PTEN, a Putative Protein Tyrosine Phosphatase Gene Mutated in Human Brain, Breast, and Prostate Cancer

See allHide authors and affiliations

Science  28 Mar 1997:
Vol. 275, Issue 5308, pp. 1943-1947
DOI: 10.1126/science.275.5308.1943


Mapping of homozygous deletions on human chromosome 10q23 has led to the isolation of a candidate tumor suppressor gene, PTEN, that appears to be mutated at considerable frequency in human cancers. In preliminary screens, mutations of PTEN were detected in 31% (13/42) of glioblastoma cell lines and xenografts, 100% (4/4) of prostate cancer cell lines, 6% (4/65) of breast cancer cell lines and xenografts, and 17% (3/18) of primary glioblastomas. The predicted PTEN product has a protein tyrosine phosphatase domain and extensive homology to tensin, a protein that interacts with actin filaments at focal adhesions. These homologies suggest that PTEN may suppress tumor cell growth by antagonizing protein tyrosine kinases and may regulate tumor cell invasion and metastasis through interactions at focal adhesions.

As tumors progress to more advanced stages, they acquire an increasing number of genetic alterations. One alteration that occurs at high frequency in a variety of human tumors is loss of heterozygosity (LOH) at chromosome 10q23. This change appears to occur late in tumor development: although rarely seen in low-grade glial tumors and early-stage prostate cancers, LOH at 10q23 occurs in ∼70% of glioblastomas (the most advanced form of glial tumor) and ∼60% of advanced prostate cancers (1, 2). This pattern of LOH, and the recent finding that wild-type chromosome 10 suppresses the tumorigenicity of glioblastoma cells in mice, suggest that 10q23 encodes a tumor suppressor gene (3).

To identify this putative tumor suppressor gene, we performed representational difference analysis (RDA) on 12 primary breast tumors (4). A probe, CY17, derived from one of the tumors was mapped to chromosome 10q23, near markers WI-9217 and WI-4264, on the Whitehead-MIT radiation hybrid map (5). To map the location of CY17 more precisely, we isolated three yeast artificial chromosomes (YACs) containing CY17 that are present on the sequence tagged site (STS)-based map of the human genome (6, 7). These YACs placed CY17 slightly centromeric to the position determined by radiation hybridization and precisely identified its location (Fig. 1A). Analysis of 32 primary invasive breast cancers revealed LOH in this region in about 50% of the samples. No homozygous deletions of CY17 were detected in a panel of 65 breast tumor cell lines (25) and xenografts (40) (8), so eight additional markers were analyzed in the 10q23 region (D10S579, D10S215, AFMA086WG9, D10S541, AFM280WE1, WI-10275, WI-8733, WI-6971). We identified homozygous deletions of AFMA086WG9 in two xenografts, Bx11 and Bx38 (Figs. 1B and 2A) and then screened a bacterial artificial chromosome (BAC) library with this marker (9). Using new STSs from four independent BAC clones, we determined that the minimal region of deletion was within BAC C (Fig. 1B) (10). Homozygous deletions of AFMA086WG9 were also detected in two of eight glioblastoma cell lines, three of 34 glioblastoma xenografts, and two of four prostate cancer cell lines (11). One of the glioblastoma samples, cell line A172, had the same deletion pattern as the original breast xenografts; the deletions in the other samples were larger (Fig. 1B).

Fig. 1.

Region of homozygous deletion on chromosome 10q23. (A) The STS-based YAC map of the region surrounding CY17. Marker locations are taken from the Whitehead STS-based map. RH indicates the radiation hybrid interval for CY17. CY17 positive YAC addresses are indicated. YAC map indicates the interval containing CY17 inferred from the YAC addresses. Cen., centromere; Tel., telomere. (B) Map of homozygous deletions on 10q23, showing the STS markers spanning the deleted region, the four BACs overlapping the region, and the location of PTEN with respect to the STS markers. STS markers Not-5′, PTPD, and ET-1 contain exonic sequences of the PTEN gene. Absence of homozygous deletion is indicated with a “+” and presence of homozygous deletion with a “−.” Numbers to the right indicate the fraction of tumor cell lines and xenografts with the deletion. The two breast cancer samples with a deletion are xenografts Bx11 and Bx38. Glioblastoma line A172 has a deletion encompassing markers JL25 through KP8 and glioblastoma line DBTRG-05MG has a deletion affecting only ET-1. The glioblastoma samples with a deletion across the entire region are the cell line U105 and xenografts 2, 3, and 11, and the samples with deletion of only PTPD, which contains the phosphatase domain, are xenografts 22, 23, 24, 25, and 32. The prostate cancer cell lines with homozygous deletions are NCI H660 and PC-3. The 5′ end of the PTEN cDNA was determined to be coincidental with the Not I site 20 kb from the centromeric end of BAC D by sequence analysis. These maps are not drawn to scale.

Fig. 2.

Homozygous deletions in tumor cell lines and xenografts. (A) A 6% polyacrylamide sequencing gel showing the products of PCR amplification of AFMA086WG9 from breast cancer cell lines (lanes 1 to 4) and xenografts (lanes 5 to 8). Lane 1, MDA-MB-330; lane 2, MDA-MB-157; lane 3, MDA-MB-134-VI; and lane 4, MDA-MB-435S; lane 5, Bx11; lane 6, Bx15; lane 7, Bx38; and lane 8, Bx39. (B) Southern blot analysis of tumor xenografts. Genomic DNA was digested with Eco RI, the fragments resolved on a 1% agarose gel, and transferred to a nylon membrane. The blot was probed with a 3-kb Eco RI fragment containing the STS marker JL25, which is within the region of homozygous deletion (top), or to a second 2-kb Eco RI fragment from chromosome 8 (bottom). Lane M, bacteriophage lambda Hind III marker. Other lanes contain DNA from breast xenografts 10, 11, 19, and 38 and brain xenografts 2, 3, and 11. Breast xenografts 10 and 19 were loaded as controls and were not expected to have homozygous deletions. (C) Homozygous deletions of exon ET-1 in glioblastoma cell lines. Genomic DNA samples were PCR amplified using intronic primers that amplify exon ET-1. The products were resolved on a 1.2% agarose gel and then stained with ethidium bromide. Lane 1 contains a DNA marker. The remaining lanes contain PCR products from control templates and seven glioblastoma cell lines: lane 2, lymphocyte DNA; lane 3, water; lane 4, U118MG; lane 5, A172; lane 6, DBTRG-05MG; lane 7, U373; lane 8, T-98G; lane 9, U-87MG; and lane 10, U138MG. Full-length products are present for all templates except water, A172, and DBTRG-05MG.

To confirm the presence of homozygous deletions, we hybridized a Southern (DNA) blot with a 3-kb probe derived from a genomic clone spanning the region of deletion (12). Xenografts anticipated to have a homozygous deletion did not hybridize to this probe; the control xenografts hybridized to the expected 3-kb band (Fig. 2B).

We identified genes within the 10q23 region by exon trap analysis of BACs C and D (Fig. 1B) (13). Two trapped exons, ET-1 and ET-2, had sequences that were perfect matches to an unmapped UNIGENE assembly of expressed sequence tags (ESTs) as well as several unassembled ESTs (6). Clones containing the ESTs were sequenced and used to assemble an open reading frame (ORF) of 403 amino acids (Fig. 3A). To verify the location of this cDNA, we obtained the intronic sequence around ET-1 by directly sequencing BAC C. An STS primer pair (ET-1) was generated that mapped back to BACs A, B, and C (Fig. 1B). In addition, we screened the Map Panel #2 monochromosome human-rodent hybrid panel to confirm the unique location of this exon on chromosome 10 (14).

Fig. 3.

(A) Predicted amino acid sequence of P-TEN. The putative phosphatase domain is underlined. The nucleotide sequence has been deposited in GenBank (accession number U93051). Abbreviations for amino acids are A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr. (B) Homology of P-TEN to protein tyrosine phosphatases. The sequence alignment was performed by ClustalW ( The National Center for Biotechnology Information (NCBI) ID numbers are P53916 (Y50.2), M61194 (CDC14), A56059 (PRL1), 1246236 (PTP-IV1), 1125812 (CPTPH), and P24656 (BVP). Black boxes indicate amino acid identities and gray boxes indicate similarities. (C) Homology of P-TEN to chicken tensin and bovine auxilin. Alignment was performed as in (B) over the region of highest homology. NCBI ID numbers are A54970 (tensin) and 485269 (auxilin).

Our entire panel of tumor xenografts and cell lines was screened with this primer pair, and we identified an additional glioblastoma cell line (DBTRG-05MG) with a deletion of 180 base pairs (bp) (Fig. 1B) (Fig. 2C). Sequence analysis revealed that the deletion had removed 180 bp of exonic sequence and the splice donor site from this 225-bp exon. This deletion was not present in 52 normal blood samples or in more than 125 other primary tumors, xenografts, and cell lines tested.

Sequence analysis of the ORF revealed a protein tyrosine phosphatase domain (Fig. 3B) and a large region of homology (∼175 amino acids) to chicken tensin and bovine auxilin (Fig. 3C). We therefore call the gene PTEN for Embedded Imagehosphatase and Embedded Imagesin homolog deleted on chromosome Embedded Image. The phosphatase domain of the P-TEN protein contained the critical (I/V)-H-C-X-A-G-X-X-R-(S/T)-G motif found in tyrosine and dual-specificity phosphatases (15). The phosphatase domain exon mapped within all four BACs and was deleted in all of the samples with homozygous deletions except for DBTRG-05MG. These results thus placed this exon within the region of homozygous deletion near JL25 and AFMA086WG9 (Fig. 1B). We then screened the remaining xenografts and cell lines for additional homozygous deletions and identified five more glioblastoma xenografts lacking this exon. These data indicate that the phosphatase domain encoded by PTEN was targeted for mutations in tumor xenografts and cell lines.

The phosphatase domain of P-TEN is most related in sequence to those of CDC14, PRL-1 (phosphatase of regenerating liver), and BVP (baculovirus phosphatase) (Fig. 3B). CDC14 and BVP are dual-specificity phosphatases that remove phosphate groups from tyrosine as well as serine and threonine (16). These phosphatases can be distinguished from the better characterized VH1-like enzymes by sequence differences outside of the core conserved domain. Both PRL-1 and CDC14 are involved in cell growth, and CDC14 appears to play a role in the initiation of DNA replication (17). In contrast to P-TEN, these phosphatases do not have extensive homology to tensin and auxilin. P-TEN is also homologous to the protein tyrosine phosphatase domains of three ORFs (Y50.2, PTP-IV1, CPTPH) whose protein products have not been characterized. Of these hypothetical proteins, only the putative yeast phosphatase Y50.2 has significant homology to tensin. Although tensin and auxilin are not expected to have phosphatase activity, they both contain elements of the protein tyrosine phosphatase signature sequence (18), which suggests that they may share a tertiary structure with these enzymes (19).

If PTEN is a tumor suppressor gene, the PTEN allele retained in tumor cells with LOH should contain inactivating mutations. To search for such mutations, we performed a protein truncation test on 20 breast, six glioblastoma, and two prostate tumor cell lines (20). Two truncating mutations in PTEN were identified in the breast samples (Table 1). BT549 cells had a 1-bp deletion of a G, leading to the formation of a stop codon TAA (Fig. 4A), and MDA-MB-468 cells had a deletion of 44 bp at codon 70, which resulted in a frameshift on the amino terminal side of the tyrosine phosphatase domain. Mutations in PTEN were also identified in three of the six glioblastoma cell lines: DBTRG-05MG cells had an in-frame deletion of 204 bp caused by the genomically deleted exon ET-1 (Fig. 4B), U373MG had a 2-bp insertion at codon 242, and U87MG had a frameshift at codon 54. Both of the prostate tumor cell lines had PTEN mutations: LNCaP cells had a 2-bp deletion at codon 6, leading to a frameshift (Fig. 4C), and DU145 cells had a Met → Leu substitution at codon 134, within the phosphatase domain. The latter mutation was detected by a change in the pattern of in vitro translation initiation and was not found in >50 other alleles tested. However, Met-134 is not required for phosphatase activity (Fig. 3B), so this alteration could be a polymorphism. With one exception (DU145), all of the cell lines retained a mutant PTEN allele and lost the other allele, indicating that these cells are null for PTEN.

Table 1.

Summary of PTEN mutations in tumor cell lines and primary tumors.

View this table:
Fig. 4.

Mutations of PTEN in cancer cell lines and primary tumors. (A) Mutation in breast cancer cell line BT549. Sequence of nucleotides 831 to 785 (bottom to top) using an antisense primer shows a deletion of a C (arrow) in sample on the right (BT549) but not in the sample on the left (breast cancer cell line ZR-75-30). (B) Mutation in glioblastoma cell line DBTRG-05MG. Sequence of nucleotides 1039 to 1010 in the antisense orientation from prostate cancer cell lines DU145 (left), LNCaP (middle), and the glioblastoma cell line DBTRG-05MG (right). Arrow indicates the in-frame deletion of nucleotides 822 to 1025 in DBTRG-05MG. (C) Mutation in prostate cancer cell line LNCaP. Sequence of nucleotides 34 to 2 of the prostate cancer cell line LNCaP (left) and the glioblastoma cell line DBTRG-05MG (right) using an antisense primer. Arrow indicates the deletion of two T nucleotides in LNCaP. (D) Mutation in primary glioblastoma 534. Sequence of nucleotides 26 to 63 of genomic DNA from blood (left) and primary tumor 534 (right) from the same patient using a sense primer. Arrow indicates insertion of AG in the tumor DNA. A, C, G, and T lanes are loaded next to each other to allow better detection of mutations.

To determine whether PTEN mutations are present in primary tumors, we screened genomic DNA from 18 primary glioblastomas for mutations in three exons (21). Mutations in PTEN were found in three of these tumors: a 2-bp insertion at codon 15 (534 T), a point mutation resulting in a Gly → Arg change at codon 129 (132T), and a 4-bp frameshift mutation at codon 337 (134T) (Table 1 and Fig. 4D). The mutation at codon 129 is within the signature sequence for tyrosine phosphatases (Fig. 3B). All three tumors appeared to have LOH in the PTEN region since the wild-type allele was substantially reduced in intensity. In addition, the tumor mutations were not detected in paired blood DNA.

In summary, we detected homozygous deletions, frameshift, or nonsense mutations in PTEN in 63% (5/8) of glioblastoma cell lines, 100% (4/4) of prostate cancer cell lines, and 10% (2/20) of breast cancer cell lines. These frequencies are likely to be underestimates since the cell lines were not systematically screened for point mutations. We screened xenografts only for homozygous deletions in PTEN and detected them in 24% (8/34) of glioblastoma xenografts and 5% (2/40) of breast cancer xenografts. Finally, we detected PTEN mutations in 17% (3/18) of primary glioblastomas; this frequency is also likely to be an underestimate since the entire coding sequence was not analyzed. The results of these preliminary screens suggest that a large fraction of glioblastomas and advanced prostate cancers may harbor PTEN mutations, whereas the mutation frequency in breast cancer may be lower. Future systematic analysis of all tumor types will be of interest.

The likely function of the P-TEN tumor suppressor as an enzyme that removes phosphate from tyrosines is intriguing, given that many oncoproteins function in the reverse process—to phosphorylate tyrosines (22). P-TEN and tyrosine kinase oncoproteins may share substrates and the tight control of these substrates through phosphorylation is likely to regulate a critical pathway that is altered late in tumor development. The homology of P-TEN to tensin is also of interest. Tensin appears to bind actin filaments at focal adhesions—complexes that contain integrins, focal adhesion kinase (FAK), Src, and growth factor receptors (23). Integrins have been implicated in cell growth regulation (24) and in tumor cell invasion, angiogenesis, and metastasis (25), so it is conceivable that PTEN regulates one or more of these processes. Finally, the identification of P-TEN as a likely tumor suppressor raises the possibility that this protein and its substrates will be useful targets for the development of new therapeutics for cancer.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
    BACs C and D were digested with Bam HI, Bgl II, or both enzymes and then ligated into the trapping vector pSPL3. Libraries were transfected into COS1 cells with lipofectamine and polyadenylated RNA was extracted after 2 days. An exon trapping kit was purchased from GIBCO/BRL.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
    Randomly primed cDNA was prepared from each of the cell lines studied. Reverse transcription (RT)-PCR reactions were performed with two overlapping primer pairs to screen the entire ORF. Primer pairs were as follows: 5′-GGATCCTAATACGACTCACTATAGGGAGACCACCATGGAGTCGCCTGTCACCATTT-C-3′ and 5′-TTCCAGCTTTACAGTGAATTG-3′; 5′-GGATCCTAATACGACTCACTATAGGGAGACCACCATGGGATTTCCTGCAGAAAG-3′and 5′-TTTTTTCATGGTGTTTTATCCCTC-3′. In vitro transcription and translation were performed with the T7 TNT kit (Promega, Madison, WI) and the translation products resolved by electrophoresis. The RT-PCR products that generated truncated proteins were directly cycle sequenced (20 ng each) to identify potential mutations. All mutations were verified by repeating the RT-PCR and mutation analysis. Complementary DNA primer sequences are available upon request.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.

Stay Connected to Science

Navigate This Article