| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
1 Department of Medicine, Section of Infectious Diseases, Baylor College of Medicine; and University of Texas Health Science Center, School of Public Health, Houston, Texas
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Stat3, previously termed acute phase response factor (APRF; refs. 2 , 8, 9, 10 ), is activated within cells by binding to the cell surface of >40 ligands; it also is constitutively activated in many human cancers (11, 12, 13) including 82% of prostate cancers (14) , 69% of breast cancers (15) , 82 to 100% of squamous cell carcinoma of head and neck (16) , and 71% of nasopharyngeal carcinoma (17) . Activated Stat3 up-regulates the expressions of antiapoptosis proteins, such as Bcl-xL and Mcl-1, thereby decreasing spontaneous apoptosis in cancer cells (18 , 19) . Targeting of Stat3 directly with agents, such as antisense oligonucleotides (20, 21, 22) , a decoy oligonucleotide (23) , or indirectly with Janus-activated kinase inhibitors, AG490 and JSI-124 (24 , 25) , demonstrated that Stat3 activation contributes to tumor cell growth and resistances to apoptosis, providing the strongest rationale for therapeutic approaches aimed at reducing levels or activity of Stat3 in cancers where it is activated.
Stat3, similar to all of the STAT proteins, is composed of several domains: a tetramerization domain, a coil-coil domain, a DNA-binding domain, a linker domain, an SH2 domain, a critical tyrosine reside (Y705), and a COOH-terminal transactivation domain. Resolution of the crystal structure of Stat3 revealed the structural basis for DNA-binding and dimer formation in sufficient detail to provide primary targets for novel drug development (26) . One class of drug in early development as a novel agent targeting Stat3 is G-rich oligodeoxynucleotides (ODNs) that form G-quartet structures intracellularly (27) .
G-rich DNA and RNA have the ability to form inter- and intramolecular four-stranded structures, referred to as G-quartets (28 , 29) . G-quartets arise from the association of four G-bases into a cyclic Hoogsteen H-bonding arrangement, and each G-base makes two H-bonds with its neighbor G-base (N1 to O6 and N2 to N7). G-quartets stack on top of each other to give rise to tetrad-helical structures. The stability of G-quartet structures depends on several factors: the presence of the monovalent cations, the concentration of the G-rich oligonucleotides present, and the sequence of the G-rich oligonucleotides under study. Potassium with the optimal size to interact within a G-octamer greatly promotes the formation of G-quartet structures and increases their stability. G-quartet oligodeoxynucleotides (GQ-ODNs) have been suggested to play a critical role in several biological processes including modulation of telomere activity (30) , inhibition of human thrombin (31) , HIV infection (32) , HIV-1 integrase activity (33, 34, 35) , human nuclear topoisomerase 1 activity (36) , and DNA replication in vitro (37) . On the basis of the structure and mechanism of Stat3 activation, G-quartet–forming oligonucleotides were developed recently to block Stat3 activity within cancer cells (27) . GQ-ODNs directly target Stat3 protein and inhibit their ability to bind DNA thereby decreasing transcriptional activation of genes important for apoptosis resistance, such as Bcl-x and Mcl-1.
In the present report, we demonstrated that i.v. administration of GQ-ODNs dramatically reduces the growth in nude mice of xenografts of prostate and breast cancer cells in which Stat3 is constitutively activated. GQ-ODNs inhibited tumor cell growth by markedly enhancing tumor cell apoptosis and representing a promising agent for treatment of metastatic tumors in which Stat3 is constitutively activated.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
25 kDa polymer, Sigma-Aldrich, Inc., St. Louis, MO) was generously provided by Dr. Charles Densmore (Baylor College of Medicine). Interleukin (IL) 6 and antibodies against Stat3, Stat1, Bcl-xL, and Bcl-2 were purchased from Santa Cruz Biotechnology Biotechnology (Santa Cruz, CA). Antibody against caspase 3 was obtained from Cell Signaling Technology (Beverly, MA).
Assay of Inhibition by GQ-ODNs of Stat3 DNA-Binding Activity in Cancer Cells.
GQ-ODN plus polyethyleneimine at a weight ratio of polyethyleneimine/ODN of 2:1 was added to cells (5–7 x 105). After incubation for 3 hours, the cells were washed three times with fresh medium without polyethyleneimine/ODN and the incubation continued. After 6 to 72 hours, cells were incubated without or with IL-6 (25 ng/mL) at 37°C for 20 minutes before extraction and analysis by electrophoretic mobility shift assay. The cell extraction and electrophoretic mobility shift assay were performed as described previously (27)
. In some experiments, cells were scraped and harvested for isolation of nuclear proteins as described previously (38)
. Briefly, 50 µL of ice-cold Buffer A [10 mmol/L Tris-HCL (pH 9.), 2 mmol/L MgCl2, 5 mmol/L KCl, 10% glycerol, 1 mmol/L EDTA, and 1 mmol/L DDT] plus 1% NP40 was added and the cells allowed to swell on ice for 15 minutes, then vortexed vigorously. The lysate was centrifuged at 700 x g for 5 minutes at 4°C and the resulting nuclear pallet resuspended in 50 µL of cold buffer (20 mmol/L HEPES, 0.4 mol/L NaCl, 1 mmol/L EDTA, 1 mmol/L EGTA, and 1 mmol/L DDT). Nuclear proteins were extracted by vigorous shaking for 30 minutes at 4°C and cell debris pelleted by centrifugation at 18,000 x g for 5 minutes at 4°C. Protein concentrations of whole-cell and nuclear extracts were measured by Bradford assay.
Statistical Docking Calculations.
The structures of the designed G-quartet ODNs, such as T40214 and its analogues, were built up through modification of the nuclear magnetic resonance structure of GQ-ODN T30923 (39)
and optimized under AMBER force field by INSIGHTII/DISCOVER. The optimization of the molecular structures proceeded as follows: (1) 100 steps of conjugate gradient energy minimization, (2) 1,000 steps of restrained MD equilibration with a time step of 0.33 fs at 1,000 K, (3) 1,000 steps of restrained MD equilibration with a time step of 0.1 fs at 300 K, and (4) 1,000 steps of conjugate gradient energy minimization. The intramolecular H-bonds of G-quartets were used as constraints for the molecular optimization.
We docked each GQ-ODN 1,000 times onto the available structure of the dimer of Stat3 SH2 (26) using the GRAMM docking program without placing any restrictions on binding sites. This program uses a geometry-based algorithm for predicting the structures of complexes between molecules of known structure. It can provide quantitative data related to the quality of the contact between the molecules. The intermolecular energy calculation relies on the well-established correlation and Fourier transformation techniques used in the field of pattern recognition. The distribution of H-bond formation between each Stat3/GQ-ODN complex was calculated and analyzed.
In vivo Delivery of Fluorescently Labeled GQ-ODN.
Fluorescent-labeled T40214 plus polyethyleneimine (each 2.5 mg/kg) was injected into the tail vein of male and female mice weighing 20 g. Twenty-four hours after the infusion, the mice were sacrificed and tissue harvested and frozen. Frozen tissues were sectioned using cryostat microtome sections, lightly fixed, and viewed microscopically.
Tumor Xenograft Models.
Athymic nude mice (Balb/nu/nu, 4 weeks old and weighing 20 g obtained from Charles River Labs) were injected s.c. into the right (or left) flank with one million cancer cells (MDA-MD-468 or PC-3) in 200 µL of PBS. After tumors were established at 7 to 14 days postinjection, 16 nude mice with breast or prostate tumors were randomly assigned into three groups. Mice in group 1 (n = 7) served as placebo and received only polyethyleneimine (2.5 mg/kg), whereas mice in groups 2 and 3 (n = 7) received T40214 (5.0 mg/kg) plus polyethyleneimine (2.5 mg/kg) and T402314 (5.0 mg/kg) plus polyethyleneimine (2.5 mg/kg), respectively. Treatments, administered by tail-vein injection, and sizing of tumors occurred every 2 days. The unpaired two-sample t test, {t = (X1 – X2)/[Sp2(1/n1+1/n2)]1/2}, was used to determine differences in tumor sizes between the placebo and the drug-treated groups.
Immunoblotting.
Tumor xenografts were harvested at the end of treatment period, cut into small pieces, homogenized on ice for 2 minutes, subjected to one round of freeze-thawing, and centrifuged (12,000 rpm for 2 min at 4°C). The supernatants were harvested, assayed for protein concentration by Bradford assay, and immunoblotted as described elsewhere (20)
.
Terminal Deoxynucleotidyl Transferase-Mediated dUTP-Biotin Nick End-Labeling Assay.
The terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end-labeling (TUNEL) assay was performed as described by the manufacturer (Roche Applied Science, Indianapolis, IN). Briefly, prostate (or breast) tumors were harvested from mice within 24 hours of their last treatment. The paraffin-embedded tissues were pretreated using Xylene and EtOH, washed with PBS, and incubated with proteinase K working solution. TUNEL reaction mixture was prepared by adding 50 µL of enzyme solution into 450 µL of the labeling solution. TUNEL reaction mixture (50 µL) was placed on slides followed by 50 µL of 3,3'-diaminobenzidine substrate. Slides were incubated for 5 minute at 24°C, rinsed three times with PBS, dehydrated with xylene, and mounted under a glass coverslip.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
; Table 1
). Each of the G-rich ODN demonstrated the ability to inhibit Stat3 DNA-binding activity when the concentration of ODN was increased from 7 to 285 µmol/L. T40214 and T40231 are the most potent inhibitors of Stat3 activation. Nonspecific ODNs incapable of forming the G-quartet structure demonstrated no activity against Stat3.
|
|
15Å width and 15Å length (39
, 40)
. T40214 and other GQ-ODNs destabilize Stat3 dimers by inserting between their SH2 domains. We proposed that GQ-ODNs interact in the region of the SH2 domain from amino acid residues 638 to 650 (ref. 27
; Fig. 2A
). To establish a structure–activity relationship between Stat3 and G-quartet inhibitors, we generated the G-quartet structure predicted to be formed by each G-rich ODN (Table 1)
and randomly docked each of them onto the known structure of the dimer of Stat3 SH2 domains (26)
1,000 times. Analysis of the histogram of H-bonds formed at each residue within the Stat3 SH2 domain demonstrated that T40214 interacts with SH2 predominantly in the region from residues 638 to 650, as expected, with 35% of total H-bonds formed located in this region (Fig. 2B)
. The percentages of H-bonds formed within the region from residues 638 to 650 of Stat3 for other GQ-ODNs were listed in Table 1
. Composite analysis of all of the histograms revealed that there is an inverse linear correlation between the percentage of GQ-ODN H-binding within this region of the Stat3 SH2 domain and the IC50 of inhibition of Stat3 DNA binding activity (Fig. 2C)
. Thus, these findings indicate that the higher the percentage of GQ-ODN H-bond formation within this region of Stat3 SH2, the greater its ability to inhibit Stat3 activation within cells.
|
confirming the selectivity of GQ-ODN for Stat3 within cells. No inhibition of Stat3 DNA binding activity was observed when HepG2 cells were preincubated with T40214 in the absence of polyethyleneimine (Fig. 3B)
confirming previous findings by us (40)
that a vehicle is required for delivery of GQ-ODN into cells.
|
, lanes 4 to 7). polyethyleneimine alone as vehicle has no effect on Stat3 activation (Fig. 4
, lane 3).
|
indicating that GQ-ODN enters and accumulates within tumor cells.
|
, were selected from among the designed GQ-ODNs to assess if GQ-ODN uptake is accompanied by inhibition of tumor growth. Nude mice were injected s.c. with MDA-MB-468 or PC3, both of which demonstrate constitutive Stat3 activity, and i.v. treated every other day after tumors were established (Fig. 5, C and D)
. The drug treatment groups received T40214 or T40231 (5.0 mg/kg) plus polyethyleneimine (2.5 mg/kg), and the placebo groups each received polyethyleneimine (2.5 mg/kg) alone by tail-vein injection. The mean size of the breast tumor xenografts of placebo-treated mice increased from 7- to 12-fold, whereas the mean sizes of both T40214 and T40231-treated mice remained unchanged (Fig. 5C)
. The growth rate of placebo-treated breast tumors ranged from 11% to 15%/day, and those of T40214 and T40231-treated breast tumors were –0.4%/day (P = 0.001) and 0.6%/day (P = 0.001), respectively. The mean size of the prostate tumor xenografts of placebo-treated mice increased 9- to 10-fold, whereas the mean sizes of T40214-treated and T40231-treated mice increased only 2.2- and 2.6-fold, respectively (Fig. 5D)
. The growth rate of placebo-treated prostate tumors ranged from 21% to 23%/day, whereas the growth of T40214- and T40231-treated prostate tumors was 9.1%/day (P = 0.001) and 8.8%/day (P = 0.001), respectively.
To gain insight into the mechanism of inhibition of tumor growth by GQ-ODN, we harvested the prostate tumor xenografts from placebo-treated and drug-treated mice after five treatments and extracted proteins to assess for levels of Stat3-pY705, Bcl-xL, Bcl-2, and activated caspase 3 protein. Levels of Stat3-pY705, Bcl-xL, and Bcl-2 were decreased by 9-, 4.3-, and 10-fold, respectively, in the tumors from drug-treated animals compared with tumors from placebo-treated mice. These changes were accompanied by a 3-fold increase in caspase 3 cleavage products in the tumors from drug-treated animals compared with tumors from placebo-treated mice (Fig. 6A)
. We have harvested the 4 tumor samples from placebo-treated mice and 3 tumor samples from drug-treated mice and then performed TUNEL assay on the samples. TUNEL staining was performed on 4 tumor samples from placebo-treated mice and 3 tumor samples from drug-treated mice to assess if GQ-ODN treatment increased tumor cell apoptosis. The percentage of apoptotic cells was increased nearly 8-fold in the tumors of drug-treated mice (83.6 ± 1.0%) compared with the tumors of placebo-treated mice (11.2 ± 10.1%; P < 0.001; Fig. 6B
).
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Our previous studies showed that the G-quartet structure of G-rich ODN is essential for the inhibition of Stat3 DNA-binding activity in vitro (27) . The G-quartet structure of T40214 closely resembles a perfect cylinder 15 Å in width and length. This conformation increases the thermal stability of the structures, reduces the capacity of ODN to form molecular aggregates, and increases the probability that each GQ-ODN will target Stat3 homodimer in cells. The finding that T40214-mediated inhibition of ligand-induced Stat3 activation within cells persisted for 72 hours after GQ-ODN exposure can be attributed, in part, to this compact configuration, which contributes to its thermal stability and resistance to nuclease digestion. The intramolecular G-quartet structure prevents single-strand endonucleases from accessing their cleavage sites (41) .
Effective delivery of GQ-ODNs into cancer cells is a key factor for success of inhibitors that target oncogenic signaling intermediates. On the basis of the property of potassium-dependent formation of G-quartet structure, a novel intracellular delivery system has been developed for GQ-ODNs (40)
. Using the novel delivery system, GQ-ODNs were delivered efficiently into the cytoplasm and nucleus of cells. The results reported here showed that GQ-ODNs penetrated poorly into cells without vehicle as evidenced by low drug activity, whereas use of an effective intracellular delivery system such as polyethyleneimine increased the drug activity of GQ-ODNs in cancer cells (Fig. 3, A and B)
. When G-rich ODNs form intramolecular G-quartet structures within the cytoplasm, they are able to diffuse freely through pores into the nucleus, bind to Stat3, and block the transcription of Stat3-regulated genes (Fig. 4
; ref. 27
).
Because of the important role that STAT proteins play in signaling within the immune system (Stat1, 2, 4, 5A, 5B, and 6) and in negative regulation of proliferation (Stat1; ref. 42
), it is highly desirable that an agent targeting Stat3 not have activity against other STAT proteins to avoid problems with immunosuppression and inadvertent stimulation of tumor cell growth. On the basis of their structures (14
, 43)
, the ligand-receptor pairs that lead to their activation (1)
and the phosphotyrosine sites that lead to their recruitment and activation (44)
, the STAT protein family member that most closely resembles Stat3 is Stat1. Stat3 and Stat1 share >50% amino acid sequence homology in the region of the SH2 domain (13)
. Therefore, to address the issue of specificity of GQ-ODN inhibition, we focused on Stat1. In previous studies, we demonstrated that the IC50 of T40214 for the inhibition of Stat3 DNA-binding activity in vitro was 4-fold less than for its inhibition of Stat1 DNA-binding activity (27)
. Here, we demonstrated that the specificity of T40214 for Stat3 versus Stat1 extends in vivo. The concentration of T40214 that inhibited 50% of Stat3 activation after preincubation of T40214 in cells was 10 µmol/L (Fig. 3C)
, whereas 50% inhibition of Stat1 under the same conditions could not be achieved with concentrations of T40214 up to142 µmol/L. The basis for this selectivity was explained, in part, by the analysis of histograms, which were generated by docking the structure of T40214 onto the structures of Stat3ß and Stat1 homodimers 2,000 times without setting any binding site restrictions (26)
. These studies demonstrated that interactions of GQ-ODN with Stat3 were concentrated within the SH2 and DNA-binding domains; in contrast, interactions of GQ-ODN with Stat1 were distributed more broadly over the whole structure of Stat1.
Intravenous administration of GQ-ODN plus polyethyleneimine was well tolerated by mice at the dose used in this study. Detailed toxicity studies of T40214 and T40231 are in progress; however, toxicity studies of GQ-ODN T30177, an analogue of T40214 that inhibits HIV-1 integrase, has been performed previously (45) . T30177 did not induce genetic mutations in three assays, the Ames/Salmonella mutagenesis assay, the Chinese hamster ovary/hypoxanthine-guanine phosphoribosyltransferase mammalian cell mutagenesis assay, and the mouse micronucleus assay. Acute toxicity studies in mice revealed an LD50 for T30177 of 1.5g/kg body weight; chronic toxicity studies in mice after multiple doses did not cause delayed mortality or changes in serum chemistry, hematologic parameters, or organ histology until the dose of T30177 reached 600 mg/kg, 120 times the dose (5 mg/kg) used in our studies.
The results of in vivo drug testing revealed that GQ-ODN T40214 dramatically suppressed the growth of xenografts of prostate and breast cancer cells in which Stat3 is constitutively activated. GQ-ODN T40214 markedly decreased the level of phosphorylated Stat3 in tumor xenografts; this decrease was associated with decreased levels of Bcl-2 and Bcl-xL protein. Bcl-2 and Bcl-xL are antiapoptosis proteins that have ion channel activity, which inhibit the release from mitochondria of cytochrome C (46 , 47) . Reduced levels of Bcl-2 and Bcl-xL in tumor xenografts would result in increased release of cytochrome C resulting in activation of the caspase cascade, which includes caspase 3, leading to cell apoptosis. There was a striking increase in caspase 3 cleavage products in the tumors of T40214-treated animals that was accompanied by an increase in tumor cell apoptosis of similar magnitude. Our results indicate that GQ-ODNs such as T40214 block tumor xenograft growth by targeting Stat3 and inhibiting its phosphorylation, which blocks the transcription of the antiapoptosis proteins and triggers the apoptosis of cancer cells. Thus, GQ-ODNs represent a novel and potentially promising class of drug for treatment of metastatic tumors in which Stat3 is constitutively activated as either a single agent or part of a combination regimen.
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Naijie Jing, Department of Medicine, Section of Infectious Diseases, Baylor College of Medicine, Houston, TX. Phone: 713-798-3685; Fax: 713-798-8948; E-mail: njing{at}bcm.tmc.edu
Received 12/24/03. Revised 7/ 7/04. Accepted 7/13/04.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Dagvadorj, R. A. Kirken, B. Leiby, J. Karras, and M. T. Nevalainen Transcription Factor Signal Transducer and Activator of Transcription 5 Promotes Growth of Human Prostate Cancer Cells In vivo Clin. Cancer Res., March 1, 2008; 14(5): 1317 - 1324. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. J. Patel, A. T. Phan, and V. Kuryavyi Human telomere, oncogenic promoter and 5'-UTR G-quadruplexes: diverse higher order DNA and RNA targets for cancer therapeutics Nucleic Acids Res., December 3, 2007; 35(22): 7429 - 7455. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. S. Redell, A. Tsimelzon, S. G. Hilsenbeck, and D. J. Tweardy Conditional overexpression of Stat3{alpha} in differentiating myeloid cells results in neutrophil expansion and induces a distinct, antiapoptotic and pro-oncogenic gene expression pattern J. Leukoc. Biol., October 1, 2007; 82(4): 975 - 985. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Germain and D. A. Frank Targeting the Cytoplasmic and Nuclear Functions of Signal Transducers and Activators of Transcription 3 for Cancer Therapy Clin. Cancer Res., October 1, 2007; 13(19): 5665 - 5669. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Huang Regulation of Metastases by Signal Transducer and Activator of Transcription 3 Signaling Pathway: Clinical Implications Clin. Cancer Res., March 1, 2007; 13(5): 1362 - 1366. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Qi, C.-P. Lin, X. Fu, L. M. Wood, A. A. Liu, Y.-C. Tsai, Y. Chen, C. M. Barbieri, D. S. Pilch, and L. F. Liu G-Quadruplexes Induce Apoptosis in Tumor Cells Cancer Res., December 15, 2006; 66(24): 11808 - 11816. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Shao, X. Xu, N. Jing, and D. J. Tweardy Unique structural determinants for stat3 recruitment and activation by the granulocyte colony-stimulating factor receptor at phosphotyrosine ligands 704 and 744. J. Immunol., March 1, 2006; 176(5): 2933 - 2941. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Jing, Q. Zhu, P. Yuan, Y. Li, L. Mao, and D. J. Tweardy Targeting signal transducer and activator of transcription 3 with G-quartet oligonucleotides: a potential novel therapy for head and neck cancer. Mol. Cancer Ther., February 1, 2006; 5(2): 279 - 286. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Proietti, M. Salatino, C. Rosemblit, R. Carnevale, A. Pecci, A. R. Kornblihtt, A. A. Molinolo, I. Frahm, E. H. Charreau, R. Schillaci, et al. Progestins Induce Transcriptional Activation of Signal Transducer and Activator of Transcription 3 (Stat3) via a Jak- and Src-Dependent Mechanism in Breast Cancer Cells Mol. Cell. Biol., June 15, 2005; 25(12): 4826 - 4840. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
) or with polyethyleneimine alone (
; top) or with T40231/polyethyleneimine complex or with polyethyleneimine alone (bottom). Data plotted are fold increase in tumor volume of each group; bars, ±SE.