Telomerase stabilizes chromosomes by maintaining telomere length, immortalizes mammalian cells, and is expressed in more than 90% of human tumors. However, the expression of human telomerase reverse transcriptase (hTERT) is not restricted to tumor cells. We have previously shown that a subpopulation of human mammary epithelial cells (HMEC) in tumor-adjacent, histologically normal (TAHN) breast tissues expresses hTERT mRNA at levels comparable with levels in breast tumors. In the current study, we first validated a reporter for measuring levels of hTERT promoter activity in early-passage HMECs and then used this reporter to compare hTERT promoter activity in HMECs derived from tumor and paired TAHN tissues 1, 3, and 5 cm from the tumor (TAHN-1, TAHN-3, and TAHN-5, respectively). Cell sorting, quantitative real-time PCR, and microarray analyses showed that the 10% of HMECs with the highest hTERT promoter activity in both tumor and TAHN-1 tissues contain more than 95% of hTERT mRNA and overexpress many genes involved in cell cycle and mitosis. The percentage of HMECs within this subpopulation showing high hTERT promoter activity was significantly reduced or absent in TAHN-3 and TAHN-5 tissues. We conclude that the field of “normal tissue” proximal to the breast tumors contains a population of HMECs similar in hTERT expression levels and in gene expression to the HMECs within the tumor mass and that this population is significantly reduced in tissues more distal to the tumor. Mol Cancer Res; 9(9); 1209–21. ©2011 AACR.

This article is featured in Highlights of This Issue, p. 1163

The presence of histologically abnormal tissue proximal to multifocal cancerous lesions was first described in 1953 by Slaughter and colleagues. The process leading to this abnormal tissue was termed “field cancerization” (1). Since this original work, investigations of field cancerization have evolved to include molecular analyses of seemingly normal tissue proximal to the tumors of the lung, skin, prostate, and breast (reviewed in ref. 2). Our own studies of breast field cancerization have showed that many molecular characteristics of the tumor tissue are also present in paired histologically normal tissues 1 cm from the tumors' margins (TAHN-1) but are absent in paired histologically normal tissues obtained 5 cm from the tumors' margins (TAHN-5) or from reduction mammoplasty (RM; ref. 3). First, telomere lengths in tumor or paired TAHN-1 tissues are significantly shorter than in histologically normal tissues from paired TAHN-5 tissues or RM (3). Telomeres are the protein nucleic acid structures that stabilize chromosome ends, thereby preventing chromosome fusion, amplification, loss, and structural rearrangements. Accordingly, the frequencies of allelic imbalance in tumor and paired TAHN-1 tissues are 15 and 5 times higher, respectively, than in either paired TAHN-5 or RM tissues. This is consistent with the interpretation that cells within TAHN-1 tissue have a telomere-dependent mechanism of generating genomic instability, an early event in breast tumorigenesis (4–6). In addition, histologically normal tissue surrounding tumors exhibits a gene expression signature consistent with a wound healing microenvironment (7, 8), which also supports tumor initiation and progression (reviewed in refs. 9, 10).

We previously reported that telomerase, the reverse transcriptase expressed in more than 90% of human tumors, is also expressed in a subpopulation of tumor adjacent human mammary epithelial cells (HMEC) at levels comparable with those in most breast tumors (11). Expression of telomerase confers extended replicative capacity through elongation of critically short telomeres (12). Combined with knockdown of p53, ectopic telomerase expression immortalizes primary human ovarian surface epithelial cells (13), and when further combined with SV40 large T oncoprotein and an oncogenic allele of Ras (H-rasV12), results in tumorigenic transformation of HMECs (14,15). Smith and colleagues also have shown that ectopic expression of telomerase in early-passage HMECs modulates the expression of growth-controlling genes and enhances cell proliferation (16). A provocative implication of these previous studies is that a significant number of cells within TAHN-1 tissues potentially possess limitless replicative capacity and a telomere-dependent mechanism of generating chromosomal rearrangements, overexpress genes involved in proliferation, and are exposed to a tumorigenic microenvironment. Our prior investigation showing hTERT expression in TAHN tissues did not determine the locations and distribution of the HMECs-expressing hTERT, that is, if they are distributed uniformly throughout the breast or, like shortened telomeres and allelic imbalance, restricted to sites proximal to the tumor. Likewise, it is not known how endogenous (vs. ectopic) telomerase expression is related to the expression of growth-controlling genes and enhances cell proliferation in HMECs. To investigate these questions, we developed an hTERT promoter–EGFP reporter vector to differentiate and separate primary HMECs based on their levels of hTERT promoter activity. In this study, we first report experiments to validate the efficacy of the system using a prototype adenovirus containing only the EGFP gene construct under the control of the hTERT promoter (Green vector, Supplementary Data). We then report experiments using the StopLight vector, which consists of the Green vector with an additional constitutively activated DsRed gene to control for differences in infection. To evaluate the efficacy of the viral construct and cell-sorting strategy, we show data on tumor HMECs infected with the StopLight vector. These data showed that this vector and cell-sorting strategy were effective in identifying and isolating a subpopulation of HMECs expressing high levels of telomerase. Last, we used the StopLight vector and cell-sorting strategy to analyze this subpopulation of hTERT-expressing cells in TAHN-1, TAHN-3, and TAHN-5 HMECs. Using this strategy, we show that the subpopulation of hTERT-expressing cells (a) exists in TAHN-1 tissue, (b) is greatly reduced or absent in paired TAHN-3 and TAHN-5 tissues, (c) differentially expresses several genes that function in cell division, particularly mitosis, and (d) is more similar to tumor than either TAHN-3 or TAHN-5 in all parameters measured.

Tissue acquisition

Fresh breast tissue was provided by the University of New Mexico Human Tissue Repository and was derived from surgical remnant tissues of patients undergoing mastectomy at the University of New Mexico Hospital. Samples (100–200 mg) were excised from tumor and tumor adjacent tissue at defined distances (1, 3, and 5 cm) from the visible tumor margin. Tissues were stored in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 200 U/mL penicillin and 200 μg/mL streptomycin until processing (typically within 1–2 hours of surgery). For each sample, a portion was used for histologic analysis by sectioning and hematoxylin and eosin staining. The remaining tissue was used to establish short-term cell cultures. Tissue was acquired from 8 patients. A full set of samples (tumor, TAHN-1, TAHN-3, and TAHN-5) was available for 5 patients, whereas a partial set (TAHN-1, TAHN-3, and TAHN-5) was available for 3 patients. All tissues used in this study were anonymous, and experiments were carried out in accordance with all federal guidelines as approved by the University of New Mexico Health Science Center Human Research Review Committee.

Cell cultures

HEK, MDA-MB-231, and WI-38 cell lines were obtained from American Type Culture Collection and cultured as recommended by the providers. WI-38 hTERT cells were created by infecting WI-38 cells with a pBABE-puro retrovirus containing the hTERT-coding sequence and were selected by puromycin resistance. Short-term, early-passage primary cultures of mammary cells were derived from “organoid” preparations of breast tissues, as previously described (17). Briefly, tissue samples were minced and enzymatically dissociated using 0.1% w/v collagenase I in Mammary epithelial growth medium (MEGM; Lonza) at 37°C for 12 to 18 hours. Small tissue fragments (organoids) remaining after digestion were collected by centrifugation at 100 × g for 2 minutes. These organoids were seeded directly into MEGM. Differential trypsinization and differential centrifugation were carried out, as previously described for maintenance of the epithelial population (17).

Cloning of the hTERT–EGFP reporter construct

A 378-bp core promoter region of hTERT (18) was amplified by PCR and cloned upstream of the EGFP-coding sequence, as previously described (19). DNA sequencing verified that the sequence of the construct matched the hTERT promoter–EGFP recombinant DNA sequence exactly as designed.

Adenoviral vectors

Green vector” was constructed by cloning the hTERT promoter–EGFP cassette into the multiple cloning site of the AdenoX Ad5 adenoviral vector (Clontech). “Red vector” is an Ad5 adenoviral vector that contains the DsRed monomer reporter gene under transcriptional control of the constitutive cytomegalovirus (CMV) promoter. “StopLight vector” was constructed by cloning the hTERT promoter–EGFP cassette into an Ad5 adenoviral vector that contains the DsRed monomer reporter gene under transcriptional control of the constitutive CMV promoter (Vector Biolabs). Adenoviral DNA was transfected into HEK293 (the packaging cell line), resulting in the formation of infectious adenovirus particles.

Adenoviral infection and cell sorting

Replicates of cultured cells were infected in parallel with either the Green vector or StopLight vector (multiplicity of infection of 100). At 48 hours postinfection, the cells were analyzed/sorted on a MoFlo high speed cell sorter (DakoCytomation). EGFP was excited with a 488-nm laser and DsRed was excited with a 532-nm laser. Fluorescent gain was set so that uninfected control samples produced a histogram peak within the first log decade of fluorescence intensity. A gate was established in the red fluorescence channel that excluded 99% of uninfected HMECs (1% gate; Fig. 3A). This gate was used subsequently to exclude uninfected cells from analysis. Data were analyzed using Summit software, version 4.3.

Quantitative real-time PCR

RNA was extracted from sorted cells using a DNA/RNA/protein isolation kit (Qiagen) according to manufacturers' instructions. cDNA was synthesized using the Retroscript RT Kit (Ambion), in a 20 μL reaction according to manufacturer's instructions. cDNA samples were diluted 1:5 prior to qPCR experiments. hTERT mRNA levels were quantified using the Light Cycler SYBR Green I Master Real-Time PCR Mix (Roche Applied Science) in a 25 μL reaction. The forward primer for hTERT was 5′-CTT TTA TGT CAC GGA GAC CAC-3′ (300 nmol/L) and the reverse primer was 5′-TAG TCC ATG TTC ACA ATC GG-3′ (600 nmol/L). Amplification of the hTERT gene was normalized against the housekeeping gene TATA-binding protein (TBP). For TBP, the forward primer was 5′-CACGAACCACGGCACTGATT-3′ (600 nmol/L) and the reverse primer was 5-TTTTCTTGCTGCCAGTCTGGAC-3′ (900 nmol/L). qPCR reactions were carried out in technical quadruplicate using the following cycling parameters: 95°C for 10 minutes, 40 cycles: (95°C for 15 seconds, 60°C for 30 seconds, 72°C for 15 seconds) on the LightCycler 480 Real-Time PCR System (Roche Applied Science). The second derivative max method was used to determine crossing points (Cp) of each amplification curve. Serial dilutions of pGEM-based plasmids that contain the sequences of hTERT or TBP were used to generate a standard curve for each respective gene. The concentrations of hTERT and TBP mRNAs were calculated by plotting the Cp value for each unknown sample to its respective standard curve. The relative expression level of hTERT mRNA, reported in arbitrary units (AU), was calculated as the concentration value of hTERT mRNA divided by concentration value of TBP mRNA. Standard error in technical quadruplicates ranged from 0.9% to 3.0% (average: 2.2%).

Microarray analysis

All procedures were conducted according to the instructions from Affymetrix. Briefly, 100 ng of total RNA from the sorted cells was used to generate double-stranded cDNA using an oligo(dT) primer containing the T7 RNA polymerase promoter site and the 3′ IVT Express Labeling Kit (Affymetrix). Next, in vitro transcription was conducted to synthesize biotin-modified aRNA with the IVT labeling master mix which generates multiple copies of biotin-modified aRNA from the double-stranded cDNA templates. The aRNA was then purified to remove unincorporated nucleotide triphosphates, salts, enzymes, and inorganic phosphate prior to fragmentation and hybridization onto Human Genome U133 Plus 2.0 GeneChip expression arrays. Genome arrays were hybridized for 16 hours at 45°C following the Affymetrix protocol. Following washing and staining, the GeneChips were scanned using the Affymetrix GeneChip Scanner 3000. Data were analyzed using Genespring GX version 11. To determine thresholds for reliability, the analysis of each of the 4 RNAs was repeated independently. Significance analysis of microarrays (SAM; ref. 20) was also conducted to determine significant differences between the hTERT high and hTERT low populations in tumor and TAHN-1 tissues.

TRAP assay

Telomerase activity was measured using the SYBR Green-telomerase repeat amplification protocol (SYBR TRAP) and a gel-based TRAP assay (Millipore; TRAPeze kit). The gel-based TRAP assay was conducted according to manufacturer's instructions and quantitated by densitometry analysis. SYBR TRAP was conducted as described by Wege and colleagues (21). Briefly, cell extracts were prepared by dissolving cells in 200 μL of 1× CHAPS lysis buffer (Chemicon) and incubating them on ice for 25 minutes. After this incubation, samples are centrifuged for 5 minutes (14,500 × g) at 4°C. The resulting supernatant was transferred to a prechilled microcentrifuge tube and was used immediately or was stored at −70°C for up to 12 months. The protein concentration within the supernatant was determined by modified Lowry assay (DC protein assay; BioRad). Telomerase activity was determined by using 2 μg total protein within 40 μL reactions containing 1× SYBR PCR master mix (Applied Biosystems), 0.16 μg TS primer (5′-AATCCGTCGAGCAGAGTT-3′), and 0.08 μg ACX return primer, 5′-GCGCGG[CTTACC]3CTACC-3′. Reactions are incubated for 20 minutes at 25°C, heated to 95°C for 10 minutes, and then amplified and detected by 40 cycles of 95°C (30 seconds), 60°C (1 minute) using the Applied Biosystems SDS 7000. Negative controls were prepared by heating 2 μg of CHAPS extract to 80°C for 10 minutes prior to the assay. Relative levels of telomerase activity are determined by comparing results with a standard curve generated by measuring telomerase within dilutions of telomerase-positive cells (HEK293).

hTERT promoter–reporter activity correlates with hTERT mRNA levels and telomerase activity in primary tumor HMECs

We developed an experimental system (the “StopLight” vector) that utilizes an adenovirus containing the enhanced green fluorescent protein (EGFP) gene under the control of the hTERT promoter to assess hTERT promoter activity and the DsRed fluorescent protein (RFP) gene under the control of the constitutive CMV promoter to distinguish between infected and noninfected cells. A prototype adenovirus containing only the EGFP gene construct under the control of the hTERT promoter (Green vector) was used in initial experiments to validate the efficacy of the system.

The underlying principle of this experimental system was tested by measuring and comparing the levels of both hTERT mRNA and telomerase activity within the subpopulations of primary tumor-derived HMECs infected with the Green vector and sorted on the basis of green fluorescence. The green fluorescence within these cells is governed by the activity of the hTERT promoter–GFP construct, so differences in green fluorescence presumably reflect underlying variation in promoter activity. Differences in promoter activity should, in turn, be correlated with altered hTERT mRNA levels and telomerase activity within these subpopulations of cells. Initial experiments, described subsequently, confirmed these predictions.

The efficacy of the hTERT promoter–GFP construct was first evaluated with the prototypical Green vector in 2 telomerase-positive cell lines, HEK293 and MDA-MB-231, and a telomerase-negative cell line, WI-38 (Supplementary Fig. S1). Eighty-nine percent and 14% of HEK293 and MDA-MB-231 cells, respectively, exceeded thresholds of green fluorescence that excluded 99% of uninfected cells (1% gate). In contrast, green fluorescence in infected WI-38 cells was indistinguishable from uninfected cells. Thus, green fluorescence driven by the hTERT promoter–GFP construct was positively correlated with telomerase expression. Next, a gel-based TRAP assay was used to confirm that green fluorescence in cells infected hTERT promoter–GFP construct could be used to separate mixed cell populations based on telomerase activity. HEK293 cells (telomerase positive) and WI-38 cells (telomerase negative) were infected separately with Green vector, mixed in a 1:1 ratio, and sorted on the basis of “high” or “low” green fluorescence. Fluorescence-activated cell-sorting (FACS) analysis of the mixed population showed that the percentages of cells in the high and low green fluorescence populations were 35% and 65%, respectively. This indicated that approximately 70% of the HEK293 cells (35%/50%) were infected. Uninfected HEK293 cells comprise 23% (15%/65%) of the low green fluorescence population. On the basis of these data, we predicted that telomerase activity would be 4.3-fold greater in sorted cells with high green fluorescence. This prediction was tested by measuring telomerase activity in the sorted populations by a gel-based TRAP assay (TRAPeze telomerase detection kit; Millipore). The results were in excellent agreement with the predicted enrichment; telomerase activity in sorted cells with high green fluorescence was 4.2-fold greater than in sorted cells with low green fluorescence when compared with the low population.

The system was further validated using short-term HMEC cultures of organoid preparations from human breast tumors. The cells were expanded for 3 passages (Fig. 1A), infected with the Green vector (Fig. 1B), and again analyzed by flow cytometry (Fig. 1C). In HMEC cultures from 2 independent tumors, 34% and 19% of infected cells exceeded thresholds of green fluorescence that excluded 99% of uninfected cells.

Figure 1.

Performance of hTERT promoter reporter (Green vector) in primary HMECs derived from human breast tumors. A, HMECs from short-term cultures of human breast tumors. B, tumor HMECs infected with Green vector. C, cell counts (y-axis) versus log green fluorescence (x-axis) from uninfected and infected tumor HMECs analyzed by flow cytometry. Region bars were set to exclude 99% of uninfected cells (“1% gate”) as in Supplementary Figure S1.

Figure 1.

Performance of hTERT promoter reporter (Green vector) in primary HMECs derived from human breast tumors. A, HMECs from short-term cultures of human breast tumors. B, tumor HMECs infected with Green vector. C, cell counts (y-axis) versus log green fluorescence (x-axis) from uninfected and infected tumor HMECs analyzed by flow cytometry. Region bars were set to exclude 99% of uninfected cells (“1% gate”) as in Supplementary Figure S1.

Close modal

FACS was used to separate tumor HMECs by log decades in to 3 groups on the basis of presumptive hTERT promoter activity, inferred from EGFP fluorescence (Fig. 2A). The hTERT mRNA levels in the 3 sorted cell populations were measured by quantitative real-time PCR (qRT-PCR). As shown in Figure 2B, hTERT mRNA levels were 22-fold higher in the tumor HMECs with the highest green fluorescence than in the total HMEC population, indicating a significant enrichment of hTERT mRNA. Similarly, telomerase activity was approximately 5-fold higher in the tumor HMECs with the highest green fluorescence than in the HMECs with the lowest green fluorescence (Fig. 2C). Several primary cultures of tumor-derived fibroblasts were similarly infected and analyzed. In contrast to HMECs, neither hTERT promoter activity nor hTERT mRNA was detectable in the tumor-derived fibroblasts (data not shown). Thus, variation in hTERT promoter activities revealed by the hTERT promoter–EGFP reporter reflects differences in hTERT mRNA levels and telomerase activity and provides a means to separate subpopulations of HMECs differing in hTERT expression.

Figure 2.

hTERT mRNA levels and telomerase activity in primary tumor HMECs infected with Green vector and sorted on the basis of green fluorescence. A, HMECs from a third independent human breast tumor were infected with Green vector and sorted by log decades of green fluorescence using FACS. The low category contains HMECs within the first log decade, the medium category contains HMECs within the second log decade, and the high category contains HMECs within the third log decade and above. B, hTERT mRNA levels in the total HMEC population and HMECs in the high fluorescence category were measured by qRT-PCR. hTERT mRNA levels are shown on the y-axis. WI-38 is a telomerase-negative cell line and serves as a negative control. The WI-38 cell line transfected with pBABE-hTERT and ectopically expressing hTERT serves as a positive control. C, relative telomerase activity was measured in HMECs in the low, medium, and high categories by the SYBR TRAP method (20). Telomerase activity is shown on the y-axis. The WI-38 and WI-38 hTERT cell lines serve as positive and negative controls.

Figure 2.

hTERT mRNA levels and telomerase activity in primary tumor HMECs infected with Green vector and sorted on the basis of green fluorescence. A, HMECs from a third independent human breast tumor were infected with Green vector and sorted by log decades of green fluorescence using FACS. The low category contains HMECs within the first log decade, the medium category contains HMECs within the second log decade, and the high category contains HMECs within the third log decade and above. B, hTERT mRNA levels in the total HMEC population and HMECs in the high fluorescence category were measured by qRT-PCR. hTERT mRNA levels are shown on the y-axis. WI-38 is a telomerase-negative cell line and serves as a negative control. The WI-38 cell line transfected with pBABE-hTERT and ectopically expressing hTERT serves as a positive control. C, relative telomerase activity was measured in HMECs in the low, medium, and high categories by the SYBR TRAP method (20). Telomerase activity is shown on the y-axis. The WI-38 and WI-38 hTERT cell lines serve as positive and negative controls.

Close modal

A limitation of the Green vector is that it does not discriminate between uninfected cells and infected cells that lack hTERT promoter activity. To correct this problem, the Green vector was modified to include a selectable marker, the DsRed RFP gene under the control of the constitutive CMV promoter. The resulting “StopLight” vector utilizes green fluorescence from the hTERT promoter–EGFP reporter to assess hTERT promoter activity and red fluorescence from the CMV promoter–RFP reporter to distinguish infected from noninfected cells. Median red fluorescence in uninfected tumor cells was 6 units, and less than 1% of uninfected tumor cells had red fluorescence above 16 units (Fig. 3A). In contrast, median red fluorescence in the subset of tumor cells infected with the StopLight vector was 889 units (Fig. 3C). Only infected cells with red fluorescence in excess of the 1% gate were included in subsequent analyses (e.g., 94% of cells in C).

Figure 3.

FACS of HMECs infected with StopLight vector. Short-term cultures derived from independent breast tumor tissue were infected with StopLight vector and analyzed by flow cytometry. A, a gate was established in the red fluorescence channel (x-axis, FL2) that excluded 99% of uninfected HMECs. This gate was used subsequently to exclude uninfected cells from analysis. B, histogram of green fluorescence in uninfected cells. C, 94% of cells infected with Stoplight vector were infected. D, tumor HMECs infected with StopLight vector were sorted into with low, medium, and high groups on the basis of green fluorescence. The low, medium, and high groups comprised 10%, 80%, and 10%, respectively, of the total tumor HMECs. E, hTERT mRNA levels in the high, medium, and low fluorescence groups sorted from tumor and TAHN-1 HMECs. hTERT mRNA levels are shown on the y-axis. WI-38 is a telomerase-negative cell line and serves as a negative control. The WI-38 cell line transfected with pBABE-hTERT and ectopically expressing hTERT serves as a positive control.

Figure 3.

FACS of HMECs infected with StopLight vector. Short-term cultures derived from independent breast tumor tissue were infected with StopLight vector and analyzed by flow cytometry. A, a gate was established in the red fluorescence channel (x-axis, FL2) that excluded 99% of uninfected HMECs. This gate was used subsequently to exclude uninfected cells from analysis. B, histogram of green fluorescence in uninfected cells. C, 94% of cells infected with Stoplight vector were infected. D, tumor HMECs infected with StopLight vector were sorted into with low, medium, and high groups on the basis of green fluorescence. The low, medium, and high groups comprised 10%, 80%, and 10%, respectively, of the total tumor HMECs. E, hTERT mRNA levels in the high, medium, and low fluorescence groups sorted from tumor and TAHN-1 HMECs. hTERT mRNA levels are shown on the y-axis. WI-38 is a telomerase-negative cell line and serves as a negative control. The WI-38 cell line transfected with pBABE-hTERT and ectopically expressing hTERT serves as a positive control.

Close modal

Tumor cells infected with the StopLight vector were fluorescence activated cell sorted on the basis of presumptive hTERT promoter activity, inferred from green fluorescence, and their hTERT mRNA levels were again measured by qRT-PCR. Consistent with the independent results obtained with the Green vector in HMECs from different tumor specimens, the 10% of tumor HMECs with the highest green fluorescence had an hTERT mRNA level of 573 AU, as compared with 1.0 and 7.9 AU for cells with the middle (80%) and lowest (10%) of green fluorescence, respectively (Fig. 3E, tumor). Thus, greater than 95% of the hTERT mRNA resided in a subset of tumor cells with the highest green fluorescence and comprising no more than 10% of the cultured HMECs.

Differential gene expression in tumor-derived HMECs expressing high and low levels of hTERT

To further evaluate the efficacy of the cell-sorting strategy, we compared the spectrum of gene expression in tumor HMECs with the highest and lowest hTERT promoter activities using the Affymetrix Human Genome U133 Plus 2.0 GeneChip Arrays. We predicted that if the sorting strategy was, in fact, separating populations that are phenotypically distinct, then the spectra of gene expression would differ between HMECs with high and low hTERT promoter activities. A total of 19,350 transcripts had expression values that exceeded the threshold of reliability. Of these, 1,142 transcripts were overexpressed at least 4-fold in the 10% of tumor HMECs with the highest hTERT promoter activities relative to the 10% of HMECs from the same culture with lowest hTERT promoter activities. Gene Ontology analysis (http://biit.cs.ut.ee/gprofiler/) revealed highly significant relationships between the genes overexpressed in cells expressing high levels of hTERT and gene categories related to proliferation and mitosis (Table 1). These results provide further evidence that differential hTERT promoter activity identifies discrete subsets of HMECs.

Table 1.

Ontology categories that are statistically overrepresented in the list of genes overexpressed more than 4-fold in the HMECs with the highest 10% of green fluorescence in tumor tissues

GO terms overrepresented in hTERT high HMECs in tumor
# in gene list/Total genes in categoryaPbOntology categoriesc
  Biological process 
57/808 l.70e-09 Cell cycle (1) 
35/383 5.02e-09 Mitotic cell cycle (2) 
46/596 4.50e-09  Cell-cycle process (2) 
41/430 6.33e-11   Cell-cycle phase (3) 
36/344 7.20e-11   M phase (4) 
28/231 3.43e-10   M phase of mitotic cell cycle (5) 
76/1455 1.25e-06 Organelle organization (1) 
27/235 2.34e-09 Organelle fission (2) 
27/225 8.92e-10  Nuclear division (3) 
27/225 8.92e-10   Mitosis (4) 
61/1086 1.60e-06 Regulation of multicelluar organismal process (1) 
15/84 2.53e-08 Chromosome segregation (1) 
34/299 2.56e-11 Cell division (1) 
7/25 6.21e-06 Chromosome condensation (1) 
29/346 6.51e-07 Regulation of cell cycle (1) 
17/156 4.52e-06 Regulation of mitotic cell cycle (2) 
  Cellular compartment 
37/548 3.81e-06 Chromosome (1) 
19/128 8.67e-09  Condensed chromosome (2) 
18/131 7.26e-08 Chromosome, centromeric region (1) 
12/65 4.34e-07  Condensed chromosome, centromeric region (2) 
13/82 9.03e-07 Kinetochore (1) 
11/57 8.26e-07  Condensed chromosome, kinetochore (2) 
  Reactome 
9/43 4.97e-06 G2/M checkpoints (1) 
31/308 1.46e-09 Cell cycle, mitotic (1) 
14/88 3.64e-07 M phase (2) 
14/84 1.99e-07  Mitotic prometaphase (3) 
14/78 7.52e-08   Kinetochore capture of astral microtubules (4) 
GO terms overrepresented in hTERT high HMECs in tumor
# in gene list/Total genes in categoryaPbOntology categoriesc
  Biological process 
57/808 l.70e-09 Cell cycle (1) 
35/383 5.02e-09 Mitotic cell cycle (2) 
46/596 4.50e-09  Cell-cycle process (2) 
41/430 6.33e-11   Cell-cycle phase (3) 
36/344 7.20e-11   M phase (4) 
28/231 3.43e-10   M phase of mitotic cell cycle (5) 
76/1455 1.25e-06 Organelle organization (1) 
27/235 2.34e-09 Organelle fission (2) 
27/225 8.92e-10  Nuclear division (3) 
27/225 8.92e-10   Mitosis (4) 
61/1086 1.60e-06 Regulation of multicelluar organismal process (1) 
15/84 2.53e-08 Chromosome segregation (1) 
34/299 2.56e-11 Cell division (1) 
7/25 6.21e-06 Chromosome condensation (1) 
29/346 6.51e-07 Regulation of cell cycle (1) 
17/156 4.52e-06 Regulation of mitotic cell cycle (2) 
  Cellular compartment 
37/548 3.81e-06 Chromosome (1) 
19/128 8.67e-09  Condensed chromosome (2) 
18/131 7.26e-08 Chromosome, centromeric region (1) 
12/65 4.34e-07  Condensed chromosome, centromeric region (2) 
13/82 9.03e-07 Kinetochore (1) 
11/57 8.26e-07  Condensed chromosome, kinetochore (2) 
  Reactome 
9/43 4.97e-06 G2/M checkpoints (1) 
31/308 1.46e-09 Cell cycle, mitotic (1) 
14/88 3.64e-07 M phase (2) 
14/84 1.99e-07  Mitotic prometaphase (3) 
14/78 7.52e-08   Kinetochore capture of astral microtubules (4) 

NOTE: HMECs with the highest hTERT promoter activity overexpress genes related to cell cycle and mitosis. Microarray analysis was conducted on HMECs from tumor tissues infected with the StopLight vector and sorted on the basis of green fluorescence. HMECs with the highest and lowest 10% of green fluorescence from Figure 3 were analyzed.

aThe number of overexpressed genes within each ontology category as a fraction of the total number of genes that comprise that category.

bP value is calculated by the gProfiler software. For a P value to be considered statistically significant, it must be lower than 95% of P values generated by a random data set of the same size.

cNumbers in parentheses denote the level of the category within the ontology hierarchy.

hTERT promoter activity varies as a function of distance from the tumor in primary HMECs infected with the StopLight vector

Our prior investigation of breast field cancerization showed that telomere length and allelic imbalance in TAHN-1 tissues are indistinguishable from those in the adjacent tumor and statistically different from those in more distal paired TAHN-5 tissues or tissues derived from RM (3). We therefore investigated whether HMECs with high and low hTERT expression had a similar spatial distribution. To address this question, short-term cultures of HMECs from paired tumor and TAHN tissues 1, 3, and 5 cm from the tumor's margin (TAHN-1, TAHN-3, and TAHN-5, respectively) were infected with StopLight vector and promoter activities were again compared by flow cytometry. Figure 4 shows results (dot plot) from a representative set of patient-matched tumor, TAHN-1, TAHN-3, and TAHN-5 tissues, an RM tissue, and the telomerase-negative WI-38 cell line (Fig. 4). Quadrants were set to exclude 99% of uninfected cells from quadrants R3 and R4 and 99% of WI-38 and HMECs from the RM tissue from quadrant R4 (not shown). The fraction of cells that were infected and presumptively hTERT positive (quadrant R4) was 47.32% and 41.28% for tumor and TAHN-1, respectively. In contrast, the fraction of cells that were infected and presumptively hTERT positive (R3) was 4.76% and 6.03% for TAHN-3 and TAHN-5, respectively. This implies that the fraction of HMECs expressing telomerase decreases with distance from the tumor.

Figure 4.

Flow cytometric analysis of HMECs from a representative set of patient-matched tumor, TAHN-1, TAHN-3, and TAHN-5 tissues, an RM tissue, and the hTERT-negative cell line WI-38. Short-term cultures of HMECs derived from patient-matched breast tumor, TAHN-1, TAHN-3, and TAHN-5 tissues, an RM tissue, and the hTERT-negative cell line WI-38 were infected with StopLight vector. Green and red fluorescence (FL1, x-axis and FL2, y-axis, respectively) were analyzed by flow cytometry and dot plots are shown. Quadrants were set by excluding 99% of uninfected HMECs from quadrants R3 and R4 and excluding 99% of infected WI-38 cells and HMECs from RM from quadrant R4.

Figure 4.

Flow cytometric analysis of HMECs from a representative set of patient-matched tumor, TAHN-1, TAHN-3, and TAHN-5 tissues, an RM tissue, and the hTERT-negative cell line WI-38. Short-term cultures of HMECs derived from patient-matched breast tumor, TAHN-1, TAHN-3, and TAHN-5 tissues, an RM tissue, and the hTERT-negative cell line WI-38 were infected with StopLight vector. Green and red fluorescence (FL1, x-axis and FL2, y-axis, respectively) were analyzed by flow cytometry and dot plots are shown. Quadrants were set by excluding 99% of uninfected HMECs from quadrants R3 and R4 and excluding 99% of infected WI-38 cells and HMECs from RM from quadrant R4.

Close modal

This conclusion was confirmed by comparing hTERT promoter activity in primary cultures derived from 7 additional independent sets of paired breast tissue samples. On the basis of the prior results, we established gates in the green fluorescence channel (R5) that excluded 99% of HMECs in the paired TAHN-5 tissue (e.g., Fig. 5A) and then determined the percentage of the total HMECs within each paired culture that exceeded the threshold defined by the R5 gate. Representative data from 1 of the independent sets of paired tissues are shown in Figure 5. The histograms of HMECs obtained from tumor, TAHN-3, and TAHN-5 tissues are compared with the histogram of HMECs derived from the paired TAHN-1 tissue, which is denoted by crosshatching (Fig. 5A–C). For this specific set of paired tissues, more than 13% of HMECs from TAHN-1 tissue had green fluorescence within the R5 gate (Fig. 5A), compared with 10.8% for the paired tumor (Fig. 5B) and 0.9% for TAHN-3 tissue (Fig. 5C).

Figure 5.

hTERT promoter activity decreases as a function of distance in cultured HMECs from TAHN tissues. A, the data from the tumor and TAHN HMECs in Figure 4 were reanalyzed to establish a more stringent gate (R5) that excluded 99% of HMECs in TAHN-5 tissue and approximates the “high” category shown in Figure 3. A–C, pairwise comparisons of green fluorescence between HMECs from TAHN-1 (crosshatched histogram) and: (A) TAHN-5, (B) TAHN-3, and (C) tumor tissues. The percentage of HMECs within the R5 gate for tumor and TAHN samples. D, short-term cultures of HMECs derived from 5 complete sets (tumor, TAHN-1, TAHN-3, and TAHN-5) and 3 partial sets (TAHN-1, TAHN-3, and TAHN-5) of independent patient-matched breast tissues were infected separately with the StopLight vector and analyzed by flow cytometry. The R5 gate for each pair was established with HMECs from the paired TAHN-5 tissue, so that each tumor, TAHN-1, and TAHN-3 sample is compared with its patient-matched TAHN-5 sample. The percentage of HMECs in the R1 gates from the 29 individual cultures is plotted. Patient-paired sets are connected by solid lines. Asterisks denote statistically significant differences from HMECS derived from TAHN-5 tissues (P ≤ 0.05) determined by Student's t test.

Figure 5.

hTERT promoter activity decreases as a function of distance in cultured HMECs from TAHN tissues. A, the data from the tumor and TAHN HMECs in Figure 4 were reanalyzed to establish a more stringent gate (R5) that excluded 99% of HMECs in TAHN-5 tissue and approximates the “high” category shown in Figure 3. A–C, pairwise comparisons of green fluorescence between HMECs from TAHN-1 (crosshatched histogram) and: (A) TAHN-5, (B) TAHN-3, and (C) tumor tissues. The percentage of HMECs within the R5 gate for tumor and TAHN samples. D, short-term cultures of HMECs derived from 5 complete sets (tumor, TAHN-1, TAHN-3, and TAHN-5) and 3 partial sets (TAHN-1, TAHN-3, and TAHN-5) of independent patient-matched breast tissues were infected separately with the StopLight vector and analyzed by flow cytometry. The R5 gate for each pair was established with HMECs from the paired TAHN-5 tissue, so that each tumor, TAHN-1, and TAHN-3 sample is compared with its patient-matched TAHN-5 sample. The percentage of HMECs in the R1 gates from the 29 individual cultures is plotted. Patient-paired sets are connected by solid lines. Asterisks denote statistically significant differences from HMECS derived from TAHN-5 tissues (P ≤ 0.05) determined by Student's t test.

Close modal

The results from these and the other 7 independent sets of paired breast tissue are summarized in Figure 5D. In 4 of the pairs, there was a 5-fold or greater enrichment of HMECs within the R5 gate in both tumor and TAHN-1, but not TAHN-3, relative to TAHN-5 samples. The percentage of cells in the R5 gate in individual tumor samples varied greatly from patient to patient. A 90-fold range in the percentage of HMECs in the R1 gate was measured in the 5 independent tumor tissues. Similarly, a 14-fold range in the percentage of HMECs in excess of the R5 gate was measured in the 7 TAHN-1 tissues. However, the average fold difference within the 5 pairs of tumor and patient-matched TAHN-1 tissues was only 1.9-fold, and the difference within 3 of the 5 pairs was less than 1.2-fold. Thus, in most instances, hTERT promoter activity in HMECs in TAHN-1 tissues mirrors the activity in the HMECs in the paired tumor. In contrast, hTERT promoter activities were statistically different (P ≤ 0.05) and up to 35-fold lower in HMECs from TAHN-3 and TAHN-5 tissues than from the paired tumor and TAHN-1 tissues.

Gene expression in the subpopulation of TAHN-1 HMECs expressing high levels of hTERT mimic gene expression in matched tumor HMECs

This study, along with several published studies, shows several similarities between tumor and TAHN tissues (3, 7, 8, 11, 22, 23). These similarities beg the question whether the subpopulations of tumor and TAHN-1 HMECs with high hTERT levels share similar gene expression profiles. FACS was used to isolate StopLight virus–infected TAHN-1 cells and hTERT mRNA levels were measured by qRT-PCR, each as carried out previously for the paired tumor cells. The hTERT mRNA level in the 10% of TAHN-1 cells with the highest promoter activities was 1,526 AU. hTERT mRNA levels in the 80% of HMECs with intermediate promoter activities and the 10% of HMECs with the lowest promoter activities were 2.4 and 1.6 AU, respectively (Fig. 3B, TAHN-1). Thus, as observed in primary HMECs derived from paired tumor tissues, greater than 95% of the hTERT mRNA resided in a subset of cells comprising no more than 10% of the cultured HMECs in TAHN-1 sample.

The spectra of gene expression in HMECs from TAHN-1 tissues with the highest and lowest hTERT promoter activities also mimicked those in the paired tumor sample. Of the19,350 transcripts with expression values that exceeded the threshold of reliability, 363 transcripts were overexpressed at least 4-fold in the 10% of TAHN-1 HMECs with highest hTERT promoter activities relative to the 10% of TAHN-1 HMECs from the same culture with lowest hTERT promoter activities. Gene Ontology analysis again revealed highly significant relationships between the genes overexpressed in cells expressing high levels of hTERT and Gene Ontology categories related to proliferation and mitosis (Table 2). To evaluate the statistical significance of genes that were differentially expressed in cells with high hTERT promoter activity in both tumor and TAHN-1 tissues, SAM analysis was conducted (Tables 3 and 4). “Telomerase high” data sets from both tumor and TAHN-1 tissues were compared with “telomerase low” data sets from tumor and TAHN-1 tissues in a paired analysis. Twelve genes were identified that were upregulated greater than 3-fold and had a false discovery rate lower than 3.85%. These genes are listed in Table 3. There were no statistically significant downregulated genes. The 12 genes were run through the g-Profiler program to determine Gene Ontology terms that are over-represented. The most significant Gene Ontology terms (P < 1.0e-10) are listed in Table 4. We conclude from these results that a population of HMECs with the same phenotype as HMECs within the tumor mass exists within the field of “normal tissue” proximal to the breast tumors, and this population is significantly reduced in tissues more distal to the tumor.

Table 2.

Ontology categories that are statistically overrepresented in the list of genes overexpressed more than 4-fold in the HMECs with the highest 10% of green fluorescence in TAHN-1 tissues

GO terms overrepresented in hTERT high HMECs in TAHN-1
#No. in gene list/Total genes in categoryPOntology categories
  Biological process 
32/808 3.07e-10 Cell cycle (1) 
23/383 4.37e-11 Mitotic cell cycle (2) 
27/596 4.08e-10 Cell-cycle process (2) 
25/430 1.19e-11  Cell-cycle phase (3) 
22/344 3.52e-11   M phase (4) 
18/231 9.54e-11    M phase of mitotic cell cycle (5) 
36/1455 3.59e-06 Organelle organization (1) 
18/235 1.26e-10 Organelle fission (2) 
18/225 6.18e-11  Nuclear division (3) 
18/225 6.18e-11   Mitosis (4) 
6/52 2.20e-05 Spindle organization (1) 
10/179 3.01e-05 DNA replication (1) 
20/299 1.30e-10 Cell division (1) 
  Cellular compartment 
19/588 2.86e-05 Microtubule cytoskeleton (1) 
91/5365 1.05e-06 Nucleus (1) 
19/548 1.08e-05 Chromosome (1) 
17/476 2.19e-05 Chromosomal part (2) 
16/131 1.18e-12  Chromosome centromeric region (3) 
10/82 2.25e-08  Kinetochore (3) 
3/4 5.21e-06  Ndc80 complex (3) 
12/128 1.78e-08 Condensed chromosome (2) 
9/65 3.73e-08  Condensed chromosome centromeric region (3) 
9/57 1.13e-08   Condensed chromosome kinetochore (4) 
  Reactome 
20/308 1.08e-09 Cell cycle, mitotic (1) 
10/88 2.27e-07 M phase (2) 
10/84 1.44e-07  Mitotic prometaphase (3) 
10/78 7.00e-08   Kinetochore capture of astral microtubules (4) 
GO terms overrepresented in hTERT high HMECs in TAHN-1
#No. in gene list/Total genes in categoryPOntology categories
  Biological process 
32/808 3.07e-10 Cell cycle (1) 
23/383 4.37e-11 Mitotic cell cycle (2) 
27/596 4.08e-10 Cell-cycle process (2) 
25/430 1.19e-11  Cell-cycle phase (3) 
22/344 3.52e-11   M phase (4) 
18/231 9.54e-11    M phase of mitotic cell cycle (5) 
36/1455 3.59e-06 Organelle organization (1) 
18/235 1.26e-10 Organelle fission (2) 
18/225 6.18e-11  Nuclear division (3) 
18/225 6.18e-11   Mitosis (4) 
6/52 2.20e-05 Spindle organization (1) 
10/179 3.01e-05 DNA replication (1) 
20/299 1.30e-10 Cell division (1) 
  Cellular compartment 
19/588 2.86e-05 Microtubule cytoskeleton (1) 
91/5365 1.05e-06 Nucleus (1) 
19/548 1.08e-05 Chromosome (1) 
17/476 2.19e-05 Chromosomal part (2) 
16/131 1.18e-12  Chromosome centromeric region (3) 
10/82 2.25e-08  Kinetochore (3) 
3/4 5.21e-06  Ndc80 complex (3) 
12/128 1.78e-08 Condensed chromosome (2) 
9/65 3.73e-08  Condensed chromosome centromeric region (3) 
9/57 1.13e-08   Condensed chromosome kinetochore (4) 
  Reactome 
20/308 1.08e-09 Cell cycle, mitotic (1) 
10/88 2.27e-07 M phase (2) 
10/84 1.44e-07  Mitotic prometaphase (3) 
10/78 7.00e-08   Kinetochore capture of astral microtubules (4) 

NOTE: HMECs with the highest hTERT promoter activity overexpress genes related to cell cycle and mitosis. Microarray analysis was conducted on HMECs from TAHN-1 tissues infected with the StopLight vector and sorted on the basis of green fluorescence. HMECs with the highest and lowest 10% of green fluorescence from Figure 3 were analyzed. Other details as in Table 1.

Table 3.

SAM analysis was conducted on microarray data

Gene IDSymbolGene descriptionFold changeq (%)
228729 at CCNB1 Cyclin B1 9.91 3.85 
228069 at FAM54A Family with sequence similarity 54, member A 9.30 3.85 
203362 s at MAD2L1 MAD2 mitotic arrest deficient-like 1 (yeast) 5.44 0.00 
209408 at KIF2C Kinesin family member 2C 5.03 3.85 
218009 s at PRC1 Protein regulator of cytokinesis 1 4.99 0.00 
218875 s at FBXO5 F-box protein 5 4.95 3.85 
219148 at PBK PDZ-binding kinase 4.85 0.00 
219918 s at ASPM asp (abnormal spindle) homolog, microcephaly associated (Drosophila) 4.44 3.85 
203213 at CDK1 Cyclin-dependent kinase 1 4.13 3.85 
214710 s at CCNB1 Cyclin B1 3.88 0.00 
208808 s at HMGB2 High-mobility group box 2 3.81 0.00 
208079 s at AURKA Aurora kinase A 3.55 0.00 
Gene IDSymbolGene descriptionFold changeq (%)
228729 at CCNB1 Cyclin B1 9.91 3.85 
228069 at FAM54A Family with sequence similarity 54, member A 9.30 3.85 
203362 s at MAD2L1 MAD2 mitotic arrest deficient-like 1 (yeast) 5.44 0.00 
209408 at KIF2C Kinesin family member 2C 5.03 3.85 
218009 s at PRC1 Protein regulator of cytokinesis 1 4.99 0.00 
218875 s at FBXO5 F-box protein 5 4.95 3.85 
219148 at PBK PDZ-binding kinase 4.85 0.00 
219918 s at ASPM asp (abnormal spindle) homolog, microcephaly associated (Drosophila) 4.44 3.85 
203213 at CDK1 Cyclin-dependent kinase 1 4.13 3.85 
214710 s at CCNB1 Cyclin B1 3.88 0.00 
208808 s at HMGB2 High-mobility group box 2 3.81 0.00 
208079 s at AURKA Aurora kinase A 3.55 0.00 

NOTE: SAM identifies 12 genes that are statistically upregulated in both tumor and TAHN-1 cells with the highest hTERT promoter activity. “Telomerase high” data sets from both tumor and TAHN-1 were compared with “telomerase low” data sets from tumor and TAHN-1 in a paired analysis. The arrays were median centered, a minimum fold change of 3 was set, and a delta value of 6.23 was set to establish a false discovery rate of 3.85% (q value). Using these values, 12 genes were identified that were statistically significantly overexpressed in the cells with high telomerase promoter activity in both tumor and TAHN-1 cells. No genes were found to be significantly underexpressed in these cells.

Table 4.

Gene Ontology terms overrepresented in the list of 12 overexpressed genes

PaTerm nameTerm domain
2.25E-14 M phase BP 
1.34E-13 Mitotic cell cycle BP 
1.67E-13 Nuclear division BP 
1.67E-13 Mitosis BP 
2.12E-13 M phase of mitotic cell cycle BP 
2.29E-13 Organelle fission BP 
2.42E-13 cell-cycle phase BP 
7.77E-13 Spindle CC 
1.07E-12 Cell division BP 
3.98E-12 Cell-cycle process BP 
3.51E-11 Organelle organization BP 
6.66E-11 Microtubule cytoskeleton CC 
7.30E-11 Cell cycle BP 
1.20E-10 Spindle pole CC 
PaTerm nameTerm domain
2.25E-14 M phase BP 
1.34E-13 Mitotic cell cycle BP 
1.67E-13 Nuclear division BP 
1.67E-13 Mitosis BP 
2.12E-13 M phase of mitotic cell cycle BP 
2.29E-13 Organelle fission BP 
2.42E-13 cell-cycle phase BP 
7.77E-13 Spindle CC 
1.07E-12 Cell division BP 
3.98E-12 Cell-cycle process BP 
3.51E-11 Organelle organization BP 
6.66E-11 Microtubule cytoskeleton CC 
7.30E-11 Cell cycle BP 
1.20E-10 Spindle pole CC 

NOTE: SAM identifies 12 genes that are statistically upregulated in both tumor and TAHN-1 cells with the highest hTERT promoter activity. The most significant Gene Ontology terms (P < 1.0e-10) are listed.

Abbreviations: CC, cellular compartment; BP, biological process.

aP value is calculated by the gProfiler software and represents the probability of the Gene Ontology term to be equally represented in a random data set of the same size.

We have developed an experimental reporter system, the StopLight vector, composed of the EGFP reporter under control of the hTERT promoter, to differentiate and viably sort primary HMECs based on their levels of hTERT promoter activity. Experiments in established cell lines positively correlated expression of the reporter with expression of telomerase. Similarly, the reporter was differentially expressed between HMECs derived from human breast tumor tissues, more than 90% of which express telomerase, and normal RM tissues, which typically do not. Cell sorting, qRT-PCR, and TRAP assay showed that hTERT promoter activity was positively correlated with the cellular concentrations of hTERT mRNA and telomerase activity. Ectopic hTERT expression has been shown to alter expression of growth-promoting genes in human breast cells (16), and our results similarly showed that several genes involved in cell cycle and mitosis were overexpressed in HMECs with high hTERT promoter activity. We conclude from these several independent data that HMECs derived from human breast tumors contain discrete subpopulations that can be differentiated and sorted on the basis of hTERT promoter activity.

We have reported previously that some early molecular changes that occur in malignant transformation of the breast, including telomere length abnormalities and genomic instability, also occur in TAHN-1 but not TAHN-5 or RM tissues (3). We used the StopLight vector to compare similarly hTERT promoter activity in HMECs derived from tumor and paired TAHN tissues 1, 3, and 5 cm from the tumor (TAHN-1, TAHN-3, and TAHN-5, respectively). Although the percentage of HMECs with high hTERT promoter activities varied 90- and 14-fold, respectively, between independent tumor and TAHN-1 tissues, the average difference in their prevalence within paired tumor and TAHN-1 tissues from the same patient was only 1.9-fold and 3 of the 5 pairs differed by less than 20%. This indicates that the factors modulating hTERT expression are remarkably constant between paired tumor and TAHN-1 tissues and independent of the absolute magnitude of the effect. In contrast, the population of HMECs with high hTERT mRNA expression is reduced by more than 90% at greater distances from the tumor margin. Only about 1% of HMECs in TAHN-3 or TAHN-5 tissues had hTERT promoter activities suggestive of high hTERT mRNA expression. The observation of consistent, substantial differences in hTERT promoter activity in HMECs from patient-matched tissues, grown in parallel, in which proximity to the tumor was the only variable, argues strongly against the possibility that the results are due to variations in short-term culture conditions. Furthermore, the spatial relationship between hTERT promoter activity and proximity to the tumor is identical to that reported previously for telomere length and allelic imbalance in bulk breast tissues (3).

By sorting cells based on hTERT promoter activity, we further showed that more than 95% of hTERT mRNA reside within a subpopulation comprising no more than 10% of the HMECs in the culture. This was true of both HMECs derived from tumor and TAHN-1 tissues. Moreover, the subpopulation of HMECs with the highest hTERT promoter activity also was significantly enriched in genes that function in cell cycle and mitosis, suggesting a functional difference, rather than arbitrary changes in gene expression. Approximately 15% (55 of 363) of the genes overexpressed in the HMECs derived from TAHN-1 tissues with the highest hTERT promoter activities are also overexpressed in HMECs derived from tumor tissues with the highest hTERT promoter activities, emphasizing the similarities between tumor and TAHN-1 tissues. These results imply that there is a significant reservoir of HMECs with extended proliferative potential within histologically normal breast tissues proximal to tumors.

It is important to emphasize that only cells that contained DsRed were included in the analyses. Therefore, all results are based on cells that contained StopLight vector and not affected by variation in infection efficiency between independent cultures. It is striking that these cells, once dispersed from the confines of the microenvironment and passaged several times in culture, have retained a phenotype that can be traced back to their origin and proximity to the tumor. This indicates that the cellular phenotypes defining the “field effect” are stable and persist even after removal from breast microenvironment.

We and others have also shown that the histologically normal tissues adjacent to tumors have a wound healing microenvironment (7), which is known to support tumorigenesis (9). In view of our results reported here and our previous results described above, it is increasingly likely that genetically unstable cell clones with extended proliferative capacity in a tumorigenic microenvironment exist within histologically normal breast tissues proximal to tumors and that these cells provide fertile ground for tumor development. In this context, several authors have proposed that some of the molecular changes necessary to develop an invasive phenotype may be acquired early in tumor progression or even preexist in seemingly normal cells (5, 6). For example, Kuperwasser and colleagues have reported that HMECs obtained from RM can give rise to “outgrowths that are indistinguishable from invasive human breast carcinomas” when implanted into humanized murine mammary fat pads (24). The authors concluded that these seemingly normal HMECs had already undergone one or several premalignant changes before they were implanted in the mice. Consistent with the possibility that precancerous epithelial cells exist in seemingly normal breast tissues, we have previously reported that a substantial fraction of the sites of allelic imbalance in TAHN-1 tissues are conserved in the patient-matched tumor, implying cellular clonal evolution (3). Our present experiments do not allow us to determine whether the tumor was spawned by this reservoir of genetically unstable cells or the reservoir developed in response to the tumor. We also do not know how telomere length and allelic imbalance segregate between HMECs with different levels of hTERT expression. However, the method of isolating hTERT-expressing cells from TAHN tissue described here provides a foundation for future investigations of the tumorigenic potential of these hTERT-expressing cells.

The potential existence of genetically unstable cells with extended replicative capacity, within a tumorigenic microenvironment existing beyond the recognizable tumor margin, has important clinical implications. These HMECs may represent immortalized cells in the early stages of tumorigenesis that, through subsequent clonal selection, spawn new tumors. Furthermore, our recent study shows that TAHN-1 tissue exhibits a wound healing microenvironment (8), which has previously been shown to drive tumor progression (9) and may increase the tumorigenic potential of the hTERT-expressing cells. Therefore, the presence of these cells after surgical resection of primary tumors may represent an ongoing risk factor for cancer recurrence or formation of secondary lesions, which occurs in up to 22% of women undergoing breast conservation therapies for small invasive and noninvasive breast cancers. Thus, this investigation also has practical and significant implications for the assessment of appropriate surgical margins.

No potential conflicts of interest were disclosed.

The authors thank UNM Flow Cytometry and Keck-UNM Genomics Resources for contributions to this investigation.

The study was supported by grants from American Cancer Society PF-08-022-01, DOD BCRP DAMD 17-02-1-0514, NCRR RR0164880, and NCI P30CA118110. The shared resources are supported by the University of New Mexico Health Sciences Center and the University of New Mexico Cancer Center

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.

1.
Slaughter
DP
,
Southwick
HW
,
Smejkal
W
. 
Field cancerization in oral stratified squamous epithelium; clinical implications of multicentric origin
.
Cancer
1953
;
6
:
963
8
.
2.
Chai
H
,
Brown
RE
. 
Field effect in cancer-an update
.
Ann Clin Lab Sci
2009
;
39
:
331
7
.
3.
Heaphy
CM
,
Bisoffi
M
,
Fordyce
CA
,
Haaland
CM
,
Hines
WC
,
Joste
NE
, et al
Telomere DNA content and allelic imbalance demonstrate field cancerization in histologically normal tissue adjacent to breast tumors
.
Int J Cancer
2006
;
119
:
108
16
.
4.
Meeker
AK
,
Hicks
JL
,
Gabrielson
E
,
Strauss
WM
,
De Marzo
AM
,
Argani
P
. 
Telomere shortening occurs in subsets of normal breast epithelium as well as in situ and invasive carcinoma
.
Am J Pathol
2004
;
164
:
925
35
.
5.
Meeker
AK
,
Hicks
JL
,
Iacobuzio-Donahue
CA
,
Montgomery
EA
,
Westra
WH
,
Chan
TY
, et al
Telomere length abnormalities occur early in the initiation of epithelial carcinogenesis
.
Clin Cancer Res
2004
;
10
:
3317
26
.
6.
Meeker
AK
,
Argani
P
. 
Telomere shortening occurs early during breast tumorigenesis: a cause of chromosome destabilization underlying malignant transformation?
J Mammary Gland Biol Neoplasia
2004
;
9
:
285
96
.
7.
Troester
MA
,
Lee
MH
,
Carter
M
,
Fan
C
,
Cowan
DW
,
Perez
ER
, et al
Activation of host wound responses in breast cancer microenvironment
.
Clin Cancer Res
2009
;
15
:
7020
8
.
8.
Trujillo
KA
,
Heaphy
CM
,
Mai
M
,
Vargas
KM
,
Jones
AC
,
Vo
P
, et al
Markers of fibrosis and epithelial to mesenchymal transition demonstrate field cancerization in histologically normal tissue adjacent to breast tumors
.
Int J Cancer
2010
Nov 23
.
[Epub ahead of print]
.
9.
Schafer
M
,
Werner
S
. 
Cancer as an overhealing wound: an old hypothesis revisited
.
Nat Rev Mol Cell Biol
2008
;
9
:
628
38
.
10.
Dvorak
HF
. 
Tumors: wounds that do not heal. Similarities between tumor stroma generation and wound healing
.
N Engl J Med
1986
;
315
:
1650
9
.
11.
Hines
WC
,
Fajardo
AM
,
Joste
NE
,
Bisoffi
M
,
Griffith
JK
. 
Quantitative and spatial measurements of telomerase reverse transcriptase expression within normal and malignant human breast tissues
.
Mol Cancer Res
2005
;
3
:
503
9
.
12.
Bodnar
AG
,
Ouellette
M
,
Frolkis
M
,
Holt
SE
,
Chiu
CP
,
Morin
GB
, et al
Extension of life-span by introduction of telomerase into normal human cells
.
Science
1998
;
279
:
349
52
.
13.
Yang
G
,
Rosen
DG
,
Mercado-Uribe
I
,
Colacino
JA
,
Mills
GB
,
Bast
RC
 Jr
, et al
Knockdown of p53 combined with expression of the catalytic subunit of telomerase is sufficient to immortalize primary human ovarian surface epithelial cells
.
Carcinogenesis
2007
;
28
:
174
82
.
14.
Elenbaas
B
,
Spirio
L
,
Koerner
F
,
Fleming
MD
,
Zimonjic
DB
,
Donaher
JL
, et al
Human breast cancer cells generated by oncogenic transformation of primary mammary epithelial cells
.
Genes Dev
2001
;
15
:
50
65
.
15.
Hahn
WC
,
Counter
CM
,
Lundberg
AS
,
Beijersbergen
RL
,
Brooks
MW
,
Weinberg
RA
. 
Creation of human tumour cells with defined genetic elements
.
Nature
1999
;
400
:
464
8
.
16.
Smith
LL
,
Coller
HA
,
Roberts
JM
. 
Telomerase modulates expression of growth-controlling genes and enhances cell proliferation
.
Nat Cell Biol
2003
;
5
:
474
9
.
17.
Speirs
V
,
Green
AR
,
Walton
DS
,
Kerin
MJ
,
Fox
JN
,
Carleton
PJ
, et al
Short-term primary culture of epithelial cells derived from human breast tumours
.
Br J Cancer
1998
;
78
:
1421
9
.
18.
Takakura
M
,
Kyo
S
,
Kanaya
T
,
Hirano
H
,
Takeda
J
,
Yutsudo
M
, et al
Cloning of human telomerase catalytic subunit (hTERT) gene promoter and identification of proximal core promoter sequences essential for transcriptional activation in immortalized and cancer cells
.
Cancer Res
1999
;
59
:
551
7
.
19.
Gu
J
,
Kagawa
S
,
Takakura
M
,
Kyo
S
,
Inoue
M
,
Roth
JA
, et al
Tumor-specific transgene expression from the human telomerase reverse transcriptase promoter enables targeting of the therapeutic effects of the Bax gene to cancers
.
Cancer Res
2000
;
60
:
5359
64
.
20.
Tusher
VG
,
Tibshirani
R
,
Chu
G
. 
Significance analysis of microarrays applied to the ionizing radiation response
.
Proc Natl Acad Sci U S A
2001
;
98
:
5116
21
.
21.
Wege
H
,
Chui
MS
,
Le
HT
,
Tran
JM
,
Zern
MA
. 
SYBR Green real-time telomeric repeat amplification protocol for the rapid quantification of telomerase activity
.
Nucleic Acids Res
2003
;
31
:
E3
3
.
22.
Lewis
CM
,
Cler
LR
,
Bu
DW
,
Zöchbauer-Müller
S
,
Milchgrub
S
,
Naftalis
EZ
, et al
Promoter hypermethylation in benign breast epithelium in relation to predicted breast cancer risk
.
Clin Cancer Res
2005
;
11
:
166
72
.
23.
Tripathi
A
,
King
C
,
de la Morenas
A
,
Perry
VK
,
Burke
B
,
Antoine
GA
, et al
Gene expression abnormalities in histologically normal breast epithelium of breast cancer patients
.
Int J Cancer
2008
;
122
:
1557
66
.
24.
Kuperwasser
C
,
Chavarria
T
,
Wu
M
,
Magrane
G
,
Gray
JW
,
Carey
L
, et al
Reconstruction of functionally normal and malignant human breast tissues in mice
.
Proc Natl Acad Sci U S A
2004
;
101
:
4966
71
.