Search

Background The nucleoside analog, ARC (NSC 188491) is a recently characterized transcriptional inhibitor that selectively kills cancer cells and has the ability to perturb angiogenesis in vitro. In this study, the mechanism of action of ARC was further investigated by comparing in vitro and in vivo activity with other anti-neoplastic purines. Methods Structure-based homology searches were used to identify those compounds with similarity to ARC. Comparator compounds were then evaluated alongside ARC in the context of viability, cell cycle and apoptosis assays to establish any similarities. Following this, biological overlap was explored in detail using gene-expression analysis and kinase inhibition assays. Results Results demonstrated that sangivamycin, an extensively characterized pro-apoptotic nucleoside isolated from Streptomyces, had identical activity to ARC in terms of 1) cytotoxicity assays, 2) ability to induce a G2/M block, 3) inhibitory effects on RNA/DNA/protein synthesis, 4) transcriptomic response to treatment, 5) inhibition of protein kinase C, 6) inhibition of positive transcription elongation factor b (P-TEFb), 7) inhibition of VEGF secretion, and 8) activity within hollow fiber assays. Extending ARC activity to PKC inhibition provides a molecular basis for ARC cancer selectivity and anti-angiogenic effects. Furthermore, functional overlap between ARC and sangivamycin suggests that development of ARC may benefit from a retrospective of previous sangivamycin clinical trials. However, ARC was found to be inactive in several xenograft models, likely a consequence of rapid serum clearance. Conclusion Overall, these data expand on the biological properties of ARC but suggest additional studies are required before it can be considered a clinical trials candidate.

Antibodies can be conjugated to effector molecules to derive targeted therapeutics with properties such as cell-specific cytotoxicity. The murine anti-CD22 antibody RFB4 linked to a member of the ribonuclease A superfamily, Onconase (Onc), becomes a potential drug candidate for non-Hodgkin's lymphoma. Onc is currently in Phase III clinical trials for unresectable malignant mesothelioma but conjugation to RFB4 considerably enhances its specificity for CD22+ lymphomas. RFB4-targeted Onc is effective in preclinical models, causes little non-specific toxicities in mice, and has favorable formulation properties. Derivatization and conjugation of RFB4 and Onc have been optimized.

The alpha subunits of five members of the Gi family of bovine brain proteins were reconstituted with adenylate cyclase in phospholipid vesicles. Our results support both the views that much of the inhibition of the enzyme by Gi is due to the action of its liberated beta gamma subunits, and that the alpha subunits themselves interact with the enzyme. Inhibition of basal or Gs-stimulated adenylate cyclase activity is small or undetectable by alpha subunits of Go, Go* and Gi-2B. On the other hand, adenylate cyclase activity is stimulated by the alpha subunits of Gi-1 and Gi-2A. The G proteins act in the absence of added GTP when reconstituted with phospholipids of low but not high fluidity.

Purpose: Activities distinct from inhibition of Bcr/abl have led to adaphostin (NSC 680410) being described as “a drug in search of a mechanism.” In this study, proteomic analysis of adaphostin-treated myeloid leukemia cell lines was used to further elucidate a mechanism of action. Experimental Design: HL60 and K562 cells treated with adaphostin for 6, 12, or 24 h were analyzed using two-dimensional PAGE. Differentially expressed spots were excised, digested with trypsin, and analyzed by liquid chromatography–tandem mass spectrometry. The contribution of the redox-active hydroquinone group in adaphostin was also examined by carrying out proteomic analysis of HL60 cells treated with a simple hydroquinone (1,4-dihydroxybenzene) or H2O2. Results: Analysis of adaphostin-treated cells identified 49 differentially expressed proteins, the majority being implicated in the response to oxidative stress (e.g., CALM, ERP29, GSTP1, PDIA1) or induction of apoptosis (e.g., LAMA, FLNA, TPR, GDIS). Interestingly, modulation of these proteins was almost fully prevented by inclusion of an antioxidant, N-acetylcysteine. Validation of the proteomic data confirmed GSTP1 as an adaphostin resistance gene. Subsequent analysis of HL60 cells treated with 1,4-dihydroxybenzene or H2O2 showed similar increases in intracellular peroxides and an almost identical proteomic profiles to that of adaphostin treatment. Western blotting of a panel of cell lines identified Cu/Zn superoxide dismutase (SOD) as correlating with adaphostin resistance. The role of SOD as a second adaphostin resistance gene was confirmed by demonstrating that inhibition of SOD using diethyldithiocarbamate increased adaphostin sensitivity, whereas transfection of SOD I attenuated toxicity. Importantly, treatment with 1,4-dihydroxybenzene or H2O2 replicated adaphostin-induced Bcr/abl polypeptide degradation, suggesting that kinase inhibition is a ROS-dependent phenomenon. Conclusion: Adaphostin should be classified as a redox-active–substituted dihydroquinone.

The ability of early passage patient-derived tumor xenografts (PDXs) to recapitulate human tumor biology makes them invaluable tools for pre-clinical drug development (1, 2). Similarly, clinically-annotated human tumor cell lines of low passage have potential as scalable biosimilar reagents for in vitro study. Here, we describe an integrated platform for the purification, expansion, and quality control of tumor cells and cancer-associated fibroblasts (CAFs) from patient biopsies, surgical resections, and mouse PDX models. Following dissociation of solid tumors, cells are cultured on Matrigel® in fibroblast conditioned media containing Y-27632 (a Rho-associated kinase inhibitor). Cells are then analyzed by FACS using a panel of species- or cell type-specific antibodies. This panel (anti-mCD9, hCD9, hCD24, hCD90, MHC, HLA, and hEPCAM) permits quantitation of the percentage of mouse/human cells, and the percentages of tumor cells and CAFs in each patient specimen. From these data, a cell sorting strategy can be selected (e.g. mCD9 and hCD90 with hEpCAM or CD24) that will permit purification of both tumor cells and CAFs, along with removal of any mouse fibroblast contamination (for PDX tumors). Samples are sorted multiple times until target populations reach 99.99% purity (typically 2-3 rounds). Purified cells are then returned to culture, expanded over several weeks, and frozen at low passage. The purity and identity of these cultures is quality-assured using 1) mouse-specific QPCR, 2) a cDNA array designed to differentiate fibroblasts from tumor cells, and 3) STR genotyping. Tumor lines and CAFs are then scaled and re-qualified at passages 10 and 20, an essential step given the propensity of low-frequency mouse and human fibroblasts to ‘overtake’ impure cultures with time. Lastly, cancer lines and CAF cultures undergo genomic characterization and are evaluated for in vivo tumorigenicity (or the lack thereof for CAFs). To date, the platform has been used to generate a growing number of purified cancer and CAF pairs from patient materials. In conclusion, our goal is to provide useful reagents to the extramural community that will accompany current efforts within the Division of Cancer Treatment and Diagnosis (DCTD, NCI) aimed at generating a large and extensively characterized repository of PDX models. Funded by NCI Contract No. HHSN261200800001E. References 1.Hidalgo M, Amant F, Biankin AV, Budinska E, Byrne AT, Caldas C, et al. Patient-derived xenograft models: an emerging platform for translational cancer research. Cancer Discov. 2014;4:998-1013. 2.Rosfjord E, Lucas J, Li G, Gerber HP. Advances in patient-derived tumor xenografts: from target identification to predicting clinical response rates in oncology. Biochem Pharmacol. 2014;91:135-43. Citation Format: Luke H. Stockwin, Michael Mullendore, Carrie Bonomi, Kelly Dougherty, James H. Doroshaw, Melinda G. Hollingshead, Dianne L. Newton. Cell sorting strategies for the isolation of cancer cells and associated fibroblasts from tumor biopsies, surgical resections, and patient-derived xenografts. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 5139. doi:10.1158/1538-7445.AM2015-5139

It is well established that cancers are heterogeneous diseases with biological, genetic, and histopathological differences among patients and within individual tumors. Consequently, it is not surprising that animal models based on established cell lines with their clonal nature and adaptations to defined media have limited preclinical value. Recent advances in the generation of patient-derived xenografts (PDX), where patient material is engrafted into immunosuppressed mice, suggests that this approach may provide a more representative surrogate for therapeutic development. In this study, a mixed cell culture derived from a SCLC PDX model was analyzed for clonal heterogeneity. Following two rounds of growth in soft agar, seven clones were derived; three were further characterized. The 1st clone, 7-7-2, similar in morphology to the parental culture was a mixture of cuboidal and spindle-shaped cells. FACs analysis confirmed two populations of cells; EpCAM+/CD90 low and EpCAM-/CD90 high population. The 2nd clone, 2-2-1, consisted of a single population of spindle-shaped EpCAM-/CD90+ cells; the 3rd clone, 7-7-1, contained a homogeneous population of cuboidal cells that were EpCAM+/CD90 low. DNA analysis confirmed all 3 clones were derived from the same patient. Furthermore, all 3 clones were tumorigenic in NSG mice. To exclude fibroblast contamination, the clones were analyzed using a quantitative RT-PCR array capable of identifying human fibroblasts and other stromal constituents. None of the 3 clones had a classical fibroblast profile. Comparison of 2-2-1 relative to the other 2 clones revealed higher expression of VIM mRNA and lower KRT5, CDH1 and EpCAM. Results were confirmed by Western blot. Additionally, ITGAV and TWIST were up-regulated. These results suggest that clone 2-2-1 had undergone epithelial-mesenchymal transition (EMT). Further confirmation was provided by microarray and proteomic analysis where GNG11, MMP2, ZEB, DDR2, TESK2, DESM and MOES were elevated consistent with EMT. Next, clones were tested for sensitivity against a panel of clinically relevant agents (trametinib, everolimus, temozolomide, ABT-888, carboplatin, MK 1775) either alone or in combination. Clone 2-2-1 proved to be more sensitive than parental cells or clones 7-7-1 or 7-7-2 to the majority of these drugs. Further studies showed the combination of temozolomide/ABT-888 was synergistic against each of the clones while the combination of MK 1775/carboplatin lacked synergism. In vivo, the parental tumor responded to the temozolomide/ABT-888 combination with tumor regressions. In summary, these data serve to highlight 1) heterogeneity within this PDX model 2) profound biological differences between clones and 3) EMT is implicated and alters drug responses. Funded by NCI Contract No. HHSN261200800001E. Citation Format: Michael Mullendore, Luke H. Stockwin, Carrie Bonomi, Kelly Dougherty, Howard Stotler, Adrienne Kimmel, Biswajit Das, Vivekananda Datta, Jason Lih, Mickie Williams, James Doroshow, Melinda Hollingshead, Dianne Lynn Newton. Clonal heterogeneity in patient-derived xenografts: The SCLC model LG0904F1496M23 contains stable clones with epithelial or mesenchymal characteristics and differential drug sensitivities. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 3236. doi:10.1158/1538-7445.AM2015-3236

Summary We report on the generation and functional characterization of a humanized immunoenzyme comprising a stable humanized single chain Fv (scFv) with grafted specificity of the anti-CD22 murine monoclonal antibody RFB4 and the human ribonuclease angiogenin (ANG). The fusion protein produced from transiently transfected mammalian Chinese hamster ovary cells could easily be purified to homogeneity, retained full ribonucleolytic activity, and efficiently killed CD22+ tumour cells with an IC50 of 56 nmol/l. In contrast, incubation of tumour cells with either ANG or scFv alone did not result in any cytotoxicity. Potent receptor-mediated killing of target cells, expected lack of extracellular toxicity, predictable low immunogenic potential, and ease of production, suggest that this novel immunoenzyme has potential for the immunotherapy of CD22+ malignancies.

Selective cytotoxicity is an important goal of specific drug targeting. Toward this end, toxins isolated primarily from higher plants and bacteria have been coupled to monoclonal antibodies (MAbs) and evaluated for their clinical efficacy in cancer, AIDS, and immunological diseases (1,2). Immune responses against murine monoclonal antibodies MAbs (3,4) and antitoxin antibodies have been detected in both animals and humans treated with immunotoxins (ITs) (5-7) and present a major obstacle to the successful application of this technology. Although development of humanized antibodies have alleviated some of these effects (8, and references therein), the toxins themselves remain a problem. Consequently, the identification of human proteins to be used as components of immunoconjugates is highly desirable.

Immunotoxins based on human and humanized ribonuclease may have potential for cancer therapy while exhibiting less toxic side effects and stimulating less of an immune response in humans than immunotoxins based on plant and bacterial toxins (1). Both recombinant RNase fusion proteins (2-4 see also Chapter 6 , this volume) and chemical RNase conjugates have been made and characterized. The cytotoxic potential of targeted ribonuclease was first demonstrated with bovine RNase conjugated to transferrin or an antibody directed against the human transferrin receptor (5). Antibody RNase conjugates have also been shown to have potent anti-tumor activity against human glioma cells in athymic mice (6) and to enhance the activity of vincristine in mdr1 multidrug-resistant colon cancer cells in vitro and in vivo (7). Recently, RNase chemically conjugated to an antibody against CD22 was found to specifically kill Daudi lymphoma cells in cell culture at picomolar concentrations (IC(50), 10-50 pM) and to exhibit potent antitumor activity in SCID mice with disseminated Daudi lymphoma (unpublished data). Methods for linking RNase to specific cell binding ligands are described.

Here we report that B16F10 murine melanoma cells mimic endothelial cell behavior and the angiogenic process in vitro and in vivo. Cord formation in vitro by tumor cells is stimulated by hypoxia and vascular endothelial growth factor (VEGF) and inhibited by antibodies against VEGF and the VEGF KDR receptor (VEGF receptor 2). We define regulation of tumor cell-derived vascular space formation by these vasoactive compounds as "vasocrine" stimulation. ICRF 159 (Razoxane; NSC 129943) prevents tumor cell but not endothelial cell cord formation in vitro, and the antiangiogenic drug TNP-470 (NSC 642492) inhibits endothelial but not tumor cell cord formation in vitro. Both drugs inhibit formation of blood-filled vascular spaces in vivo. These results bear on the anticipated action of ICRF 159 in human clinical trials and novel strategies for targeting tumor blood supplies.

Targeting CD22 on human B-cells with a monoclonal antibody conjugated to a cytotoxic RNAse causes potent and specific killing of the lymphoma cells in vitro. This translates to anti-tumor effects in human lymphoma models in SCID mice. RNA damage caused by RNAses could be an important alternative to standard DNA-damaging chemotherapeutics. A second generation construct with an improved recombinant cytotoxic RNAse is described. Targeted RNAses may overcome problems of toxicity and immunogenicity associated with plant or bacterial toxin-containing immunoconjugates.

This article describes the construction, expression, and purification of RNase single-chain antibody fusion proteins. To construct a fusion protein, the gene for each moiety, the RNase and the binding ligand, is modified separately to contain complementary DNA encoding a 13 amino acid spacer that separates the RNase from the binding moiety. Appropriate restriction enzyme sites for cloning into the vector are also added. The modified DNA is combined and fused using the PCR technique of splicing by overlap extension (1). The resulting DNA construct is expressed in inclusion bodies in BL21(DE3) bacteria that are specifically engineered for the expression of toxic proteins (2). After isolation and purification of the inclusion bodies, the fusion protein is solubilized, denatured, and renatured. The renatured RNase fusion protein mixture is purified to homogeneity by two chromatography steps. The first column, a CM-Sephadex C-50 or a heparin Sepharose column, eliminates the majority of contaminating proteins while the second column, an affinity column (Ni2+-NTA agarose), results in the final purification of the RNase fusion protein.

Generally, the new strategy in cancer drug discovery and development is to define and validate the most promising cancer related molecular targets to which new drugs can be designed. For small molecule drugs, the focus is on interfering with molecular targets that contribute to deregulated cell growth and signaling pathways, predominantly within the cancer cell. Antigens aberrantly expressed on the cancer cell surface also afford possible valid molecular targets. Thus antibodies against tumor related antigens are a class of drugs that fit the new paradigm for cancer drug development and treatment.

Targeting CD22 on human B-cells with a monoclonal antibody conjugated to a cytotoxic RNAse causes potent and specific killing of the lymphoma cells in vitro. This translates to anti-tumor effects in human lymphoma models in SCID mice. RNA damage caused by RNAses could be an important alternative to standard DNA damaging chemotherapeutics. Moreover, targeted RNAses may overcome problems of toxicity and immunogenicity associated with plant or bacterial toxin containing immunotoxins.

Toxins isolated from plants and bacteria have been coupled to monoclonal antibodies to produce potent selective cytotoxic agents (immunotoxins) for the treatment of cancer, AIDS, and immunological diseases (Vitetta 1993) (Brinkmann 1994). However a major obstacle to the successful use of these reagents has been toxic side effects and immunogenicity (Sawler 1985) (Schroff 1985) (Harkonen 1987) (Rybak 1991) (Soler-Rodriguez 1993) (Vitetta 1993). While much effort has been devoted to the humanization of the antibody portion of the reagent (See Chapters 6, 7, 14, 39, 40 and references therein), the toxin moiety still remains a problem. Human enzymes are now being used in lieu of plant and bacterial enzymes (reviewed in (Rybak 1999)). This chapter describes the preparation of RNasebased antibody fusion proteins. Non-cytotoxic human RNases such as eosinophil-derived RNase (EDN), pancreatic RNase, or angiogenin are equally or more cytotoxic than plant and bacterial toxins when injected into Xenopus oocytes (Saxena 1991). RNases linked chemically or fused genetically to cell surface binding ligands have potent anti-tumor effects both in vitro and in vivo. Furthermore therapeutic RNases do not cause non-specific toxic side-effects associated with the classical plant and bacterial toxins in mice (reviewed in (Rybak 1999)) or in humans (Mikulski 1993).

Antibodies fused to human enzymes offer an alternative to specifically targeting tumors with antibodies linked to plant or bacterial toxins. Since large amounts of these reagents can be administered without eliciting non-specific toxicities, efficient methods of production are needed. The goal of this work was to express a complex immunoenzyme fusion protein (immunotoxin) in the mammary gland of transgenic mice. A chimeric mouse/human antibody directed against the human transferrin receptor (E6) was fused at its CH2 domain to the gene for a human angiogenic ribonuclease, angiogenin (Ang). It was expressed in the mammary gland of mice and secreted into mouse milk. Expression levels in milk were approximately 0.8 g/l. The chimeric protein retained antibody binding activity and protein synthesis inhibitory activity equivalent to that of free Ang. It was specifically cytotoxic to human tumor cells in vitro.

Angiogenesis defines the many steps involved in the growth and migration of endothelial cell-derived blood vessels. This process is necessary for the growth and metastasis of tumors, and considerable effort is being expended to find inhibitors of tumor angiogenesis. This usually involves screening of potential anti-angiogenic compounds on endothelial cells. To this end, two candidate anti-angiogenic RNA-damaging agents, onconase and (-4)rhEDN, were screened for their effects on endothelial cell proliferation using three distinct types of endothelial cells in culture: HPV-16 E6/E7-immortalized human umbilical vein endothelial cells (HUVECs), a Kras-transformed HPV-16 E6/E7 HUVEC (Rhim et al., Carcinogenesis 4, 673-681, 1998), and primary HUVECs. Onconase similarly inhibited proliferation in all three cell lines (IC(50) = 0.3-1.0 microM) while (-4)rhEDN was more effective on immortalized HUVEC cell lines (IC(50) = 0.02-0.06 microM) than on primary HUVECs (IC(50)0.1 microM). Differential sensitivity to these agents implies that more than one endothelial cell type must be used in proliferation assays to screen for novel anti-angiogenic compounds.

Efforts are underway to improve the treatment of recalcitrant, rare and neglected cancers through the discovery of potential therapeutic targets, the identification of possible therapeutic combinations, and the identification of genomic vulnerabilities, using state-of-the-art drug discovery, molecular characterization, and mechanism-of-action techniques. Our current focus is on sarcoma and small cell lung cancer. Sarcoma comprises approximately 1% of cancers inclusive of 50 subtypes and occurs in patients of all ages with a frequency spread evenly over the human age range. Small cell lung cancer (SCLC) constitutes approximately 15% of lung cancers, is extremely aggressive, has a high mortality rate and frequently recurs after conventional cytotoxic chemotherapy. We have established comprehensive panels of human sarcoma and small cell lung cancer cell lines (approximately 80 per tumor type). Cell lines are being screened for response to all of the FDA approved anticancer agents and to a library of investigational agents with a goal of identifying existing agents that may be suitable for novel clinical trials, plus identifying potential points of vulnerability for drug discovery. Analysis of selected somatic mutations showed that ATM is frequently mutated in both soft tissue and bone sarcoma. In addition, gene expression profiles were measured in sarcoma cell lines using Affymetrix Exon 1.0 ST arrays in an effort to facilitate identification of splice variants and fusion genes. The gene-level data from the Exon ST1 array compares well with published U133plus 2 expression profiles. Principal components analysis of genes in 48 sarcoma lines indicates that expression varies primarily by disease subtype. For instance, Ewing's Sarcoma overexpressed genes including PRKCB, NPY5R and NPY1R, ITM2A plus adrenergic receptors B1, B3 and A1D, while ACTC1, IGF2 and CHRNA1 were dysregulated in rhabdomyosarcoma. Predicted associations between gene and miRNA expression included CSF1 with miR-128, thioredoxin reductase with miR-324-5p, MDM4 with miR-152, and PODXL with miR-199a-5p, suggesting potential regulatory relationships that might be exploited as cancer targets. In these lines, exome expression was found to be altered more than 10% in 232 genes, including known splice variant genes such as KLK11, UBE2C and sarcoma related fusion genes EWSR1 and ETV. Calculated fusion scores based on EWSR1-FLI1 expression profiles successfully segregated Ewing's sarcoma from all other sarcoma cell lines. Association analysis between gene and microRNA expression and sensitivity to specific anticancer agents is ongoing. Our mission is to improve the treatment of these neglected cancers by providing a comprehensive public database of molecular signatures and sensitivities that can be leveraged by the scientific community at large. Funded by NCI Contract no. HHSN261200800001E. Citation Format: Anne Monks, Annamaria Rapisarda, Kazimierz O. Wrzeszczynski, E. Michael August, Eric C. Polley, Sudhir B. Kondapaka, Gurmeet Kaur, Dianne Newton, Beverley A. Teicher. Target and drug discovery for recalcitrant, rare and neglected cancers. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 4361. doi:10.1158/1538-7445.AM2013-4361

Onconase, a ribonuclease isolated from Rana pipiens oocytes and early embryos, is a member of the RNase A superfamily. Onconase has anti-neoplastic properties both in vitro and in vivo, and is undergoing clinical evaluation. In the present study, Onconase was combined with or conjugated to MRK16, an anti-P-glycoprotein (Pgp) monoclonal antibody. The interaction of these combinations with vincristine (VCR) against parental and multidrug resistant (MDR), Pgp expressing, human colon carcinoma cells caused increased VCR cytotoxicity in vitro and enhanced survival of athymic nude mice given transplants of drug resistant HT-29(mdr1) cells in vivo. The results suggest that combination treatment with Onconase and other agents that modulate the chemosensitivity of Pgp-expressing human tumor cells has the potential to overcome MDR.

The gene for human angiogenin (Ang), a member of the ribonuclease superfamily, was fused to a gene encoding a single-chain antibody (sFv) against the human transferrin receptor. Three Ang single-chain immunofusion proteins (AngsFvs) were constructed with variations in the type of linker connecting the VL and VH chain [EGKSSGSGSESKEF, L1 or (GGGGS)3, L2] as well as with or without a spacer (FB) connecting the Ang and sFv (AngFBsFvL1 or L2; AngsFv(L2)]. Although the nature of the linker did not affect the enzymatic activity of the FB-containing fusion proteins, the fusion protein containing the L2 linker was 2.3-fold more effective than the L1 linker in competing with the labeled monoclonal IgG1 antibody for binding to the transferrin receptor. The fusion protein containing the L2 linker without the FB spacer exhibited a 13-fold decrease in binding to the transferrin receptor as well as a decrease in its capacity to degrade tRNA and to inhibit translation in the rabbit reticulocyte lysate compared to its counterpart containing the FB spacer. Binding of placental ribonuclease inhibitor (PRI) to Ang also was affected by the nature of the linker and by the presence or absence of a spacer. PRI bound to Ang and AngFBsFv(L2) and inhibited their ribonuclease activity. A 3-fold greater concentration of PRI, however, did not affect the activity of AngFBsFv(L1) or AngsFv(L2), suggesting that the conformation of these fusion proteins was altered. Binding of monoclonal and polyclonal anti-Ang antibodies to AngsFvs was also used to investigate conformational alterations of the fusion proteins. AngFBsFv(L2) was the least altered while AngFBsFv(L1) exhibited the greatest change in structure. Yet maximal concentrations of all AngsFvs elicited angiogenesis in the chick chorioallantoic membrane assay, demonstrating that Ang in all three fusion proteins remained functionally active. Consistent with all the activities, the fusion protein containing the FB spacer and L2 linker was the most cytotoxic to three different human tumor cell lines. The fusion protein lacking the FB spacer exhibited the least cytotoxicity. These data demonstrate that the linker connecting the VH-VL chains can affect the binding and cellular cytotoxicity of Ang immunofusions and that placement of a spacer between the antibody binding domains and Ang is necessary for optimal activity. Thus, a new class of targeted therapeutic agents containing Ang as the toxic moiety can be designed that potentially will be less immunogenic and less toxic than immunotoxins available currently.

Several indole-3-carbinol analogs are under preclinical evaluation at the National Cancer Institute. One analog, NSC743380 (1-[(3-chlorophenyl)-methyl]-1H-indole-3-carbinol), is selectively toxic to a subset of NCI 60 cell lines in vitro and has undergone extensive in vivo testing. In vivo, NSC743380 produced complete regressions in A498 renal xenograft models at doses as low as 45 mg/kg when administered intraperitoneally. Additional studies demonstrated that NSC743380 is orally bioavailable. To extend knowledge regarding the mechanism of action, two pairs of resistant and sensitive cell lines [A498/ACHN (renal) and NCI-H226/A549 (NSCLC), sensitive/resistant, respectively] were used. Results showed that 5-10 min following NSC743380 treatment, phosphorylation of p38 and JNK were enhanced in sensitive but not resistant lines. Two hrs of treatment resulted in an inhibition of transcription and translation (45% and 75% inhibition, respectively) and by 4 hrs NSC743380 induced caspase-dependent apoptosis with loss of c-FLIP. Pathway-specific inhibitors were then used to gain mechanistic insight. Diverse antioxidants and NSAIDs, inhibitors of JNK, RAS/RAF, PPAR, lipid 2nd messenger signaling, transcription and actin polymerization were shown to completely inhibit NSC743380 activity. Microarray analysis of three sensitive cell lines using Affymetrix U133 Plus 2 chipset identified several trends in the transcriptome notably the upregulation of diverse immediate-early genes (IEGs) including; transcriptional regulators [EGR1, FOS/FOSB, HES1, MAFF and SOX9], secreted factors [HBEGF, IL-8 and GDF-15] and several general IEGs [ARC, ERRFI1, GADD45A/B, GEM and IER2]. A second noteworthy trend involved increased expression of the IL-6 family members, IL-6, IL-11 and LIF. These data led to an exploration as to whether enhanced JAK/STAT signaling was responsible for the above effects. Short time course lysates from 3 sensitive (A498, CAKI-1 and NCI-H226) cell lines and 1 resistant (A549) cell line showed that in sensitive lines only, NSC743380 rapidly (5 minutes) enhanced phosphorylation of the JAK family member Tyk2 at Y1054/55 and STAT3 at Y705. Additionally several phosphostate changes were observed in a subset of sensitive cell lines such as JAK1 phosphorylation in A498 cells and transient increase in pSTAT1 in A498 and CAKI-1 cell lines. Furthermore, two inhibitors of JAK/STAT signaling, AG490 and resveratrol, completely inhibited NSC743380 activity and p38/JNK phosphorylation. Subsequent siRNA knockdown experiments showed that Tyk2 siRNA inhibited NSC743380 activity whereas STAT3 and JAK1 siRNA along with scrambled siRNA had no effect. These data suggest that signaling through the JAK family member Tyk2 contributes towards NSC743380 activity and the data are being evaluated to determine how best to move forward. Funded by NCI Contract No. HHSN261200800001E. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 2810. doi:1538-7445.AM2012-2810

Onconase, a protein with anti-tumor activity, causes potent inhibition of protein synthesis in the rabbit reticulocyte lysate (IC50 10(-11) M) and when microinjected into Xenopus oocytes (IC50 10(-10) M). Onconase is a member of the RNase A superfamily; however, unlike RNase A, the mechanism of protein synthesis inhibition does not involve apparent degradation of lysate or cellular ribosomal RNAs. Rather, reticulocyte and oocyte tRNA is hydrolyzed after Onconase treatment. Furthermore, re-addition of tRNA to Onconase pretreated lysates or oocytes restores the translational capacity of the system. Taken together these results suggest that Onconase causes potent protein synthesis inhibition by a mechanism involving inactivation of cellular tRNA.

Purkinje cell toxicity is one of the characteristic features of the Gordon phenomenon, a syndrome manifested by ataxia, muscular rigidity, paralysis, and tremor that may lead to death (Gordon, 1933). Two members of the RNase superfamily found in humans, EDN (eosinophil- derived neurotoxin) and ECP (eosinophil cationic protein), cause the Gordon phenomenon when injected intraventricularly into guinea pigs or rabbits. We have found that another member of the RNase superfamily, an antitumor protein called onconase, isolated from Rana pipiens oocytes and early embryos, will also cause the Gordon phenomenon when injected into the cerebrospinal fluid of guinea pigs at a dose similar to that of EDN (LD50, 3–4 micrograms). Neurologic abnormalities of onconase- treated animals were indistinguishable from those of EDN-treated animals, and histology showed dramatic Purkinje cell loss in the brains of onconase-treated animals. The neurotoxic activity of onconase correlates with ribonuclease activity. Onconase modified by iodoacetic acid to eliminate 70% and 98% of the ribonuclease activity of the native enzyme displays a similar decrease in ability to cause the Gordon phenomenon. In contrast, the homologous bovine pancreatic RNase A injected intraventricularly at a dose 5000 times greater than the LD50 dose of EDN or onconase is not toxic and does not cause the Gordon phenomenon. A comparison of the RNase activities of EDN, onconase, and bovine pancreatic RNase A using three pancreatic RNA substrates demonstrates that onconase is orders of magnitude less active enzymatically than EDN and RNase A. Thus, another member of the RNase superfamily in addition to EDN and ECP can cause the Gordon phenomenon.(ABSTRACT TRUNCATED AT 250 WORDS)

Members of the pancreatic ribonuclease (RNase) family have diverse activities toward RNA that could cause them to function during host defense and physiological cell death pathways. This activity could be harnessed by coupling RNases to cell binding ligands for the purpose of engineering them into cell-type specific cytotoxins. Therefore, the cytotoxic potential of RNase was explored by linking bovine pancreatic ribonuclease A via a disulfide bond to human transferrin or antibodies to the transferrin receptor. The RNase hybrid proteins were cytotoxic to K562 human erythroleukemia cells in vitro with an IC50 around 10(-7) M, whereas10(-4) M of native RNase was required to inhibit protein synthesis. Cytotoxicity required both components of the conjugate since excess transferrin or ribonuclease inhibitors added to the medium protected the cells from the transferrin-RNase toxicity. Importantly, the RNase conjugates were found to have potent antitumor effects in vivo. Chimeric RNase fusion proteins were also developed. F(ab')2-like antibody-enzyme fusions were prepared by linking the gene for human RNase to a chimeric antitransferrin receptor heavy chain gene. The antibody enzyme fusion gene was introduced into a transfectoma that secreted the chimeric light chain of the same antibody, and cell lines were cloned that synthesized and secreted the antibody-enzyme fusion protein of the expected size at a concentration of 1-5 ng/mL. Culture supernatants from clones secreting the fusion protein caused inhibition of growth and protein synthesis toward K562 cells that express the human transferrin receptor but not toward a nonhuman derived cell line. Since human ribonucleases coupled to antibodies also exhibited receptor mediated toxicities, a new approach to selective cell killing is provided. This may allow the development of new therapeutics for cancer treatment that exhibit less systemic toxicity and, importantly, less immunogenicity than the currently employed ligand-toxin conjugates.

Treatment of all-trans-retinal with a series of 1,3-diketones using Knoevenagel conditions gave the expected condensation products. These retinylidene 1,3-diketones were characterized and their biological activities in the hamster tracheal organ culture test measured. It was found that the cyclohexane- 1,3-dione derivatives are highly active in this in vitro assay, while other 1,3-diketones are less active. Retinylidenedimedone has been chosen for further evaluation.

Structure-function relationships of 19 retinoids were studied in three in vitro systems in defined, serum-free medium. These systems measure effects on differentiation and growth, as well as toxic effects. The ring, side chain, and polar terminal group of natural retinoids all may be modified synthetically to yield highly active compounds. Several retinoid ethers were found to be significantly less toxic than the corresponding carboxylic acids or alcohols.

Highly purified tryptic fragments of calmodulin were tested for their ability to stimulate adenylate cyclase activity of Bordetella pertussis spheroplast membranes and were compared to their activities on brain Ca2+/calmodulin-dependent cyclic nucleotide phosphodiesterase. The C-terminal fragment, consisting of residues 78-148, was a full agonist for the cyclase with 0.1-0.15 the potency of calmodulin but did not stimulate phosphodiesterase. Fragments 1-77, 1-90, and 107-148 stimulated adenylate cyclase (and not phosphodiesterase) at low potency; this was not due to calmodulin contamination, but contamination by fragment 78-148 could not be excluded with certainty. An adduct of norchlorpromazine isothiocyanate and calmodulin showed full agonist activity for adenylate cyclase at 0.01-0.02 the potency of calmodulin. Stimulation of adenylate cyclase by a number of the fragments occurred in the absence of Ca2+, but stimulator potency was enhanced 20-60-fold in its presence. The similarity of Ca2+ requirements of fragment 78-148 and calmodulin suggests that occupancy of the two C-terminal Ca2+ binding sites of calmodulin accounts for most of the Ca2+ enhancement of calmodulin stimulation of adenylate cyclase.

Polypeptides characterized by their ability to confer a transformed phenotype on an untransformed indicator cell have been isolated directly from tumor cells growing both in culture and in the animal, by using an acid/ethanol extraction procedure. Assay of these polypeptides is based on their ability to induce normal rat kidney fibroblasts to form colonies in soft agar. Peptides from murine sarcoma virus-transformed mouse 3T3 cells grown in culture had the highest specific activity in this assay; peptides from sarcomas produced from these cells or from chemically induced transplantable bladder carcinomas of mice were one-third as active; and peptides from a chemically induced rat tracheal carcinoma had only one-tenth the activity. Treatment with either trypsin or dithiothreitol destroyed the activity of all of these materials. The properties of these intracellular polypeptides from both virally and chemically transformed cells are similar to those described for the sarcoma growth factors (SGFs) previously isolated from the conditioned medium of sarcoma virus-transformed mouse 3T3 cells, suggesting the definition of a class of transforming growth factors common to tumor cells of different origins. The transforming peptides from the cultured sarcoma virus-infected cells were separately by gel filtration into two fractions of apparent molecular weight 7000 and 10,000. The major fraction at molecular weight 7000 represented approximately 0.1% of the original cell protein and had a specific activity 50 times that of the original acid/ethanol extract.

FIFTY years after the discovery that vitamin A controls cel differentiation in epithelial tissues the mechanism involved is still unknown. The recent introduction of two-cell and organ culture systems, involving serum-free medium, for the assay of vitamin A activity in vitro may help. One assay uses epidermal cell cultures derived from mouse skin1 and the other uses tracheal organ cultures from vitamin A-deficient hamsters2. In both, the observed cellular response depends on the addition of vitamin A, which induces a marked increase in the cellular RNA of the skin cultures and a change from the production of keratin to the production of cilia and mucus in the tracheal organ cultures. We have now investigated the Biological activity of seven analogues of natural vitamin A ester (β-retinyl acetate) and vitamin A acid (β-retinoic acid), in line with the common practice of assaying physiological and pharmacological actions of vitamins, hormones and drugs using structural analogues of a parent compound. In the analogues we used, the 5,6-cyclohexenyl ring system of natural vitamin A is modified substantially. Although data obtained in vivo show that modification of the ring portion of the molecule can reduce growth-promoting activity3–5, we found that several vitamin A analogues with alterations in the ring, including a shift of the 5,6-double bond of the cyclohexene ring to the 4,5-position and replacement of the cyclohexene ring with substituted aromatic and cyclopentene rings, have substantial activity in the skin and tracheobronchial assays. Because of the sensitivity of the two assays it is now possible to assay vitamin A activity in the 10−9–10−10M range (30–300 pg ml−1). Moreover, there is excellent correlation between the results of the two assays in terms of the evaluation of Biol.ogical activity of new analogues.

Occupancy of one of the two phenothiazine-binding sites on calmodulin does not significantly decrease the affinity of calmodulin for its target proteins; however, it does affect the ability of calmodulin to activate some enzymes. Previously we demonstrated that a covalent adduct of calmodulin with one molecule of phenothiazine (CAPP 1 -calmodulin) is an antagonist for the calmodulin-dependent enzymes, cAMP phosphodiesterase and myosin kinase, and a partial agonist for calcineurin. We now show that CAPP 1 -calmodulin is a full agonist for glycogen synthase kinase and phosphorylase kinase. Unlike phenothiazines, CAPP 1 -calmodulin is specific for calmodulin-regulated proteins; it has no effect on protein kinase C. With the exception of phosphorylase kinase, occupancy of two phenothiazine-binding sites completely eliminates the ability of calmodulin to activate these proteins. Thus, the study of the interaction of CAPP 1 -calmodulin with calmodulin target proteins demonstrates that calmodulin interacts differently with different proteins. This is confirmed by studies of the effect of calmodulin fragments, 1–77 and 78–148, on calmodulin-regulated enzymes.

A rapid large scale purification procedure based on three conventional purification steps, ammonium sulfate fractionation, DEAE cellulose and hydroxylapaptite chromotography yields gram amounts of calmodulin. The protein is more than 98% pure by SDS gel electrophoresis and amino acid composition. It is free of contaminating EGTA or EDTA and the omission of heat treatment or denaturing solvents during the preparat-ion avoids possible dena-turation of the protein and minimizes partial proteolysis.

Limited proteolysis of calmodulin with trypsin in the presence of ethylene glycol bis(beta-aminoethyl ether)-N, N,N',N'-tetracetic acid (EGTA) or Ca2+ was performed according to a modification of the method of Drabikowski et al. (Drabikowski, W., Kuznicki, J., and Grabarek, Z. (1977) Biochim. Biophys. Acta 485, 124-133). The resulting peptides were purified by reverse-phase high performance liquid chromatography. Tryptic digests in EGTA yielded peptides 1-106, 1-90, and 107-148 with yields of 9, 47, and 61%, respectively. The digests performed with Ca2+ yielded peptides 1-77 and 78-148 in 35 and 45% yield. Analysis by high performance liquid chromatography indicated that the purified fragments contained less than 0.1% contamination by calmodulin, thus allowing a definitive study of the ability of these fragments to activate, or interact with, calmodulin-regulated enzymes and anti-calmodulin drugs. Each of the fragments, except 107-148, bound to a phenothiazine affinity column in a Ca2+-dependent manner. Thus, calmodulin contains two interaction sites for phenothiazines: one on the NH2-terminal half (fragment 1-77) and one on the COOH-terminal half (fragment 78-148). None of the fragments activates the protein phosphatase, calcineurin, or prevents its stimulation by calmodulin, nor does any of the fragments stimulate Ca2+-dependent cAMP phosphodiesterase. A single cleavage in the middle of the calmodulin molecule results in the rapid dissociation of the two resultant fragments and a loss of ability to activate cAMP phosphodiesterase. One fragment, 78-148, interacts with phosphodiesterase and prevents its activation by calmodulin (Ki: 1.5 +/- 0.4 X 10(-6) M). The same fragment, 78-148, can fully activate phosphorylase kinase but with a lower affinity than calmodulin (Kuznicki, J., Grabarek, Z., Brzeska, H., Drabikowski, W., and Cohen, P. (1981) FEBS Lett. 130, 141-145). Thus, peptide 78-148 behaves as a calmodulin agonist or antagonist or as neither, depending on the enzyme under study.

Incubation of [3H]retinoic acid in the presence of hamster liver 10000g supernatant produces several metabolites that are more polar than the parent compound. Two of these metabolites are identical with synthetic all-trans-4-hydroxyretinoic acid and all-trans-4-oxoretinoic acid both in ultraviolet absorption and mass spectral characteristics and in migration rates on two different reverse-phase high-pressure liquid chromatographic systems. The metabolites produced in a cell-free liver incubation reaction also migrate on a high-pressure liquid chromatography column together with metabolites isolated from a tracheal organ culture system. Both the metabolites and the synthetic standards show less biological activity than the parent all-trans-retinoic acid in a tracheal organ culture assay.

A covalent adduct of norchlorpromazine (CAPP) and calmodulin is a very potent antagonist of calmodulin activation of several enzymes. The phenothiazine-calmodulin complex (CAPP-calmodulin) acts as a pure antagonist with phosphodiesterase and myosin kinase or a partial agonist with the phosphoprotein phosphatase, calcineurin. Because of its potency and the selectivity inherent to its calmodulin moiety, CAPP-calmodulin should be a uniquely useful probe of calmodulin actions.

Calmodulin forms a covalent, one to one, complex with 3H-labeled norchlorpromazine isothiocyanate. Complex formation was monitored by high-performance liquid chromatography using a CN reverse-phase column which resolves calmodulin, the calmodulin-norchlorpromazine adduct, and norchlorpromazine isothiocyanate. Formation of the adduct requires Ca2+ and is not observed with norchlorpromazine. The one to one calmodulin-norchlorpromazine complex does not activate phosphodiesterase but can interact with the enzyme and competitively inhibit its stimulation by calmodulin. High concentrations of trifluoperazine inhibit whereas low concentrations stimulate complex formation. This apparent potentiation of the interaction of calmodulin with norchlorpromazine by another phenothiazine suggests that calmodulin contains at least two phenothiazine binding sites and that the binding of phenothiazine to calmodulin is cooperative.

Methods were developed for the isolation and peptide mapping of calmodulin and its chemically modified derivatives by reversed-phase high-performance liquid chromatography (HPLC). Calmodulin and its guanidinated, iodinated, and performic acid-oxidized derivatives can be isolatedon alkylphenyl columns by using gradients of acetonitrile in 10 mM potassium phosphate, pH 6.0, 2mM EGTA. Peptide mapping by HPLC, following complete digestion of the proteins with clostripain, allows identification of the modified amino acids residues. Clostripain peptides are eluted in the order 87–90, 75–86, 91–106, 107–126, 127–148, 107–148, 1–37, and 38–74. Performic acid oxidation of methionines decreases the retention times of the modified peptides, whereas iodination of tyrosines or guanidination of lysines increases retention times of modified peptides. These HPLC methods are applicable to the identification of specific modifications of calmodulin, allowing the assessment of the role of individual amino acid residues in determining the unique physical, chemical, and spectroscopic properties of this ubiquitous intracellular calcium-binding protein.