SIGNAL TRANSDUCTION BY TYROSINE
KINASES/MOLECULAR MECHANISMS OF CELLULAR
RESISTANCE TO CYTOTOXIC DRUGS

GARY D. KRUH, M.D., Ph.D., Associate Member

The research in our laboratory focuses on two distinct areas--signal transduction by protein kinases and the molecular mechanisms of cellular resistance to chemotherapeutic agents. In the area of signal transduction, we are interested in Arg and c-Abl, which are ubiquitously expressed proteins that represent the mammalian members of the Abelson family of nonreceptor protein tyrosine kinases (TKs). As such, Arg and c-Abl are part of a large family of proteins, which are known to be involved in critical cellular events such as growth differentiation, and in tumorigenesis. c-Abl plays a central role in pre-B lymphomas induced by Abelson murine leukemia virus (AMLV) in mice, and chronic myelogenous leukemia and some forms of acute lymphocytic leukemia in humans. Arg is located in the cytoplasm, whereas c-Abl is located in the nucleus and cytoplasm, suggesting that the two proteins are likely to have distinct, but possibly overlapping, functions. However, neither the normal functions of Arg and c-Abl, nor the biochemical mechanisms that regulate their activities, are well understood. To improve our understanding of these two proteins, our laboratory is currently exploring the biological function of Arg.

Cellular resistance to chemotherapeutic agents, our second area of focus, represents a major obstacle to the treatment of disseminated malignancies. Understanding the molecular basis of drug resistance is essential to improvements in cancer treatment. Our laboratory is interested in membrane-based mechanisms of cytotoxic drug resistance, and we are particularly interested in understanding the role of membrane proteins that function to efflux drugs from the cell.

Several strategies are being employed to define Arg protein sequences that are important in signaling. The first involves a structure-function analysis of transformation activity by Arg and c-Abl. These two proteins contain a series of known motifs including src-homology (SH) 3, SH2 and TK motifs, as well as large carboxyl terminal domains. We previously found that Arg is not activated by fusion of AMLV Gag sequences at the junction of its SH2 and SH3 domains. This is unlike c-Abl for which fusion of the Gag protein with c-Abl results in c-Abl activation. The basis for this difference in transforming activity was explored by analyzing chimeric Arg/c-Abl proteins. Substitution of the SH2 and TK domains with the Arg counterparts did not alter Abl transforming activity; however, substitution of the c-Abl Cterminus completely abolished transforming activity. Identification of the Cterminal Arg and c-Abl sequences that distinguish the transforming activities of the two proteins is under exploration. This information may lead to the identification of protein motifs that also distinguish the normal functions of the two proteins. Several known interaction domains reside within this ~650 amino acid domain in Arg and c-Abl, including three functional SH3 domain binding sites located just downstream of the TK domains, as well as RNA polymerase II binding sites at the extreme C-terminus. We are currently identifying the specific Arg Cterminal sequences that suppress c-Abl oncogenic activity.

The identification of Arg interacting proteins is a direct method for elucidating the pathways through which Arg signals. To accomplish this, we used the yeast two-hybrid system; the bait comprised the segment of the Arg C-terminal domain that contains three SH3 binding sites. Using this approach, we identified two novel Arg interacting proteins, Arg binding protein 1 (ArgBP1; Wang et al., Oncogene 12:1921, 1996) and ArgBP2 (Wang et al., J. Biol. Chem. 272:17542, 1997). Analysis of ArgBP1 revealed that: 1) the protein contains a Cterminal SH3 domain, several PEST sequences, a serine-rich domain and an SH3 binding site, 2) the protein is found to be ubiquitously expressed as two transcripts of ~2.2 kb and ~8 kb, with highest levels in brain, heart and testis, 3) ArgBP1 is located in the cytoplasm, 4) the ArgBP1 SH3 domain binds to a C-terminal Arg SH3-binding site, and 5) an N-terminal ArgBP1 proline-rich sequence binds to the Arg SH3 domain. The association of ArgBP1 with Arg in living cells was confirmed by coimmunoprecipitation in cotransfected COS cells. The similarity of the Arg and ArgBP1 expression pattern and subcellular localization, and the potential for a highly specific and strong association mediated by two pairs of SH3 domain/proline-rich motif interactions, suggests that ArgBP1 is likely to be a physiological regulator and/or effector of Arg function. In addition, a Xenopus homologue of ArgBP1 has been described, and this protein has a developmental pattern that is restricted to the CNS. Together these data suggest a possible role for Arg and c-Abl in CNS development. Further analysis of signaling by ArgBP1 is underway.

The second Arg interacting protein, ArgBP2, is a novel protein that harbors three C-terminal SH3 domains, two of which bind to the Arg and c-Abl proline-rich sequences. ArgBP2 interacts with Arg and c-Abl and is phosphorylated by these kinases in cotransfection studies. The ArgBP2 transcript is widely expressed, but is most abundant in heart. In most cells, ArgBP2 is located in stress fibers and focal adhesions; however, in cardiac cells, ArgBP2 is located at Z discs. Together these observations suggest that ArgBP2 is an adapter protein that functions to transduce signals from Abl family kinases, and possibly other kinases, to the cytoskeleton. In addition, ArgBP2 is the first signaling protein that has been reported to be localized in Z discs (Wang et al., J. Biol. Chem. 272:17542, 1997). This observation indicates that the Z disc, which is unique to striated muscle, is a target of signal transduction cascades. Studies designed to elucidate the function of ArgBP2 in signaling to the cytoskeleton and in cardiac myocytes are underway.

Peutz-Jegher Syndrome (PJS) is a rare, dominantly inherited condition characterized by intestinal polyposis and increased risk of cancers of the colon, breast, testis and ovary. Recently, LKB1, the gene defective in PJS was reported. LKB1, encodes a serine threonine kinase; little is known about this protein aside from its loss in PJS. Our laboratory has undertaken fundamental studies of this kinase. The ultimate goal is to use LKB1 as a probe to elucidate pathways involved in inherited and sporadic colon cancer, and to understand the nature of the G0 (resting state) of the cell cycle. These studies are designed to define the type of cell cycle perturbations induced by LKB1 expression, and to identify potential interacting proteins and relevant signal transduction pathways.

FUNCTION OF THE MRP/cMOAT SUBFAMILY OF TRANSPORTERS. BELINSKY, LEE, ZENG, BAIN, KRUH

The paradigm for membrane-based mechanisms of cytotoxic drug resistance is Pglycoprotein (Pgp), an ATP-binding cassette (ABC) transporter that functions to efflux natural product cytotoxic drugs such as doxorubicin, vinca alkaloids and Taxol from the cell. It has become clear in the past few years that there are, in addition to Pgp, other ABC transporters, which also confer resistance to this class of clinically important chemotherapeutic agents. The multidrug resistance-associated protein (MRP) is a widely expressed ABC transporter that shares surprisingly little amino acid identity with Pgp. Previous work by our laboratory and others has demonstrated that MRP also functions as an ATP-dependent cellular efflux pump for natural product cytotoxic drugs. However, while the two pumps confer a similar drug resistance phenotype, their substrate specificities are distinct. In contrast to Pgp, which pumps natural product drugs in the absence of other cellular components, MRP appears to cotransport these agents with free glutathione. In addition, MRP is capable of transporting a wide spectrum of anionic organic conjugates, including glutathione, glucuronide, and sulfate conjugates.

Understanding that MRP is, in fact, an organic anion transporter quickly led to the identification by other laboratories of the canalicular multi-specific organic ion transporter (cMOAT), which is highly related to MRP, and functions in the hepatobiliary transport of anionic conjugates and lipophilic cytotoxic drugs. Increasing evidence has also linked cMOAT with cellular drug resistance. Based upon the important biological functions of MRP and cMOAT, our laboratory sought to identify additional MRP/cMOAT subfamily members, with the view that they might also play important roles in cellular and hepatobiliary/renal excretion. We have now identified three additional subfamily members, which we designated MOAT-B, MOAT-C and MOAT-D (1). Amino acid identities among MRP/cMOAT family members is shown in Table 1. Of the three new family members, MOAT-D is most closely related to MRP and cMOAT. A comparison of hydropathy profiles of the MRP/cMOAT subfamily is shown in Figure 1. Most mammalian ABC transporters have a 6+6 configuration of transmembrane (TM) segments in which each set of six TM segments is followed by an ATP-binding fold. However, MRP and cMOAT have an additional N-terminal hydrophobic domain, which is thought to harbor five additional TM segments. Of the three new subfamily members, only MOAT-D has this distinctive N-terminal structural feature. Thus, based upon amino acid identity and overall protein topology, MOAT-D most closely resembles MRP and cMOAT. This similarity suggests that it may share some of the same substrates as MRP and cMOAT. We are currently using molecular and biochemical approaches to understand the functions of MOAT-B, C and D.

Retinoic acid is a differentiating agent that is effective in the treatment of a subclass of adult leukemias, and is currently being tested as a treatment for other malignancies and as a cancer preventive agent. As with other chemotherapeutic agents, the development of cellular resistance is a major obstacle to the use of retinoids. To explore cellular mechanisms of resistance to this class of agents, we developed a highly resistant HL60 cell line by step-wise selection in all-trans retinoic acid. We are currently using a variety of approaches to define the resistance mechanisms using this cellular model. One approach involves the identification, using comparative genomic hybridization (CGH), of stable chromosomal alterations. Using this approach, we have identified several chromosomal alterations in the resistant cells, and this information is being used to identify and test candidate resistance genes.

PUBLICATIONS

1.   BELINSKY, M.G., BAIN, L.J., BALSARA, B.B., TESTA, J.R., KRUH, G.D. Characterization of MOAT-C and MOAT-D, new members of the MRP/cMOAT subfamily of transporter proteins. J. Natl. Cancer Inst. 90: 1735-1741, 1998.

BELINSKY, M.B., KRUH, G.D. MOAT-E (ARA) is a full length MRP/cMOAT subfamily transporter expressed in liver and kidney. Br. J. Cancer (in press).

SHEN, H., SCHULTZ, M., KRUH. G.D. Increased expression of DNA-dependent protein kinase confers resistance to adriamycin. BBA 138:131-138, 1998.

^§    Fox Chase researcher

^a   L.J. Bain: Present address-Clemson University, Clemson, SC 29631

Illustrations or unpublished data in these reports should not be used without permission of the author.

Fox Chase Cancer Center

Scientific Report 1998