The telomerase enzyme is a reverse transcriptase capable of replacing the telomeric DNA sequences that are lost at each cell division. Telomerase activation permits extended cell proliferation beyond normal senescence checkpoints, and accordingly, telomerase activity has been detected in a wide range of malignant cells and tissues but is absent in terminally differentiated somatic cells. To date, the majority of cancer-related telomerase analyses have been performed on carcinomas that originate from epithelial cells, and few reports have included tumors originating from nonepithelial cells. In this study, we used the PCR-based telomeric repeat amplification protocol (TRAP) to assay telomerase activity in nuclear protein extracts obtained from a range of malignant and benign connective tissue lesions. In total, 62 histologically diagnosed specimens were analyzed including 37 sarcomas, 7 benign mesenchymal tumors, 12 normal tissue samples, and 6 carcinoma metastases obtained from bone. Thirty (81%) of the 37 primary sarcoma samples contained telomerase activity, and four of the six carcinoma metastases were also positive. Conversely, telomerase activity was detectable in only one of seven benign lesions and in none of the 12 normal connective tissue controls. Tumors of connective tissue origin can sometimes be difficult to categorize and to evaluate microscopically with regard to clinical management. As is the case in carcinomas, the presence of telomerase activity appears to be indicative of malignancy in mesenchymal tumor biopsy material and therefore may be useful as an adjunct to the pathologist in the assessment of borderline cases.
Sarcomas are malignant mesenchymal tumors. Mesenchymal tissue is defined as the complex of nonepithelial structures of the body exclusive of reproductive tissue, glia, and hematopoietic and lymphoid tissue. These are derived embryonically from mesoderm and, to a lesser degree, from neuroectoderm and are grouped into fibrous tissue, adipose tissue, skeletal muscle, blood and lymph vessels, and peripheral nervous tissue.
Sarcomas constitute a large and heterogeneous group of neoplasms. Accurate assessment of these lesions is crucial because a misdiagnosis of sarcoma can lead to serious consequences for the patient, such as unnecessary amputation. The availability of a test using a molecular biomarker to aid reliability in interpretation of mesenchymal tumors would be clinically very valuable. Specimens used to investigate suspicious soft-tissue lesions by histology are usually obtained by needle core biopsy or after surgical excision, but it is possible to obtain specimens with the less invasive procedure of fine-needle aspiration biopsy (1) . Less invasive sampling is clearly preferable, but any potential molecular analyte would then need to be reliably detectable in minimal amounts of material.
The telomerase enzyme is a promising candidate marker for the detection of a wide spectrum of neoplasias (2, 3, 4, 5) because activation of the enzyme is thought to be necessary for cells to become immortal, or at least capable of extended proliferation (6 , 7) . The termini of eukaryotic chromosomes are capped with up to 30 kb of DNA composed of a simple hexanucleotide repeat (TTAGGG; Ref. 8 ). These telomeric regions function to maintain chromosome structural integrity (8, 9, 10) and act as a buffer zone, preventing the loss of vital coding sequences during the gradual erosion of terminal DNA that occurs at each cell division (11 , 12) . It has been suggested that telomere shortening may be part of a “mitotic clock” mechanism that limits uncontrolled cell proliferation by initiating a signal for cells to exit from the cell cycle and enter a senescent state (13 , 14) . Avoidance of normal proliferative control is a defining characteristic of tumor growth, and the restoration of telomeric repeats by the activation of the telomerase enzyme may be one way in which tumor cells avoid senescence.
The development of a highly sensitive PCR-based assay, i.e., TRAP,2 allows the detection of telomerase activity in samples obtained from as little as 10 cells (15) . The use of this assay has shown that telomerase activity is detectable in all but a few human cancer cell lines and in the majority of human carcinomas tested (5 , 16, 17, 18) , whereas it is rare in normal tissues (15 , 18) . The analysis of telomerase activity status has potential clinical utility for diagnosis, screening, and monitoring treatment, but it is yet to be validated in specimens from solid tumors of nonepithelial origin, i.e., sarcomas. In this study, we have evaluated whether the presence of telomerase activity is indicative of malignant tumor cells in a variety of soft-tissue sarcomas and other nonepithelial tissue lesions.
MATERIALS AND METHODS
Tissue samples were obtained from surgical specimens resected in the Royal Orthopaedic Hospital (Birmingham, United Kingdom). Tissues were snap-frozen immediately after resection and stored in liquid nitrogen until analysis. Histological diagnosis of all samples was determined by consensus opinion of three independent pathologists of H&E-stained, cryostat-cut sections, and all soft-tissue sarcomas were graded using the system of Trojani et al. (19) . A total of 62 samples were obtained from 37 primary bone and soft-tissue sarcomas (see Table 1⇓ for clinical and histological details) in patients of ages 10–82 years (average age, 40 years), of whom 62% were female. For comparison, protein extracts from 7 benign lesions, 6 carcinomatous bone metastases, and from 12 normal mesenchymal tissue biopsies were also evaluated.
Twenty, 10-μm cryostat sections were mixed with 50 μl of ice-cold lysis buffer [10 mm Tris-HCl (pH 7.5), 1 mm MgCl2, 1 mm EGTA, 0.1 mm phenylmethylsulfonyl fluoride, 5 mm β-mercaptoethanol, 0.5% 3-[(3-cholamidopropyl)dimethylamino]-1-propanesulfonate, and 10% glycerol] and extracted for 30 min on ice. The lysate was centrifuged at 10,000 × g for 20 min at 4°C, and the supernatant was immediately frozen at −70°C. The protein content in each lysate was measured and adjusted to 3 μg/μl.
Telomerase activity was assayed by the TRAP method as described by Kim et al. (15) . Briefly, 6 μg of cell extract protein (2 μl) were incubated in 50 μl of reaction mixture [20 mm Tris-HCl (pH 8.3), 1.5 mm MgCl2, 63 mm KCl, 0.005% Tween 20, 1 mm EGTA, 50 μm deoxynucleotide triphosphates including 32P-labeled dCTP (Amersham, Arlington Heights, IL), 0.1 μg of TS oligonucleotide (5′-AATCCGTCGAGCAGAGTT-3′), and 1μ g of T4 gene-32 protein (Boehringer Mannheim, Indianapolis, IN), 0.1 μg/ml BSA (Sigma Chemical Co., St. Louis, MO; Boehringer Mannheim, Indianapolis, IN)] at 20°C for 30 min for telomerase-mediated extension of TS primers. After heating the mixture at 90°C for 3 min to inactivate telomerase, 1 μg of CX primer[ 5′-(CCCTTA)3 CCCTAA-3′], 2 units of Taq DNA polymerase, and 10 attograms of an ITAS (20) were added, and the mixture was subjected to 31 cycles of: 94°C for 45 s, 50°C for 45 s, and 72°C for 90 s. Sixteen μl of each PCR reaction were analyzed by electrophoresis on 12% polyacrylamide nondenaturing gels. The gel was dried and processed for autoradiographic exposure. To test the specificity of telomerase-positive assays, 1 μg of DNase-free RNase (Promega Corp., Madison, WI) was added to 5 μl of each sample, and after incubation for 20 min at 37°C, 2 μl of the treated sample was applied to the TRAP assay. The inclusion of an ITAS, a cDNA that is amplified with the TS and CX primers to generate a 150-bp product, aids the detection of false-negatives, which can result from the presence of PCR inhibitors in the extracts (20) . The criterion for a positive TRAP assay was a hexanucleotide ladder of three or more bands that were absent from the negative controls of: (a) no sample protein; and (b) the RNase-digested sample.
Malignant cases in this study represent some of the more common histological types of sarcomas (Table 1)⇓ . A representative TRAP assay is depicted in Fig. 1⇓ . The presence of active telomerase in a sample is revealed by a characteristic ladder of products, created by PCR amplification of the DNA synthesized by the enzyme that is entirely composed of 6-bp TTAGGG tandem repeats. All samples were evaluated using 6 and 0.6 μg of total protein to confirm telomerase status but also to reveal by dilution any interference of the assay by polymerase inhibitors. RNase treatment of extracts prior to the assay reaction always negated positive assays, thereby confirming a ribonucleoprotein activity.
Thirty (81%) of the 37 sarcoma samples contained telomerase activity as evaluated by the TRAP assay (Table 1)⇓ . The seven telomerase-negative samples consisted of two of nine myxoid malignant fibrous histiocytomas tested (both being grade 3 lesions), one of three malignant schwannomas, one grade 2 lesion of nine leiomyosarcomas, a chondrosarcoma, a low-grade fibromyxoid sarcoma, and the lipoma-like liposarcoma. No particular multicase subgroup had a significant number of negative samples within it, and among all of the sarcoma samples, no correlation between the presence of telomerase activity and the stage or grade of lesion was revealed.
The overall proportion of telomerase-positive samples (81%) in this broad range of sarcoma subtypes is similar to the 70–90% values obtained in studies of carcinoma lesions (20) . Accordingly, telomerase activity was also detected in four (66%) of the six bone metastases derived from primary carcinomas. These secondary lesions are by far the commonest tumor found in bone, most often originating from breast, kidney, or prostate tumors. In contrast to the sarcoma lesion sampling, only one (14%) of the seven benign lesion samples contained telomerase activity, and none of the 12 samples from healthy muscle tissue were positive. Although no normal fat tissue was obtained from the liposarcoma cases evaluated in this study, we have previously evaluated normal mammary tissue that is primarily fatty tissue and found it to be invariably telomerase negative (2) . The single positive benign sample was obtained from a schwannoma case, a classification subgroup that also gave a negative assay in one case diagnosed as malignant. Schwannomas are truly encapsulated neoplasms of neural origin and are benign. Malignant schwannomas (malignant peripheral nerve sheath tumors) arise either de novo or, especially in patients with neurofibromatosis type 1, within neurofibromas.
The development of the sensitive TRAP assay (15 , 20) has enabled the evaluation of telomerase activity in many types of human cancers. The reported frequencies of positive TRAP assays of solid tumor samples are ∼90% (21) , and it is now widely accepted that, except for a few specialized cell types, telomerase activity in cells of somatic origin is indicative of immortal transformation. One of the major advantages of using this analyte is the ability to use minimal amounts of clinical sample material. This allows samples obtained by noninvasive collection methods to be analyzed for telomerase expression. Exfoliated cells collected from urine (3) , the colonic lumen (22) , or oral rinses (23) have been reliably analyzed. Furthermore, the use of samples obtained by fine needle aspiration of the thyroid (17) suggests that the TRAP assay would be applicable to material now being obtained from lesions of connective tissues using needle biopsy or aspiration (1) .
To date, few reports on telomerase activity in cancer have included connective tissue tumors. A study of 14 skeletal sarcoma specimens revealed telomerase activity in 57% of samples, but with the majority of osteosarcomas being telomerase negative (24) . More recently, Schneider-Stock et al. (25) investigated 60 soft-tissue sarcomas of different histology and six benign tumors for telomerase activity using the TRAP assay. They found telomerase activity in 38.3% of cases, with activity correlating with a more aggressive behavior of soft-tissue sarcoma according histologically determined grade of malignancy. Only one of the benign tumors showed telomerase activity, and they concluded that telomerase activity is a potential prognostic factor in malignant soft-tissue tumors. In another report, the same group examined a panel of myxoid liposarcomas from 19 patients (26) . They found telomerase expression in nine of nine myxoid/round cell liposarcomas and in 3 of 10 pure myxoid tumors, suggesting that high levels of telomerase expression and activity are associated with tumor progression from low-grade pure myxoid to higher-grade malignant round cell liposarcoma. In a study on the status of telomere shortening in sarcomas, the majority (70%) of leiomyosarcomas analyzed were telomerase positive (27) . In this study, the percentage of telomerase-positive sarcomas was high at 81% overall, but this may be a reflection of the predominance of high-grade tumors specimens procured and evaluated. Grade 3 lesions represented 64% of sarcomas sampled, and grades 2 and 3 constituted 83% of the total. The high telomerase-positive rate may reflect a trend of telomerase induction or activation with increasing tumor progression, as suggested by previous reports (25 , 26) .
Although telomerase activity is perhaps the single best molecular analyte for malignancy currently available, the assay can be problematical. The presence of inhibitors of either the telomerase enzyme or the polymerase used in PCR has been reported to be a potential problem in the analysis of a total protein cell extract (3 , 20) , resulting in the generation of false-negative assays. However, the inclusion of an internal TRAP assay standard (20) and the susceptibility of the enzyme to RNase degradation of its RNA moiety have helped to validate telomerase activity assay data. The fact that the telomerase holoenzyme has an essential RNA component is also problematical. Handling and homogenization of solid tissue specimens may lead to partial degradation of labile RNA and coupled with the low abundance of hTERT transcripts, any loss can be a serious concern, especially if attempting quantitation. False-positive assays can also be misleading. Telomerase is known to be active in stem cell lineages (15 , 28) , and the high sensitivity of the PCR-based TRAP assay can produce a positive signal if just a few stem cells “contaminate” the sample. There are currently attempts to overcome this problem by developing more quantitative versions of the assay.
Light microscopic evaluation of H&E-stained sections remains the standard technique for the diagnosis of these tumors and is usually sufficient. More specialized techniques have been successful in improving diagnostic accuracy. The use of tissue-related markers in immunohistochemistry is used to refine classification of these neoplasms, and the use of cytogenetics has shown the existence of nonrandom chromosomal alterations that can be used to identify and classify a subset of sarcoma lesions (29) . The critical evaluation of benign or malignant is of the greatest importance and is largely based on morphological criteria. Some tumors defy precise classification and are referred to as “borderline” tumors; it is the evaluation of these tumors that may benefit from the extra information obtained from molecular pathological techniques.
A number of potential markers have been studied, and differing results are obtained within the range of sarcoma subtypes. For example, bcl-2 expression is frequent in synovial and rhabdomyosarcoma but is negative in leiomyosarcomas (30) , and p53 accumulation appears to be more prevalent in high-grade tumors (31) . However, a study by Jensen et al. (32) found that such molecular markers were not as consistent as measurement of the proliferative index for the prognostic evaluation of a wide range of soft-tissue sarcomas.
The absence of telomerase activity may be a limiting factor for cellular proliferation. Conversely, the activity of the p53 gene product has been implicated as a barrier to extended proliferation, and alterations in cellular regulation associated with p53 have been described in up to 30% of specific sarcomas (33 , 34) . Because the signal for p53-mediated growth arrest is damaged DNA, it is conceivable that p53 could also signal cell growth arrest in response to critical telomere reduction, in which case there would be an association between the p53 status and telomerase activity in a cell. Aberrant expression of mdm-2, a protein thought to interfere with growth arrest by binding p53 in an inactive complex (35 , 36) , has been implicated in sarcomas (37 , 38) . However, re-expression of wild-type p53 in a human sarcoma cell line containing a missense p53 mutation did not alter telomerase activity (39) , indicating that there is no direct relationship between the functions of these genes. Murine fibroblast cell lines established from tissues of p53+/+ and p53−/− mice both expressed telomerase activity (39) , confirming that activation of telomerase can occur in cells regardless of p53 status.
Sarcomas are relatively rare tumors of the connective tissues that occur over a wide age range, do not have a morphologically recognizable in situ phase, and are conventionally thought to metastasize via the blood. Carcinomas originating from epithelial cells are far more common, they do have an in situ stage, and usually occur over the age of 50 years. However, the criteria for the distinction between benign and malignant are common to both types of lesion. Malignant tumors are highly proliferative, with irregular nuclei and poor differentiation. The lesion is often invasively endophytic with irregular borders and is therefore judged to be prone to metastasize. Even so, it can sometimes be difficult to recognize invasive behavior in a mesenchymal tumor because of the less precisely organized cellular alignment in these tissues, and nonmorphological markers therefore could be very helpful in some circumstances. The high mitotic activity and rapid growth suggest that the tumor cells have overcome the normal replicative checkpoints and become “immortal,” a situation requiring telomerase activation, and this is why telomerase is a particularly valuable marker for malignancy because it is applicable to so many tumor types.
This study demonstrates the feasibility of detecting malignancy in mesenchymal tumor samples by applying the TRAP assay, but further investigations are necessary to determine whether it will be reliable enough to be used as a routine clinical diagnostic assay for the presence of malignancy in these tumor specimens. In this study, specificity is good, with only one benign lesion testing positive, and the sensitivity is similar to that obtained for carcinoma samples (80–90%). Telomerase appears to be at least as good as any molecular analyte available for the diagnosis of malignancy in human tissues, and in combination with other markers, promises to be an invaluable adjunct to the pathologist in a wide range of tumor types.
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 To whom requests for reprints should be addressed, at University of California San Diego Cancer Center, 9500 Gilman Drive, La Jolla, CA, 92093-0912. Phone: (858) 822-2083; Fax: (858) 822-2084; E-mail:
↵2 The abbreviations used are: TRAP, telomeric repeat amplification protocol; ITAS, internal telomerase assay standard.
- Received July 12, 2000.
- Revision received September 18, 2000.
- Accepted September 20, 2000.