In vivo selection of retrovirally transduced hematopoietic stem cells.

One of the main impediments to effective gene therapy of blood disorders is the resistance of human hematopoietic stem cells to stable genetic modification. We show here that a small minority of retrovirally transduced stem cells can be selectively enriched in vivo, which might be a way to circumvent this obstacle. We constructed two retroviral vectors containing an antifolate-resistant dihydrofolate reductase cDNA transcriptionally linked to a reporter gene. Mice were transplanted with transduced bone marrow cells and then treated with an antifolate-based regimen that kills unmodified stem cells. Drug treatment significantly increased the percentage of vector-expressing peripheral blood erythrocytes, platelets, granulocytes, and T and B lymphocytes. Secondary transplant experiments demonstrated that selection occurred at the level of hematopoietic stem cells. This system for in vivo stem-cell selection provides a means to increase the number of genetically modified cells after transplant, and may circumvent an substantial obstacle to successful gene therapy for human blood diseases.

Human myeloid plasma membrane glycoprotein CD13 (gp150) is identical to aminopeptidase N.

To determine the primary structure of CD13, a 150-kD cell surface glycoprotein originally identified on subsets of normal and malignant human myeloid cells, we isolated the complete sequences encoding the polypeptide in overlapping complementary DNA (cDNA) clones. The authenticity of our cDNA clones was demonstrated by the ability of the coding sequences, subcloned in a retroviral expression vector, to mediate expression of bona fide CD13 molecules at the surface of transfected mouse fibroblasts. The nucleotide sequence predicts a 967 amino acid integral membrane protein with a single, 24 amino acid hydrophobic segment near the amino terminus. Amino-terminal protein sequence analysis of CD13 molecules indicated that the hydrophobic segment is not cleaved, but rather serves as both a signal for membrane insertion and as a stable membrane-spanning segment. The remainder of the molecule consists of a large extracellular carboxyterminal domain, which contains a pentapeptide consensus sequence characteristic of members of the zinc-binding metalloprotease superfamily. Sequence comparisons with known enzymes of this class revealed that CD13 is identical to aminopeptidase N, a membrane-bound glycoprotein thought to be involved in the metabolism of regulatory peptides by diverse cell types, including small intestinal and renal tubular epithelial cells, macrophages, granulocytes, and synaptic membranes prepared from cells of the central nervous system.

Transfer and expression of the gene encoding a human myeloid membrane antigen (gp150).

DNA from the human myeloid cell line HL-60 was cotransfected with the cloned thymidine kinase (tk) gene of herpes simplex virus into tk-deficient mouse L cells. tk-positive recipients expressing antigens detected on HL-60 cells were isolated with a fluorescence-activated cell sorter by use of a panel of monoclonal antibodies that detect epitopes on both normal and malignant myeloid cells. Independently sorted populations of transformed mouse cells showed concordant reactivities with four of the monoclonal antibodies in the panel (DU-HL60-4, MY7, MCS.2, and SJ-D1), which suggested that these antibodies reacted to products of a single human gene. A second round of DNA transfection and cell sorting was performed with donor DNA from primary transformants. Two different dominant selection systems were used to isolate secondary mouse L cell and NIH/3T3 cell transformants that coexpressed the same epitopes. Analysis of cellular DNA from secondary mouse cell subclones with a probe specific for human repetitive DNA sequences revealed a minimal human DNA complement containing a characteristic set of restriction fragments common to independently derived subclones. Two glycoproteins, of 130,000 (gp130) and 150,000 (gp150) mol wt, were specifically immunoprecipitated from metabolically labeled lysates of mouse cell transformants and were shown to contain [35S]methionine-labeled tryptic peptides identical to those of analogous glycoproteins expressed in the donor human myeloid cell line. Kinetic and biochemical analyses established that gp130 is a precursor that differs in its carbohydrate moiety from gp150, the mature form of the glycoprotein detected on the cell surface. The isolation of human gene sequences encoding gp150 in a mouse cell genetic background provides the possibility of molecularly cloning the gene and represents a general strategy for isolating human genes encoding differentiation-specific cell surface antigens.

Molecular cloning, expression, and chromosomal localization of the gene encoding a human myeloid membrane antigen (gp150).

DNA from a tertiary mouse cell transformant containing amplified human sequences encoding a human myeloid membrane glycoprotein, gp150, was used to construct a bacteriophage lambda library. A single recombinant phage containing 12 kilobases (kb) of human DNA was isolated, and molecular subclones were then used to isolate the complete gp150 gene from a human placental genomic DNA library. The intact gp150 gene, assembled from three recombinant phages, proved to be biologically active when transfected into NIH 3T3 cells. Molecular probes from the gp150 locus annealed with a 4.0-kb polyadenylated RNA transcript derived from human myeloid cell lines and from tertiary mouse cell transformants. The gp150 gene was assigned to human chromosome 15, and was subchromosomally localized to bands q25-26 by in situ hybridization. The chromosomal location of the gp150 gene coincides cytogenetically with the region assigned to the c-fes proto-oncogene, another human gene specifically expressed by myeloid cells.

Macrophage lineage switching of murine early pre-B lymphoid cells expressing transduced fms genes.

fms genes encoding either wild-type or constitutively activated colony-stimulating factor 1 receptors (CSF-1R) were introduced by retroviral infection into long-term mouse lymphoid cultures. Four early pre-B-cell lines transformed by the feline v-fms oncogene underwent spontaneous and irreversible differentiation to macrophages when transferred from RPMI 1640 to Iscove modified Dulbecco medium. Expression of wild-type human CSF-1R in early pre-B cells conferred no proliferative advantage unless human CSF-1 was added to the culture medium. A clonal, factor-dependent early pre-B-cell line (D1F9), selected for continuous growth on NIH 3T3 cell feeder layers producing human CSF-1, could be maintained in RPMI 1640 medium containing interleukin-7 (IL-7) but also differentiated to macrophages when grown in Iscove modified Dulbecco medium containing human CSF-1. The macrophages retained parental immunoglobulin gene rearrangements and proviral insertions, lost B-cell antigens, expressed butyrate esterase and MAC-1, were actively phagocytic, and no longer survived in IL-7. Unlike factor-independent v-fms transformants, the irreversible commitment of D1F9 cells to differentiate in the macrophage lineage could be suppressed by IL-7, depended on human (but not mouse) CSF-1, and was inhibited by an antibody to human CSF-1R. Signals mediated by transduced CSF-1R can therefore play a deterministic role in cell differentiation.

Transduction of human colony-stimulating factor-1 (CSF-1) receptor into interleukin-3-dependent mouse myeloid cells induces both CSF-1-dependent and factor-independent growth.

A retroviral vector encoding the receptor for human colony-stimulating factor-1 (CSF-1) was introduced into murine myeloid FDC-P1 cells which require interleukin-3 (IL-3) for their proliferation and survival in culture. Cells expressing the CSF-1 receptor (CSF-1R), selected by fluorescence-activated cell sorting in the continued presence of murine IL-3, formed colonies in semisolid medium and were able to proliferate continuously in liquid cultures containing human recombinant CSF-1. Thus, although they do not synthesize endogenous murine CSF-1R, FDC-P1 cells express the downstream components of the CSF-1 mitogenic pathway necessary for its signal-response coupling. After receptor transduction, slowly proliferating factor-independent variants that produced neither CSF-1 nor growth factors able to support the proliferation of parental FDC-P1 cells also arose. When the human CSF-1R was expressed in FDC-P1 cells under the control of an inducible metallothionein promoter, the frequencies of both CSF-1-responsive and factor-independent variants increased after heavy-metal treatment. In addition, a monoclonal antibody to human CSF-1R arrested colony formation by both the CSF-1-dependent and factor-independent cells but did not affect their growth in response to IL-3. Therefore, the induction of both the CSF-1-dependent and factor-independent phenotypes depended on expression of the transduced human CSF-1R.

Novel INK4 proteins, p19 and p18, are specific inhibitors of the cyclin D-dependent kinases CDK4 and CDK6.

Cyclin D-dependent kinases act as mitogen-responsive, rate-limiting controllers of G1 phase progression in mammalian cells. Two novel members of the mouse INK4 gene family, p19 and p18, that specifically inhibit the kinase activities of CDK4 and CDK6, but do not affect those of cyclin E-CDK2, cyclin A-CDK2, or cyclin B-CDC2, were isolated. Like the previously described human INK4 polypeptides, p16INK4a/MTS1 and p15INK4b/MTS2, mouse p19 and p18 are primarily composed of tandemly repeated ankyrin motifs, each ca. 32 amino acids in length, p19 and p18 bind directly to CDK4 and CDK6, whether untethered or in complexes with D cyclins, and can inhibit the activity of cyclin D-bound cyclin-dependent kinases (CDKs). Although neither protein interacts with D cyclins or displaces them from preassembled cyclin D-CDK complexes in vitro, both form complexes with CDKs at the expense of cyclins in vivo, suggesting that they may also interfere with cyclin-CDK assembly. In proliferating macrophages, p19 mRNA and protein are periodically expressed with a nadir in G1 phase and maximal synthesis during S phase, consistent with the possibility that INK4 proteins limit the activities of CDKs once cells exit G1 phase. However, introduction of a vector encoding p19 into mouse NIH 3T3 cells leads to constitutive p19 synthesis, inhibits cyclin D1-CDK4 activity in vivo, and induces G1 phase arrest.

Synthesis of membrane-bound colony-stimulating factor 1 (CSF-1) and downmodulation of CSF-1 receptors in NIH 3T3 cells transformed by cotransfection of the human CSF-1 and c-fms (CSF-1 receptor) genes.

NIH 3T3 cells cotransfected with the human c-fms proto-oncogene together with a 1.6-kilobase cDNA clone encoding a 256-amino-acid precursor of the human mononuclear phagocyte colony-stimulating factor CSF-1 (M-CSF) undergo transformation by an autocrine mechanism. The number of CSF-1 receptors on the surface of transformed cells was regulated by ligand-induced receptor degradation and was inversely proportional to the quantity of CSF-1 produced. A tyrosine-to-phenylalanine mutation at position 969 near the receptor carboxyl terminus potentiated its transforming efficiency in cells cotransfected by the CSF-1 gene but did not affect receptor downmodulation. CSF-1 was synthesized as an integral transmembrane glycoprotein that was rapidly dimerized through disulfide bonds. The homodimer was externalized at the cell surface, where it underwent proteolysis to yield the soluble growth factor. Trypsin treatment of viable cells cleaved the plasma membrane form of CSF-1 to molecules of a size indistinguishable from that of the extracellular growth factor, suggesting that trypsinlike proteases regulate the rate of CSF-1 release from transformed cells. The data raise the possibility that this form of membrane-bound CSF-1 might stimulate receptors on adjacent cells through direct cell-cell interactions.

Inhibition of cell proliferation by the Mad1 transcriptional repressor.

Mad1 is a basic helix-loop-helix-leucine zipper protein that is induced upon differentiation of a number of distinct cell types. Mad1 dimerizes with Max and recognizes the same DNA sequences as do Myc:Max dimers. However, Mad1 and Myc appear to have opposing functions. Myc:Max heterodimers activate transcription while Mad:Max heterodimers repress transcription from the same promoter. In addition Mad1 has been shown to block the oncogenic activity of Myc. Here we show that ectopic expression of Mad1 inhibits the proliferative response of 3T3 cells to signaling through the colony-stimulating factor-1 (CSF-1) receptor. The ability of over-expressed Myc and cyclin D1 to complement the mutant CSF-1 receptor Y809F (containing a Y-to-F mutation at position 809) is also inhibited by Mad1. Cell cycle analysis of proliferating 3T3 cells transfected with Mad1 demonstrates a significant decrease in the fraction of cells in the S and G2/M phases and a concomitant increase in the fraction of G1 phase cells, indicating that Mad1 negatively influences cell cycle progression from the G1 to the S phase. Mutations in Mad1 which inhibit its activity as a transcription repressor also result in loss of Mad1 cell cycle inhibitory activity. Thus, the ability of Mad1 to inhibit cell cycle progression is tightly coupled to its function as a transcriptional repressor.

DNA-binding specificity and trans-activating potential of the leukemia-associated E2A-hepatic leukemia factor fusion protein.

Hybrid transcription factors, resulting from gene fusions in the wake of chromosomal translocations, have been implicated in leukemogenesis, but their precise contributions to oncogenic conversion remain unclear. The E2A-HLF fusion gene, formed by a t(17;19)(q22;p13) in childhood pro-B-cell acute lymphoid leukemia, encodes a hybrid protein that contains the trans-activation domain of E2A (E12/E47) linked to the bZIP DNA-binding and dimerization domain of hepatic leukemia factor (HLF). Here we report that both HLF and E2A-HLF bind to a 10-bp consensus sequence, 5'-GTTACGTAAT-3', with a core dyad-symmetric motif characteristic of the bZIP scissors-grip model of DNA binding. A probe containing this sequence bound chimeric E2A-HLF proteins in nuclear extracts of a leukemic cell line (UOC-B1) containing the t(17;19), as demonstrated by complexes supershifted with antibodies specific for amino-terminal epitopes of E2A or carboxyl-terminal eptiopes of HLF. E2A-HLF functioned as a potent trans activator of reporter gene expression from a plasmid that contained the consensus DNA-binding sequence. Interestingly, wild-type HLF was restricted in its capacity to act as a trans activator, functioning in human fetal kidney cells but not HepG2 hepatocarcinoma cells or NIH 3T3 mouse fibroblasts. The ability of the E2A-HLF hybrid protein to bind DNA in a sequence-specific manner and trans activate the expression of artificial reporter genes suggests that it could subvert transcriptional programs that normally control the growth, differentiation, and survival of lymphoid progenitor cells.

Complementation of growth factor receptor-dependent mitogenic signaling by a truncated type I phosphatidylinositol 4-phosphate 5-kinase.

Substitution of phenylalanine for tyrosine at codon 809 (Y809F) of the human colony-stimulating factor 1 (CSF-1) receptor (CSF-1R) impairs ligand-stimulated tyrosine kinase activity, prevents induction of c-MYC and cyclin D1 genes, and blocks CSF-1-dependent progression through the G1 phase of the cell cycle. We devised an unbiased genetic screen to isolate genes that restore the ability of CSF-1 to stimulate growth in cells that express mutant CSF-1R (Y809F). This screen led us to identify a truncated form of the murine type Ibeta phosphatidylinositol 4-phosphate 5-kinase (mPIP5K-Ibeta). This truncated protein lacks residues 1 to 238 of mPIP5K-Ibeta and is catalytically inactive. When we transfected cells expressing CSF-1R (Y809F) with mPIP5K-Ibeta (delta1-238), CSF-1-dependent induction of c-MYC and cyclin D1 was restored and ligand-dependent cell proliferation was sustained. CSF-1 normally triggers the rapid disappearance of CSF-1R (Y809F) from the cell surface; however, transfection of cells with mPIP5K-Ibeta (delta1-238) stabilized CSF-1R (Y809F) expression on the cell surface, resulting in elevated levels of ligand-activated CSF-1R (Y809F). These results suggest a role for PIP5K-Ibeta in receptor endocytosis and that the truncated enzyme compensated for a mitogenically defective CSF-1R by interfering with this process.

Differentiation and the multiple drug resistance phenotype in human leukemic cells.

A common feature of mammalian cell lines selected for multiple drug resistance is the overexpression of a high-molecular-weight surface membrane glycoprotein(s). While its amount has been shown to be related to the degree of resistance of such cells, its function in this phenomenon remains obscure. Because there are some biochemical and functional similarities between drug-resistant cells and differentiated cells, we asked if resistance-associated glycoproteins were also associated with cellular differentiation. Using three monoclonal antibodies against antigens known to be associated with differentiation and three monoclonal antibodies that distinguish our multiple drug-resistant human leukemic CEM/VLB100 cells from their drug-sensitive counterparts, we found that the resistant cells were neither altered in their apparent state of differentiation nor were they altered in their ability to respond to a differentiation stimulus with the phorbol ester, 12-O-tetradecanoylphorbol-13-acetate. We did find, however, that one antibody that recognizes the resistance-associated glycoprotein, Mr 180,000 glycoprotein (gp180), was only minimally altered in amount bound after treatment with the phorbol ester, but two other resistance-associated glycoproteins, Mr 155,000 glycoprotein (gp155) and, to a lesser extent, Mr 130,000 glycoprotein (gp130), were reduced in expression after this treatment. We suggest that the function of the previously described "marker" glycoprotein associated with multiple drug resistance remains unknown, but that the expression of two other resistance-associated glycoproteins also appears to be related to cellular differentiation or maturation in these cells.

Monoclonal antibodies to glycoproteins of Vinca alkaloid-resistant human leukemic cells.

Drug resistance is a major problem in the successful treatment of cancer. Resistance to one drug is often associated with cross-resistance to other anticancer agents. This is commonly seen with the "natural product" anticancer drugs such as the Vinca alkaloids and anthracyclines. In experimental systems, specific changes in plasma membranes characterize this "multiple drug resistance." The most prominent of these is the enhanced expression in several systems of high-molecular-weight glycoproteins ranging from Mr approximately equal to 150,000 to approximately equal to 180,000, the amount of which has been shown to be related to the degree of drug resistance. We report here the production of three monoclonal antibodies that bind preferentially to the surfaces of cultured human leukemic lymphoblasts resistant to the Vinca alkaloid vinblastine. Each antibody recognizes a surface membrane glycoprotein with molecular weights of 180,000 to 210,000. Additionally, two of the antibodies also recognize a second surface glycoprotein with molecular weights of either approximately equal to 155,000 or approximately equal to 130,000. All of these glycoproteins are overexpressed in the alkaloid-resistant cells. While a Mr approximately equal to 180,000 protein has been shown to be associated with multiple drug resistance, the other two glycoproteins have not been described previously in these cells. These antibodies may be useful in studies of the mechanisms of drug resistance, as well as in screening cells from drug-resistant patients for these resistance-associated glycoproteins.

Reversal of Vinca alkaloid resistance but not multiple drug resistance in human leukemic cells by verapamil.

We examined the ability of verapamil, a Ca2+ channel blocker, to overcome Vinca alkaloid and multiple drug resistance in our CEM/VLB100 and CEM/DOX human leukemic lymphoblasts. Compared with the parent CCRF-CEM cells, CEM/VLB100 cells are approximately equal to 200- to 800-fold resistant to vinblastine and express cross-resistance to vincristine, doxorubicin, and other "natural product" drugs, as determined by comparing 50% inhibitory concentrations in a 48-h growth inhibition assay. Verapamil (10 microM) decreased the 50% inhibitory concentrations for Vinca alkaloids in CEM/VLB100 cells by approximately equal to 75- to 85-fold but caused only slight (approximately equal to 2- to 5-fold) decreases in 50% inhibitory concentrations for anthracyclines, epipodophyllotoxins, and other tubulin-binding drugs (colchicine and podophyllotoxin). Qualitatively similar results were obtained with doxorubicin-resistant cells, termed CEM/DOX; verapamil caused a 19-fold increase in doxorubicin toxicity but 67- and 3500-fold increases in the toxicities of vinblastine and vincristine, respectively. These results indicate that the effect of verapamil is relatively greater for Vinca alkaloids, with less pronounced effects for the other natural product drugs against which these cells express multiple drug resistance. In flow cytometric studies, individually nontoxic or minimally toxic concentrations of vinblastine plus verapamil caused measurable accumulation in the G2 + M phase as early as 4 h after the drug combination was added to cultures of CEM/VLB100 cells; this finding correlated with a comparable increase in the number of cells in mitosis and measurable decreases in the total number of cells. Since similar effects on cell cycle distribution, percentage of cells in mitosis, and cell number were seen when CEM/VLB100 cells were treated with toxic concentrations of vinblastine alone, we conclude that the primary toxicity of the vinblastine-verapamil combination stems from the alkaloid. Further, the rapid lytic effect of the drug combination was associated with cellular vacuolization. The vacuoles did not stain for lipids, and increases in their number were evident within 2 to 4 h after drug treatment. We suggest that verapamil enhances Vinca alkaloid cytotoxicity by altering "cryptic" cytotoxic targets, possibly related to these vacuoles, through some as yet undefined membrane effect.

Effect of antitumor diarylsulfonylureas on in vivo and in vitro mitochondrial structure and functions.

Diarylsulfonylureas are novel oncolytic agents shown to have therapeutic activity against both rodent solid tumors and xenografts of human tumors in mice. Previous studies have shown that diarylsulfonylureas localize in mitochondria and cause morphological changes in these organelles. We have investigated the mechanism of action of diarylsulfonylureas, namely, N-(5-indanylsulfonyl)-N'-(4-chlorophenyl)urea (ISCU) and the N-4-methyl analogue (MPCU), by studying their effect on mitochondrial morphology and uptake of rhodamine 123 in GC3/c1 cells in culture and the oxidative phosphorylation in isolated mitochondria from mouse liver, using pyruvate-malate and succinate as substrates. Morphometric analysis of mitochondria in GC3/c1 cells exposed to ISCU showed that ISCU (165 microM) doubled the mitochondrial size after 24-h exposure in culture. Also, ISCU (100 microM), like 40 microM carbonylcyanide p-trifluoromethoxyphenylhydrazone, significantly reduced the rhodamine 123 uptake by GC3/c1 cells studied by flow cytometry. In isolated mitochondria both ISCU and MPCU uncoupled oxidative phosphorylation at 50 microM, with pyruvate-malate as substrate, as was indicated by a significant increase in the State 4 oxygen consumption. This resulted in the loss of ADP phosphorylation and, therefore, the ADP/oxygen ratio was reduced to zero and the respiratory control ratio to one. The succinate oxidation was also significantly impaired by ISCU, causing some decrease in ADP phosphorylation. On the other hand, MPCU did not exhibit any significant effect on the oxidation of succinate. At concentrations of lower than 50 microM, both of these compounds exhibited a deleterious effect, causing damage to mitochondrial functions in the presence of pyruvate-malate as substrates. These data confirm, through morphometric analysis, our previous qualitative observations of abnormal mitochondrial morphology observed in GC3/c1 cells grown in the presence of high concentrations of ISCU and MPCU and further suggest that diarylsulfonylureas, by uncoupling mitochondrial oxidative phosphorylation, may lower cellular ATP. It is probable that this mechanism contributes, at least partially, to cytotoxicity in GC3/c1 cells exposed to high concentrations of ISCU for relatively brief periods (2 to 4 h) and possibly contributes to cytotoxicity at drug concentrations that can be achieved in rodents.

Effects of cytotoxicity of 2-chloro-2'-deoxyadenosine and 2-bromo-2'-deoxyadenosine on cell growth, clonogenicity, DNA synthesis, and cell cycle kinetics.

The cytotoxic effects of the adenosine deaminase resistant analogues 2-bromo-2'-deoxyadenosine (2-BrdAdo) and 2-chloro-2'-deoxyadenosine (2-CldAdo) have been compared with those of deoxyadenosine (dAdo). Like 2-CldAdo, 2-BrdAdo is highly effective in inhibiting the growth of many T-lymphoblastoid, B-lymphoblastoid, and myeloid cell lines in culture. Concentrations required to inhibit growth of CCRF-CEM human T-lymphoblastoid cells by 50% (IC50) are: 2-CldAdo, 0.045 microM; 2-BrdAdo, 0.068 microM; dAdo, 0.9 microM in the presence of 5 microM erythro-9-(2-hydroxy-3-nonyl)adenine. Like dAdo, 2-BrdAdo causes a much greater decrease in DNA synthesis than in RNA and protein synthesis. For each of the nucleosides the concentration required to cause 50% inhibition of DNA synthesis (as measured by thymidine incorporation) in an 18-h exposure is very similar to the IC50 for growth and to the concentration required to decrease viability (clonogenicity) over 18 h by 50% (EC50). A fraction of CCRF-CEM cells (approximately equal to 30%) is resistant to killing by exposure to 2-BrdAdo or 2-CldAdo for 4 h at concentrations 100 times the EC50, but 3% of cells are resistant to exposure for 4 h to a concentration of dAdo 3 times the EC50. Each of the three nucleosides causes accumulation of cells in S phase, the accumulation becoming more marked with longer periods of exposure and with higher concentrations of nucleoside. During exposures for 18-24 h at a concentration of nucleoside near the EC50 most cells accumulate in S, with most in early S, whereas exposure to concentrations greater than EC95 accumulates cells at the G1/S border. This suggests that loss of viability is associated with a blockade of some process specifically occurring at the initiation of S phase. At an optimum dosage schedule, 2-BrdAdo and 2-CldAdo have similar therapeutic effects against L1210 in vivo, both producing over 99% cell kill, but the optimum dosage of 2-CldAdo is lower.

Aminopeptidase N is a receptor for tumor-homing peptides and a target for inhibiting angiogenesis.

Phage that display a surface peptide with the NGR sequence motif home selectively to tumor vasculature in vivo. A drug coupled to an NGR peptide has more potent antitumor effects than the free drug [W. Arap et al., Science (Washington DC), 279: 377-380, 1998]. We show here that the receptor for the NGR peptides in tumor vasculature is aminopeptidase N (APN; also called CD13). NGR phage specifically bound to immunocaptured APN and to cells engineered to express APN on their surface. Antibodies against APN inhibited in vivo tumor homing by the NGR phage. Immunohistochemical staining showed that APN expression is up-regulated in endothelial cells within mouse and human tumors. In another tissue that undergoes angiogenesis, corpus luteum, blood vessels also expressed APN, but APN was not detected in blood vessels of various other normal tissues stained under the same conditions. APN antagonists specifically inhibited angiogenesis in chorioallantoic membranes and in the retina and suppressed tumor growth. Thus, APN is involved in angiogenesis and can serve as a target for delivering drugs into tumors and for inhibiting angiogenesis.

Rapamycin causes poorly reversible inhibition of mTOR and induces p53-independent apoptosis in human rhabdomyosarcoma cells.

The mammalian target of rapamycin (mTOR) has been shown to link growth factor signaling and posttranscriptional control of translation of proteins that are frequently involved in cell cycle progression. However, the role of this pathway in cell survival has not been demonstrated. Here, we report that rapamycin, a specific inhibitor of mTOR kinase, induces G1 cell cycle arrest and apoptosis in two rhabdomyosarcoma cell lines (Rh1 and Rh30) under conditions of autocrine cell growth. To examine the kinetics of rapamycin action, we next determined the rapamycin sensitivity of rhabdomyosarcoma cells exposed briefly (1 h) or continuously (6 days). Results demonstrate that Rh1 and Rh30 cells were equally sensitive to rapamycin-induced growth arrest and apoptosis under either condition. Apoptosis was detected between 24 and 144 h of exposure to rapamycin. Both cell lines have mutant p53; hence, rapamycin-induced apoptosis appears to be a p53-independent process. To determine whether induction of apoptosis by rapamycin was specifically due to inhibition of mTOR signaling, we engineered Rh1 and Rh30 clones to stably express a mutant form of mTOR that was resistant to rapamycin (Ser2035-->Ile; designated mTOR-rr). Rh1 and Rh30 mTOR-rr clones were highly resistant (>3000-fold) to both growth inhibition and apoptosis induced by rapamycin. These results are the first to indicate that rapamycin-induced apoptosis is mediated by inhibition of mTOR. Exogenous insulin-like growth factor (IGF)-I protected both Rh1 and Rh30 from apoptosis, without reactivating ribosomal p70 S6 kinase (p70S6K) downstream of mTOR. However, in rapamycin-treated cultures, the response to IGF-I differed between the cell lines: Rh1 cells proliferated normally, whereas Rh30 cells remained arrested in G1 phase but viable. Rapamycin is known to inhibit synthesis of specific proteins but did not inhibit synthesis or alter the levels of mTOR. To examine the rate at which the mTOR pathway recovered, the ability of IGF-I to stimulate p70S6K activity was followed in cells treated for 1 h with rapamycin and then allowed to recover in medium containing > or =100-fold excess of FK506 (to prevent rapamycin from rebinding to its cytosolic receptor FKBP-12). Our results indicate that, in Rh1 cells, rapamycin dissociates relatively slowly from FKBP-12, with a t1/2 of approximately 17.5 h. in the presence of FK506, whereas there was no recovery of p70S6K activity in the absence of this competitor. This was of interest because rapamycin was relatively unstable under conditions of cell culture having a biological t1/2 of approximately 9.9 h. These results help to explain why cells are sensitive following short exposures to rapamycin and may be useful in guiding the use of rapamycin analogues that are entering clinical trials as novel antitumor agents.

Constitutive c-myc expression in an IL-3-dependent myeloid cell line suppresses cell cycle arrest and accelerates apoptosis.

In the murine interleukin 3 (IL-3)-dependent myeloid cell line 32D, down-regulation of c-myc and ornithine decarboxylase (ODC) expression is an immediate response to IL-3 deprivation. This is followed by an accumulation of cells in the G1 phase of the cell cycle, and eventual cell death. However, clones of 32D cells harboring an expression vector which constitutively expresses murine c-myc did not down-regulate ODC transcripts in response to IL-3 withdrawal, and they failed to G1 arrest. Moreover, in contrast to control cultures in which the majority of death occurred following G1 arrest, c-myc clones rapidly initiated a program of cell death characteristic of apoptosis following IL-3 deprivation, and their subsequent loss of viability occurred with accelerated kinetics. The premature induction of apoptosis in cells harboring a deregulated c-myc gene suggests that apoptosis may be an important mechanism in the elimination of hematopoietic cells harboring mutations, such as constitutive c-myc expression, which imbalance normal cell cycle regulatory controls.

Cloning and characterization of murine p16INK4a and p15INK4b genes.

Progression through the G1 phase of the cell cycle is regulated in part by the D-type cyclin-dependent kinases, cdk4 and cdk6. Genes encoding two specific inhibitors of these kinases, human p16(INK4a/MTS1) and p15(INK4b/MTS2), map to a region of common cytogenetic abnormalities on chromosome 9p21. The murine cognates of these genes were isolated and identified as mouse p16INK4a and p15INK4b based on their homology to their human counterparts and their selective transcriptional induction by SV40T-antigen and TGF-beta, respectively. Both genes map to position C3-C6 on mouse chromosome 4, in a region syntenic with human chromosome 9p. Amplification of polyadenylated mRNA by polymerase chain reactions revealed no expression of mouse p16INK4a in many normal tissues, whereas p15INK4b was expressed ubiquitously. Like human p16INK4a, mouse p16INK4a binds specifically to cdk4 and cdk6 in vitro and inhibits the phosphorylation of the retinoblastoma protein, pRb, by each of these cyclin D-dependent kinases. In mouse MEL erythroleukemia cells, p16INK4a associates preferentially with cdk6 under conditions where cdk4 and cdk6 are coexpressed at equivalent levels. Expression vectors encoding human or mouse p16INK4a caused G1 phase arrest in NIH3T3 fibroblasts, and cyclin D1- and cdk4-dependent pRb kinase activities were inhibited in the p16INK4a-arrested cells.