The Philadelphia chromosome (Ph) translocation generates a chimeric tyrosine kinase oncogene, BCR/ABL, which causes chronic myelogenous leukemia (CML) and a type of acute lymphoblastic leukemia (ALL). In primary samples from virtually all patients with CML or Ph ؉ ALL, the CRKL adapter protein is tyrosine phosphorylated and physically associated with p210 BCR/ABL . CRKL has one SH2 domain and two SH3 domains and is structurally related to c-CRK-II (CRK) and the v-Crk oncoprotein. We have previously shown that CRKL, but not the related adapter protein c-CRK, is tyrosine phosphorylated in cell lines transformed by BCR/ABL, and that CRKL binds to BCR/ABL through the CRKL-SH3 domains. Furthermore, the CRKL-SH2 domain has been shown to bind one or more cellular proteins, one of which is p120CBL . Here we demonstrate that another cellular protein linked to BCR/ABL through the CRKL-SH2 domain is p130 CAS . p130 CAS was found to be tyrosine phosphorylated and associated with CRKL in BCR/ABL expressing cell lines and in samples obtained from CML and ALL patients, but not in samples from controls. In both normal and BCR/ABL transformed cells, p130 CAS was detected in focal adhesion-like structures, as was BCR/ABL. In normal cells, the focal adhesion proteins tensin, p125 FAK , and paxillin constitutively associated with p130 CAS . However, in BCR/ABL transformed cells, the interaction between p130 CAS and tensin was disrupted, while the associations between p130 CAS , p125 FAK , and paxillin were unaffected. These results suggest that the BCR/ABL oncogene could alter the function of p130 CAS in at least three ways: tyrosine phosphorylation, inducing constitutive binding of CRKL to a domain in p130 CAS containing Tyr-X-X-Pro motifs (substrate domain), and disrupting the normal interaction of p130 CAS with the focal adhesion protein tensin. These alterations in the structure of signaling proteins in focal adhesion like structures could contribute to the known adhesion abnormalities in CML cells.

Tensin is a focal-adhesion molecule that binds to actin filaments and interacts with phosphotyrosine-containing proteins. To analyse tensin's function in mammals, we have cloned tensin cDNAs from human and cow. The isolated approx. 7.7-kb human cDNA contains an open reading frame encoding 1735 amino acid residues. The amino acid sequence of human tensin shares 60% identity with chicken tensin, and contains all the structural features described previously in chicken tensin. This includes the actin-binding domains, the Src homology domain 2, and the region similar to a tumour suppressor, PTEN. Two major differences between human and chicken tensin are (i) the lack of the first 54 residues present in chicken tensin, and (ii) the addition of 34- and 38-residue inserts in human and bovine tensin. In addition, our interspecies sequencing data have uncovered the presence of a glutamine/CAG repeat that appears to have expanded in the course of evolution. Northern-blot analysis reveals a 10-kb message in most of the human tissues examined. An additional 9-kb message is detected in heart and skeletal muscles. The molecular mass predicted from the human cDNA is 185 kDa, although both endogenous and recombinant human tensin migrate as 220-kDa proteins on SDS/PAGE. The discrepancy is due to the unusually low electrophoretic mobility of the central region of the tensin polypeptide (residues 306-981). A survey of human prostate and breast cancer cell lines by Western-blot analysis shows a lack of tensin expression in most cancer cell lines, whereas these lines express considerable amounts of focal-adhesion molecules such as talin and focal-adhesion kinase. Finally, tensin is rapidly cleaved by a focal-adhesion protease, calpain II. Incubation of cells with a calpain inhibitor, MDL, prevented tensin cleavage and induced morphological change in these cells, suggesting that cleavage of tensin and other focal-adhesion constituents by calpain disrupts maintenance of normal cell shape.

Typically, multiple cytokines act in concert to mediate a desired immunological response, and thus more effective therapeutics may be achieved by combining several cytokines with potentially synergistic activities. We have developed a series of bi-functional cytokine fusion proteins which, when additionally linked to an intact antibody (or the Fc portion of an antibody) in a variety of configurations, can be specifically targeted. We focus here mainly on the synergizing cytokine combination interleukin-2/interleukin-12 (IL-2/IL-12), but also demonstrate the utility of this approach with interleukin-4/granulocyte-macrophage colony-stimulating factor (IL-4/GM-CSF). Cytokine activity was retained in constructs where the cytokines were fused in tandem at the carboxyl terminus of the Fc or antibody heavy (H) chain, as well as in constructs where one cytokine was fused at the carboxyl terminus of the H chain while the second cytokine was fused to the amino terminus of either the H or light (L) chain variable region. Even in such constructs, antigen binding of the antibody-cytokine fusion proteins could be maintained. In the context of bi-functional fusion proteins, hetero-dimeric IL-12 could be expressed either in a single-chain form, or maintained as a heterodimer in which the p40 subunit was fused to IL-2. These IL-12/IL-2 bi-functional fusion proteins were shown to induce extremely high levels of interferon-gamma (IFN-gamma), similar to the synergy normally seen with the combined application of the individual cytokines. In addition, these bifunctional molecules were shown to have striking anti-tumor activity as either gene therapy or as an antibody cytokine(s) fusion protein, and may provide a useful approach to the treatment of cancer.