Transfection with Anti-P65 Intrabody Suppresses Invasion And

Cancer Therapy: Preclinical

Transfection with Anti-p65 Intrabody Suppresses Invasion and Angiogenesis in Glioma Cells by Blocking Nuclear Factor-KB Transcriptional Activity Liang Li,1Christopher S. Gondi,1Dzung H. Dinh,2 William C. Olivero,2 Meena Gujrati,3 andJastiS.Rao1, 2

Abstract Purpose: The strategy of intracellular antibodies to neutralize the function of target proteins has been widely developed for cancer research. This study used an intrabody against p65 subunit to prevent nuclear factor nB(NF-nB) transcriptional activity in glioma cells and to inhibit the expression of its target genes involved in the invasion and angiogenesis of human gliomas. Experimental Design: A single-chain fragment of antibody variable region (scFv) against p65 was prepared using phage display technique.We then prepared an anti-p65 intrabody construct (pFv/nu) by cloning the scFv-encoding sequence into the mammalian nuclear-targeting vector, pCMV/myc/nuc. Results: p65 expression in human glioma cells (U251and U87) transfected with pFv/nu was significantly decreased.We showed that NF-nB nuclear translocation and its DNA binding activity were blocked via intrabody transfection in electrophoretic mobility shift assays and the inhibition of NF-nB activity in nucleus resulted in the decreasing expression and bioactivity of matrix metalloproteinase-9, urokinase-type plasminogen activator receptor, urokinase-type plasminogen activator, and vascular endothelial growth factor. The intrabody transfected glioma cells showed a markedly lower level of invasion in Matrigel invasion assay.The capillary-like structure formation of endothelial cells was also repressed by coculture with the intrabody transfected glioma cells or exposure to their conditional medium. Intrabody transfection neither induced apoptosis nor altered cell proliferation in U251and U87 cells as compared with the control vector pCMV/nu. After the injection of pFv/nu-transfected glioma cells, preestablished tumors were almost completely regressed when compared with mock, pCMV/nu, and pGFP/nu. Conclusion: Blocking NF-nB activity via the nuclear intrabody expression might be a potential approach for cancer therapy.

Malignant gliomas are the most common primary tumors Nuclear factor nB (NF-nB) is a sequence-specific transcription arising in the human brain, accounting for >60% of all primary factor that belongs to the reticuloendotheliosis (Rel) family brain tumors (1). Glioblastoma multiforme (WHOgrade 4 proteins (3, 4). Constitutive activation of NF-nB plays an glioma) is the most frequent and refractory malignant brain important role in the tumorigenesis of diffuse gliomas and in tumor to affect adults, with a median survival rate of <1 year promoting the growth of high-grade gliomas, as observed by and a 5-year survival rate of V5%, despite surgery, chemother- in vitro and in vivo studies (5, 6). The autocrine activity of apy, and radiotherapy (2). Therefore, gene therapy against the tumor necrosis factor-a seems likely to sustain NF-nB activa- characteristic invasive behavior of glioblastoma might offer tion in human gliomas. Once activated by tumor necrosis novel treatment strategies when combined with conventional factor-a, NF-nB initiates a cascade of signaling events that therapeutic agents. trigger phosphorylation, ubiquitination, and subsequent pro- teasomal degradation of InB. Consequently, NF-nB is released to translocate into the nucleus and activate the transcription

1 of a variety of genes, including a gene encoding one of its Authors’Affiliations: Program of Cancer Biology, Department of Cancer Biology cytosolic inhibitors, InBa, as well as other genes that encode and Pharmacology, 2Department of Neurosurgery, and 3Department of Pathology, University of Illinois College of Medicine at Peoria, Peoria, Illinois cytokines, cytokine receptors, adhesion molecules, and proteins Received 7/13/06; revised 11/21/06; accepted 1/23/07. involved in tumorigenesis and progress. NF-nB subunits, Grant support: National Cancer Institute grants CA75557, CA92393, CA95058, including RelA (p65), RelB (p50), and RelC, contain two and CA116708; National Institute of Neurological Disorders and Stroke grant essential domains involved in NF-nB activity. One is a NS47699; Caterpillar, Inc.; and OSF Saint Francis, Inc., Peoria, IL (J.S. Rao). The costs of publication of this article were defrayed in part by the payment of page conserved region in their NH2 terminus, known as the Rel charges. This article must therefore be hereby marked advertisement in accordance homology domain, which is responsible for dimerization, with 18 U.S.C. Section 1734 solely to indicate this fact. nuclear localization, and DNA binding; the other serves as Requests for reprints: Jasti S. Rao, Department of Cancer Biology and transactivation domain within their COOH terminus, of which Pharmacology, University of Illinois College of Medicine at Peoria, One Illini Drive, n @ the activated nuclear NF- B is phosphorylated after its Peoria, IL 61605. Phone: 309-671-3445; E-mail: jsrao uic.edu. n F 2007 American Association for Cancer Research. translocation from cytoplasm and modulates NF- B trans- doi:10.1158/1078-0432.CCR-06-1711 criptional activity. Growing evidence has shown that the

Clin Cancer Res 2007;13(7) April 1, 20 07 2178 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2007 American Association for Cancer Research. Intrabody against NF-kB Suppresses Glioma Invasion transcriptional activity of NF-nB can be regulated both by phages, rescued by the helper phage, and finally the antigen-bound cytosolic sequestration of InB and by inducible phosphoryla- phages were prepared for the further selection after enrichment, E. coli tion of p65 (7–9). The molecular mechanisms and biological amplification, and purification. After four rounds, TG-1 was consequences of p65 phosphorylation are currently another infected with the final phage preparations and individual ampicillin- resistant colonies (phage clones) were selected for further analysis. crucially important determinant for NF-nB mechanism–based ELISA. To measure antigen-binding specificity of scFvs, polyclonal therapies against cancer. and monoclonal phage ELISA were used to screen phage populations Due to the presence of a strong transcriptional activation produced during each round of selection and phages from single domain, RelA (p65) is responsible for most of the transcrip- colonies, respectively. The antigen, p65 bioactive peptide coated in 96- tional activity of NF-nB (10). However, intrabody strategies well plates was incubated with either polyethylene glycol–precipitated based on p65 phosphorylation during NF-nB transcriptional phage from the end of each round of selection or single-phage activation have not yet been done. Therefore, we prepared a suspension for 3 h, with bovine serum albumin or PBS as negative single-chain antibody against human p65 (RelA subunit of controls. The amount of bound phages was determined using NF-nB), which recognizes a COOH-terminal epitope in the horseradish peroxidase/anti-M13 monoclonal conjugate in polyclonal transactivation region using human single-fold single-chain ELISA, and the incorporation of a small peptide epitope (myc-tag) at variable fragment (scFv) libraries (Tomlinson I + J). We then the COOH-terminal end of scFv allowed detection with the primary antibody [anti–c-myc monoclonal antibody (mAb)] to determine its constructed a pFv/nu vector, designed to prevent phosphory- specific affinity to p65 in monoclonal ELISA. After visualization with lation of transactivated p65, which could express anti-p65 the 3,3¶,5,5¶-tetramethylbenzidine substrate (GE Healthcare, Waukesha, intrabodies in the nucleus of human glioma cells, resulting in WI), the colorimetric evaluation of binding activity was measured in inhibitory expression of NF-nB–dependent genes, such as an ELISA reader (Benchmark, Bio-Rad, Hercules, CA) at 450 nm, with matrix metalloproteinase (MMP-9), urokinase-type plasmino- a reference wavelength of 650 nm. gen activator (uPA), uPA receptor (uPAR), and vascular To determine the affinity of scFv to p65 bioactive peptide, endothelial growth factor (VEGF). competitive ELISA was done. In brief, the supernatants of total lysates Our findings indicate that the anti-p65 nuclear-targeted from HB2151 bacteria culture were extracted to isolate soluble scFv and intrabodies significantly decreased the expression of phospho- measure the concentration using a BCA assay. The different scFvs were incubated with p65 bioactive peptide (competitor) at serial concen- p65, p65, p50, and InBa proteins in U251 and U87 glioma À À trations ranging from 10 11 to 10 5 for 2 h. Then, ELISA was done as cells, whereas the nuclear translocation and sequence-specific described above. PBS and scFvs without preincubation with competitor DNA binding activities of NF-nB were blocked. In addition, the n were also used as negative and positive controls, respectively. The blockade of NF- B transcriptional activity in nucleus led to affinity constant was defined as the reciprocal of the competitor the down-regulation of MMP-9, uPAR, uPA, and VEGF at both concentration required for 50% inhibition of maximal scFv binding mRNA and protein levels. After intrabody transfection, the (determined by the positive control, non-preincubated antibody) and ability of in vitro migration and angiogenesis of glioma cells was expressed as per molar concentration. was inhibited. However, the NF-nB functions interrupted by PCR screening and sequence alignments of p65-specific phage intrabody therapy did not alter cell proliferation or induce clones. To check individual clones for the presence of full-length VH apoptosis in U251 or U87 cells. In addition, the growth of and VL inserts, the selected clones were screened using PCR with the j j the intracranial tumor established by the injection of pFv/nu- following conditions: 35 cycles of 94 C for 1 min, 55 C for 1 min j j transfected U251 and U87 cells was almost completely (annealing), and 72 C for 2 min (extension) with a terminal 72 C, 10 min final extension cycle. The positive selected phage clones were inhibited. This finding suggests that the blockade of transcrip- ¶ n sequenced using the primers LMB3 (5 -CAGGAAACAGCTATGAC) and tional activity of NF- B by intrabody expression in the nuclei pHEN Seq (5¶-CTATGCGGCCCCATTCA) anti-p65 scFv DNA fragment. of human glioblastoma cells could contribute to the inhibition Sequence alignments to National Center for Biotechnology Information of invasion, angiogenesis, and tumor growth through down- database were carried out using Immunoglobulin BLAST. regulation of NF-nB target proteins, uPA, uPAR, MMP-9, and Construction of the nuclear-targeted anti-p65 intrabody. ScFv coding VEGF, and serve as a promising therapeutic strategy for the sequence was isolated from pIT2 by NcoI/NotI restriction and treatment of human glioblastomas. subcloned into the vector pCMV/myc/nuc (pCMV/nu) digested with the same enzymes. The anti-p65 intrabody construct (named pFv/nu) was cloned in-frame with the nuclear location signal and c-myc epitope Materials and Methods at the COOH terminus and expressed as a fusion protein, subsequently confirmed by DNA sequencing. Selection of scFv from phage display library. ScFv selection was Cell culture and transfection. Human glioma cell lines, U251 and carried out using the human single-fold scFv libraries (Tomlinson I + J), U87, were grown in DMEM (Invitrogen, Carlsbad, CA) supplemented which are being distributed by Medical Research Council HGMP with 10% fetal bovine serum in a humidified incubator with 5% CO2 at Resource Centre. In the library, >100 million different scFv fragments 37jC. For transfection with pFv/nu, pCMV/nu, and pGFP/nu, Lipofect- are cloned into an ampicillin-resistant phagemid vector pIT2 under amine reagent (Invitrogen) was used according to the manufacturer’s pelB-leader control and transformed into TG1 E. coli cells. All scFv instructions. After 48 h posttransfection, cells were harvested for fragments are composed of a single polypeptide with the VH and VL isolation of total RNA and/or whole-cell lysate for reverse transcription- domains attached to one another by a flexible glycine-serine linker, PCR (RT-PCR) or Western blotting. For zymography and in vitro fused with a myc tag and a His tag at the COOH terminus. angiogenic assays, the complete medium was replaced with serum-free To obtain anti-p65 scFv fragments from the two libraries, 10 Agof medium for 12-h incubation after transfection, and conditioned a bioactive peptide from the COOH-terminal region of the human p65 medium was collected. DMEM containing G418 (800 Ag/mL) was used subunit of transcription factor NF-nB (New England Peptide Co.; p65 for selection of stable clones of transfection, and DMEM with 400 Ag/mL bioactive peptide: amino acids 463-472, located within transactivation G418 was used for maintenance of the stable transfectants. domain 2) were coated overnight in Maxisorb tubes (Nunc Corp., Western blotting assay. Western blotting was done. Equal amounts Naperville, IL) and four rounds of selections were undertaken according of protein from cell lysates were mixed with 2Â SDS loading buffer to a standard protocol. Briefly, E. coli TG-1 was infected with the eluted and were separated by SDS-PAGE, transferred onto nitrocellulose

www.aacrjournals.org 2179 Clin Cancer Res 2007;13(7) April1,2007 Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2007 American Association for Cancer Research. Cancer Therapy: Preclinical membranes, and incubated with primary antibodies. The enhanced chemiluminescence system was used for detection of immunoreactive Table 1. Oligonucleotide primers of human uPAR, proteins with horseradish peroxidase–conjugated immunoglobulin G uPA, MMP-9, VEGF, and GAPDH used in RT-PCR (IgG) as secondary antibodies. The following antibodies were used: assay anti–c-myc mAb clone 9E10 (Sigma, St. Louis, MO), anti-p65, anti– phospho-p65, anti-p50, anti-InBa, anti– phospho-InBa,anti–InB Oligonucleotide Oligonucleotide Fragment kinase a, anti–MMP-9, anti-uPA, anti-uPAR, anti-VEGF, and anti– name primer sequence (5¶-3¶) size (bp) glyceraldehyde-3-phosphate dehydrogenase (GAPDH). All antibodies uPAR were obtained from (Biomeda, Foster City, CA) and Santa Cruz Forward primer TTACCTCGAATGCATTTCCT 455 Biotechnology (Santa Cruz, CA). Reverse primer TATGGTAAGAGGCTGTGCAA Immunofluorescence assay. Cells were fixed with 10% buffered uPA formalin for 20 min at room temperature, permeabilized with 0.1% Forward primer CACGCAAGGGGAGATGAA 341 Triton X-100 in PBS for 5 min, and incubated with primary antibodies Reverse primer ACAGCATTTTGGTGGTGACTT followed by fluorescent secondary antibody. A Leitz fluorescence MMP-9 Forward primer TGGACGATGCCTCAACGTG 455 microscope was used to acquire the images for further analysis. Reverse primer TGCCTTTGGACACGCACGAC Cell proliferation assay. Colorimetric 3-(4, 5-dimethylthiazol-2-yl)- VEGF 2, 5-diphenyl tetrasodium bromide (MTT; Chemicon, Temecula, CA) Forward primer ATGAACTTTCTGCTGTCTTGGGT 575 assay was used for the evaluation of glioma cells proliferation according Reverse primer GTGACAAGCCGAGGCGGTGA to the manufacturer’s instructions. After seeding cells into 96-well GAPDH plates, MTT was added into the plates for additional incubation for Forward primer CCACCCCATGGCAAATTCC 298

3 to 4 h in a humidified incubator with 5% CO2 at 37jC. Following Reverse primer CAGGAGGCATTGCTGATGAT addition of detergent reagent, the absorbance at 570 nm was measured on a microplate reader (Benchmark, Bio-Rad) with a reference wavelength of 630 nm. using the above-mentioned method instead of the gel-containing fibrin Flow cytometry. Samples were incubated in the presence of anti-p65 and plasminogen. antibody at 4jC overnight at various time points posttransfection. FITC- Matrigel invasion assay. The invasiveness of glioma cells was tested conjugated IgG (final concentration, 50 nmol/L) was added and after transfection as previously described. The cells (1 Â 106/mL) were incubated for 1 h at room temperature. Finally, the cells were analyzed added to the upper wells coated with Matrigel (1 mg/mL; Collaborative with a FACScan cytometer (Becton Dickinson, Mountain View, CA) Research, Inc., Boston, MA) with serum-free medium containing equipped with an argon ion laser emitting at 488 nm. 25 Ag/mL fibronectin as a chemoattractive agent in the lower wells. Electrophoretic mobility shift assay. To detect the DNA binding After a 24-h incubation period, cells that migrated through the filters activity of NF-nB, electrophoretic mobility shift assay was done into the lower chamber were counted by the number of cells on the according to the manufacturer’s instructions (Panomics, Inc., Redwood lower side of the membrane in five random fields after staining with City, CA). Briefly, nuclear proteins were prepared using a nuclear Hema-3 kit. extraction kit and their concentrations determined by protein estima- In vitro angiogenesis assay. Glioma cells (2 Â 104/mL) were tion procedure. A biotin-labeled NF-nB probe with a 5¶-AGTTGAGGG- transfected with various plasmids for 48 h and the conditioned GACTTTCCCAGGC-3¶ sequence or an unlabeled cold probe was used to medium was filtered off for future research as previously described bind nuclear proteins at 15jCto20jC for 30 min. Products were run (12). HMEC-1 endothelial cells (4 Â 104) were seeded onto eight-well on a 6% nondenaturing polyacrylamide gel in 0.5Â Tris-borate EDTA at chamber slides and the aforementioned conditioned medium was 120 V for 60 min at 4jC; the shifted bands corresponding to the added. Cells were cultured for 72 h until capillary network formation protein/DNA complexes were separated relative to the unbound was observed. The number of branch points and total number of dsDNA. The gel was then transferred onto a presoaked membrane branches per point were counted after H&E staining to quantify the (Pall Biodyne B membrane) at 300 mA for 30 min at 4jC. Following degree of angiogenesis. the immobilization of bound oligonucleotides in the membrane by a Capillary-like structure formation in coculture assay. Glioma cells UV-cross-linking oven for 5 min, the shifted bands were visualized after (2 Â 104) were seeded onto eight-well chamber slides and transfected exposure to film. with various plasmids for 72 h. Cell medium was removed and allowed RT-PCR. Total RNA was isolated from the transfected cells using the to coculture with HMEC-1 endothelial cells (4 Â 104) for 72 h. Qiagen RNeasy kit (Qiagen, Inc., Valencia, CA) according to the Capillary networks, indicators of angiogenesis, were assessed with VIII manufacturer’s protocol, and OneStep RT-PCR kit (Qiagen) was used antibody and then with FITC-conjugated secondary antibody under for detecting mRNA expression of MMP-9, uPA, and VEGF. First-strand fluorescence microscopy. cDNA was prepared using Omniscript and Sensiscript reverse tran- Apoptosis assay. Cells were cultured in six-well plates (1 Â 106 per scriptases at 50jC for 30 min. PCR amplification was then carried out well) and transfected for 72 h. The apoptotic cells were stained with under the following conditions: 95jC for 15 min, followed by 35 cycles Hoechst 33244 dye, which selectively stains nuclei of apoptotic cells. at 94jC for 1 min, at either 48jC (for amplification of uPAR) or 56jC Terminal deoxynucleotidyl transferase biotin-dUTP nick-end labeling (for MMP-9, VEGF, uPA, and GAPDH as an internal control) for 1 min, apoptosis assay was done for the DNA fragmentation fluorescence and at 72jC for 1 min. The final extension was completed at 72jC for staining under the manufacturer’s instructions (Upstate Cell Signaling 10 min. The primers used are shown in Table 1. Solution, Charlottesville, VA). Briefly, cells in eight-well chamber slides Zymography. Gelatin and fibrin zymography were done to detect were fixed with 4% paraformaldehyde-PBS for 15 min at room MMPs and uPA, respectively, as previously described (11). For gelatin temperature and then incubated with 50 AL of terminal deoxynucleo- zymography, the cell conditioned medium was collected following tidyl transferase end-labeling cocktail for 1 h, following washing in PBS protein estimation and equal amounts of protein were subjected to thrice and permeabilization with 0.05% Tween 20, 0.2% bovine serum SDS-PAGE under nonreducing conditions in 10% SDS-PAGE gel albumin in PBS for 15 min at room temperature. After the steps containing 0.1% gelatin. Gels were washed in 2.5% Triton X-100 and including immersion of cells in Tris-borate buffer to stop the reaction, incubated in Tris-CaCl2 buffer at 37jC for 18 to 22 h. The gels were washing in PBS, and blocking in blocking buffer, the slides were then stained with 0.2% Coomassie blue for 30 min and destained in subsequently incubated with avidin-FITC solution in the dark for 20% methanol and 10% acetic acid. The clear bands represented 30 min at room temperature and viewed under a fluorescent micro- gelatinase activity. Fibrin zymography was similarly done to detect uPA scope, followed by washing and mounting coverslip with gel.

Clin Cancer Res 2007;13(7) April 1, 20 07 2180 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2007 American Association for Cancer Research. Intrabody against NF-kB Suppresses Glioma Invasion

Intracranial tumor growth. Athymic male nude mice (nu/nu, Results 6-8 weeks age) were obtained from Harlan Sprague-Dawley (Indian- apolis, IN). Animal handling and experimental procedures were Preparation of anti-p65 human scFv and construction of approved by the University of Illinois College of Medicine Institutional Animal Care and Use Committee. For the intracerebral tumor model, intrabody. Two large phage display libraries, human single- 2 Â 106 U87 (or U251) cells and stable transfected cells were counted fold scFv libraries (Tomlinson I + J; Cambridge Antibody and intracerebrally inoculated into nude mice (five mice per group). Technology, Cambridge, United Kingdom), were panned to On day 45 after orthotopic tumor implantation, mouse brains were obtain anti-p65 scFv fragments. After four rounds of panning removed and fixed. Paraffin sections (3-5 Am) were prepared and by polyclonal phage ELISA and monoclonal phage ELISA, crude observed blindly to evaluate tumor size. Tumor size was measured scFv preparations bound to the target molecule were eluted. semiquantitatively in terms of maximum cross-section diameter as an The clones that showed a 3- to 5-fold higher specific affinity index of intracranial tumor size. The variation between the sections in to the target (86 of 200 randomly picked) were then used for each group was <10% for tumor regression experiments. Statistical analysis. The significance of the results was determined downstream applications. The specific affinity to antigen by the Student’s t test (two-tailed). Values are expressed as mean F SD p65 and the lower affinity binding to nonrelated bovine from at least three separate experiments and differences were considered serum albumin of 28 selected scFv clones were shown in ELISA significant at P < 0.05. (Fig. 1A).

Fig. 1. Characterization and construction of the anti-p65 scFv and intrabody. A, specificity of scFv to p65 bioactive peptide. After four rounds of panning by polyclonal ELISA, 28 scFvs from the positive clones were selected to detect their binding activity in monoclonal ELISA. Briefly,100 AL of single-phage suspension were added into 96-well plates precoated with p65 peptide or unrelated antigen, bovine serum albumin (BSA), for 3h of incubation, as well as PBS for negative control. The anti ^ c-myc mAb, recognizing myc-tag at the COOH-terminal end of scFv, was used as the primary antibody to determine its specific affinity to p65. Following visualization by 3,3¶,5,5¶-tetramethylbenzidine substrate, colorimetric evaluation was measured at 450 nm. B, structural schema of both scFv and intrabody-encoding genes. NLS, nucleotide sequence of a nuclear localization signal from SV40; arrows, restricted enzyme sites. C, the nucleotide sequence encoding cDNA of nuclear-targeted intrabody against p65 (DNA sequence, 900 bp) was obtained from DNA sequencing of the selected scFvs. D, amino acid sequence of nuclear-targeted intrabody against p65 (protein sequence, 299AA; protein MW, 31,896 g/mol).

www.aacrjournals.org 2181 Clin Cancer Res 2007;13(7) April1,2007 Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2007 American Association for Cancer Research. Cancer Therapy: Preclinical

Fig. 2. Expression and location of anti-p65 intrabody in glioma cells. A, expression of anti-p65 intrabody in U251- and U87-transfected cells was determined by Western blotting assay. During the detection of the myc-tagged fusion protein by the anti ^ c-myc (9E10) mAb, the upper bands were endogenous c-myc proteins in U251or U87 cells whereas the endogenous c-myc protein presents in both cytosolic and nuclear compartments; the lower bands showed expression of GFP or scFv fusion proteins. As described in Materials and Methods, the cytosolic and nuclear fractions from the 48-h posttransfected cells was extracted and loaded into each well of both cytosolic and nuclear lanes of the 12% SDS-PAGE gel with 30 Ag of total proteins per well. Following transfer, the nitrocellulose membrane was incubated with anti ^ c-myc mAb (clone 9E10) and goat anti-mouse horseradish peroxidase ^ conjugated IgG as the primary and secondary antibodies, respectively. Consequently, the band of anti-p65 intrabody was visualized with the enhanced chemiluminescence system. B, location of the intrabody in U251and U87 cells as detected by immunofluorescence assay. Briefly, cells with 48-h posttransfection of either pFv/nu or empty control vector, pCMV, were fixed, permeabilized, and incubated with primary antibody, anti ^ c-myc mAb, followed by fluorescent secondary antibody, goat anti-mouse FITC-conjugated IgG.With a fluorescence microscope, the images of the expressed intrabody in nucleus were acquired for the further analysis. C, colocalization of p65 and anti-p65 intrabody in U251 (top)andU87(bottom) transfected cells was determined in immunofluorescence assay, as described in (B). Left, distribution of p65 (red) detected with polyclonal anti-p65 antibody, followed by Texas red ^ conjugated antirabbit IgG. Middle, localization of anti-p65 intrabody (green) detected with c-myc (9E10) mAb, followed by FITC-conjugated antimouse IgG. Right, merged signal by p65 (red) and intrabody (green) in yellow.

À The scFv affinity constant was f107 (mol/L) 1 according to For nuclear-targeted expression of intrabodies in mammalian the result from competitive ELISA. Then the eluted clones were cells, the scFv-encoding gene was PCR-modified and cloned into digested with BstNI to analyze the restriction patterns of the the NcoI and NotI sites of an intrabody plasmid pCMV/nu, samples in agarose gel assays. The results (data not shown) which is a 5.0-kb expression vector designed to express recom- indicate that the selected positive clones were derived from a binant protein as a fusion to a SV40 nuclear location signal and single group and possessed the same pattern of restriction sites. c-myc epitope. The nuclear-targeted anti-p65 intrabody was After putative binders were isolated depending on their affinity obtained after cloned in-frame with the nuclear location signal in ELISA and were analyzed for diversity by their BstNI restric- and c-myc epitope at the COOH terminus and expressed as a tion enzyme mapping pattern, scFv-encoding inserts of phage fusion protein in the nucleus. The scheme in Fig. 1B showed clones were amplified by PCR using the primers LMB3 (CAGG- the structure of the intrabody-encoding gene, confirmed by AAACAGCTATGAC) and pHEN (CTATGCGGCCCCATTCA). restriction endonuclease digestion analysis (data not shown). The PCR products were visualized at 935 bp in agarose gel Because the clinical efficacy of the intrabodies depends on analysis, which further confirmed the expected size of the entire their stability rather than their epitope affinity, we did scFv cDNA insert (data not shown). The structure of scFv- immunoblotting and immunofluorescence assays to analyze encoding gene derived from pIT2 vector shown in the scheme their longevities when expressed in glioblastoma cells with (Fig. 1B) was confirmed using both DNA sequencing and recombinant transfection. Compared with the other experi- restriction endonuclease digestion analysis (data not shown). mental clones, the recombinant derived from clone 1 presented The results from DNA sequencing revealed that of 28 clones the longest half-life, indicating its efficient and steady-state selected, only 4 clones (named clone 1, clone 2, clone 3, and accumulation in the nucleus after transit through the cytosol. clone 4) had intact, full-length cDNA without TAG stop Therefore, we chose the recombinant, which we named pFv/nu, codons; the others had internal amber (TAG) stop codons, for further research due to its intracellular stability and activity. probably as a result of the synthetic assembly process. The four The construct pFv/nu was confirmed by double digestion with intact scFv-encoding clones contained 732-bp full-length frag- NcoI and NotI in agarose gel analysis (data not shown) and ments, coding for 244 amino acids, without any premature stop DNA sequencing (the sequence of its nucleic acid and amino codons, and their molecular weights were f27 kDa. acid was shown in Fig. 1C and D, respectively).

Clin Cancer Res 2007;13(7) April 1, 20 07 2182 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2007 American Association for Cancer Research. Intrabody against NF-kB Suppresses Glioma Invasion

Expression and location of anti-p65 intrabody in glioma nu and pFv/nu. The results (Fig. 2B) concur with those from cells. Western blotting analysis indicated that the anti-p65 the immunoblotting analysis. In addition, colocalization of intrabody (the below band) was expressed primarily in the p65 and anti-p65 intrabody is shown in Fig. 2C. The results nuclei of U251 and U87 cells and only minutely in the suggest that anti-p65 intrabody fusion protein presented exactly cytosolic compartments, whereas samples from mock-trans- in the nuclei of human glioblastoma cells. Furthermore, the fected and empty vector (pCMV/nu)–transfected cells did subcellular localization of anti-p65 intrabody was consistent not show any bands of the appropriate molecular weight with the distribution of p65 in the nucleus. besides the above band of endogenous c-myc protein in Kinetic analysis of intrabody expression. The result from each lane (Fig. 2A). Because the nuclear location signal and Western blotting assay (Fig. 3A) shows that a gradual increase the c-myc epitope added f5 kDa to scFv at its COOH in the nuclear expression of the anti-p65 intrabody in U251 terminus, the size of the total scFv-fusion protein was f32 and U87 cells was observed on days 0, 1, 2, 3, 5, and 7 after kDa; the bands of GFP proteins were also detected as positive transfection with pFv/nu. Meanwhile, the nuclear p65 expres- control. sion in the intrabody-transfected cells was reduced by the The nuclear location of anti-p65 intrabody was confirmed by maximal inhibition rate at 40% (U251) or 51.1% (U87) on day immunofluorescence assay after transfection with both pCMV/ 3 or 5 compared with mock cells (Fig. 3B and C), determined

Fig. 3. Kinetic analysis of intrabody expression. A, increased expression of anti-p65 intrabody from nuclear fractions of U251and U87 cells was detected on days 0, 1, 2, 3, 5, and 7 after pFv/nu transfection, as determined by Western blotting assay.Tomonitor the expression of the intrabody, the nuclear extract from the pFv/nu-transfected cells was collected forWestern blotting assay with anti ^ c-myc mAb and goat anti-mouse horseradish peroxidase ^ conjugated IgG as the primary and secondary antibodies, respectively, whereas anti-GAPDH antibody served as an internal control. B, densitometry results of immunublots described in (A), plotted as change from control amount on day 0 after normalization to GAPDH expression level of each lane. C, depletion of p65 nuclear expression in U251and U87 cells with anti-p65 intrabody transfection was monitored for 7 d by flow cytometry.The nuclear fractions were incubated with anti-p65 antibody as a primary antibody and then with FITC-conjugated goat anti-rabbit IgG antibody as a secondary antibody. D, quantitation of nuclear p65 expression on days 0, 1, 2, 3, 5, and 7 after transfection. The formula used for calculation of expression inhibition was percent inhibition = (p65mock À p65Fv) / p65 mock, which is shown in (C). Points, mean from three separate experiments; bars, SD (P < 0.001).

www.aacrjournals.org 2183 Clin Cancer Res 2007;13(7) April1,2007 Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2007 American Association for Cancer Research. Cancer Therapy: Preclinical by flow cytometry. p65 expression level inversely correlated contribute to the reduced p65 expression in tumor cells after with the increasing expression of the intrabody on days 1, 2, 3, pFv/nu transfection. and 5, which was consistent with the hypothesis that the Effect of anti-p65 intrabody on NF-kB activity. There was no increased expression of anti-p65 intrabody in nucleus might apparent reduction in both fractions of the control samples

Clin Cancer Res 2007;13(7) April 1, 20 07 2184 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2007 American Association for Cancer Research. Intrabody against NF-kB Suppresses Glioma Invasion and cytosolic fraction of intrabody-transfected samples, as of U251 and U87 cells by 88.6% and 84%, respectively, determined by Western blotting (Fig. 4A), whereas a significant as compared with mock-transfected and control vector– decrease in expression of p50, p65, and their phosphorylated transfected cells. Furthermore, the results (Fig. 7B) showed forms was detected in the nuclear fraction of intrabody- that HMECs treated with conditioned media from mock- and transfected samples. This implicated that the translocation pCMV/nu-transfected glioblastoma cells were able to form ability of p65/p50 dimers into the nuclei was apparently capillary-like structures. It shows that HMECs treated with inhibited via the intrabody transfection, leading to their down- conditioned media from pFv/nu-transfected U251 and U87 regulation mainly in the nucleus of intrabody-transfected cells. cells showed less capillary-like networks (60.5% and 62.1%, res- In addition, the result (Fig. 4B) from Western blotting showed pectively), as compared with the controls. Additionally, capillary- that anti-p65 intrabody significantly down-regulated InBa and like structure formation was observed using immunofluorescent phospho-InBa by blocking nuclear translocation of NF-nB and staining for factor VIII in cocultured assay. HMECs cultured transcription of its target genes. However, the expression of InB with mock- and pCMV/nu-transfected cells exhibited capillary kinase a as the upstream protein of NF-nB activation signal networks, whereas the formation of capillary-like structures in pathway did not significantly alter. These findings were further pFv/nu-transfected U251 and U87 cells was almost completely shown to be consistent with the results shown in Fig. 4A. suppressed (87.8% and 78.6%, respectively; Fig. 7C). The results (Fig. 4C) from electrophoretic mobility shift assay Furthermore, results from the invasion and angiogenesis shows that the ability of NF-nB to bind to probe DNA was assay indicated that the abilities of local invasiveness and almost completely inhibited in the nuclei of U251 and U87 neovascularization of glioma cells showed cell line–specific cells transfected with pFv/nu as compared with mock cells and differences associated with intrabody expression. As we control vector–transfected cells, which showed that NF-nB detected, the invasive ability of U87 cells was significantly transcriptional activity was significantly inhibited by the impaired, as well as that of U251 cells in which the expression intrabody transfection. of uPA and MMP-9 showed less reduction. Although the VEGF Effect of transfection on cell proliferation and apoptosis of expression in U87 cells was not inhibited as much as that in glioma cells. Figure 5A shows that the growth rate of all the U251 cells, the formation of capillary-like structure was almost cells of 72-h posttransfection was very similar in MTT assay, completely suppressed in U87 cells with the intrabody suggesting that transfection does not alter cell proliferation rate expression. It suggested that the proposed molecular basis in these cells. for endothelial tube formation and local invasion is a very The result (Fig. 5B and C) indicates that transfection might, to complicated process, which involved interactions of multiple a small degree, induce apoptosis in glioma cells. However, there molecules such as adhesive molecules, epidermal growth factor was no significant difference between pCMV/nu (control vector) receptor family, and other transcription cofactors. and pFv/nu (experimental construct), which implies that the Inhibition of anti-p65 intrabody on intracranial tumor growth intrabody does not seem to induce NF-nB–mediated apoptosis. in nude mice. Having shown that anti-p65 intrabody decreased Effect of anti-p65 intrabody on invasion and angiogenesis via invasion and angiogenesis by down-regulation of MMP-9, uPA, blocking NF-kB bioactivity. Western blotting assay (Fig. 6A) and VEGF in U251 and U87 cells in vitro, we tested the growth and RT-PCR assay (Fig. 6B) showed reduced expression of of the tumors accomplished by pFv/nu-transfected cells protein and mRNA of MMP-9, uPAR, uPA, and VEGF after using an orthotopic intracranial glioma model. As shown transfection with anti-p65 intrabody. Enzymatic activity of in Fig. 8A and B, the tumors intracranially injected with pFv/ MMP-9 and uPA was significantly inhibited as compared with nu-transfected U251 or U87 cells showed a significant reduction the controls by gelatin and fibrin zymography (Fig. 6C). The as compared with control groups, whereas U251 (or U87) cells expression level in related target proteins and mRNAs of each developed obvious intracranial tumors in nude mice. cell line showed a cell line–specific difference (Fig. 6A and B), in which VEGF decreased dramatically with intrabody expres- Discussion sion in U251 but not in U87 cells, whereas MMP-9 expression showed more reduction in U87 than in U251 cells. It implied With the advent of phage display libraries derived from the that decreased NF-nB–mediated protein expression by intra- human IgG repertoire, an scFv against the target of interest can body expression is attributed to the cell line–specific difference be selected in a relatively short time. Over the past decade, and the effect of other transcriptional cofactors besides those there has been growing interest in the use of scFv fragments controlled by NF-nB. aimed at intracellular targets to inhibit protein expression and The result (Fig. 7A) from Matrigel invasion assay indicates trafficking as a gene therapeutic strategy against cancer, that intrabody transfection significantly inhibited the invasion commonly referred to as intrabody. Because intrabodies can

Fig. 4. Effect of anti-p65 intrabody on NF-nBactivity.A, reduced expression of p65, p50, and their phosphorylated forms in cell nucleus. Expression of p65/p50 heterodimer in nuclear extractions and cytosolic fractions of U251and U87 tumor cells after transfection with anti-p65 intrabody recombinant was determined by Western blot assay, as described in Materials and Methods. GAPDH antibody was used to verify that the same amount of protein has been loaded in each lane. B, the expressions of NF-nB activity ^ related proteins in human glioma U251and U87 cells after transiently transfection with recombinants were analyzed byWestern blot assay, as described in Materials and Methods. GAPDH antibody was used as an internal control.Western blot results are representative of three independent preparations. A (right)andB(bottom), quantitation analysis of the respective data. Columns, mean of three observations followed by normalization with GAPDH protein amount. The relative protein expression in each lane was compared with untreated mock cells, arbitrarily as one unit (AU, arbitrary unit). C, effect of anti-p65 intrabody expression on the DNA binding activity of NF-nB in human glioma U251 (left)andU87(right) cells after transient transfection with recombinant pFv/nu in electrophoretic mobility shift assay, detected as described in Materials and Methods. First, nuclear proteins were incubated with a biotin-labeled NF-nB probe or an unlabeled cold probe after preparation of nuclear extracts. Subsequently, the shifted bands of the protein/DNA complexes were separated from the unbound dsDNA by running on a 6% nondenaturing polyacrylamide gel. Following Pall Biodyne B membrane transfer, immobilization of bound oligonucleotides, and exposure to film, the shifted bands were visualized by enhanced chemiluminescent system. For supershift assay, anti-p65 and anti-p50 antibodies were added to the nuclear extracts before incubation with the DNA probe.

www.aacrjournals.org 2185 Clin Cancer Res 2007;13(7) April1,2007 Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2007 American Association for Cancer Research. Cancer Therapy: Preclinical

Fig. 5. Effect of transfection on cell proliferation and apoptosis of glioma cells. A, effect of transfection on cell proliferation. In the MTTassay, U251and U87 cells were trypsinized after 48-h posttransfection with mock, pCMV/nu, pGFP/nu, and pFv/nu, respectively, and continually cultured for 48 h. The number of metabolically active cells (5 Â 10 3) was detected following the addition of MTT for 3-h incubation at 37jC followed by detergent reagent for additional 1-h incubation. Absorbance of living cells in 96-well plates was measured on an ELISA reader at 570 nm with a reference wavelength of 630 nm. Columns, mean from three separate experiments; bars, SD (P < 0.001). B, effect of transfection on apoptosis. U251and U87 cells after 48-h transfection were stained with Hoechst 33244. The apoptotic cells were counted under a fluorescent microscope in five random 200-fold fields, as compared with the control, arbitrarily chosen as one unit. Right, columns, mean from three separate experiments; bars, SD (P < 0.001). C, terminal deoxynucleotidyl transferase biotin-dUTP nick-end labeling apoptosis assay was done for DNA fragmentation/fluorescence staining. Briefly, pFv/nu- or pCMV/nu-transfected cells in eight-well chamber slides were fixed, washed, permeabilized, and incubated with terminal deoxynucleotidyl transferase end-labeling cocktail. Following immersion inTris-borate buffer to stop the reaction and blocking, the cells were subsequently incubated with avidin-FITC solution.Finally,theslides were observed to count the apoptotic cells stained in green under a fluorescent microscope by the same method as in (B). Right, columns, mean from three separate experiments; bars, SD (P < 0.001). be directed to specific subcellular compartments and precise the human single-fold scFv libraries (Tomlinson I + J) to epitopes on target proteins, they have the potential to block prepare for the humanized anti-p65 intrabody precisely only one of several functions of an expressed protein at the expressed in the nucleus of glioma cells as determined by posttranslational level. This potential has been exploited by immunoblotting and immunofluorescent results and also several laboratories in gene therapeutic strategies as well as in proved that the recombinant presented a longer half-life and revealing the biological function of proteins in cancer, AIDS, more effective intracellular activity. and other research areas (13–15). Intrabodies present a potent In the past, scFvs derived from high-affinity antibodies were alternative to methods of gene inactivation that allows for thought to be most desirable for use as intrabodies due to their targeting at the level of DNA or mRNA, such as antisense inherent antigen-binding properties. However, growing evi- oligonucleotides (16), zinc finger proteins (17), and RNA dence suggests that the intracellular stability of an intrabody, not interference (18). In the present study, we succeeded in using its affinity, is most critical for its efficacy. A useful description

Clin Cancer Res 2007;13(7) April 1, 20 07 2186 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2007 American Association for Cancer Research. Intrabody against NF-kB Suppresses Glioma Invasion of an effective intrabody should include specificity, affinity, and intrabody in the in vivo environment, it is proposed that we intracellular stability, where stability is the determining factor could fuse the scFv with other stabilized fragments such as Fc of for intrabody efficacy. Stability seems to be especially important human IgG, or establish other additional intracellular screen- when intrabodies are expressed in the strongly reducing ing steps such as ribosome display, mRNA display, and yeast environment of cytosol, which disrupts the formation of two-hybrid screening after the primary phage display (25–28). intrachain disulfide bridges that can be essential for the proper Local invasive infiltration and growth are key features in folding, solubility, and inherent stability of scFvs. ScFvs with glioblastoma, which are accompanied by remodeling of the extended half-lives have been shown to achieve higher steady- vasculature and by the destruction of the surrounding tissues. state expression levels, a by-product of which seems to be a more Several reports in human gliomas showed overexpression of effective target molecule inactivation. Removal from the harsh uPA, uPAR, MMP-9, and VEGF, which are involved in the cytosolic environment by directing scFvs expression to the ER or adhesion, invasion, metastasis, and angiogenesis of human nucleus can also increase half-life, as long as the desired glioblastomas. Numerous studies on interference with the uPA- localization was detectable for interaction with the target pro- uPAR system and the MMP and VEGF pathways have shown the tein (19–24). Our construct pFv/nu, which displayed a long successful inhibition of neovascularization and tumor growth half-life and accumulated at steady-state level in the nucleus in preclinical or clinical trials (29–40). It has been shown that even after transit through the reducing cytosolic environment, the nB binding site of NF-nB was identified in the promoter proved to be suitable for the interaction with the target mole- regions of uPA, uPAR, MMP-9, and VEGF genes. As such, their cule owing to its intracellular stability and activity. In future transcriptional activation could be controlled by NF-nBinthe research, to enhance intracellular stability and activity of the nucleus where it recognizes and functionally binds to their

Fig. 6. The effects of mock-, pCMV/nu-, pGFP/nu-, and pFv/nu-transfected cells on the expression of MMP-9, uPAR, uPA, andVEGF at protein and mRNA levels byWestern blotting assay (A) and RT-PCR assay (B), respectively, as described in Materials and Methods.The used primers for mRNA expression detection of MMP-9, uPAR, uPA,VEGF, and GAPDH were shown inTable1.Results fromWestern blots and RT-PCR are representative of three independent preparations. A and B, right, quantitation of the respective data. Columns, mean of three observations and normalized to untreated mock cells, arbitrarily chosen as one unit. C, effect on enzymatic activities of MMP-9 and uPA in transfected cells. Gelatin zymography for MMP-9 and fibrin zymography for uPA in U251and U87 cells after 48-h transfection with mock, pCMV/nu, pGFP/nu, and pFv/nu were done. Briefly, the conditioned medium from transfected cells was collected, followed by incubation with serum-free medium and protein estimation. Equal amount of protein was loaded into nondenaturing SDS-PAGE gel containing relevant substrates. Subsequently, the gel was incubated at 37jC overnight, stained, and destained; the clear bands in the gels represent enzymatic activity of MMP-9 or uPA. Representative of three separate experiments.

www.aacrjournals.org 2187 Clin Cancer Res 2007;13(7) April1,2007 Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2007 American Association for Cancer Research. Cancer Therapy: Preclinical

Fig. 7. Effect of anti-p65 intrabody on invasion and angiogenesis. A, invasiveness of U251and U87 cells transfected with recombinants in Matrigel invasion assay. Cells were transfected with empty vector pCMV/nu or anti-p65 intrabody construct pFv/nu. After 2 d, 1 Â10 5 cells were allowed to invade through transwell inserts (8 Am) coated with Matrigel. The cells on lower surface of chambers were stained, counted, and photographed under a light microscope. The in vitro inhibition of invasiveness was calculated in five random 200-fold fields using the following formula: percent inhibition = (mock À Fv) / mock. Right, columns, mean from three separate experiments; bars, SD (P < 0.001). B, HMEC endothelial cells treated with conditional medium from U251and U87 glioma cells after intrabody transfection by in vitro angiogenesis assay. The conditioned medium of glioma cells was collected after 48-h transfection following filtering of medium. HMEC-1endothelial cells seeded in eight-chamber slides were cultured with the above medium for 72 h until the formation of capillary network was observed. In the end of the experiment, angiogenesis was assessed by H&E staining and photographed under a microscope. Right, columns, mean from three separate experiments; bars, SD (P < 0.001). C, formation of capillary-like structures in coculture of endothelial HMEC cells and glioma U251/U87 cells after intrabody transfection. The mock or 48-h posttransfected glioma cells were plated onto eight-chamber slides following seeding of HMEC-1endothelial cells to coculture constitutively for 72 h until the formation of capillary network. The angiogenesis was assessed by the immunofluorescent staining of factor VIII and quantified under a fluorescent microscope. Right, columns, mean from three separate experiments; bars, SD (P < 0.001).

Clin Cancer Res 2007;13(7) April 1, 20 07 2188 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2007 American Association for Cancer Research. Intrabody against NF-kB Suppresses Glioma Invasion

Fig. 8. Evaluation of intracranial tumor growth in the brains of nude mice by intrabody transfection. A, intracranial tumor assay.The growth of the tumor accomplished glioma cells was tested using an orthotopic intracranial glioma model. Briefly, U251 (or U87) glioma mock cells (2 Â 10 6 in serum-free medium), control vector (pCMV/nu and pGFP/nu), and intrabody (pFv/nu) ^ transfected cells were injected intracranially, respectively. After 5 wk, mice from each group (five per group) were killed and their brains removed, fixed in formaldehyde, sectioned, and stained (H&E). B, semiquantification of tumor volume in control, pCMV/nu-, pGFP/nu-, and pFv/nu-treated groups. Sections were observed blindly and scored semiquantitatively for tumor size under a microscope (Â100; Olympus BX16). Columns, mean from five animals in each group; bars, SD (P < 0.001). promoters and further triggers their transcription (29–33). explain why the depletion of nuclear p65 in U87 cells could Moreover, aberrant or constitutive activated NF-nB has been maintain longer than in U251 cells. To overcome this transient detected in human high-grade gliomas (3–5). Based on the depletion of p65, the selection of a stably transfected clone could important role of NF-nB and its mediated genes in tumor be a helpful approach to obtain a stable nuclear expression of the growth, angiogenesis, invasion, and metastasis, blocking NF-nB intrabody. In addition, we are planning to screen an intrabody activity to inhibit target gene expression could then serve as a able to directly degrade its target protein (p65) after binding promising approach for cancer therapy. According to reports, reaction, which could be another more effective way to constitutive NF-nB activity in cancer cells can be inhibited completely inhibit the activity of NF-nB in nucleus. The result through down-regulation of the NF-nB intrinsic inhibitor InB; from intracranial tumor experiment showed remarkable inhibi- its specific upstream kinases such as phosphatidylinositol tion of tumor growth after injection with stable clones of pFv/ 3-kinase, protein kinase C, mitogen-activated protein kinase, nu-transfected U251 or U87 cells compared with obvious and InB kinase; and a nonspecific proteasome inhibitor, PS-341 intracranial tumors established by control vector–transfected (40–43). We therefore hypothesized that direct blocking U251 (or U87) cells in nude mice, which further showed that the NF-nB transcriptional activity by intrabody expression could constant inhibition of tumor growth by pFv/nu transfection contribute to the inhibitory expression and activity of MMP-9, could be observed after stable expression of the intrabody. uPA, uPAR, and VEGF, resulting in inhibition of tumor inva- Because most of studies on NF-nB indicate that it mediates sion, metastasis, and angiogenesis of human gliomas. the expression of antiapoptotic and proapoptotic proteins, we For the first time, we reported that the inhibition of p65 detected whether apoptotic cells were induced after transfection. transactivation by anti-p65 intrabody expression in the nucleus However, there was no apparent difference in the number of might suppress the transcriptional activity of NF-nB, resulting in apoptotic cells (Fig. 5B and C), suggesting that anti-p65 intrabody the inhibition of the expression and activity of its target proteins expression did not induce the NF-nB–mediated apoptosis path- associated with human glioma invasiveness and angiogenesis. way. Although the expression of its target antiapoptotic pro- Based on the kinetic analysis (Fig. 3), we showed that nuclear teins, such as FLIP and Bcl-XL, was decreased by the inhibition of p65 level decreased maximally with inhibition rates of 40% NF-nB transcription activity in our experiment (data not shown), (U251) and 50% (U87) on day 3 (in U251) or day 5 (in U87), the caspase expression associated with apoptotic phenotype did respectively. After that time point, the expression level of nuclear not increase. This implies that there perhaps exists an alternative- p65 started to be restored although the intrabody amount signaling pathway to prevent apoptosis of glioma cells. maintained a raised trend until day 7 after transfection, sug- In conclusion, blocking the transcriptional activity of NF-nB gesting that the action mechanism of the intrabody herein, using anti-p65 nuclear-targeting intrabody inhibited glioblas- which is able to neutralize its target protein (p65), change toma cell angiogenesis, invasion, and intracranial tumor its space conformation by forming the antigen-antibody (p65- growth via the down-regulation of MMP-9, uPAR, uPA, and intrabody) complex, and then inhibit the nuclear p65 expres- VEGF. As such, this approach may be useful for the treatment of sion, was a reversible and time-limited reaction based on a gliomas as well as other tumors. Furthermore, the intrabody certain level of expression of the intrabody in nucleus. As such, strategy provides a novel way to interfere with NF-nB activity in the depleted nuclear p65 could be free partly once the nuclear the nuclear compartment and to study the biological signifi- anti-p65 intrabody was degraded to release p65 protein from the cance of the NF-nB signaling pathway in cancer biology. complex due to the transient expression of the intrabody after transient pFv/nu transfection into the glioma cells. In addition, Acknowledgments the expression of the intrabody showed a cell line–specific difference that the increasing ratio of the intrabody in U87 cells We thank Shellee Abraham for manuscript preparation and Diana Meister and was significantly more than that in U251 cells, which might Sushma Jasti for manuscript review.

www.aacrjournals.org 2189 Clin Cancer Res 2007;13(7) April1,2007 Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2007 American Association for Cancer Research. Cancer Therapy: Preclinical

References 1. Cavenee WK, Furnari FB, Nagane M. Diffusely infil- cross-reactive anti-VEGF receptor 2 recombinant antigen activator system: a multifunctional role in tumor trating astrocytomas. In: Kleihues P,CaveneeWK, edi- bodies selected from an immune b9 allotype rabbit progression and metastasis. Clin Orthop Relat Res tors.Tumors of the nervous system. Lyon: IARCPress; antibody library. J Immunol Methods 2004;288: 2003;(415 Suppl):S46 ^58. 2000.p.9^51. 14 9 ^ 6 4. 31. Machein MR, Plate KH.VEGFin brain tumors. J Neu- 2. Burger PC,Vogel FS, Green SB, StrikeTA. Glioblasto- 15. Wheeler YY,Chen SY,Sane DC. Intrabody and intra- rooncol 2000;50:109 ^ 20. ma multiforme and anaplastic astrocytoma. Pathologic kine strategies for molecular therapy. Mol Ther 2003; 32. HasanJ, Jayson GC.VEGF antagonists. Expert Opin criteria and prognostic implications. Cancer 1985;56: 8:355^66. Biol Ther 2001;1:703^ 18. 110 6 ^ 11. 16. Wagner RW, Flanagan WM. Antisense technology 33. IkedaE,AchenMG,BreierG,RisauW.Hypoxia- 3. KarinM,CaoY,GretenFR,LiZW.NF-nBincancer: and prospects for therapy of viral infections and can- induced transcriptional activation and increased from innocent bystander to major culprit. Nat Rev cer. Mol Med Today 1997;3:31 ^ 8. mRNA stability of vascular endothelial growth fac- Cancer 2002;2:301 ^ 10. 17. Beerli RR, Barbas CF III. Engineering polydactyl tor in C6 glioma cells. J Biol Chem 1995;270: 4. Ravi R, Bedi A. NF-nB in canceröa friend turned foe. zinc-finger transcription factors. Nat Biotechnol 19761 ^ 6. Drug Resist Updat 2004;7:53^ 67. 2002;20:135^ 41. 34. Kerr DJ. Targeting angiogenesis in cancer: clinical 5. Kanzawa T, Ito H, Kondo Y, Kondo S. Current and 18. Hannon GJ. RNA interference. Nature 2002;418: development of bevacizumab. Nat Clin Pract Oncol future gene therapy for malignant gliomas. J Biomed 244 ^51. 2004;1:39 ^43. Biotechnol 2003;2003:25 ^ 34. 19. Caron dF, Gruel N,Venot C, et al. Restoration of tran- 35. Kim KJ, Li B,WinerJ, et al. Inhibition of vascular en- 6. Wang H, Wang H, Zhang W, Huang HJ, Liao WS, scriptional activity of p53mutants in human tumour dothelial growth factor-induced angiogenesis sup- Fuller GN. Analysis of the activation status of Akt, cells by intracellular expression of anti-p53single presses tumour growth in vivo. Nature 1993;362: NFnB, and Stat3in human diffuse gliomas. Lab Invest chain Fv fragments. Oncogene 1999;18:551 ^ 7. 841^4. 2004;84:941 ^ 51. 20. Mary MN, Venot C, Caron de Fromentel C, et al. A 36. Lin P, Sankar S, Shan S, et al. Inhibition of tumor 7. Buss H, Dorrie A, Schmitz ML, et al. Phosphorylation tumor specific single chain antibody dependent gene growth by targeting tumor endothelium using a solu- of serine 468 by GSK-3h negatively regulates basal expression system. Oncogene 1999;18:559 ^ 64. ble vascular endothelial growth factor receptor. Cell p65 NF-nB activity. J Biol Chem 2004;279:49571 ^ 4. 21. Rajpal A, Turi TG. Intracellular stability of anti-cas- Growth Differ 1998;9:49 ^ 58. 8. Egan LJ, Mays DC, Huntoon CJ, et al. Inhibition of pase-3intrabodies determines efficacy in retargeting 37. Plate KH, Breier G,Weich HA, Risau W.Vascular en- interleukin-1-stimulated NF-nB RelA/p65 phosphory- the antigen. J Biol Chem 2001;276:33139 ^ 46. dothelial growth factor is a potential tumour angio- lation by mesalamine is accompanied by decreased 22. Ramm K, Gehrig P, Pluckthun A. Removal of the genesis factor in human gliomas in vivo.Nature1992; transcriptional activity. J Biol Chem 1999;274: conserved disulfide bridges from the scFv fragment 359:845^8. 26448^53. of an antibody: effects on folding kinetics and aggre- 38. PulukuriSM,GondiCS,LakkaSS,etal.RNAinter- 9. Wang D,Westerheide SD, HansonJL, Baldwin AS, Jr. gation. J Mol Biol 1999;290:535 ^ 46. ference-directed knockdown of urokinase plasmino- Tumor necrosis factor a-induced phosphorylation of 23. Wu Y, Duan L, Zhu M, et al. Binding of intracellular gen activator and urokinase plasminogen activator RelA/p65 on Ser529 is controlled by casein kinase II. anti-Rev single chain variable fragments to different receptor inhibits prostate cancer cell invasion, survival, J Biol Chem 2000;275:32592 ^ 7. epitopes of human immunodeficiency virus type 1 and tumorigenicity in vivo. J Biol Chem 2005;280: 10. Schmitz ML, dos Santos Silva MA, Baeuerle PA. rev: variations in viral inhibition. J Virol 1996;70: 36529^ 40. Transactivation domain 2 (TA2) of p65 NF-nB. Simi- 3290 ^ 7. 39. Lakka SS, Gondi CS,Yanamandra N, et al. Inhibition larity to TA1and phorbol ester-stimulated activity and 24. Zhu Q, Zeng C, Huhalov A, et al. Extended half-life of cathepsin B and MMP-9 gene expression in glio- phosphorylation in intact cells. J Biol Chem 1995;270: and elevated steady-state level of a single-chain Fv blastoma cell line via RNA interference reduces tumor 15576 ^ 8 4. intrabody are critical for specific intracellular retarget- cell invasion, tumor growth and angiogenesis. Onco- 11. Gondi CS, Lakka SS, Dinh D, Olivero W, Gujrati M, ing of its antigen, caspase-7. J Immunol Methods gene 2004;23:4681 ^ 9. Rao JS. Down-regulation of uPA, uPAR and MMP-9 1999;231:207 ^ 22. 40. Huang S, Pettaway CA, Uehara H, Bucana CD, using small, interfering, hairpin RNA (siRNA) inhibits 25. Amstutz P, Forrer P, Zahnd C, Pluckthun A. In vitro Fidler IJ. Blockade of NF-nB activity in human pros- glioma cell invasion, angiogenesis and tumor growth. display technologies: novel developments and appli- tate cancer cells is associated with suppression of an- Neuron Glia Biology 2004;1:165 ^ 76. cations. Curr Opin Biotechnol 2001;12:400 ^ 5. giogenesis, invasion, and metastasis. Oncogene 2001; 12. Lakka SS, Gondi CS,Yanamandra N, et al. Synergis- 26. der MaurAA, Zahnd C, Fischer F, et al. Direct in vivo 20:4188 ^ 97. tic down-regulation of urokinase plasminogen activa- screening of intrabody libraries constructed on a high- 41. Sliva D, Rizzo MT, English D. Phosphatidylinositol tor receptor and matrix metalloproteinase-9 in SNB19 ly stable single-chain framework. J Biol Chem 2002; 3-kinase and NF-nB regulate motility of invasive glioblastoma cells efficiently inhibits glioma cell inva- 277:45075^85. MDA-MB-231 human breast cancer cells by the se- sion, angiogenesis, and tumor growth. Cancer Res 27. Strube RW, Chen SY. Enhanced intracellular stabil- cretion of urokinase-type plasminogen activator. J 2003;63:2454 ^ 61. ity of sFv-Fc fusion intrabodies. Methods 2004;34: Biol Chem 2002;277:3150 ^ 7. 13. Heng BC, CaoT. Making cell-permeable antibodies 17 9 ^ 8 3. 42. Sliva D, English D, Lyons D, Lloyd FP, Jr. Pro- (Transbody) through fusion of protein transduction 28. Worn A, Pluckthun A. Stability engineering of anti- tein kinase C induces motility of breast cancers by domains (PTD) with single chain variable fragment body single-chain Fv fragments. JMol Biol 2001;305: up-regulating secretion of urokinase-type plasmino- (scFv) antibodies: potential advantages over antibod- 9 89 ^ 1010. gen activator through activation of AP-1 and NF- ies expressed within the intracellular environment 29. Chintala SK,Tonn JC, Rao JS. Matrix metalloprotei- nB. Biochem Biophys Res Commun 2002;290: (Intrabody). Med Hypotheses 2005;64:1105 ^ 8. nases and their biological function in human gliomas. 552^ 7. 14. Popkov M, Jendreyko N, Gonzalez-Sapienza G, Int J Dev Neurosci 1999;17:495 ^ 502. 43. Adams J. Development of the proteasome inhibitor Mage RG, Rader C, Barbas CF III. Human/mouse 30. Choong PF, Nadesapillai AP. Urokinase plasmino- PS-341. Oncologist 2002;7:9^ 16.

Clin Cancer Res 2007;13(7) April 1, 20 07 2190 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2007 American Association for Cancer Research. Transfection with Anti-p65 Intrabody Suppresses Invasion and Angiogenesis in Glioma Cells by Blocking Nuclear Factor-κB Transcriptional Activity

Liang Li, Christopher S. Gondi, Dzung H. Dinh, et al.

Clin Cancer Res 2007;13:2178-2190.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/13/7/2178

Cited articles This article cites 42 articles, 13 of which you can access for free at: http://clincancerres.aacrjournals.org/content/13/7/2178.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/13/7/2178.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/13/7/2178. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.