Search

Hepatocyte growth factor (HGF) has been shown to induce angiogenesis in vivo and has potential as a candidate gene for 'therapeutic angiogenesis'. In vivo, two isoforms of HGF, HGF₇₂₃ and HGF₇₂₈, consisting of 723 and 728 amino acids, are generated through alternative splicing between exons 4 and 5, but the biological effects of their coexpression have not yet been elucidated. In this study, we generated a series of genomic-complementary DNA (cDNA) hybrids of the HGF gene by inserting various truncated intron 4 into the junction of exons 4 and 5 of HGF cDNA and analyzed the biological activities of these hybrid constructs. We showed that: (1) the hybrid called HGF-X7, which contained 1502 base pairs of intron 4, could drive a higher level of HGF expression than other hybrid constructs and cDNAs of each isoform alone; (2) the pCK vector was most efficient for the gene expression of HGF-X7; (3) coexpression of both isoforms of HGF could more efficiently induce the migration of human umbilical vein endothelial cell (HUVEC) and of the mouse myoblast cell line C₂C₁₂ myoblasts than a single isoform of HGF and human vascular endothelial growth factor (VEGF)₁₆₅ at a given concentration; (4) intramuscular administration of pCK-HGF-X7 resulted in transient and localized HGF expression in the injected muscle without an increase in the HGF protein levels in other tissues including serum; and (5) intramuscular injection of pCK-HGF-X7 could more efficiently increase the number of angiographically recognizable collateral vessels, as well as improve an intra-arterial Doppler wire-measured blood flow in the rabbit model of hindlimb ischemia when compared with the identical vector encoding VEGF₁₆₅ gene. These results showed that transfer of the genomic-cDNA hybrid of the HGF gene could be used as a potential therapeutic approach to human vascular diseases.

Her-2/neu is a well-characterized tumor-associated antigen, the overexpression of which in human carcinomas correlates with a poor prognosis. Here, we evaluated Her-2/neu-specific humoral and cellular immune responses in immunized monkeys after immunization with nonreplicating adenovirus (AdHM) expressing the extracellular and transmembrane domain of human Her-2/neu (HM) and/or naked DNA vaccine (pHM-hGM-CSF) expressing human granulocyte-macrophage colony-stimulating factor together with HM. Priming of monkeys with AdHM generated Her-2/neu-specific long-lasting antibody production. Furthermore, these Her-2/neu-specific antibodies produced by AdHM immunization, some of which shared epitope specificity with Herceptin, were able to induce antibody-dependent cellular cytotoxicity against Her-2-expressing target cells. Cellular immune responses were elicited in all monkeys immunized with Her-2/neu-expressing vaccine; interferon-gamma was secreted when these splenocytes were restimulated with Her-2/neu-expressing autologous cells, and immunization with AdHM induced Her-2/neu-specific lymphoproliferative responses. Further, immunization with pHM-hGM-CSF before AdHM immunization noticeably enhanced cytotoxic T-lymphocyte activity. In addition, we observed no abnormalities that would indicate that the genetic vaccines had toxic effects in the immunized monkeys. Thus, we can conclude that our genetic vaccines efficiently elicited Her-2/neu-specific humoral and cellular immune responses without causing severe adverse effects in nonhuman primates and that as such they warrant further clinical investigation.

Gene therapy represents a possible alternative to the chronic delivery of recombinant antiangiogenic proteins to cancer patients. We have constructed retroviral and adenoviral vectors that express murine N-terminal fragments of thrombospondin-2 (NfTSP2), a potent endogenous inhibitor of tumor growth and angiogenesis. To test the possibility of anticancer gene therapy using NfTSP2, we tested whether an ex vivo retrovirus-mediated procedure could be used for the treatment of tumors. The treatment of tumor-bearing mice with syngenic immortalized cell lines expressing NfTSP2 led to a tumor volume reduction up to 70% as compared with the controls (P<0.005). In addition, the established tumors were eradicated in 40% of the mice treated with NfTSP2-expressing cells. Furthermore, the intratumoral injection of the NfTSP2-expressing adenoviral vector to the human squamous cell carcinoma in nude mice resulted in a significant reduction of the growth rates and the volumes of the carcinoma (P<0.05). Immunohistochemical staining of the tumors indicated that the total area and the average size of tumor vessels were significantly reduced in the treatment group versus the controls (P<0.05). In conclusion, the present study clearly demonstrates that the viral vector-mediated transfer of the NfTSP2 gene could inhibit the growth of tumors by perturbing tumor-associated angiogenesis.

The interleukin-1 receptor antagonist (IL-1Ra) is an endogenous protein that can prevent the binding of IL-1 to its cell-surface receptors. Among a number of techniques for gene transfer in vivo, the direct injection of naked DNA into muscle is simple, inexpensive and safe. In this study, we evaluated the potential of intramuscular gene therapy with plasmid DNA containing the cDNA for IL-1Ra in the prevention of murine collagen-induced arthritis (CIA). DBA/1 mice were immunized with bovine type II collagen. At 4 weeks after the initial immunization, expression plasmid for IL-1Ra was injected into four selected sites in the thigh and calf muscles of DBA/1 mice. Control mice received the same plasmid, but lacking the IL-1Ra coding sequence. Macroscopic analysis of paws for redness, swelling and deformities showed that the onset of moderate to severe CIA in the paws of mice injected with IL-1Ra DNA was significantly prevented (P<0.05). In addition, both the synovitis and the cartilage erosion in knee joints were dramatically reduced in mice treated with IL-1Ra DNA (P<0.05). The expression of IL-1beta was significantly decreased in the ankle joints of mice treated with IL-1Ra (P<0.01). Interestingly, the levels of IL-1Ra in sera and joints after intramuscular injection of IL-1Ra DNA were significantly lower than when protein had been used in previous reports, suggesting that the therapeutic effect may be achieved by an alternative mechanism(s) rather than by systemic elevation of IL-1Ra. These observations provide the first evidence that direct intramuscular injection of expression plasmid for IL-1Ra may effectively suppress the inflammatory pathology in arthritis.

Tumor necrosis factor (TNF) is a proinflammatory cytokine involved in the pathogenesis of rheumatoid arthritis, and antagonism of TNF may reduce the activity of the disease. Among a number of techniques for gene transfer in vivo, the direct injection of plasmid DNA into muscle is simple, inexpensive, and safe. In this study, we attempted to treat collagen-induced arthritis (CIA) with anti-TNF gene therapy by transferring the plasmid encoding soluble p75 TNF receptor linked to the Fc portion of human IgG1 (sTNFR:Fc) using in vivo electroporation. DBA/1 mice were immunized with bovine type II collagen and boosted with the same antigen. At 2 days after boosting, the plasmid vector containing cDNA for the sTNFR:Fc was injected into one selected site in the gastrocnemius muscle followed by electroporation. Serum levels of sTNFR:Fc reached 2.3 ng/ml on day 5 when gene expression reached its peak. Macroscopic analysis of paws for redness, swelling and deformities showed that the onset of moderate-to-severe CIA in mice treated with sTNFR:Fc was prevented on a significant level compared with the control mice (P<0.05). The beneficial effect of sTNFR:Fc DNA transfer lasted for at least 18 days following treatment. In addition, both the synovitis and the erosion of cartilage in the knee joints were dramatically reduced in mice treated with sTNFR:Fc (P<0.05). The expression of IL-1beta and IL-12 in the paw was also decreased by sTNFR:Fc treatment (P<0.01) while there was little change in the levels of IL-17 and vWF. These data showed that sTNFR:Fc expression plasmid was effective in the prevention of CIA, and in vivo electroporation-mediated gene transfer may provide a new approach to cytokine therapy in autoimmune arthritis.

Although polyethylenimine (PEI) has been widely used as a nonviral vector, there is little mechanistic understanding on PEI-mediated delivery. Here, we studied whether the expression of murine interleukin-2 (mIL-2) plasmids could be improved by complexation with PEI at various N/P ratios, and whether the cellular uptake, nuclear translocation, and retention of plasmids could be affected by the N/P ratios. Compared with the naked mIL-2, PEI/mIL-2 complexes showed at least two orders of magnitude higher expression at Raw264 cells in the N/P ratio-dependent manner. PEI-mediated cellular uptake and nuclear trafficking of plasmids, quantitated by competitive polymerase chain reaction, also depended on the N/P ratios showing the highest cell and nuclear levels of plasmids at 10/1. The higher cellular levels of plasmid DNA after PEI-mediated delivery were also observed in other cell lines. Unlike naked plasmids, PEI/mIL-2 complexes (N/P ratios >/=4/1) showed prolonged cellular and nuclear retention of mIL-2 plasmids. The nuclear translocation and higher cellular level of plasmids given in PEI complexes were similarly observed by fluorescence microscopy. Moreover, PEI/mIL-2 complexes revealed high stability against DNase I, partly explaining the prolonged subcellular retention. These results indicate that the expression of plasmid mIL-2 might be highly enhanced by complexation with PEI and that such increased expression could be attributed by the higher cellular uptake, nuclear translocation and prolonged retention.

Almost all currently available retroviral vectors based on murine leukemia virus (MLV) contain one or more viral coding sequences. Because these sequences are also present in the packaging genome, it has been suggested that homologous recombination may occur between the same nucleotide sequence in the packaging genome and the vector, resulting in the production of replication competent retrovirus (RCR). Up until now, it has been difficult to completely remove viral coding sequences since some were thought to be involved in the optimum function of the retroviral vector. For example, the gag coding sequence present in almost all available retroviral vectors has been believed to be necessary for efficient viral packaging, while the pol coding sequence present in the highly efficient vector MFG has been thought to be involved in achieving the high levels of gene expression. However, we have now developed a series of retroviral vectors that are absent of any retroviral coding sequences but produce even higher levels of gene expression without compromising viral titer. In these vectors, the intron and exon sequences from heterologous cellular or viral genes are present. When compared with the well-known MLV-based vectors, some of these newly developed vectors have been shown to produce significantly higher levels of gene expression for a longer period. In an experimental system that can maximize the production of RCR, our newly constructed vectors produced an absence of RCR. These vectors should prove to be safer than other currently available retroviral vectors containing one or more viral coding sequences.

The internal ribosome entry site (IRES) from the picornavirus family has frequently been used to express multiple genes from a polycistronic message in retroviral vectors. While examining factors affecting levels of gene expression in IRES-containing retroviral vectors, it was found that retroviral vectors expressing the two genes linked by IRES, the reporter gene and the selectable marker neo, produced significantly lower levels of protein than those containing a reporter gene alone. This observation has been made with various cDNA sequences. However, when the neo was replaced with a different cDNA, the level of gene expression was increased, often to the level achieved with a vector expressing a single gene, suggesting that the bacterial neo sequence has a negative effect on expression. Analysis of the steady-state RNA levels isolated from transfected packaging cells showed that the neo-containing retroviral vectors produced significantly lower levels of RNA than those lacking this bacterial sequence indicating that neo interferes with expression of the neighboring gene at the level of RNA. Furthermore, the order of genes in the IRES-neo-containing vectors appeared to be more important than in the vector lacking the neo sequence. Our results suggest that neo has to be used in the retroviral vector with care, especially when a high level gene expression is needed.

We have developed a simple and rapid method for the determination of retroviral titer by G418 selection which is both accurate and reproducible. NIH3T3 cells were transduced with recombinant retroviral vectors harboring the neo gene and selected in the presence of the antibiotic G418 for 3 days. Cells were then trypsinized, harvested, and counted on a hemacytometer. The control NIH3T3 cells were completely dead 3 days after treatment with G418, while cells transduced with retroviral vectors containing the neo gene produced an average of 15 progeny cells per colony. To estimate viral titer, therefore, the total number of trypsinized G418-resistant cells was divided by 15. Retroviral titers measured by this method did not differ significantly from those determined by the conventional method based on counts of G418-resistant colonies 10-14 days after G418 selection. The method worked well with various retroviral constructs and its efficacy was not influenced by cDNA insert. This method not only reduces the amount of work but also shortens the time needed for determining viral titer.

We have analyzed the relative level of gene expression and viral titer from different types of retroviral vectors used for gene therapy, the LTR-based MFG vector and the internal promoter-containing vectors, LNCX, LNSX and LXSN. The CAT gene was used for comparison of retroviral vector gene expression in both transfected and transduced cells, while the neo gene was used to evaluate viral liter. In transfected cells, MFG-CAT expressed higher levels of CAT then the other vectors, LNC-CAT was next, while L-CAT-SN and LNS-CAT produced much lower levels. CAT expression from MFG-CAT was particularly high in the human T lymphoid cell lines CEM-SS and H9. In nonselected transduced cells. CAT expression from MFG was 10- to 50-fold higher than with the other vectors. Similar observations were made with retroviral constructs expressing human EPO and murine GM-CSF. In transient transfection assays, the titer of MFG was at least five-fold higher than the other vectors as determined by Southern analysis and G418 resistance. Analysis of the steady-state RNAs produced after transfection of the packaging cell lines showed that MFG expressed a significantly higher level of genomic RNA, which contains the packaging signal, than the other vectors while still expressing a high level of the subgenomic RNA encoding CAT. The high level of genomic RNA most likely contributes directly to the higher titer of MFG. We also compared viral titers from subcloned PA317 producer lines containing LNC-CAT and MFG-CAT-Neo, and confirmed that the titer of the MFG virus was higher than that of the LNCX. In selected subcloned transduced NIH3T3 cells, average levels of CAT activity were nine-fold higher from MFG-based vector. Our results suggest that there are significant differences in both the titer and the level of gene expression between retroviral vectors which are currently being used in gene therapy clinical trials.