFB2 gene transcription, three different fragments of the TGFB2 promoter (−1729/+64, −778/+64, and −531/+64) carrying either the −246ins sequence or the wild-type sequence were inserted upstream of the luciferase gene in the pGL3-Basic vector. These constructs were transiently transfected into MCF-7, MDA-MB-231, and MDA-MB-435 breast cancer cells. In all cell lines, reporter constructs containing the −246ins sequence showed significantly higher luciferase activities than reporter constructs containing the wild-type sequence, indicating that the −246ins polymorphism enhances transcriptional activity of the TGFB2 promoter (Fig. 1). The results suggest that the −246ins acts as a general enhancer of TGFB2 promoter activity. The −246ins influences transcriptional activity of the TGFB2 promoter probably by facilitating DNA binding of a transcription factor with enhancing potential.

Nuclear factor Sp1 binds to TGFB2−246ins allele. To identify transcription factors that might bind to the −246ins allele and thus lead to enhanced transcriptional activity of the allele, we did electrophoretic mobility shift assays (EMSA). When nuclear extracts from MCF-7 cells were incubated with a ³²P-labeled oligonucleotide containing the wild-type sequence of the TGFB2 promoter, one protein complex was identified binding specifically to the wild-type sequence (Fig. 2A,, lane 2). Binding was inhibited by competition with increasing amounts of unlabeled wild-type probe (Fig. 2A,, lane 3-5). When nuclear extracts from MCF-7 cells were incubated with the −246ins oligonucleotide, two distinct DNA/protein complexes were observed (Fig. 2B,, lane 2): one complex with a mobility similar to the one of the complex formed with the wild-type oligonucleotide and one complex with lower mobility. Formation of both DNA/protein complexes was inhibited by competition with increasing amounts of unlabeled probe, indicating that the two complexes represent sequence-specific DNA/protein interactions (Fig. 2B,, lanes 3-5). Similar results were obtained with nuclear extracts from MDA-MB-231 and MDA-MB-435 breast cancer cells (Fig. 2C).

Analysis of the TGFB2 promoter region using a computer algorithm (28) revealed a putative Sp1 binding motif in the −246ins sequence. EMSAs were done to characterize the potential Sp1 binding site. MCF-7 nuclear extracts were incubated with ³²P-labeled oligonucleotide containing the −246ins sequence. Two DNA/protein complexes were observed as described above. Competition with increasing amounts of an unlabeled Sp1 consensus oligonucleotide inhibited formation of the upper protein complex and formation of the lower protein complex to a lesser extent (Fig. 3A,, lanes 3-5). These results suggest that GC-box binding proteins are involved in the formation of both upper and lower DNA/protein complex. To further analyze if the complexes formed with the −246ins oligonucleotide and MCF-7 nuclear extract contain Sp1 protein, supershift assays were carried out. The addition of a Sp1-specific antibody inhibited the formation of the upper DNA/protein complex (Fig. 3B,, lane 2), whereas nonspecific IgG had no effect on formation of both DNA/protein complexes (lane 1), indicating that the upper complex contains Sp1 protein. When the Sp1 consensus oligonucleotide was incubated with MCF-7 nuclear extract, two protein complexes were formed showing the same mobility as the complexes formed with the −246ins oligonucleotide (Fig. 3B , lane 4). Thus, we propose that both the upper and the lower protein complex contain GC-box binding proteins. However, the lower DNA/protein complex might contain another GC-box binding protein different from Sp1, which was not further characterized in our experiments. Taken together, our results indicate that the transcription factor Sp1 specifically binds to the −246ins allele and might function as a transcriptional activator of the mutated TGFB2 allele.

Overexpression of Sp1 increases transcriptional activity of the TGFB2 −246ins allele. To investigate the role of Sp1 in transactivation of the TGFB2 promoter, reporter constructs containing either the TGFB2 promoter wild-type sequence (p-531) or the −246ins sequence (p-531-246ins) were cotransfected with a Sp1 expression vector or an empty control vector into MCF-7 cells. Sp1 (200 ng) stimulated luciferase expression of the −246ins reporter construct 2-fold (Fig. 3C), whereas cotransfection of Sp1 did not activate luciferase expression of the wild-type reporter construct, suggesting that −246ins is the target sequence for Sp1. Overexpression of Sp1 increased TGFB2 promoter activity of the −246ins allele in a dose-dependent manner, confirming that Sp1 contributes to the promoter activity of this allele. Altogether, these results show that the transcription factor Sp1 binds to the −246ins sequence of the TGFB2 promoter and mediates transcriptional activation of the −246ins allele.

TGFB2 −246ins allele and the risk of breast cancer. DNA from healthy randomly selected German blood donors (n = 303, 160 male and 143 female) was genotyped to address the frequency of the −246ins allele in a Caucasian population. The −246ins allele was found in 16% of the German Caucasian population, 28.1% were heterozygous and 1.6% were homozygous for TGFB2 −246ins. Our data show that the −246ins allele is a relatively common variant among Caucasians. The TGFB2 −246ins allele might have an influence on the risk of developing breast cancer. Therefore, we genotyped DNA from breast cancer patients (n = 78) and compared genotype frequencies between breast cancer patients and female controls taken from the group of Caucasian blood donors (n = 143). No significant differences were found between breast cancer patients and controls (P = 0.339, Fisher's exact test; Table 1). This was also true following age adjustment (P = 0.434, Mantel-Haenszel test). In all study sets, distribution of genotype frequencies was in Hardy-Weinberg equilibrium.

Association of TGFB2 −246ins allele with breast cancer progression. To investigate whether the TGFB2 −246ins polymorphism might have an influence on breast cancer progression, we tested for an association between the TGFB2 −246ins genotype and prognostic factors, like tumor size, nodal status, metastases, estrogen receptor, and progesterone receptor status. The presence of the −246ins allele was significantly associated with lymph node metastasis (P = 0.003; Table 2). Patients having at least one TGFB2 −246ins allele had a 4.7-fold increased risk for lymph node metastases compared with patients having only the wild-type allele (P = 0.013 for pN > 0 versus pN = 0; odds ratio, 4.7; 95% confidence interval, 1.35-16.34). Multivariate logistic regression analysis revealed that the presence of the −246ins allele was a significant predictor for lymph node metastasis (P = 0.0118, odds ratio, 5.18; 95% confidence interval, 1.44-18.62) being independent of estrogen and progesterone receptor status. We also found a significant association between the presence of the −246ins allele and tumor size (P = 0.044). Furthermore, the −246ins allele was found more frequently in patients with advanced tumor (TNM) stage (P = 0.009). No correlation was observed between the −246ins allele and estrogen and progesterone receptor status. Taken together, the results suggest that the TGFB2 −246ins polymorphism is associated with more aggressive tumor behavior and contributes to the metastatic and invasive properties of the tumor.

TGFB2 −246ins allele is associated with enhanced TGF-β₂ expression in breast cancer tissue. Next, we wanted to find out if the −246ins allele was also associated with an increased TGF-β₂ expression. Therefore, we evaluated TGF-β₂ protein expression in tumor tissues by immunohistochemistry in another independent set of 78 breast cancer patients. We found that tumors that were heterozygous or homozygous for −246ins showed a stronger staining intensity than tumors with the wild-type sequence. Figure 4 shows a representative immunohistochemical staining for TGF-β₂. The proportion of tumors showing expression for TGF-β₂ was higher in patients with the −246ins allele. Staining for TGF-β₂ was positive in 79% of the tumors with at least one −246ins allele, whereas only 32% of the tumors with wild-type genotype showed positive staining for TGF-β₂ (15 of 19 versus 19 of 59, respectively, P = 0.0005). It is of note that we compared genotypes obtained from tumor tissues with those obtained from normal tissues in 22% of cases (n = 17) and observed concordance in 100%.

In the present study, we show that an insertion polymorphism in the TGFB2 promoter located at −246 bp upstream of the transcriptional start site enhances the transcription of the gene and is likely linked to lymph node metastasis in breast cancer patients. Our experiments show that the −246ins polymorphism is indeed functional. Transient transfection experiments revealed that the −246ins polymorphism increased TGFB2 promoter activity in different breast cancer cell lines. Furthermore, binding of the transcription factor Sp1 to the TGFB2 −246ins allele led to enhanced transcriptional activity of the promoter. Tumor sections from patients with the −246ins polymorphism displayed increased levels of TGF-β₂ protein, providing evidence for the functional relevance of this polymorphism in vivo.

Considering the role of TGF-β in malignant transformation, the −246ins allele might influence the risk of developing breast cancer. There was no significant difference in allele and genotype distribution between breast cancer patients and controls. However, our results show a significant association between the TGFB2 −246ins polymorphism and the presence of lymph node metastases and advanced tumor stage. The risk of lymph node metastases was 4.7-fold higher for carriers of the TGFB2 −246ins allele, suggesting that the polymorphism is associated with enhanced tumor growth and invasiveness. No association with the occurrence of distant metastases was observed, possibly due to the low numbers of patients with distant metastases reported in this collection. Our observation of an association of the −246ins allele with lymph node metastasis might be explained by enhanced transcriptional activity of the −246ins allele, which could stimulate TGF-β₂ expression in breast cancer tissue, thereby enhancing tumor progression. We therefore analyzed TGF-β₂ expression in breast cancer tissue. The proportion of tumors showing expression for TGF-β₂ was significantly higher in patients with the −246ins allele. These data support the hypothesis that increased TGF-β₂ expression level in tumors with the −246ins allele may promote metastasis in breast cancer development. Several studies have shown that enhanced expression of TGF-β is correlated with invasion and metastasis of breast cancer. TGF-β₁ expression is higher in invasive breast carcinoma than in ductal carcinoma in situ (9). Another study reported elevated levels of TGF-β₁ expression in lymph node metastases compared with primary tumor tissues (13). Although most studies analyzed TGF-β₁ expression, TGF-β₂ may also favor invasiveness of breast cancer because both isoforms share similar biological activities. This hypothesis is supported by in vitro data showing increased invasiveness of breast cancer cells treated with TGF-β₁ or TGF-β₂ (29). Recently, TGFB2 was identified as one of the genes differentially expressed in highly metastatic tumors in a mouse model of spontaneous breast cancer metastasis to bone (30). We therefore propose a possible role for TGF-β₂ in breast cancer metastasis.

Enhanced TGF-β production may contribute to the development of metastases in several different ways. Both isoforms TGF-β₁ and TGF-β₂ stimulate the metastatic potential of mammary tumor cells presumably by controlling their ability to break down and penetrate basement membrane barriers (29). TGF-β can induce epithelial-to-mesenchymal transition in tumor cells leading to enhanced invasiveness (31–33) and stimulate angiogenesis. Tumor-derived TGF-β may also act as an immune suppressor and allow tumor cells to escape from immune surveillance (4, 34, 35). Several studies have shown that TGF-β promotes tumor progression by inhibiting antitumor immune responses (35–38). TGF-β₂-mediated immune suppression has been observed in glioblastoma that frequently overexpresses TGF-β₂ (39), suggesting that TGF-β₂ is a highly potent immunosuppressor (39, 40). Tumor secreted TGF-β₂ may inhibit the function of immune effector cells at the sites of tumor implantation. We therefore suppose that the immunosuppressive role of TGF-β₂ is particularly important for evasion of tumor cells from immune surveillance in the lymph nodes and might provide a mechanism by which tumor cells escape from elimination by tumor-specific lymphocytes. Enhanced TGF-β₂ expression might consequently lead to increased tumor aggressiveness by TGF-β-mediated enhanced immunosuppression.

Our analyses suggest that the transcription factor Sp1 might play a critical role in promoter activity of the TGFB2 −246ins allele. The higher transcriptional activity of the −246ins allele can be explained by binding of the transcriptional activator Sp1. Sp1 belongs to a family of ubiquitously expressed zinc-finger transcription factors that regulate the expression of a wide variety of mammalian genes (41). Each member of the family carries a conserved COOH-terminal domain with three Cys₂-His₂ zinc fingers required for sequence-specific DNA binding to GC-rich promoter elements. Several members of the TGFB gene family are regulated by the transcription factors Sp1 and Sp3. The promoter regions of the TGFB1, TGFB3, TGFBR1, and TGFBR2 genes contain multiple Sp1 binding sites, which regulate their activity (42–45). In contrast, the TGFB2 promoter lacks consensus Sp1 binding sites (24). We show that a protein complex binding to the TGFB2 −246ins allele contains the transcription factor Sp1, pointing out that the insertion sequence serves as a Sp1 binding site. The activity of the TGFB2 promoter is unaffected by Sp1 coexpression (43), which corresponds to our results. We found that the activity of the wild-type promoter construct is not influenced by Sp1 overexpression, but Sp1 overexpression induced the promoter activity of the −246ins allele, suggesting that Sp1 is responsible for the elevated transcriptional activity of the −246ins allele.

The binding sequence does not match the typical consensus sequence of a GC-box (GGCGGG) but uses an alternative binding element containing (CTTCCTTC) as a core sequence. Despite the lack of similarity to the consensus sequence of a GC-box, the transcription factor Sp1 seems to recognize this motif and binds to it. A similar binding motif CCTCCT has been found in reverse orientation in the interleukin-10 promoter and has been proposed to be a “second consensus sequence” (46). Another study also identified a similar Sp1 binding motif in the epidermal growth factor receptor promoter consisting of two direct CT-rich repeats (TCCTCCTCC; ref. 47). We now hypothesize that different CT-rich regions can serve as Sp1 binding elements. More than one repeat of the CTTC sequence might be necessary for Sp1 binding activity in the TGFB2 promoter as the wild-type sequence containing only one repeat was not sufficient for serving as a Sp1 target sequence.

In summary, we have shown that the TGFB2 promoter comprises a novel functional polymorphism that affects transcriptional activity of the gene. Our study provides the first clinical evidence that a polymorphism that leads to enhanced TGF-β₂ expression might influence the progression of breast cancer toward lymph node metastasis without having an effect on breast cancer susceptibility. We propose that detection of the −246ins allele might be a useful prognostic marker for lymph node metastasis in breast cancer patients. Larger clinical studies are necessary to corroborate this hypothesis.

Grant support: Deutsche Forschungsgemeinschaft grant KN-228/2-1/2 and the Robert Bosch Foundation.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.