US 6,255,464 B1Grant
Mad-related genes in the human
Issue Date:2001-07-03
•3 Claims
•8 Drawing Sheets
Abstract
Five human genes related to the Mad gene of Drosophila were identified. One of these genes (Smad2) was found to reside at chromosome 18q21, adjacent to a previously described member of this family called DPC4 (Smad4). Smad2 was found to be somatically mutated in two of eighteen human colorectal cancers. Smad2 and Smad4 are important in the suppression of neoplasia by mediating the growth inhibitory effects of TGF-β-like ligands.
Metadata
Assignee
- The Johns Hopkins University
Inventors
- Bert Vogelstein
- Kenneth W. Kinzler
- Gregory R. Riggins
- Sam Thiagalingam
Application Information
Application Number:US 08/840,767
Filing Date:1997-04-16
Priority Date:1996-04-18
Art Unit:7
Classifications
IPC:
C07H 2102C07H 2104C07K 100
Field of Search:
536 231536 235530350
Patent Drawings (8 sheets)
Description
[0002] This application claims the benefit of co-pending provisional application Ser. No. 60/015,823, filed Apr. 18, 1996.
BACKGROUND OF THE INVENTION
[0004] Neoplasia involves the clonal expansion of cells which can escape the inhibitory influences that normally limit cell growth. The revolution in cancer research that has occurred over the last decade has largely involved the discovery of genes which confer these inhibitory influences and the delineation of mutations of these genes in various tumour types. TGF-β is one of the cytokines that can inhibit epithelial cell growth, and has been extensively studied for over a decade (1-5). Resistance to TGF-β is common in human cancers of many types, emphasizing the importance of this secreted polypeptide to the neoplastic process (6). However, the mechanisms by which tumour cells become resistant to TGF-β are generally unknown.
[0005] Much has been learned about TGF-β, its receptors, and its physiologic effects. TGF-β is now known to be the prototype for a large and conserved family of related polypeptides which have diverse physiologic functions in organisms as disparate as Drosophila, C. elegans,and Homosapiens. The TGF-β signal is initially received by a receptor complex containing the products of three different genes (TGF-β receptor types RI, RII, and RII). This signal is apparently mediated by the serine/threonine kinase activities of RI and RII. Though a small number of potential substrates for these receptor kinases have been identified (7-11) the biochemical pathways that are stimulated by these kinase activities are largely unknown. However, genes which appear to function downstream of these receptors, on the basis of genetic criteria, have been identified in Drosophila and C. elegans.Mutations of the C. elegansgenes sma-2, sma-3, and sma-4 confer phenotypic abnormalities identical to those observed with mutants of a TGF-β-family receptor gene (daf-4) in this organism (12). Savage et al. have proposed the name “dwarfins” for the corresponding proteins, homologues of which have been identified in the human (12). Similarly, mutations in the Drosophila Mad gene result in phenotypes like those observed in strains with mutations of the TGF-β-like ligand dpp (13,14). The Mad and sma genes are highly related by sequence, suggesting they all function similarly as mediators of TGF-β-like signaling (13,14). The Mad and sma genes have no relationship to other known signaling molecules, and encode no motifs that provide clues to their biochemical function. It is likely, however, that the final arbiters of the growth inhibition conferred by TGF-β family members include the cyclin-dependent kinase inhibitors (15).
[0006] The importance of the TGF-β pathway to colorectal tumorigenesis has recently been highlighted by two observations. A subset of colorectal tumors has been shown to harbor inactivating mutations of the TGF-β RII gene (16,17). This subset, accounting for about 15% of total colorectal cancers (18-20), comprises tumours with characteristic defects in mismatch repair (21), and the high frequency of TGF-β RII gene mutations in this tumour type results from a mutation-prone polyadenosine tract within the coding region of the gene. Though most other colorectal cancers have been shown to be resistant to the inhibitory effects of TGF-β (2,6), the cause of such insensitivity is unknown. Mutations of the TGF-β RII gene or of the cdk inhibitor genes p15, p16, and p21 are not generally found in these tumours (22,23), suggesting that the defects lie in the intermediates in the signaling pathway. This hypothesis recently received experimental support with the discovery of the DPC4 gene (24). DPC4 was identified through a positional cloning approach designed to identify a pancreatic tumour suppressor gene on chromosome 18q21. The DPC4 gene was found to be highly related to Mad and sma and genetic alterations which affected DPC4 were observed in over 50% of pancreatic cancers. Analogous studies of other human cancers, including those of the colon, revealed that DPC4 was genetically altered in only a minority of cases (25,26). There is a need in the art for identification of additional human genes homologous to DPC4 and its lower eukaryotic homologues (hereinafter referred to as the Mad gene family) which are involved in neoplastic processes.
SUMMARY OF THE INVENTION
[0007] It is an object of the invention to provide a cDNA of the human Smad2 gene.
[0008] It is another object of the invention to provide an isolated Smad2 human protein.
[0009] It is yet another object of the invention to provide a pair of primers for amplification of the Smad2 gene.
[0010] It is still another object of the invention to provide probe for detecting Smad2 coding sequences.
[0011] It is an object of the invention to provide a preparation of antibodies which specifically binds a human Smad2 protein.
[0012] It is another object of the invention to provide methods for screening test substances for the ability to suppress neoplasia.
[0013] It is still another object of the invention to provide an animal which carries in its genome a mutation in the Smad2 gene.
[0014] It is yet another object of the invention to provide methods to aid in diagnosing or prognosing neoplasias.
[0015] It is a further object of the invention to provide methods of supplying wild-type Smad2 gene function to a cell which has lost Smad2 gene function.
[0016] It is another object of the invention to provide methods of supplying wild-type Smad2 gene function to a cell which has altered Smad2 gene function.
[0017] It is still another object of the invention to provide a method of detecting neoplastic tissue.
[0018] It is a further object of the invention to provide a method for detecting genetic predisposition to cancer.
[0019] It is another object of the invention to provide a cDNA of a human Smad gene.
[0020] It is yet another object of the invention to provide an isolated human Smad protein.
[0021] These and other objects of the invention are provided by one or more of the embodiments disclosed below. In one embodiment, an isolated cDNA of the human Smad2 gene is provided.
[0022] In another embodiment of the invention an isolated human Smad2 protein is provided.
[0023] In yet another embodiment of the invention a pair of primers for amplifying Smad2 coding sequences is provided. The first primer of said pair comprises at least 12 contiguous nucleotides selected from SEQ ID NO:1 and the second primer of said pair comprises at least 12 contiguous nucleotides selected from the complement of SEQ ID NO:1.
[0024] In still another embodiment of the invention a probe for detecting Smad2 coding sequences is provided. The probe comprises an oligonucleotide consisting of at least 12 contiguous nucleotides selected from SEQ ID NO:1 or the complement thereof.
[0025] In another embodiment of the invention a preparation of antibodies is provided. The preparation of antibodies specifically binds human Smad2 protein. The antibodies are not substantially immunoreactive with other human proteins.
[0026] In one embodiment of the invention a method of screening test substances for the ability to suppress a neoplastically transformed phenotype is provided. A test substance is applied to a cell which carries a mutation in the Smad2 gene. The ability of the test substance to suppress the neoplastically transformed phenotype is then determined.
[0027] In another embodiment of the invention another method of screening test substances for the ability to suppress neoplastic growth is provided. A test substance is administered to an animal which carries a mutant Smad2 gene. The ability of the test substance to prevent or suppress the growth of tumors is then determined.
[0028] In still another embodiment of the invention an animal which carries in its genome a mutant Smad2 gene is provided.
[0029] In yet another embodiment of the invention an animal which has been genetically engineered to contain in its genome an insertion mutation which disrupts the Smad2 gene is provided.
[0030] In a further embodiment of the invention a method of aiding in the diagnosis or prognosis of a neoplastic tissue of a human is provided. Alteration of wild-type coding sequences of the Smad2 gene are detected in a tumor tissue isolated from a human. The alteration indicates neoplasia of the tissue.
[0031] In still another embodiment of the invention a method for supplying wild-type Smad2 gene function to a cell which has lost Smad2 function is provided. A wild-type Smad2 gene is introduced into a cell which has lost Smad2 gene function such that said wild-type gene is expressed in the cell.
[0032] In yet another embodiment of the invention a method of supplying wild-type Smad2 gene function to a cell which has altered Smad2 gene function by virtue of a mutation in the Smad2 gene is provided. A portion of a wild-type Smad2 gene is introduced into a cell which has lost Smad2 gene function such that the portion of the Smad2 gene is expressed in the cell. The portion of the Smad2 gene encodes a part of the protein which is required for non-neoplastic growth of the cell.
[0033] In a further embodiment of the invention another method for supplying wild-type Smad2 gene function to a cell which has altered Smad2 gene function by virtue of a mutation in the Smad2 gene is provided. Human wild-type Smad2 protein is applied to a cell which has lost the wild-type Smad2 gene function.
[0034] In another embodiment of the invention a method of detecting the presence of a neoplastic tissue in a human is provided. An alteration of a wild-type Smad2 coding sequence is detected in a body sample isolated from a human. The alteration indicates the presence of a neoplastic tissue in the human.
[0035] In yet another embodiment of the invention a method of detecting genetic predisposition to cancer in a human is provided. Germline alteration of the wild-type Smad2 gene coding sequences is detected in a human sample selected from the group consisting of blood and fetal tissue. The alteration indicates predisposition to cancer.
[0036] In another embodiment of the invention cDNAs of human Smad genes are provided. The cDNAs are selected from the group consisting of Smad1 (SEQ ID NO:9), Smad3 (SEQ ID NO:5), Smad5 (SEQ ID NO:7), and Smad6 (SEQ ID NO:3).
[0037] In a further embodiment of the invention isolated human Smad proteins are provided. The proteins are selected from the group consisting of Smad1 (SEQ ID NO:10), Smad3 (SEQ ID NO:6), Smad5 (SEQ ID NO:8), and Smad6 (SEQ ID NO:4).
[0038] These and other embodiments of the invention which are described in more detail below, provide the art with additional tools and weapons for detecting and fighting against cancers.
BRIEF DESCRIPTION OF THE DRAWINGS
[0039] FIG. 1 Mad family members. Sequences from the carboxyl-termini of proteins encoded by Mad-related genes identified in the human genome are compared. For Smad6, the homology identified by data-base searching was confined to Mad residues 249 to 344, and are therefore not shown in this figure. Blocks of significant homology between all 9 Mad homologues were identified using the MACAW multiple alignment software (v2.01) (41). Smad2, SEQ ID NO: 38; Smad1, SEQ ID NO:39; Smad6, SEQ ID NO: 40; and Smad3, SEQ ID NO:41. Amino acids in conserved blocks are capitalized and shaded based on the means of their pairwise score. The sequence listing provides the complete amino acid sequences of the Drosophila Mad (SEQ ID NO:11), and C. eleganssma4 (SEQ ID NO:12), sma3 (SEQ ID NO:13), sma2 (SEQ ID NO:14) proteins and the human DPC4 (SEQ ID NO:15) protein.
[0040] FIG. 2 Sequence of Smad2. The predicted amino acid sequence determined from the studies described in the text is given for Smad2, (SEQ ID NO:42); and compared with that of Mad. The accession number for the Smad2 nucleotide sequence is U59911 (SEQ ID NO:51). No upstream stop codon was identified, so it is possible that the coding region initiates upstream of that indicated. The amino acid sequences were aligned using the MACAW (version 2.01) program. Blocks of significant homology were identified. Amino acids in conserved blocks are capitalized and shaded based on the mean of their pairwise score.
[0041] FIG. 3 Mutational analysis of Smad2. FIG. 3A illustrates the PCR analysis of genomic DNA from tumour Mx5 (lane 3), showing a homozygous deletion of Snad2. The PCR product shown is 110 bp, while the Smad5 PCR product (136 bp) served as a control. Lanes 1, 2, and 4 contained PCR products derived from other tumours. FIG. 3B demonstrates the in vitro synthesis of protein (IVSP) analysis of tumour Mx21 (lane 5), showing a slightly smaller Smad2 polypeptide than found normally. The normal size product migrating at approximately 50 kd was found in the other tumours analyzed (lanes 1-4, 6). FIG. 3C shows a PCR analysis of DNA from tumour Smad2 (lane 2), showing a deletion of 42 bp in tumour Mx21 not found in the DNA from normal colon of the same patient (lane 1) or in the cDNA (lane 3) or genomic DNA (lane 4) of other non-neoplastic samples. FIG. 3D shows the sequence of the deleted region in tumour Mx21 SEQ ID NOS:44 and 45, amino acid and nucleotide sequences of region without deletion; SEQ ID NOS:45 and 46, amino acid and nucleotide sequences in region with deletion.
[0042] FIG. 4. Amino acid alignment of human Smad5 (SEQ ID NO:48) & 6 (SEQ ID NO:49) with Drosophila Mad (SEQ ID NO:43). The coding regions of the novel human Smads are compared to the Drosophila prototype. Amino acids in conserved blocks are capitalized and shaded based on the means of their pairwise scores.
[0043] FIG. 5. Alignment of amino acids in the six known Smad genes. Conserved domains Wed DH1a, DH1b, DH2a, and DH2b) represent sequence blocks which were highly related in at least 5 of the 6 proteins. Dark blue vertical bands within blocks reflect identical residues in at least 5 genes. Thin horizontal black lines represent gaps introduced to optimize the alignment. Domains DH1 and DH2 as previously defined (12), were each divided into an ‘a’ and ‘b’ block to maintain the optimum sequence alignment, yielding 4 highly conserved sequence blocks. Using Smad1 as reference, DH1a, DH1b, DH2a, and DH2b extended from codons 20 to 45, 68 to 145, 265 to 367, and 402 to 454, respectively. The accession numbers for the human genes include corrections to previous database entries made on the basis of sequencing cDNA clones in our laboratory. Accession numbers for murine Smad genes (full-length for murine Smad1, 2 & 5 and partial for murine Smad3, 4 & 6) are provided in the last column. The chromosomal positions shown in the second column are derived from reference 3.
[0044] FIG. 6. Examples of mutation screening. FIG. 6A shows PCR products from cDNA of 8 breast cancer samples. Lanes marked “+” and “−” represent cDNA synthesis reactions with and without reverse transcriptase, respectively. The arrow points to the full length PCR product, which includes the entire open reading frame and a T7 promotor for the IVSP assay. FIG. 6B shows polypeptides translated from the PCR products in A were separated by electrophoresis through an SDS polyacrylamide gel. Full length proteins (marked with an arrow) were observed as the major bands in each case. The minor bands represent internally initiated polypeptides generated during in vitro translation.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0045] It is a discovery of the present inventors that five previously undescribed genes which are Mad-related exist in the human genome. The genes have been dubbed Smad1, Smad2, Snad3, Smad5, and Smad6. These genes are involved in the signaling pathway which mediates the effects of binding of TGF-β or TGF-β-like ligands to their receptors. Defects in these genes may lead to resistance to the inhibitory effects of TGF-β. Such resistance has been demonstrated in tumors of many types. Mutations have been found in Smad2 and Smad4 (DPC4) alleles in tumors. Thus, Smad2 and Smad4 are important genes for tumor suppression.
[0046] Moreover, three of these genes (Smad3, Smad5, and Smad6) have been localized to chromosomes noted to be frequently lost in human cancers. Chromosome 5q is often lost in lung and esophagus tumours, for example, but no relevant suppressor gene mutations have yet been identified, despite extensive study (31-34). Similarly, a recent report suggests that chromosome 15q losses are remarkably common in metastatic cancers of the breast, colon, and lung (35). The minimally lost regions (MLRs) in these cancers have been mapped to 5q31 and 15q14-21, close to the positions of the Smad3, Smad5, and Smad6 genes (See Table 1, below).
[0047] Human SmadcDNA molecules (SEQ ID NOS:1, 3, 5, 7, and 9), according to the present invention, are provided in isolated form, free from other sequences. They are identified by the chromosome and band position to which they have been mapped by fluorescence in situ hybridization, as well as by the determined sequences of the 5′ and 3′ ends. In addition, the amino acid sequences of the encoded proteins are provided (SEQ ID NOS:2, 4, 6, 8, and 10). This information unambiguously identifies the cDNAs which are provided. The cDNA molecules can be isolated from libraries of human cDNAs, from YACs, or from any other source of human cDNA. The cDNA molecules can be inserted into vectors, as is known in the art for various uses. The cDNA molecules can be linked to other sequences, such as selectable markers for certain uses. The CDNA molecules can have transcription and/or translation signals appended to either end, as is known in the art for expression of the encoded protein or for expression of anti-sense RNA molecules. Sequences from these cDNA molecules may be used, inter alia, to make hybridization probes and primers for PCR-based assays which may in turn be used to identify additional MAD-related genes involved in TGF-β signaling and/or in tumorigenesis.
[0048] The encoded proteins of the disclosed genes can be obtained by in vitro synthesis, or by synthesis in a recombinant organism or cell. Alternatively, the encoded proteins can be purified from cells or tissues naturally producing the proteins using antibodies made to the encoded proteins, or to peptide portions thereof. Isolated proteins need not be homogeneous to be useful, but can be partially purified (i.e., 10%, 50%, 75%, 90%, or 95% pure) from other cellular components and/or other proteins. The proteins may be used, inter alia, to generate anti-Smad antibodies or to supply a wild-type Smad protein function to cells which have lost that function. Wild-type Smad proteins may also be compared with Smad proteins in tissue samples in order to detect the presence of altered Smad proteins.
[0049] Antibodies according to the present invention bind specifically to human Smad proteins and thus are not substantially immunoreactive with other human proteins. Specific binding, as used herein, means that the antibody binds with higher affinity to a Smad protein than to non-Smad protein. No substantial immunoreactivity, as used herein, means that the antibody does not bind with high affinity to a non-Smad protein. Under standard conditions, antibodies that bind specifically to human Smad proteins do not cross-react with other proteins, as evidenced by Western blotting. Anti-Smad antibodies can be raised by immunization of mammals with peptide portions of the encoded proteins, either alone or attached to another protein to render it more immunogenic. Alternatively, the entire protein can be used as an immunogen. Antiserum can be used directly or after some purification steps to render it more specific. The immunized animals can also be used to raise monoclonal antibodies, as is known in the art, through the process of fusion of antibody producing cells to myeloma cells. Any techniques known in the art for raising polyclonal or monoclonal antibodies can be used. The antibodies are useful for purifying Smad proteins and for detecting alterations in Smad proteins which result from mutations in Smad genes.
[0050] Primer pairs are typically used for amplification of the cDNAs. Particularly suitable primers are disclosed below, but any primers can be used which comprise oligonucleotides of at least 10, 12, or 14 contiguous nucleotides. Each of the primers in a pair is complementary to an opposite strand of the gene. Typically at least 15, 18, or 20 contiguous nucleotides are desirable. In some circumstances it may be desirable that additional moieties be attached to the contiguous nucleotides of the gene. For example, restriction enzyme sites may be desirable to facilitate cloning. Transcription and translation signals may be desirable to facilitate in vitro assays of RNA and protein products (e.g., IVSP).
[0051] Similarly, nucleotide probes can be useful for identifying the disclosed genes in clinical test samples. Probes can be similarly sized to primers, but can often be much larger, up to the size of the full cDNA. Fluorescent or radioactive labels may be optionally appended to the probes to facilitate recognition.
[0052] Mutations in the Smad2 gene lead to neoplasia; thus cells harboring mutations in this gene can be used to screen for candidate therapeutic agents. The mutation may occur spontaneously in test cells or may be induced by mutagenic methods. In a preferred embodiment, the cells are genetically engineered to contain the mutation. By applying a candidate test substance to such cells and observing its effect, one can determine whether such an agent suppresses the transformed phenotype of the cell. The test substance may be a pharmacologic agent already known in the art or may be a substance previously unknown to have any pharmacologic activity. The substance may be naturally occurring or may be designed in the laboratory. It may be isolated from a microorganism, plant, or animal or may be produced recombinantly or synthetically by chemical methods known in the art. Any phenotype of the cell which is indicative of neoplastic transformation can be observed. Such phenotypes are well known in the art. Cells to be used are preferably epithelial cells, but any type of cell which exhibits a transformed phenotype and carries a mutation in the Smad2 gene can be used. Suitable methods for culturing cells, such as monolayer or explant culture methods, are well known in the art. Similar assays can be performed using whole animals which carry a mutant Smad2 gene as disclosed. The test substance may be administered to the animal orally or by injection. The ability of a candidate test substance to suppress or prevent the growth of tumors provides an indication of usefulness as a therapeutic agent.
[0053] The present invention also provides animals which carry a mutation in the Smad2 gene. The animals are made using techniques which are well known in the art. The mutation may be a missense mutation, an insertion, a deletion, or a rearrangement. These animals can be used to screen for potential therapeutic agents and to study the initiation, progression, or stimulation of neoplasia, for example by exposure to potential environmental carcinogenic agents. Diagnosis and/or prognosis of neoplasia can be aided using the Smad2 gene of the present invention. The presence of a mutation in the Smad2 gene can be used to aid in diagnosing a cancer or in prognosing a clinical outcome. Mutations can be detected by assaying a Smad2 allele, mRNA, or protein. Presence of a deviation from the wild-type Smad2 sequence or from the Smad2 sequence found in the surrounding normal tissue indicates a mutation. Conventional assays for mutations as are known in the art can be used. For example, sequence analysis, Southern, Northern, or Western analysis, can be used. Immunoblotting, immunohistochemistry, and other immunological techniques can be used and are routinely practiced in the art. Alternatively, Smad2 gene sequences in the samples may be amplified using, inter alia, PCR. Mutations in the sequence can then be detected using mutation- or wild-type-specific nucleic acid probes. Samples for testing for mutations or mutant gene products can be a suspected or confirmed neoplastic tissue, or a body sample, such as serum, stool, urine, or sputum. Germline mutations can also be detected, for example, by testing blood or fetal tissue.
[0054] Replacement therapy can be used to remedy a defective gene. This can be accomplished by introducing all or a portion of a wild-type Smad2 gene into a cell which has lost responsiveness to TGF-β by virtue of a mutation in the Smad2 gene. The portion of the Snad2 gene includes at least the portion comprising codon 345-358 of SEQ ID NO:1 or a portion comprising at least 30 nucleotides which include the mutated codons. Any vectors known in the art can be used to introduce the gene or gene portion, including, but not limited to retroviral or adenoviral vectors. Other techniques such as liposome-mediated gene transfer can be used as well. The gene or gene portion can be introduced on a vector which remains extrachromosomal, such as a minichromosome or episome. It may be introduced into cells which have been removed from the body or by direct injection into suitable target organs or tumors. Alternatively, the gene or gene portion may be introduced so that it integrates in the genome. In an alternative embodiment a wild-type Smad2 protein having the sequence shown in SEQ ID NO:2 can be applied to cells which have a mutation in the Smad2 gene. The protein can be applied in a suitable vehicle for internalization within cells, such as liposomes. Targeting of the agents to particular organs or tissues can be accomplished as is known in the art, using specific targeting moieties, such as antibodies, receptors, or ligands.
[0055] The following examples are provided for exemplification purposes only and are not intended to limit the scope of the invention.
EXAMPLE 1
[0056] This example demonstrates the identification of Mad homologues.
[0057] To identify human homologues of Mad, we searched expressed-sequence tag (EST) databases. These databases contain a significant fraction of the total human genes (27), and have been previously shown to be valuable cancer gene discovery tools (28). Several sequences were used as search queries, including the Mad gene, sma genes, and Smad4. Once candidate homologues were identified, new searches were carried out with these in order to identify additional homologues or additional clones of the same genes. In toto, thirty EST clones were identified that contained statistically significant matches to the queries. Sequences of these clones were then aligned to connect clones into contigs. In some cases, clones were re-sequenced to resolve ambiguities in the EST entries. This approach allowed us to define portions of five different human genes, dubbed Smad2, Smad5, Smad3, Smad6, and Smad1. In four of the five genes, the homologies included the predicted carboxyl-termini of the gene products, and these are aligned in FIG. 1 along with other members of the Mad family. The highest conservation was noted with Smad1, which was predicted to have 58 of 59 (98%) carboxyl-terminal residues identical to that of the Mad gene product. Smad3 (85% identity) and Smad2 (83% identity) were also very similar to Mad, while Smad5 differed significantly (29% identity). In comparison, Smad4 was only 47% identical to Mad in this region and was not more highly related to Smad5 than to the other genes.
[0058] We next determined the chromosome localizations of these genes. Despite their sequence similarities, we were eventually able to design primers and conditions that specifically amplified each gene from human DNA and did not amplify related genes from the mouse or hamster genomes (Table 1). These primers were used to screen a somatic cell hybrid panel, affording localization of Smad2, Smad5, Smad3, Smad6, and Smad1 to chromosomes 18, 15, 15, 5, and 4, respectively. The same PCR conditions were then used to screen a YAC panel. This was successful in four cases, and the YAC's were used to further localize Smad2, Smad5, Smad3, Smad6 to chromosomes 18q21, 15q15-21, 15q15-21, and 5q31, respectively. One YAC was found to contain both Shad5 and Smad3, indicating that these genes were within approximately 1 Mb. These data are summarized in Table 1.
Methods
[0059] Identification of Mad homologues. Full length protein sequences of Mad and its known homologues were used to search the National Center for Biotechnology Information (NCBI) expressed sequence tag database (dbest) using the ‘tblastn’ search program (36). New matches were used for subsequent rounds of searching and to create contigs. To confirm sequence information at the 3′ ends of the newly identified genes, manual sequencing of cDNA clones obtained from the IMAGE consortium (37) was performed, and the resultant sequences recorded through the Accession Numbers listed in Table 1.
[0060] Chromosome localization. PCR primers were designed to amplify the genes identified in the data-base search. A variety of primers were synthesized for each gene, searching for pairs which would provide robust signals of the expected length from human DNA but would not amplify hamster or mouse DNA. Appropriate primer pairs (see Table 1) were then used for amplification of DNA from a somatic cell hybrid panel containing isolated human chromosomes (Coriell Institute for Medical Research, Camden, N.J.). The same primers were then used to screen the CEPH A YAC panel (Research Genetics, Atlanta, Ga.). The chromosomal positions of the identified YACs were determined from published genetic maps (38).
EXAMPLE 2
[0061] This example demonstrates the detailed analysis of Smad2.
[0062] The localization of Smad2 to chromosome 18q21 was particularly intriguing in light of previous studies of this chromosomal region in colorectal cancers. In particular, it has been shown that over 60% of such cancers lose genetic information from this chromosomal region (29). The minimally lost region (MLR) encompasses a 16 Mb interval between marker D18S535 and D18S858 and includes Smad4 (25). To determine the position of Smad2 with respect to this region, we studied somatic cell hybrids containing portions of chromosome 18q, YAC libraries, and colorectal cancer cell lines with homozygous deletions in the area (25, 30). These studies revealed that Smad2 was between D18S535 and Smad4, approximately 3 Mb from each. No YAC clone was identified which contained both Smad4 and Smad2.
[0063] Thus both Smad2 and Smad4 were located within the MLR and were candidates for the presumptive tumour suppressor gene(s) located in this region. There is much precedent for clustering of genes of similar function, thought to arise through gene duplication of an ancestral precursor. As a prelude to a mutational analysis, we therefore determined the complete coding sequence of Smad2 and compared it to that of Smad4 and other Mad homologues. The predicted amino acid sequence of Smad2 (SEQ ID NO:42) is shown in FIG. 2, and the nucleotide sequence entered in Genbank under Accession #U599 11 (SEQ ID NO:51). Smad2 encoded a 467 residue open reading frame, with maximal homology to other Mad family members at the amino- and carboxyl-termini of the protein. Smad2 was more similar to Mad (62% identity over 373 aa) than to the C. eleganshomologues sma-2 (50% identify over 365 aa), sma-3 (45% identity over 204 aa), sma-4 (25% identify over 361 aa) or to Smad4 (44% over 158 aa).
[0064] We then evaluated Smad2 in a panel of 18 early-passage colorectal cancer lines, each containing an allelic loss of the MLR on chromosome 18q. RT-PCR was performed on RNA from these xenografts, using primers which did not amplify the Smad2 mouse homologue. Smad2 was found to be expressed in the normal colon mucosa, normal brain, and in 17 of the 18 colorectal tumours. In the one case (Mx5) in which Smad2 was not expressed at detectable levels, a homozygous deletion of Smad2 sequences was identified (FIG. 3A). This deletion was verified with three separate primer pairs and included both the 5′ and 3′ ends of Smad2, but did not extend proximally to D18S535 or distally to Smad4 (not shown). The deletion was not found in DNA from the normal colon of the corresponding patient, so represented a somatic alteration.
[0065] To search for alterations of Smad2 that would alter the size of the encoded polypeptide, an in vitro synthesized protein assay was performed by incorporating recognition signals for in vitro transcription and translation in the primers used for RT-PCR. One tumour, Mx21, exhibited a smaller protein in this assay (FIG. 3B). The alteration was traced to a 42 bp deletion, extending from codon 345 to 358 (FIG. 3C and FIG. 3D). PCR amplification of normal genomic DNA across regions flanking the deletions demonstrated that it occurred completely within an exon (rather than representing a deleted exon). The deletion was somatic, as it was not found in the patient's normal cells (FIG. 3c). In addition, the deleted residues were highly conserved, with thirteen of the fourteen residues identical to the homologous region in the Mad gene.
[0066] Finally, we attempted to search for missense mutations of Smad2 by direct sequencing of RT-PCR products in the 16 tumours encoding an apparently full-length polypeptide. No mutations or polymorphisms were detected within the entire open reading frame in any of the tumours.
Methods
Sequence and Mutational Analysis of Smad2
[0067] Sequences already entered into the data-bank (Table 1), representing the 5′ and 3′ ends of Smad2, were used to design primers for amplification of cDNA via RT-PCR. Direct sequencing of the RT-PCR products allowed us to determine the entire coding sequence and correct ambiguities in the previous EST entries. The new sequence, including the entire ORF, has been recorded under Accession Number U59911. RT-PCR products were transcribed and translated in vitro as described (39). The entire coding region of the gene was amplified using primers 5′-GGA TCC TAA TAC GAC TCA CTA TAG GGA GAC CAC CAT GGG TAA GAA CAT GTC GTC CAT C-3′ (SEQ ID NO:26) (including signals for transcription and translation) and 5′-TTT CCA TGG GAC TTG ATT GG-3′ (SEQ ID NO:27). Sequencing of the RT-PCR products was performed using internal primers, available from the authors upon request, end-labeled with 32P-γ-ATP and SequiTherm Polymerase (Epicentre, Madison, Wis.), as described (40).
EXAMPLE 3
[0068] This example demonstrates the comparison of Smad genes.
[0069] Partial sequences for known Smad genes were used to make hybridization probes and primers for PCR-based assays. Hybridization probes were used to screen cDNA libraries using standard methods. Clones were manually sequenced using 3P-labelled terminator chemistry (ThermoSequenase, Amersham Life Science, Cleveland, Ohio). Contigs of the clones were used to create a full length coding sequence. The coding region of each gene, including portions originally derived from expressed sequence tags found by computer searches (National Center for Biotechnology Information, World Wide Web page), was confirmed or corrected by manual sequencing as described above. To determine the 5′ end of Smand5, the rapid amplification of cDNA ends technique was employed, using primers derived from the cDNA sequence obtained by cDNA cloning.
[0070] The amino acid sequence of the six known Smads were aligned using Macaw Version 2.0.3 optimizing for the mean pairwise score. Blocks of homology shown in FIG. 5 were produced by selecting all permissible and statistically significant blocks as determined by Macaw. Conserved domains in FIG. 5 were identified by selecting sequence that had a more stringent selection, including mean pairwise scores above 80 in at least five of the six genes.
[0071] Sequence for two novel human Smad genes (Smad5 and Smad6; FIG. 4) were determined through a combination of database searching, cDNA library screening, and PCR-based techniques. These sequences were compared to the published sequences of Smad1-4 (in the case of Smad3, corrections to the published sequence were made on the basis of independently cloning and sequencing the entire gene). Alignment of the six genes documented four highly conserved domains (D)H's, indicated by boxes in FIG. 3), with 125 residues identical in at least five of the genes (indicated by vertical blue lines in the boxes). The naming of these domains was based on a previous comparison of three C. elegansMad genes and a Drosophila Mad gene that showed two major regions of homology; corresponding to DH1and DH2. Our comparison of the six human genes show that further division of the two original domains yields a significantly better alignment. The structure of all six human genes was very similar, though Smad6 lacked DH1. Evidence that Smad6 was full length included a reasonable translation initiation consensus sequence at the presumed initiating methionine and an upstream stop codon.
EXAMPLE 4
[0072] This example demonstrates mutations in the Smad genes.
[0073] Mutation Detection. The in vitro protein synthesis assay (IVSP) was used in combination with manual sequencing. Each Smad studied was amplified with PCR primers that spanned the entire coding region. All forward primers listed below were synthesized with the following T7 promoter sequence at the 5′ end: 5′-gga tcc taa tac gac tca cta tag gga g-3′ (nucleotides 1-25 of SEQ ID NO:26). The T7 promotor and the Kozak consensus sequences shown for each forward primer allowed in vitro transcription and translation. The following gene specific primers were used: Smad1F=5′-T7-acc acc atg gea cca tat cca agg agt ata act ag (SEQ ID NO:28); Smad1R=5′-ttt m ata tga atc caa cag ttg gtc aca gag g (SEQ ID NO:29); Smad2F=acc acc atg ggt a aga aca tgt cgt cc atc (SEQ ID NO:30); Smad2R=ttt cca tgg gac ttg att gg (SEQ ID NO:31); Smad3F=gag cca gcc atg tcg tcc atc c (SEQ ID NO:32); Smad3R=ttt tcc cca agc ctg ccc tc (SEQ ID NO:33); Smad5F=5′-acc acc atg gtc tcc gaa gat ttg tgt caa (SEQ ID NO:34); Smad5R=5′-tttm ata tct gtt ttc aat gta agc tca cag (SEQ ID NO:35); Smad6F=5′ acc acc atg gaa tct ccg cca cct ccc tac (SEQ ID NO:36); Smad6R=5′-cg cca cta tct ggg gtt g (SEQ ID NO:37). PCR products were transcribed and translated in vitro, and the resultant proteins separated by SDS-polyacrylamide gel electrophoresis, as described . The RT-PCR products were also used for sequencing with internal primers (primer sequences available upon request).
[0074] Mutations of Smad2 and Smad4 have been observed in human tumors (24, 25, 43-46), but the other four genes have not previously been evaluated for genetic alterations. To address this issue, a panel of 167 cancer cell lines (passaged in vitro or as xenografts in nude mice) was assembled for mutation analysis. Cell lines rather than primary tumors were chosen so that nucleic acids from non-neoplastic cells within primary tumor specimens would not cloud interpretation. The panel included cancers of the following types: cancers of the colorectum (70), breast (22), brain (22), lung (15), pancreas (12), head and neck (8), ovary (6), esophagus (4), stomach (4), and prostate (4). These cancers represent most of the common tumor types, and several have been shown to lose heterozygosity at chromosomal positions encompassing the Smad genes.
[0075] RNA from each of the 167 lines was used as template in reverse transcriptase-polymerase chain reactions (RT-PCR), using primer sequences specific for Smad1, 2, 3, 5, and 6. These genes were ubiquitously expressed, as revealed by expression of each of the four genes at easily detectable levels in 165 of the 167 samples (examples in FIG. 6). To screen for mutations, we employed an in vitro synthesized protein (IVSP) assay. RT-PCR products containing the open reading frame of each of the four genes were transcribed and translated in vitro, and the resultant polypeptides separated by SDS-acrylamide gel electrophoresis. Full length products were observed in each of the 165 samples yielding RT-PCR products, and no truncated proteins were observed.
[0076] The RT-PCR and IVSP assays detect most types of mutations affecting potential tumor suppressor genes, such as homozygous deletions, nonsense mutations, deletions or insertions creating frameshifts, and splice site mutations. A small number of missense mutations have been identified in the DH2 domains of Smad2 and Smad4 (24, 25, 44-46), that would have been missed by an IVSP assay. In order to detect missense mutations that occurred at significant frequency, we chose 50 tumors from our panel for sequencing of the DH2a and HD2b domains. DH2b domains in the four other Smads. The RT-PCR products were manually sequenced from the same amplified cDNA used for the IVSP assay. Only the wild-type sequence was identified in each case.
[0077] In summary, the four genes analyzed here, plus the previously analyzed Smad2 and Smad4, are expressed in virtually all cancers analyzed, suggesting they are important for signal transduction in most cell types. Mutations of these genes in tumors, however, are not widespread. Smad 4 is altered in a significant fraction of pancreatic cancers and in a minority of colorectal tumors, but rarely in other tumor types, but rarely in other tumor types and Smad 2 is altered only in a small fraction of colorectal and lung tumors but in no other common tumor types. The other four genes do not appear to be frequently mutated in any of the cancer types analyzed here, and mutations of the TGF-( receptors themselves are also uncommon except in the subset of colorectal cancers with mismatch repair deficiency (16, 17). Thus the basis of TGF-β resistance in human cancers is unlikely to be due, in general, to alterations of currently known genes. Assuming that the TGF-β resistance in tumors has a genetic basis, other components of the TGF-β signaling pathway, such as the one recently identified in Xenopus laevismust be responsible.
[0078] Three additional genes, apparently not identical to those disclosed here, have been previously described (12). Though the number of Mad-related human genes may appear large, this number should be viewed in the context of other genes involved in TGF-β-like signaling. Over 20 TGF-β family members have been described in humans, including several bone morphogenetic proteins and activins, each with its own tissue distribution and function (3). Three genes encode the receptor complex that mediates signaling by TGF-β1, β2, and β3, and other genes encode receptors for additional ligands in the TGF-β superfamily. The large number of Mad-related genes may therefore parallel the large number of receptors and ligands which initiate signaling. It is possible that each receptor-ligand combination is associated with a unique Mad family member complex. Alternatively, as sma-2, sma-3, and sma-4 mutations yield identical phenotypes in C. elegans,Savage et al. have suggested that the encoded proteins are likely to either form a heteromeric complex or be activated sequentially, in a signaling cascade (12). In view of the high conservation of this family of genes, the human homologues may function in a similar manner.
[0079] The principles, preferred embodiments and modes of operation of the present invention have been described in the foregoing specification. The invention which is intended to be protected herein, however, is not to be construed as limited to the particular forms disclosed, since they are to be regarded as illustrative rather than restrictive. Variations and changes may be made by those skilled in the art without departing from the spirit of the invention.
REFERENCES
[0080] 1. Attisano, L., Wrana, J. L., Lopez-Casillas, F. & Massague, J. TGF-beta receptors and actions. Biochim. Biophys. Acta1222, 71-80 (1994).
[0081] 2. Brattain, M. G., Howell, G., Sun, L. Z. & Willson, J. K. Growth factor balance and tumor progression. Curr. Opin. Oncol.6, 77-81 (1994).
[0082] 3. Kingsley, D. M. The TGF-beta superfamily: new members, new receptors, and new genetic tests of function in different organisms. Genes Dev.8, 133-146 (1994).
[0083] 4. Roberts, A. B. & Sporn, M. B. Physiological actions and clinical applications of transforming growth factor-beta (TGF-beta). Growth Factors8, 1-9 (1993).
[0084] 5. Satterwhite, D. J. & Moses, H. L. Mechanisms of transforming growth factor-beta 1-induced cell cycle arrest. Invasion Metastasis14, 309-318 (1994).
[0085] 6. Fynan, T. M. & Reiss, M. Resistance to inhibition of cell growth by transforming growth factor-beta and its role in oncogenesis. Crit. Rev. Oncog.4, 493-540 (1993).
[0086] 7. Chen, R. H., Miettinen, P. J., Maruoka, E. M., Choy, L. & Derynck, R. A WD-domain protein that is associated with and phosphorylated by the type II TGF-beta receptor. Nature 377, 548-552 (1995).
[0087] 8. Wang, T., Donahoe, P. K. & Zervos, A. S. Specific interaction of type I receptors of the TGF-beta family with the immunophilin FKBP-12. Science265, 674-6 (1994).
[0088] 9. Atfi, A., Lepage, K., Allard, P., Chapdelaine, A. & Chevalier, S. Activation Of a Serine Threonine Kinase Signaling Pathway By Transforming Growth Factor Type Beta. Proc. Natl. Acad. Sci.92, 12110-12114 (1995).
[0089] 10. Yamaguchi, K., et al. Identification Of a Member Of the Mapkkk Family As a Potential Mediator Of Tgf-Beta Signal Transduction. Science270, 2008-2011 (1995).
[0090] 10 11. Wang, T. W., et al. The P21(Ras) Farnesyltransferase Alpha Subunit In TGF-Beta and Activin Signaling. Science271, 1120-1122 (1996).
[0091] 12. Savage, C., et al. Caenorhabditis Elegans Genes Sma2, Sma-3, and Sma-4 Define a Conserved Family Of Transforming Growth Factor Beta Pathway Components. Proc. Natl. Acad. Sci.USA 93, 790-794 (1996).
[0092] 13. Sekelsky, J. J., Newfeld, S. J., Raftery, L. A., Chartoff, E. H. & Gelbart, W. M. Genetic characterization and cloning of mothers against dpp, a gene required for decapentaplegic function in Drosophila melanogaster. Genetics139, 1347-1358 (1995).
[0093] 14. Hursh, D. A., Padgett, R. W. & Gelbart, W. M. Cross regulation of decapentaplegic and Ultrabithorax transcription in the embryonic visceral mesoderm of Drosophila. Development117, 1211-1222 (1993).
[0094] 15. Massague, J. & Polyak, K. Mammalian antiproliferative signals and their targets. Curr. Opin. Genet. Dev.5, 91-6 (1995).
[0095] 16. Markowitz, S., et al. Inactivation of the type II TGF-beta receptor in colon cancer cells with microsatellite instability. Science268, 1336-1338 (1995).
[0096] 17. Parsons, R., et al. Microsatellite Instability and Mutations Of the Transforming Growth Factor Beta Type II Receptor Gene In Colorectal Cancer. Cancer Res.55, 5548-5550 (1995).
[0097] 18. Ionov, Y., Peinado, M. A., Malkhosyan, S., Shibata, D. & Perucho, M. Ubiquitous somatic mutations in simple repeated sequences reveal a new mechanism for colonic carcinogenesis. Nature363, 558-561 (1993).
[0098] 19. Thibodeau, S. N., Bren, G. & Schaid, D. Microsateffite instability in cancer of the proximal colon. Science260, 816-9 (1993).
[0099] 20. Aaltonen, L. A., et al. Clues to the pathogenesis of familial colorectal cancer. Science260, 812-816 (1993).
[0100] 21. Marra, G. & Boland, C. R. Hereditary nonpolyposis colorectal cancer: the syndrome, the genes, and historical perspectives. J. Natl. Cancer Inst.87, 1114-25 (1995).
[0101] 22. Jen, J., et al. Deletion of p16 and p15 genes in brain tumors. Cancer Res.54, 6353-6358 (1994).
[0102] 23. Shiohara, M., et al. Absence of WAF1 mutations in a variety of human malignancies. Blood84, 3781-4 (1994).
[0103] 24. Hahn, S. A., et al. Dpc4, A Candidate Tumor Suppressor Gene At Human Chromosome 18q21.1. Science271, 350-353 (1996).
[0104] 25. Thiagalingam, S. Evaluation of Chromosome 18q in Colorectal Cancers. submitted (1996).
[0105] 26. Schutte, M. Tissue-restriction in DPC4 gene alterations. submitted (1996).
[0106] 27. Korenberg, J. R., Chen, X. N., Adams, M. D. & Venter, J. C. Toward a cDNA Map Of the Human Genome. Genomics29, 364-370 (1995).
[0107] 28. Papadopoulos, N., et al. Mutation of a mutL homolog in hereditary colon cancer. Science263, 1625-1629 (1994).
[0108] 29. Fearon, E. R. & Vogelstein, B. A genetic model for colorectal tumorigenesis. Cell61, 759-67 (1990).
[0109] 30. Rojas, K., Silverman, G. A., Hudson, J. R., Jr. & Overhauser, J. Integration of the 1993-94 Genethon genetic linkage map for chromosome 18 with the physical map using a somatic cell hybrid mapping panel. Genomics 25, 329-330 (1995).
[0110] 31. Shibagaki, I., et al. Allelotype analysis of esophageal squamous cell carcinoma. Cancer Res.54, 2996-3000 (1994).
[0111] 32. Aoki, T., et al. Allelotype study of esophageal carcinoma. Genes Chrom. Cancer10, 177-82 (1994).
[0112] 33. Ogasawara, S., et al. Common Deleted Region On the Long Arm Of Chromosome 5 In Esophageal Carcinoma. Gastroenterology110, 52-57 (1996).
[0113] 34. Wieland, I., et al. Allelic Deletion Mapping On Chromosome 5 In Human Lung Carcinomas. Oncogene12, 97-102 (1996).
[0114] 35. Wick, W. Evidence for a novel tumor suppressor gene on chromosome 15 associated with progression to a metastatic stage in breast cancer. Oncogene 12, 973-978 (1996).
[0115] 36. Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. Basic local alignment search tool. J. Mol. Biol.215, 403-10 (1990).
[0116] 37. Hillier, L. Generation and preliminary analysis of over 200,000 human expressed sequence tags. Nature, in press (1996).
[0117] 38. Hudson, T. J., et al. An STS-Based Map Of the Human Genome. Science270, 1945-1954 (1995).
[0118] 39. Powell, S. M., et al. Molecular diagnosis of familial adenomatous polyposis. N. Engl. J. Med.329, 1982-1987 (1993).
[0119] 40. Leach, F. S., et al. Mutations of a mutS homolog in hereditary nonpolyposis colorectal cancer. Cell75, 1215-1225 (1993).
[0120] 41. Schuler, G. D., Altschul, S. F. & Lipman, D. J. A workbench for multiple alignment construction and analysis. Proteins 9, 180-190 (1991).
[0121] 42. Houlgatte, R., et al. The genexpress index—A resource for gene discovery and the genic map of the human genome. Genome Research5, 272-304 (1995).
[0122] 43. Riggins, G. J., Kinzler, K. W., Vogelstein, B., and Thiagalingam, S., Mad-related genes in the human. Nature Genetics13: 347-349 (1996).
[0123] 44. Uchida, K., Nagatake, M., Osada, H., Yatabe, Y., Kondo, M., Mitsudomi, T., Masuda, A., Takahashi, T., and Takahashi, T. Somatic in vivo alterations of the WV18-1 Gene at 18q21 in human lung cancers, Cancer Research56: 5583-5585 (1996).
[0124] 45. Nagatake, M., Takagi, Y., Osada, H., Uchida, K., Misudomi, T., Saji, S., Shimokata, K., T ahash, T., and Takahash, T. Somatic in vivo alterations of the DPC4 gene at 19821 in human lung cancers, Cancer Research56: 2718-2720 (1996).
[0125] 46. Eppert, K., Scherer, S. W., Ozcelik, H., Pirone, R., Hoodless, P., Kim, H., Tsui, L.- C., Bapat, B., Gallinger, S., Andrulis, I. L., Thomsen, G. H., Wrana, L., and Attisano, L. MADR2 maps to 18q21 and encodes a TGF-beta-regulated MAD-related protein that is mutated in colorectal carcinoma. Cell86: 543-552 (19916).
Claims
What is claimed is:
1. An isolated cDNA of a human Smad5 gene which encodes a Smad5 protein having the amino acid sequence of GenBank Accession No. U59913 (SEQ ID NO:52).
2. The isolated cDNA of claim 1 which has the coding sequence of GenBank Accession No. U59913 (SEQ ID NO:53).
3. An isolated cDNA of a human Smad5 gene which has a coding sequence of nucleotides 375 to 1769 of SEQ ID NO:7.
Patent Citations (3)
Non-Patent Literature (11)
- Eppert, K. et al., Cell, vol. 86, pp. 543-552, Aug. 23, 1996.Applicant
- Yingling, J. et al., Proc. Natl. Acad. Sci. USA, vol. 93, pp. 8940-8944, Aug. 23, 1996.Applicant
- Riggins, G. et al., Nature Gen., vol. 13, pp. 347-349, Jul. 1996.Applicant
- Nishimura, R. et al., Genbank, Accession No. U73825, Feb. 4, 1997.Applicant
- Hoodless, P. et al., Genbank, Accession No. U54826, May 25, 1996.Applicant
- Zhang et al., “The tumor suppressor Smad4/DPC 4 as a central mediator of Smad function” Current Biology, Mar. 19, 1997, vol. 7, No. 4, pp. 270-276.Applicant
- Zhang et al., “Receptor-associated Mad homologs synergize as effectors of the TGF-beta response” Nature, Sep. 12, 1996, vol. 383, pp. 168-172.Applicant
- Lagna et al., “Partnership between DPC4 and SMAD proteins in TGF-beta signalling pathways” Nature, Oct. 31, 1996, vol. 383, pp. 832-836.Applicant
- Graff et al., “Xenopus Mad proteins transduce distinct subsets of signals for the TGF beta superfamily”, Cell May 17, 1996, vol. 86, pp. 479-487.Applicant
- Nakao et al., “Identification of Smad2, a human mad-related protein in the transforming growth factor beta signaling pathway” Journal of Biology Chemistry, Jan. 31, 1997, vol. 272, No. 5, pp. 2896-2900.Applicant
- Derynck et al., “Nomenclature: vertebrate mediators of TGF beta family signals” Cell, Oct. 18, 1996, vol. 87, p. 173.Applicant