Gene (2001) 8, 400–407  2001 Nature Publishing Group All rights reserved 0969-7128/01 $15.00 www.nature.com/gt RESEARCH ARTICLE Intramuscular delivery of IFN-␣ therapy for inhibition of tumor growth located at a distant site

S Li, X Zhang, X Xia, L Zhou, R Breau, J Suen and E Hanna Department of Otolaryngology/Head and Neck Surgery, University of Arkansas School of , 4001 W Capital Avenue, Little Rock, AR 72205, USA

Although electroporation has been shown in recent years to 2 or endostatin gene, also delivered by electro-injection. The be a powerful method for delivering to muscle, no increased therapeutic efficacy was associated with a high via electro-injection has been studied for the level and extended duration of IFN-␣ expression in muscle treatment of tumors. In an immunocompetent tumor-bearing and serum. We also discovered that the high level of IFN-␣ murine model, we have found that delivery of a low dose of expression correlated with increased expression levels of reporter gene DNA (10 ␮g) to muscle via electroporation the antiangiogenic genes IP-10 and Mig in local tumor under specific pulse conditions (two 25-ms pulses of 375 tissue, which may have led to the reduction of blood vessels V/cm) increased the level of by two logs of observed at the local tumor site. Delivery of increasing doses magnitude. Moreover, administration of 10 ␮g of interferon (10–100 ␮g) of IFN-␣ DNA by injection alone did (IFN)-␣ DNA plasmid using these parameters once a week not increase antitumor activity, whereas electroporation for 3 weeks increased the survival time and reduced squam- delivery of increasing doses (10–40 ␮g) of IFN-␣ plasmid ous carcinoma (SCC) growth at a distant site in the DNA did increase the survival time. Our data clearly demon- C3H/HeJ-immunocompetent mouse. IFN-␣ gene therapy strate the potential utility of electroporation for delivery of delivered to muscle using electroporation demonstrated stat- gene therapy to muscle for the treatment of residual or dis- istically significant (P Ͻ 0.05) therapeutic efficacy for treating seminated tumors. Gene Therapy (2001) 8, 400–407. SCC located at a distant site, compared with interleukin (IL)-

Keywords: gene therapy; electroporation; muscle; IFN-␣; antiangiogenesis; IL-12

Introduction Skeletal muscle is an attractive site for somatic because of its large size, good capacity for pro- ␣ Interferon (IFN)- recombinant therapy is the tein synthesis, easy accessibility for intramuscular injec- first cytokine to be used clinically for cancer patients and tion, and ability to take up plasmid after intramuscular has been shown to be effective in the treatment of several administration.13 Intramuscular injection of high-dose, cancers including renal cell carcinomaa, hairy cell leuke- but not low-dose, IFN-␣ gene using a syringe needle sig- mia, malignant , basal cell carcinoma, and mul- nificantly inhibited B16F10 melanoma and glioma 261 1–3 tiple myeloma. However, long-term daily admini- growth in mice and reduced Cloudman melanoma tumor ␣ stration of IFN- protein is required to maintain growth in vivo,14 demonstrating the high level of IFN-␣ ␣ therapeutic efficacy, and some IFN- protein clinical that is required to achieve therapeutic effect. 4 ␣ trials have been disappointing. Ex vivo IFN- gene ther- Electroporation is a highly effective method for apy studies have demonstrated tumor inhibition and an increasing gene expression by creating transient pores in 5–9 elicited tumor-specific immune . However, cell membranes through which can gain entry this approach is patient-specific and the application is into the cell. Although electroporation delivery of ␣ time-consuming. Direct intratumoral injection of IFN- to reporter and therapeutic genes to muscle has been 10 11 treat solid Renca, basal cell carcinoma, and hetero- described by several groups over the past few years,15–17 xenograft prostatic (PC-3) or hepatocellular carcinoma there are no data available to determine the therapeutic 12 (Hep3B) with the use of viral or nonviral delivery sys- efficacy of such an approach in a tumor model. tems has demonstrated effective tumor suppression in IFN-␣ has several subtypes and multiple biological mouse models in vivo, but the intratumoral approach is functions that control cell growth, modulate host immun- not practical for metastatic lesions and microscopic ity, and inhibit .18–21 The mechanisms of anti- residual lesions after surgery. tumoral cell proliferation and immunity modulation by IFN-␣ have been extensively examined, but the mech- anism for antiangiogenesis is less well studied and only Correspondence: S Li a few antiangiogenic genes regulated by IFN-␣ have been Received 7 September 2000; accepted 20 December 2000 identified.18–22 The expression of interferon inducible pro- Electroporation delivery of IFN-␣ gene into muscle for inhibition of tumor growth SLiet al 401 tein-10 (IP-10) and monokine induced by IFN-␥ (Mig) started when the tumor volume was 30 mm3, or twice have been known to be enhanced by IFN-␥ and inhibit the size as in the above experiment. Although the angiogenesis in tumors in vivo,23–28 but it is not clear reduction was not as dramatic as in the previous experi- whether IFN-␣ will have a similar effect in vivo. ment, a 50% reduction in tumor growth was obtained We used squamous cell carcinoma (SCCVII) in an with electroporation delivery (Figure 2c). These data immunocompetent mouse (C3H/HeJ) as a model for this clearly demonstrate that although electro-injection of study. We present data demonstrating that electro-injec- IFN-␣ was not able to eradicate the established SCCVII tion of the IFN-␣ gene is a very powerful method for tumor in vivo, it was able to reduce significantly the rate generating a high and durable level of gene expression. of tumor growth. We also show that this method of delivering the IFN-␣ gene into muscle induces significant anti-SCCVII tumor High-level and extended gene expression in serum after growth by increasing the level of IFN-␣ expression, electro-injection of IFN-␣ gene in muscle which may in turn inhibit tumor vascularization by To determine whether the enhanced therapeutic efficacy inducing IP-10 and Mig expression. of IFN-␣ by electroporation delivery directly correlates with IFN-␣ production, Western blot analysis was perfor- med to analyze the level of IFN-␣ gene expression in both Results injected muscle tissue and serum. As expected, high lev- els of IFN-␣ gene expression were detected in both mus- Increased gene expression in C3H/HeJ mice using cle and serum on day 3 after electro-injection. There was electro-injection in muscle no detectable level of IFN-␣ expression in either muscle In an earlier study, we observed a two- to three-fold tissue or serum following injection alone (Figure 3a). increase in the level of gene expression by electro-injec- The duration of IFN-␣ gene expression after a single tion of plasmid DNA in the muscle of the CD1 mouse, administration of IFN-␣ gene therapy was also deter- using electroporation conditions of two 25-ms pulses at mined. Serum was collected on days 1, 3 and 7. IFN-␣ 375 V/cm field strength.29 To determine whether the protein was detectable in the serum of electro-injected same results could be obtained in a murine tumor model, mice at all three sampling times by Western blot analysis. we determined enzyme activity after delivery of lucifer- However, it was not detected in mice receiving the injec- ase or secreted alkaline phosphatase (SEAP) gene to the tion without electroporation (Figure 3b). Clearly, the muscle of the C3H/HeJ mouse, an immunocompetent therapeutic efficacy of electroporation therapy could be mouse model, using 10 ␮g of DNA for each muscle, with attributed, at least in part, to the resulting high-level and or without electroporation. As shown in Figure 1, the extended expression of IFN-␣. level of luciferase expression in muscle reached 135 ng/mg total protein (Figure 1a), which was more than Inhibition of vascularization in tumor tissue after 100-fold higher than that in the injection-only group. The electroporation delivery of IFN-␣ gene in muscle maximum level of SEAP in serum was 285 ng/ml blood, To determine the mechanism by which electroporation which was reached on day 14 after delivery and was 100- gene therapy with IFN-␣ inhibited tumor growth, a cyto- fold higher than that in the injection-only group (Figure toxic T lymphocyte (CTL) assay was performed. This 1b). As expected, the magnitude of increase in gene showed no increase in tumor cell-lysing activity in either expression in C3H/HeJ mice was similar to that which the electro-injected animals or controls (data not shown). we observed in CD1 mice. The increased level of gene However, immunostaining with anti-CD8+ anti- expression by electro-injection was correlated to the body and anti-endothelial cell marker CD31 increased number of fiber cells transfected with exogen- demonstrated a slightly increased number of infiltrating ous DNA as determined by green fluorescent protein CD8+ T cells (P Ͼ 0.05) and significantly decreased vessel (GFP) expression (Figure 1c). Thus, our delivery tech- density (P Ͻ 0.05) after electroporation delivery of IFN- nique was applicable to this murine tumor model. ␣ gene in muscle (Figure 4A–C). The tumor samples for immunostaining were obtained from mice killed 3 days Inhibition of SCCVII tumor growth via electro-injection of after the second administration, with a week interval IFN-␣ gene in muscle between administrations. Tumor growth in the animal group receiving 10 ␮gof INF-␣ plasmid DNA injected by electroporation was sig- Enhanced expression of IP-10 and Mig after nificantly reduced, compared with that in either the syr- electroporation delivery of IFN-␣ gene in muscle inge needle injection group or the control groups (empty The slight increase in CD8+ T cell infiltration and decrease plasmid with or without electroporation delivery), and in vessel density in tumor observed in our study impli- the survival time was increased (P Ͻ 0.05; Figure 2a, b). cated a functionally active chemokine. It is known that This result has been demonstrated in two independent IP-10 and Mig are regulated by IFN-␥ to chemo-attract experiments. Interestingly, either low-dose (10 ␮g) or immune cells and inhibit angiogenesis.23–28 To determine high-dose (100 ␮g) IFN-␣ DNA plasmid by syringe whether IP-10 and Mig were also up-regulated by IFN-␣ needle injection alone did not enhance any antitumor expression, we analyzed the levels of both Mig and IP- activity (Figure 2a). However, intramuscular electro- 10 expression using Northern blot, quantifying the results injection of high-dose IFN-␣ DNA (40 ␮g) led to longer by phosphorimager scanner. The expression levels of survival time than electro-injection of low-dose IFN-␣ both Mig and IP-10 were increased by 1.8- to three-fold plasmid DNA (10 ␮g) (Figure 2b). after electro-injection of a high dose (40 ␮g) of INF-␣ We also determined whether IFN-␣ gene therapy using (Figure 5). At a low dose (10 ␮g), the increase was about intramuscular electroporation delivery inhibited tumor 1.3- to two-fold. Analysis of the level of vascular endo- growth with delayed treatment, wherein treatment was thelial growth factor (VEGF) in serum showed that there

Gene Therapy Electroporation delivery of IFN-␣ gene into muscle for inhibition of tumor growth SLiet al 402

Figure 1 Analysis of gene expression and efficiency with and without electroporation. (a) Luciferase activity in muscle 24 h after injection with or without electroporation (pulse). A 10 ␮g dose of luciferase DNA plasmid in a volume of 10 ␮l was injected into the left and right hind limb tibialis muscles of five mice (n = 10 muscles for each treatment or control group). A significant difference between electroporation and needle injection alone was detected (P Ͻ 0.05). The bars represent the standard error of the mean (s.e.m.). (b) Level and duration of SEAP in serum on days 3, 7 and 14 after the injection of 10 ␮g SEAP DNA plasmid in a volume of 10 ␮l to the left hind limb tibialis muscles of five mice with or without electroporation (n = 5). A significant difference was detected at all time-points (P Ͻ 0.05). The bars represent s.e.m. (c) Section of left hind tibialis muscle injected with green fluorescent protein (GFP) DNA plasmid with (right panel, C/R) or without electroporation (left panel, C/L). Three muscles were examined for each treatment, and a representative view is shown. The white and light-white cells express GFP.

was no difference with or without electroporation level of gene expression after injection of DNA plasmid delivery of the IFN-␣ DNA plasmid (data not shown). into muscle has limited the development of this nonviral gene therapy approach. Efforts to increase the level of Comparison of different therapeutic genes for SCCVII gene expression by improving the expression systems13,30 tumor treatment with or without electroporation and chemical formulations have been significant,30–32 but We compared the delivery of IFN-␣ gene with the T cell to achieve therapeutic levels, nanogram quantities are proliferation-stimulating IL-2 gene, the known Mig- and needed. Thus far, this has not been achieved with other IP-10-inducing IL-12 gene, and the known antiangiogenic plasmid-based muscle delivery systems. endostatin gene, with or without electroporation, to Electroporation has been shown to be a powerful gene determine their relative efficacy in reducing tumor delivery method for muscle, as well as other tissues,33–35 growth. Using groups of five tumor-bearing mice for and several groups have observed a two- to three-log each gene, as described in Materials and methods, we increase in gene expression level.15–17 We demonstrated were able to demonstrate that IFN-␣ and IL-12 gene ther- for the first time that electroporation delivery of as little apy using intramuscular electroporation delivery had the as 10 ␮g of IFN-␣ DNA plasmid significantly reduced most significant reduction in tumor growth (Figure 6). IL- SCCVII tumor growth in mice and increased survival 2 gene therapy via electro-injection showed some effect, time (P Ͻ 0.05) compared with syringe needle injection but endostatin gene therapy showed no tumor inhibition. only. A high dose (40 ␮g) of IFN-␣ DNA plasmid delivered Discussion via electroporation led to a longer survival time than that of the low dose (10 ␮g) using electroporation delivery Treatment of disseminated malignancies is one of the (Figure 2b). However, a high dose delivered by syringe most difficult tasks facing clinical oncologists. Thus, this needle injection alone did not improve therapeutic effi- remains the focus for optimizing existing cancer cacy (Figure 2a). This was probably due to the fact that and devising new ones. Muscle has been an attractive it was only with electroporation that a high level of IFN- somatic delivery site for gene therapy. However, the low ␣ expression (Figure 3) and, subsequently, a reduced

Gene Therapy Electroporation delivery of IFN-␣ gene into muscle for inhibition of tumor growth SLiet al 403

Figure 3 Western blot analysis of IFN-␣ expression in muscle extract and serum after injection, with or without electroporation, of IFN-␣ DNA plasmid or control empty plasmid. Ten micrograms of DNA plasmid in a volume of 10 ␮l were delivered to the hind limb tibialis muscle with or without electroporation (CP, control empty DNA plasmid). The loading sample was a pool of three individual mice, with 10 ␮g of total protein from muscle and 10 ␮l of serum used for each sample. The light chain of immunoglobin (IgL) serves as an internal control, and the 19 kD protein is IFN-␣. (a) Western blot analysis of the serum and muscle extracts obtained on day 3 after injection, with or without electroporation, of 40 ␮g of IFN-␣ DNA or control empty DNA plasmid. (b) Western blot analysis of serum samples obtained on days 1, 3 and 7 after a single injection, with or without electroporation, of 10 ␮g of IFN-␣ DNA plasmid.

number of vessels in tumor tissue (Figure 4B) were only achieved with electroporation. In our series of experiments, we discovered that Mig and IP-10 genes were up-regulated by IFN-␣ (Figure 4). Mig and IP-10 have been known to inhibit angiogenesis and to chemo-attract monocytes, NK cells, and activated T cells.23–28 Thus, inhibition of SCCVII tumor growth by intramuscular electro-injection of IFN-␣ may be due to both an immune stimulation and an antiangiogenic effect. This theory is supported by the observation that electro- injection of IL-12 gene, known to induce Mig and IP-10, Figure 2 Tumor growth and survival curves after injection, with or with- was able to inhibit SCCVII tumor growth (Figure 6). Our out electroporation, of IFN-␣ gene to hind tibialis muscle of SCCVII hypothesis is also supported by the fact that neither the tumor-bearing mice. Three administrations once a week, depicted by antiangiogenic endostatin gene nor the T cell prolifer- arrows in the figure, were performed, and tumor growth was measured ation-stimulating IL-2 gene alone demonstrated any sig- as described in Materials and methods. CP represents empty control plas- mid DNA. The bars on the curves are s.e.m. The final tumor volumes nificant therapeutic efficacy, with or without electropor- between syringe needle injection and pulse delivery are significantly differ- ation delivery. However, we plan to conduct additional ent (P Ͻ 0.05). All experiments were performed independently in dupli- studies that should further substantiate our hypothesis. cate. (a) Comparison of growth inhibition of SCCVII tumors by IFN-␣ Although it has been shown that intramuscular injec- gene administered intramuscularly with or without electroporation. Treat- tion of IFN-␣ gene using a syringe needle significantly 3 = ment was initiated when the tumor was around 15 mm (n 5). (b) inhibited B16F10 melanoma and glioma 261 growth in Comparison of Kaplan–Meier survival curves of the SCCVII tumor-bear- ␣ mice and reduced Cloudman melanoma tumor growth in ing mice treated with different doses of IFN- with or without electropor- 14 ation. Treatment was initiated when the tumor was approximately 40 vivo, we were unable to obtain similar results using the mm3. Two mice in the group treated with 40 ␮g of IFN-␣ DNA plasmid SCCVII tumor model with injection only of either a high via pulse survived more than 45 days (n = 10). (c) Inhibition of tumor or low dose of IFN-␣ DNA using a syringe needle (Figure growth by delayed IFN-␣ treatment administered by intramuscular elec- 2). Likewise, although it has been reported that intramus- troporation. Treatment was initiated when tumor was approximately 30– cular injection of endostatin gene reduced the number of 3 = 40 mm (n 5). Renca metastasis nodules in the lung, as well as primary

Gene Therapy Electroporation delivery of IFN-␣ gene into muscle for inhibition of tumor growth SLiet al 404

Figure 4 Inhibition of tumor vascularization and CD8+ T cell infiltration by intramuscular electroporation delivery of IFN-␣ gene. The SCCVII tumor samples were obtained 3 days after the second administration of 10 ␮g of IFN-␣ or empty DNA plasmid. Errors are the s.e.m. Different alphabetical letters indicate a significant difference between treatments (P Ͻ 0.05). (a) Reduction of tumor vascularization. SCCVII tumor sections were stained with an antibody reactive to CD31 and visualized under ×10 magnification. Representative tumor vasculature from control (bottom section, empty plasmid DNA) and IFN-␣ gene (top section) by intramuscular electroporation delivery are shown. The arrows point to the vessel stained by CD31 antibody. (b) Vessel density was quantified by counting vessels from eight fields in each tumor section, with three tumors used in each group. (c) Infiltrating CD8+ T cells were stained with an antibody reactive to CD8 and counted from eight fields. Three tumors were used in each group.

Figure 6 Comparison of different therapeutic genes for SCCVII tumor Figure 5 Determination of Mig and IP-10 expression with Northern blot treatment. Each gene (10 ␮g in a volume of 10 ␮l) was delivered to the analysis in tumor tissue treated with IFN-␣ gene therapy with or without hind limb tibialis muscle of five mice with or without electroporation. electroporation. The tumor samples were harvested 3 days after the third Administrations were delivered once a week for 3 weeks. Tumor volume delivery of 10 ␮gor40␮g of IFN-␣ DNA plasmid to muscle (CP, control was measured on day 24 after inoculation of tumor cells. *Significant empty plasmid). The values at the bottom demonstrate the relative level difference from the injection (P Ͻ 0.05). of expression.

ME, USA). The C3H/HeJ strain was used because this tumor growth,36 our data using the SCCVII tumor model strain can host SCCVII tumors. Animals were maintained suggested that intramuscular delivery of endostatin gene under NIH guidelines at 37°C, 12-h day/night cycle, and with or without electroporation could not slow the tumor 80% relative humidity. growth rate. Our findings might differ from those reported in the literature because the SCCVII tumor is Gene construct and plasmid manufacture more aggressive than the other tumor models tested. The reporter genes, luciferase and alkaline phosphatase, In conclusion, electroporation delivery of gene to mus- driven by the CMV () promoter, have cle is simple, economic, directly targeted, and safe. We previously been described.30 The therapeutic gene con- anticipate that the therapeutic efficacy will be significant structs, which express endostatin, IL-2, IL-12, and IFN-␣, when the appropriate gene is chosen. Thus, this will be were provided by Valentis (The Woodlands, TX, USA), the most attractive strategy for nonviral gene delivery. and the structures of the constructs are depicted in pre- vious publications.10,36,37 The control construct is the Materials and methods result of a deletion of IFN-␣ cDNA. All plasmids were manufactured with the Qiagen Giga Experimental animals Endo-Free Prep Kit (Valencia, CA, USA). Residual Female 6- to 8-week-old C3H/HeJ mice weighing 18–20 were removed from the plasmid by against ster- g were purchased from Jackson Laboratories (Bar Harbor, ile water (USP) with a Millipore dialysis tube (Bedford,

Gene Therapy Electroporation delivery of IFN-␣ gene into muscle for inhibition of tumor growth SLiet al 405 MA, USA). The purity of the plasmid was confirmed ized for another 3 min. The suspension was centrifuged using 1% agarose gel . DNA concentration at 14000 r.p.m. for 15 min, and the supernatant was was measured by absorption at 260 nm. The percentage assayed for total protein and luciferase activity, as of supercoiled DNA to total DNA and the OD 260/280 described in a previous publication.30 Total protein was ratios of these plasmid preparations ranged from 80 to determined with the BCA protein assay kit (Pierce, Rock- 90% and 1.5 to 1.7, respectively. ford, IL, USA). Luciferase was determined with the luciferase assay kit from Promega. SCCVII tumor generation in mice and tumor growth monitoring Western blot analysis SCCVII is a spontaneously arising murine squamous cell Muscle extract and serum were used for Western blot carcinoma that has been well characterized in C3H/HeJ analysis to determine IFN-␣ gene expression. Briefly, 10 mice.38,39 We obtained this cell line from Dr Candice John- ␮g of total protein from muscle extract or 10 ␮l of serum son at the University of Pittsburgh (Pittsburgh, PA, USA). was subjected to SDS-PAGE in a gradient gel, and the The cells were maintained in AIM-V growth medium protein was transferred to a nitrocellulose membrane with 10% fetal bovine serum (Gibco, Grand Island, NY, (Millipore). Immunoblotting was performed with a pri- USA). Mice were subcutaneously inoculated with 1 × 105 mary anti-IFN-␣ polyclonal antibody (PBL Biomedical SCCVII cells in a volume of 30 ␮l. When tumors reached Laboratories, New Brunswick, NJ, USA) and anti-rabbit a measurable size (2–3 mm in diamater), mice were ran- secondary antibody conjugated with horseradish peroxi- domly divided into groups of five each. Mice bearing dase (HRP). The HRP signal was generated by incubation tumors that were 30% greater or smaller than the average with a chemiluminescence detection kit (Amersham, were excluded from the experimental group. Alameda, CA, USA) and exposure on X-ray film. Seven to 9 days after tumor inoculation, when tumor volume was about 15–30 mm3, we delivered the IFN-␣ Immunostaining analysis gene or control plasmid to the left hind tibialis muscle, Fifty cubic millimeters of tumor tissue were used for with or without electroporation. Three administrations frozen-block preparation. The mold in which the tissue were performed once a week, and a different hind limb was arranged in the frozen embedding medium (Tissue- muscle was used for each administration, proceeding Tek; Sakura Finetek, Torrance, CA, USA) was plunged from the left hind tibialis to the right hind tibialis and into 2-methylbutane prechilled in liquid nitrogen, until then back to the left hind tibialis. the block almost solidified. The frozen tissue blocks were Tumor growth was monitored twice a week with a cal- stored at −70°C until sectioning. A 5-␮m section was cut iper as soon as the treatment started. Tumor volume was and fixed with cold acetone (−20°C) for 2 min. The sec- calculated with the formula: V = ␲(Dd2/4),40 where V is tions were rinsed with phosphate-buffered saline (PBS) volume, D is maximum tumor diameter, and d is diam- to remove frozen mounting media and blocked with 5% eter at 90° to D. At the completion of the study, mice goat serum. The protocol was provided by PharmaGen were killed with carbon dioxide narcosis. (San Diego, NJ, USA). The primary applied to the sections were anti-CD31 (1:200) and anti-CD8+ T cell Electroporation delivery of DNA plasmid antibody (1:50), which were removed after incubation for Purified plasmids were formulated in 150 mm sodium 1 h. After rinsing of the slides three times, the diluted chloride, as described previously.30 Animals were anes- biotinylated secondary antibody (1:400 dilution for anti- thetized by intraperitoneal administration of a mixture of CD31 staining and 1:200 for anti-CD8 staining) was ketamine (42.8 mg/ml), xylazine (8.6 mg/ml), and acep- applied and incubated for 30 min. Streptavidin/HRP was romazine (1.4 mg/ml) at a dose of 1.8–2.0 ml/kg. A 10 then applied after three rinses with PBS and incubated ␮l formulation containing 10 ␮g DNA was injected intra- for another 30 min. After rinsing of the slides, DAB was muscularly into the tibialis muscle of each animal with applied for 5–10 min, followed by three rinses in water. a 28.5-gauge needle. Immediately after injection, animals All the antibody and immunostaining reagents were pur- undergoing electroporation had two heads of a caliper chased from PharmaGen. Hematoxylin and bluing sol- electrode positioned at two sides of the leg with contact utions (Fisher Scientific, Pittsburgh, PA, USA) were used of the skin. The pulse was applied to the leg through the for counterstaining the slides. electrode with a power supply device (BTS EC830; Gene- tronics, San Diego, CA, USA). The parameters were two Probe isolation by reverse transcriptase polymerase 25-ms pulses of 375 V/cm.29 chain reaction (RT-PCR) and Northern blot analysis of Mig and IP-10 expression Bleeding and SEAP assay Three days after the third administration of gene therapy, Mice were bled from the retro-orbital plexus. The blood mice were killed and tumors were harvested for total was centrifuged at 4000 r.p.m. for 5 min, and serum was RNA preparation with RNAzol (TelTest, Friendswood, collected for the SEAP assay with the use of serum separ- TX, USA). Residual DNA was removed from RNA by ators (Becton Dickinson, Franklin Lakes, NJ, USA), as incubating the RNA extract with DNase at 37°C for 20 described in our previous publication.30 min. To obtain IP-10, Mig, and GAPDH templates for probes, 5 ␮g of total RNA was subjected to RT with oligo- Muscle protein extraction and luciferase assay dT primer. Five percent of the cDNA from the RT reac- Lyophilized muscles were homogenized with silica beads tion was used to amplify the target gene with Mig-, IP- for 2 min with the use of a Mini-BeadBeater (BioSpec Pro- 10-, or GAPDH-specific primers for 30 cycles at the fol- ducts, Bartlesville, OK, USA). One milliliter of luciferase lowing conditions: 94°C, 30 s; 65°C, 30 s; 72°C, 45 s. The cell lysis buffer (Promega, Madison, WI, USA) was added PCR product was isolated from agarose gel. The North- to the powdered muscle, and the samples were homogen- ern blot analysis was processed, as described pre-

Gene Therapy Electroporation delivery of IFN-␣ gene into muscle for inhibition of tumor growth SLiet al 406 viously,41 with the probes of Mig, IP-10, and actin. Each 11 Hottiger MO et al. -mediated gene transfer into human probe was randomly prime-labeled (DECA Primer II basal cell carcinoma. Gene Therapy 1999; 6: 1929–1935. DNA Labeling Kit; Ambion, Austin, TX, USA) with 32P- 12 Ahmed CM et al. In vivo tumor suppression by adenovirus- dCTP (3000 Ci/mmol; Amersham Pharmacia Biotech, mediated interferon alpha-2b gene delivery. Hum Gene Ther Piscataway, NJ, USA). Approximately 10 ␮g of total RNA 1999; 10: 77–84. 13 Wolff JA et al. Direct gene transfer into mouse muscle in vivo. from tumor tissues was subjected to 1% agarose-formal- 1990; 247: 1465–1468. dehyde gel electrophoresis at 100 V for 2 h. The RNA 14 Horton HM et al. A gene therapy for cancer using intramuscular was transferred to a positively charged nylon membrane injection of plasmid DNA encoding interferon-␣. Proc Natl Acad (Boehringer Mannheim, Indianapolis, IN, USA) and pre- Sci USA 1999; 96: 1553–1558. hybridized and hybridized at 42°C in a hybridization 15 Mathiesen I. Electropermeabilization of skeletal muscle buffer purchased from Ambion. Hybridization of the enhances gene transfer in vivo. Gene Therapy 1999; 6: 508–514. same filters with a probe for ␤-actin was used as an 16 Aihara H, Miyazaki J. Gene transfer into muscle by electropor- internal control for RNA loading, and methylene blue ation in vivo. Nat Biotechnol 1998; 16: 867–870. staining was used to determine the relative sizes of 17 Rizzuto G et al. Efficient and regulated erythropoietin pro- hybridized RNA.41 The Northern blot results were quant- duction by naked DNA injection and muscle electroporation. 96 ified by scanning the expression signal intensity with a Proc Natl Acad Sci USA 1999; : 6417–6422. 18 Belardelli F, Gresser E, Maury C, Maunoury MT. Antitumor PhosphorImager analyzer (Model 445 SI; Molecular effects of interferon in mice injected with interferon-sensitive Dynamics, Sunnyvale, CA, USA). and interferon-resistant Friend cells. II. Role of host mechanisms. Int J Cancer 1982; 30: 821–825. Statistical analysis 19 Gresser I. Antitumor effects of interferon. Acta Oncol 1989; 28: Experimental data were analyzed by one-way analysis of 347–353. variance (ANOVA), with measurement of tumor growth 20 Singh RK et al. Interferons ␣ and ␤ downregulate the expression or enzyme activity as the main outcome measured. of basic fibroblast growth factor in human carcinomas. Proc Natl Means of individual treatments were compared with Acad Sci USA 1995; 92: 4562–4566. ␣ Student’s t test when the main effect was significant. 21 Tuting T et al. Interferon- gene therapy for cancer: retroviral Statistical significance was defined as a P Ͻ 0.05. of fibroblasts and particle-mediated transfection of tumor cells are both effective strategies for gene delivery in murine tumor models. Gene Therapy 1997; 4: 1053–1060. Acknowledgements 22 Slaton JW et al. Interferon-alpha-mediated down-regulation of angiogenesis-related genes and therapy of bladder cancer are We would like to thank Valentis, Inc., and Gentronics, dependent on optimization of biological dose and schedule. Clin Inc., for allowing us to use their gene constructs and elec- Cancer Res 1999; 5: 2726–2734. troporator, respectively, for this study. This work was 23 Tannenbaum C et al. The CXC chemokines IP-10 and Mig are partially supported by the Sam Walton Fund and a Uni- necessary for IL-12 mediated regression of the mouse RENCA versity of Arkansas for Medical Sciences pilot grant. tumor. J Immunol 1998; 161: 927–932. 24 Sgadari C et al. Interferon-inducible protein-10 identified as a mediator of tumor necrosis in vivo. Proc Natl Acad Sci USA 1996; References 93: 13791–13796. 25 Angiolillo A et al. Human interferon-inducible protein 10 is a 1 Savaawy HME et al. Enhancement of 5-fluorouracil cytotoxicity J Exp Med 182 on human colon cancer cells by -mediated interferon- potent inhibitor of angiogenesis. 1995; : 155–162. alpha gene transfer. Int J Oncol 1999; 14: 1143–1151. 26 Coughlin C et al. Tumor cell responses to IFN-gamma affect 2 Pfeffer LM et al. Biological properties of recombinant alpha- tumorigenicity and response to IL-12 therapy and antiangiogen- interferons: 40th anniversary of the discovery of interferons. esis. Immunity 1998; 9: 25–34. Cancer Res 1998; 58: 2489–2499. 27 Asselin-Paturel C et al. Transfer of the murine interleukin-12 3 Buechner SA. Intralesional interferon alfa-2b in the treatment gene in vivo by a Semliki Forest vector induces B16 tumor of basal cell carcinoma: immunohistochemical study on cellular regression through inhibition of tumor blood vessel formation immune reaction leading to tumor regresion. J Am Acad Dermatol monitored by Doppler ultrasonography. Gene Therapy 1999; 6: 1999; 24: 731–734. 606–615. 4 Gutterman JU. Cytokine therapeutics: lessons from interferon- 28 Horton MR et al. Hyaluronan fragments synergize with inter- ␣. Proc Natl Acad Sci USA 1994; 91: 1198–1205. feron-gamma to induce the C-X-C chemokines mig and inter- 5 Ogura H et al. Implantation of genetically manipulated fibro- feron-inducible protein-10 in mouse macrophages. J Biol Chem blasts into mice as antitumor ␣-interferon therapy. Cancer Res 1998; 273: 35088–35094. 1990; 50: 5102–5106. 29 MacLaughlin FC et al. Plasmid gene delivery: advantages and 6 Belldegrun A et al. Human renal carcinoma line transfected with limitations. In: Gregoriadis G, McCormack B (eds). Targeting of interleukin and/or interferon-␣ gene(s): implication for live Drugs: Strategies for Gene Constructs and Delivery. IOS, 2000. cancer . J Natl Cancer Inst 1993; 85: 207–216. 30 Li S et al. Increased level and duration of expression in muscle 7 Ferrantini M et al. Interferon alpha-1 gene transfer into meta- by co-expression of a transactivator using plasmid systems. Gene static Friend leukemia cells abrogated tumorigenicity in immu- Therapy 1999; 6: 2055–2011. nocompetent mice: antitumor therapy by means of interferon- 31 Barnhart KM et al. Enhancer and promoter chimeras in plasmids producing cells. Cancer Res 1993; 53: 4607–4615. designed for intramuscular injection: a comparative in vivo and 8 Zhang JF et al. Gene therapy with an adeno-associated virus car- in vitro study. Hum Gene Ther 1998; 9: 2545–2553. rying an interferon gene results in tumor growth suppression 32 Mumper RJ et al. Protective interactive noncondensing (PINC) and egression. Cancer Gene Ther 1996; 3: 31–38. for enhanced plasmid distribution and expression in 9 Hiroishi K, Tuting T, Lotze MT. IFN-alpha-expression tumor rat skeletal muscle. J Control Rel 1998; 52: 10876–10880. cells enhance generation and promoter survival of tumor- 33 Heller R et al. In vivo gene electroinjection and expression in rat specific CTLs. J Immunol 2000; 164: 567–572. liver. FEBS Lett 1996; 389: 225–228. 10 Coleman M et al. Nonviral interferon-alpha gene therapy 34 Nishi T et al. High-efficiency in vivo gene transfer using intrarter- inhibits growth of established tumors by eliciting a systemic ial plasmid DNA injection following in vivo electroporation. immune response. Hum Gene Ther 1998; 9: 2223–2230. Cancer Res 1996; 56: 1050–1055.

Gene Therapy Electroporation delivery of IFN-␣ gene into muscle for inhibition of tumor growth SLiet al 407 35 Mir LM et al. High-efficiency gene transfer into skeletal muscle 39 O’Malley B, Cope K, Johnson C, Schwartz M. A new immuno- mediated by electric pulses. Proc Natl Acad Sci USA 1999; 96: competernt murine model for oral cancer. Arch Otolaryngol Head 4262–4267. Neck Surg 1997; 123: 20–24. 36 Blezinger P et al. Systemic inhibition of tumor growth and tumor 40 Snyderman C, Abbas M, Wagner R, D’Amico F. Inhibition of metastases by intramuscular administration of the endostatin growth of a murine squamous cell carcinoma by a cyclooxygen- gene. Nat Biotechnol 1999; 17: 343–348. ase inhibitior increases leukotriene B4 production. Arch Otola- 37 Blezinger P et al. Intratracheal administration of interleukin 12 ryngol Head Neck Surg 1995; 121: 1017–1020. plasmid–cationic lipid complexes inhibits murine lung meta- 41 Li S, Wilkinson MF. A -like and nuclear-associated stases. Hum Gene Ther 1999; 10: 723–731. mechanism dictate the nonsense codon-mediated TCR mature 38 Grandis J, Chang MJ, Yu WD, Johnson C. Antitumor activity of transcript decay. J Exp Med 1997; 185: 985–992. interleukin-1-alpha and cisplatin in a murine model system. Arch Otolaryngol Head Neck Surg 1999; 121: 197–200.

Gene Therapy