Title: Receptor-mediated Transfection of Murine and Ovine Mammary Glandsin Vivo
Abstract: Transfection of HC-11 murine epithelial mammary cells as well as murine and sheep mammary glands were carried out using insulin-containing constructs that deliver DNA by receptor-mediated endocytosis to receptor-expressing cells. In vivotransfection of mammary gland tissue with the luciferase gene was carried out by introducing the DNA constructs into the mammary ducts of both mice and sheep. The successful transfection of ewe mammary glands was demonstrated by the detection of luciferase activity in mammary gland biopsy material up to a month after a single administration of the construct. To test whether products of expression of transfected genes could be secreted into the milk in this system, the N-terminal secretory signal sequences of bovine β-lactoglobulin or the entire coding sequence of human α-lactalbumin were fused to the N terminus of the luciferase gene. After transfection with the modified luciferases, both murine and sheep milk could be shown to contain luciferase activity, whereas mice, which had been transfected with the nonmodified luciferase gene, did not secrete any activity in the milk. This approach demonstrates for the first time the possibility of gene transfer in vivo into mammary gland epithelial cells using constructs delivering DNA via receptor-mediated endocytosis. Transfection of HC-11 murine epithelial mammary cells as well as murine and sheep mammary glands were carried out using insulin-containing constructs that deliver DNA by receptor-mediated endocytosis to receptor-expressing cells. In vivotransfection of mammary gland tissue with the luciferase gene was carried out by introducing the DNA constructs into the mammary ducts of both mice and sheep. The successful transfection of ewe mammary glands was demonstrated by the detection of luciferase activity in mammary gland biopsy material up to a month after a single administration of the construct. To test whether products of expression of transfected genes could be secreted into the milk in this system, the N-terminal secretory signal sequences of bovine β-lactoglobulin or the entire coding sequence of human α-lactalbumin were fused to the N terminus of the luciferase gene. After transfection with the modified luciferases, both murine and sheep milk could be shown to contain luciferase activity, whereas mice, which had been transfected with the nonmodified luciferase gene, did not secrete any activity in the milk. This approach demonstrates for the first time the possibility of gene transfer in vivo into mammary gland epithelial cells using constructs delivering DNA via receptor-mediated endocytosis. Three major approaches are currently being used to achieve gene delivery into recipient organs: (i) animal virus-based vectors (retroviruses and adenoviruses), (ii) liposomal transfer, and (iii) receptor-mediated endocytosis (for review, see Refs. 1Cooper M.J. Semin. Oncol. 1996; 23: 172-187PubMed Google Scholar, 2Ali M. Lemoine N.R. Ring C.J. Gene Ther. 1994; 1: 367-384PubMed Google Scholar, 3Felgner P.L. Tsai Y.J. Sukhu L. Wheeler C.J. Manthorpe M. Marshall J. Cheng S.H. Ann. N. Y. Acad. Sci. 1995; 772: 126-139Crossref PubMed Scopus (253) Google Scholar, 4Cotten M. Curr. Top. Microbiol. Immunol. 1995; 199: 283-295PubMed Google Scholar, 5Guy J. Drabek D. Antoniou M. Mol. Biotechnol. 1995; 3: 237-248Crossref PubMed Scopus (94) Google Scholar, 6Perales J.C. Ferkol T. Molas M. Hanson R.W. Eur. J. Biochem. 1994; 226: 255-266Crossref PubMed Scopus (158) Google Scholar). In terms of wide application in human gene therapy, the first approach is limited by safety concerns with respect to the use of viral vectors, whereas the second approach has disadvantages associated with the cytotoxicity of Lipofectin for some organs and in vivo instability of liposomes. Since gene delivery via receptor-mediated endocytosis does not make use either of viral vectors or liposomes, it appears to be the method of choice for gene transfer in vivo, although its application in human gene therapy thus far has been largely restricted by problems of high variability and lower efficiency of gene transfer, compared with the first two methods. These problems associated with receptor-mediated gene delivery are attributable to the fact that most parameters determining the efficiency of this process have not been characterized or optimized. In particular, the methods of preparation of DNA-containing complexes, the specificity of the receptor selected for delivery, the specific properties of target tissue, etc. may all be critical. To date very few receptors have been used for DNA delivery via receptor-mediated endocytosis, including asialoglycoprotein (7Wu G.Y. Wu C.H. J. Biol. Chem. 1987; 262: 4429-4432Abstract Full Text PDF PubMed Google Scholar, 8Ferkol T. Lindberg G.L. Chen J. Perales J.C. Crawford D.R. Ratnoff O.D. Hanson R.W. FASEB J. 1993; 7: 1081-1091Crossref PubMed Scopus (50) Google Scholar, 9Plank C. Zatloukal K. Cotten M. Mechtler K. Wagner E. Bioconjugate Chem. 1992; 3: 533-539Crossref PubMed Scopus (313) Google Scholar, 10Perales J.C. Grossmann G.A. Molas M. Liu G. Ferkol T. Harpst J. Oda H. Hanson R.W. J. Biol. Chem. 1997; 272: 7398-7407Abstract Full Text Full Text PDF PubMed Scopus (127) Google Scholar), transferrin (11Wagner E. Zenke M. Cotten M. Beug H. Birnstiel M.L. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 3410-3414Crossref PubMed Scopus (660) Google Scholar), polymeric immunoglobulin (12Ferkol T. Perales J.C. Eckman E. Kaetzel C.S. Hanson R.W. Davis P.B. J. Clin. Invest. 1995; 95: 493-502Crossref PubMed Google Scholar), and several other (13Cotten M. Wagner E. Curr. Opin. Biotechnol. 1993; 4: 705-710Crossref PubMed Scopus (126) Google Scholar, 14Buschle M. Cotten M. Kirlappos H. Mechtler K. Schaffner G. Zauner W. Birnstiel M.L. Wagner E. Hum. Gene Ther. 1995; 6: 753-761Crossref PubMed Scopus (107) Google Scholar) receptors, whereas in vivo studies have been restricted to only two organs, liver and lung (8Ferkol T. Lindberg G.L. Chen J. Perales J.C. Crawford D.R. Ratnoff O.D. Hanson R.W. FASEB J. 1993; 7: 1081-1091Crossref PubMed Scopus (50) Google Scholar, 12Ferkol T. Perales J.C. Eckman E. Kaetzel C.S. Hanson R.W. Davis P.B. J. Clin. Invest. 1995; 95: 493-502Crossref PubMed Google Scholar, 15Wu G.Y. Wu C.H. J. Biol. Chem. 1988; 263: 14621-14624Abstract Full Text PDF PubMed Google Scholar, 16Wu C.H. Wilson J.M. Wu G.Y. J. Biol. Chem. 1989; 264: 16985-16987Abstract Full Text PDF PubMed Google Scholar, 17Wilson J.M. Grossman M. Cabrera J.A. Wu C.H. Wu G.Y. J. Biol. Chem. 1992; 267: 11483-11489Abstract Full Text PDF PubMed Google Scholar, 18Wu G.Y. Wilson J.M. Shalaby F. Grossman M. Shafritz D.A. Wu C.H. J. Biol. Chem. 1991; 266: 14338-14342Abstract Full Text PDF PubMed Google Scholar, 19Wilson J.M. Grossman M. Wu C.H. Chowdhury N.R. Wu G.Y. Chowdhury J.R. J. Biol. Chem. 1992; 267: 963-967Abstract Full Text PDF PubMed Google Scholar, 20Chowdhury N.R. Wu C.H. Wu G.Y. Yerneni P.C. Bommineni V.R. Chowdhury J.R. J. Biol. Chem. 1993; 268: 11265-11271Abstract Full Text PDF PubMed Google Scholar, 21Gao L. Wagner E. Cotten M. Agarwal S. Harris C. Romer M. Miller L. Hu P.C. Curiel D. Hum. Gene Ther. 1993; 4: 17-24Crossref PubMed Scopus (94) Google Scholar, 22Perales J.C. Ferkol T. Beegen H. Ratnoff O.D. Hanson R.W. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 4086-4090Crossref PubMed Scopus (297) Google Scholar, 23Stankovics J. Crane A.M. Andrews E. Wu C.H. Wu G.Y. Ledley F.D. Hum. Gene Ther. 1994; 5: 1095-1104Crossref PubMed Scopus (102) Google Scholar, 24Bommineni V.R. Chowdhury N.R. Wu G.Y. Wu C.H. Franki N. Hays R.M. Chowdhury J.R. J. Biol. Chem. 1994; 269: 25200-25205Abstract Full Text PDF PubMed Google Scholar). A serious problem with respect to many methods of gene delivery is the degradation of DNA in cellular lysosomes before it reaches the nuclei of transfected cells. Wu and co-workers (16Wu C.H. Wilson J.M. Wu G.Y. J. Biol. Chem. 1989; 264: 16985-16987Abstract Full Text PDF PubMed Google Scholar, 17Wilson J.M. Grossman M. Cabrera J.A. Wu C.H. Wu G.Y. J. Biol. Chem. 1992; 267: 11483-11489Abstract Full Text PDF PubMed Google Scholar) showed that the long term persistence and expression of a foreign gene in rat liver can occur when partial hepatectomy is performed after gene delivery, which is suggested to be through partial hepatectomy, disrupting the hepatocyte microtubular network which in turn blocks the progression of DNA-containing endosomes to lysosomes for degradation (24Bommineni V.R. Chowdhury N.R. Wu G.Y. Wu C.H. Franki N. Hays R.M. Chowdhury J.R. J. Biol. Chem. 1994; 269: 25200-25205Abstract Full Text PDF PubMed Google Scholar). To rescue DNA from cellular lysosomes, some authors have suggested the use of DNA-delivering constructs coupled with mutant adenoviruses (25Curiel D.T. Agarwal S. Wagner E. Cotten M. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 8850-8854Crossref PubMed Scopus (413) Google Scholar,26Curiel D.T. Wagner E. Cotten M. Birnstiel M.L. Agarwal S. Li C.M. Loechel S. Hu P.C. Hum. Gene Ther. 1992; 3: 147-154Crossref PubMed Scopus (215) Google Scholar) or the addition of amphipathic peptides (27Plank C. Oberhauser B. Mechtler K. Koch C. Wagner E. J. Biol. Chem. 1994; 269: 12918-12924Abstract Full Text PDF PubMed Google Scholar) or a lysosomotropic agent such as chloroquine (28Cotten M. Langle-Rouault F. Kirlappos H. Wagner E. Mechtler K. Zenke M. Beug H. Birnstiel M.L. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 4033-4037Crossref PubMed Scopus (312) Google Scholar), although the latter may have only limited application because of the known toxicity of the inhibitor. Although impressive data on direct gene transfer to airway lung epithelium employing adenovirus·polylysine·DNA complexes have been obtained (21Gao L. Wagner E. Cotten M. Agarwal S. Harris C. Romer M. Miller L. Hu P.C. Curiel D. Hum. Gene Ther. 1993; 4: 17-24Crossref PubMed Scopus (94) Google Scholar), it is not yet clear how far this approach may be extended to other systems of gene delivery via receptor-mediated gene uptake. Thus, the scarcity of reports appearing to date on in vivo receptor-mediated transfection makes their generalization impossible. We have previously investigated the use of insulin and the insulin receptor for gene delivery and demonstrated in vitro gene transfer by an (Ins-pLys)-DNA 1The abbreviations used are: Ins-pLys, insulin-polylysine conjugate; Ad5, human adenovirus type 5; EDS-76, duck egg drop syndrome virus; LA, α-lactalbumin; pLys, poly-l-lysine; Str-pLys, streptavidin-polylysine conjugate; CMV, cytomegalovirus; luc, luciferase; AU, arbitrary units. 1The abbreviations used are: Ins-pLys, insulin-polylysine conjugate; Ad5, human adenovirus type 5; EDS-76, duck egg drop syndrome virus; LA, α-lactalbumin; pLys, poly-l-lysine; Str-pLys, streptavidin-polylysine conjugate; CMV, cytomegalovirus; luc, luciferase; AU, arbitrary units. construct via the receptors of cultured cells (29Sobolev, A. S. (1988) International Conference on Medical Biochemistry, October 17–21, 1988, USSR Ministry of Health, Moscow.Google Scholar, 30Rosenkranz A.A. Yachmenev S.V. Sobolev A.S. Dokl. Akad. Nauk. SSSR (Russia). 1990; 312: 493-494PubMed Google Scholar, 31Rosenkranz A.A. Yachmenev S.V. Jans D.A. Serebryakova N.V. Murav'ev V.I. Peters R. Sobolev A.S. Exp. Cell Res. 1992; 199: 323-329Crossref PubMed Scopus (98) Google Scholar, 32Sobolev, A. S., Rosenkranz, A. A., and Nikitin, V. A. (1994) Russian Patent 2025487.Google Scholar). 2Patents pending in U. S. A., Canada, Australia, Brazil, Japan, and 17 European countries. 2Patents pending in U. S. A., Canada, Australia, Brazil, Japan, and 17 European countries. The insulin receptor is known to be widely represented in almost all tissues of mammals, which makes it promising for local gene transfection to virtually any organ. In the present report, we demonstrate for the first time in vivo insulin receptor-mediated gene transfer to mammary glands of mice and ewes. We analyze the effect of co-transfection of the mammary gland with avian and attenuated human adenoviruses on the level of gene expression (luciferase) and its persistence for several weeks. In addition, we demonstrate the feasibility of using receptor-mediated transfection of mammary gland tissue to enable the products of foreign genes to be secreted in the milk of transgenic animals. Our results have relevance both to gene therapy approaches and the potential production of proteins of importance in the milk of large mammals such as dairy cattle. Murine mammary epithelial HC-11 cells were grown (33Doppler W. Groner B. Ball R.K. Proc. Natl. Acad. Sci. U. S. A. 1989; 86: 104-108Crossref PubMed Scopus (217) Google Scholar) at 37 °C in RPMI 1640 medium with 10% fetal calf serum and 50 μg/ml gentamycin using a humidified atmosphere with 5% CO2 and supplemented with 5 μg/ml insulin and 10 ng/ml epidermal growth factor (all Sigma). 18–20 g of female BALB/c mice and 3-year-old Romanov ewes were used. Bovine insulin (Sigma) was125I-iodinated with the use of IODO-GEN (Pierce). 90-kDa poly-l-lysine (pLys; Sigma) was labeled with 1-fluoro-2,4-dinitro-[U-14C]benzene (Amersham Pharmacia Biotech). The preparation and purification of the conjugate were carried out as described previously (29Sobolev, A. S. (1988) International Conference on Medical Biochemistry, October 17–21, 1988, USSR Ministry of Health, Moscow.Google Scholar); briefly, bovine [125I]iodoinsulin conjugated with citraconic anhydride (Serva) to protect its terminal amino groups was covalently linked to [14C]pLys with the use of the bifunctional cross-linking reagent, N-succinimidyl-3-(2-pyridyldithio) propionate (Sigma) according to Jung et al. (34Jung G. Kohnlein W. Luders G. Biochem. Biophys. Res. Commun. 1981; 101: 599-606Crossref PubMed Scopus (33) Google Scholar). The resultant conjugate was purified on a Sephacryl S-200 (Amersham) column, and the citraconyl groups were then removed. The insulin/pLys molar ratio in different batches of the conjugate was between 5 and 10 insulin molecules per molecule of pLys. About 80% of the insulin reacted with the pLys, with the yield after purification procedure being ∼50%. Streptavidin (Molecular Probes) was linked to 90-kDa pLys with the use of N-succinimidyl-3-(2-pyridyldithio) propionate using the same procedure as that described above without the terminal amino group protection, and the conjugate was purified on a Sephacryl S-300 column. The pLys/streptavidin molar ratio was between 1 and 2 pLys molecules per molecule of streptavidin. The initial plasmid was pRSVL, kindly provided by Dr. D. Helinski (35de Wet J.R. Wood K.W. Deluca M. Helinski D.R. Subramani S. Mol. Cell. Biol. 1987; 7: 725-737Crossref PubMed Scopus (2474) Google Scholar). In this vector, the luciferase gene is under control of the long terminal repeat of Rous sarcoma virus. To prepare plasmid pRSV-spBLG-luc and all other constructions, site-directed mutagenesis was used (36Drutsa V.L. Kaberdin V.R. Koroleva O.N. Shilov I.A. Bioorgan. Khim. 1991; 17: 1487-1493PubMed Google Scholar). Briefly, plasmids were digested with a restriction endonuclease, cutting at a unique site in the vicinity of the site selected for mutation and then treated with T4 DNA polymerase in the absence of dNTP to digest one strand of each end of the cleaved DNA. A mutagenic primer and two adaptor primers were then annealed to the single-stranded tails thus formed. The adaptors were sequences complementary to the opposite strands of DNA, which were needed to restore the sticky ends at the initial site of plasmid cleavage. The adaptor complementary to the strand to be mutagenized remained unphosphorylated at the 5′-end. The subsequent addition of T4 DNA ligase and T4 DNA polymerase in the presence of dNTP restored the initial site of plasmid cleavage, filled the gaps, and closed the mutant strand into a circle. The wild type strand remained open because of the nonphosphorylated adaptor. Finally, the second mutant strand was formed by nick translation using Escherichia coli DNA polymerase. After transformation, the yield of mutants was normally about 100%. To construct the plasmid with a signal peptide, the HindIII site of pRSVL was used with the adaptors 5′-GTGTGCACCTCCA-3′ and 5′-GTACCGGAATGCTA-3′ for the upper and bottom DNA strands, respectively. The following mutagenic oligonucleotide was used to prepare pRSV-spBLG-luc: 5′-CATCCTCTAG AGGATAGAAT GGCGCCGGGC CTTTCTTTAT GTTTTTGGCG TCTTCGACGA TGATGGCC-3′. To construct the plasmids of the pCMV series,HindIII-SmaI fragments of plasmid pRSVL (comprising the whole luciferase gene together with its 5′- and 3′-untranslated regions) were subcloned into the HindIII and EcoRV sites of plasmid pCRTM3 (Invitrogen). This vector permits expression of a protein from the strong cytomegalovirus (CMV) promoter. This generated the pCMV-luc plasmid. To generate the construct encoding the LA luciferase fusion protein, an NheI restriction site was introduced immediately upstream of the luciferase gene in pCMV-luc using the mutagenic primer 5′-GTTTTTGGCGT CTTCCGCTAG CATGGTTTAC CAACA-3′ by the same method of mutagenesis (36Drutsa V.L. Kaberdin V.R. Koroleva O.N. Shilov I.A. Bioorgan. Khim. 1991; 17: 1487-1493PubMed Google Scholar), resulting in an NheI site containing pCMV-luc vector. In parallel, the human LA gene was synthesized by polymerase chain reaction using human genomic DNA as a template. The primers for polymerase chain reaction contained BglII and NheI restriction sites for subsequent subcloning of the polymerase chain reaction product. The human LA gene thus obtained was then cloned into pGEM7Zf(+) vector (Promega), and the sequence was confirmed. Finally, the LA gene was excised using HindIII and NheI restriction sites and inserted into the modified NheI site containing pCMV-luc vector to generate plasmid pCMV-LA-luc. Duck egg drop syndrome (EDS-76) virus was grown as described (37Rosenkranz A.A. Antonenko Y.N. Smirnova O.A. Yurov G.K. Naroditsky B.S. Sobolev A.S. Biochem. Biophys. Res. Commun. 1997; 236: 750-753Crossref PubMed Scopus (11) Google Scholar). Chicken embryo lethal orphan virus was grown in 9-day-old chicken embryos; after a 3-day incubation, virus-containing allantois fluid was harvested (37Rosenkranz A.A. Antonenko Y.N. Smirnova O.A. Yurov G.K. Naroditsky B.S. Sobolev A.S. Biochem. Biophys. Res. Commun. 1997; 236: 750-753Crossref PubMed Scopus (11) Google Scholar). Human adenovirus type 5 (Ad5) dl312 strain was grown in the 293 cell line. After incubation, infected cells were harvested, centrifuged, homogenized, mixed with an equal volume of Freon 113, and then recentrifuged. The aqueous phase was used for virus purification. Virus-containing material was centrifuged through a discontinuous CsCl gradient. The viral bands were collected and centrifuged through a preformed linear gradient of CsCl. The viral bands were again collected and stored at 4 °C. Virion concentration was determined by spectrophotometric analysis (38Chardonnet Y. Dales S. Virology. 1970; 40: 462-477Crossref PubMed Scopus (126) Google Scholar). Virions (5 × 1011 virions/ml in 1.36 g/ml CsCl solution containing 150 mm NaCl, 25 mm HEPES, pH 7.8) were biotinylated (3 h at room temperature) with 7 μm biotinamidocaproate N-hydroxysuccinamide ester (Sigma) followed by dialysis against 25 mm HEPES, pH 7.5, containing 0.15 m NaCl. Complex formation of Ins-pLys conjugate with plasmid DNA was carried out as described previously (31Rosenkranz A.A. Yachmenev S.V. Jans D.A. Serebryakova N.V. Murav'ev V.I. Peters R. Sobolev A.S. Exp. Cell Res. 1992; 199: 323-329Crossref PubMed Scopus (98) Google Scholar). Briefly, Ins-pLys conjugate was added dropwise to the plasmid solution in 0.25 mm EDTA, pH 8.0, with mixing. Formation of virus-containing complexes was accomplished at pH 7.5 in 25 mm HEPES buffer with 150 mm NaCl and 0.25 mm EDTA by the sequential addition of Str-pLys, plasmid, and Ins-pLys to biotinylated viruses. Internalization of fluorescein isothiocyanate-labeled (31Rosenkranz A.A. Yachmenev S.V. Jans D.A. Serebryakova N.V. Murav'ev V.I. Peters R. Sobolev A.S. Exp. Cell Res. 1992; 199: 323-329Crossref PubMed Scopus (98) Google Scholar) Ins-pLys plasmid construct was assessed by video-intensified microscopy using an AT200 cooled CCD camera (Photometrics) and Axioplan microscope (Zeiss). HC-11 cells were plated 2 days before an experiment. After washing with fresh RPMI 1640 medium, transfecting complexes (1–12.5 nm DNA) in RPMI 1640 medium with 2 mg/ml bovine serum albumin buffered with HEPES, pH 7.5, were added to the cells and incubated for 3 or 18 h at 37 °C. A 5-fold excess of RPMI 1640 medium supplemented with 5% fetal calf serum and bicarbonate was then added, and the cells were placed in a CO2 incubator for 2 days unless otherwise indicated. A special device to introduce material into the milk duct of a mammary gland was constructed (32Sobolev, A. S., Rosenkranz, A. A., and Nikitin, V. A. (1994) Russian Patent 2025487.Google Scholar).2 Briefly, the device consisted of a body with a nipple-shaped hollow, a channel for delivery of negative pressure, a special sphincter dilator with an axial channel in the bottom of the hollow; it also had a special cannula, with a cavity to measure volumes that passed through the axial channel (32Sobolev, A. S., Rosenkranz, A. A., and Nikitin, V. A. (1994) Russian Patent 2025487.Google Scholar).2 25–75 μl and 75 ml of transfecting complexes in Hanks' solution containing 50 μg/ml gentamycin were introduced into the milk ducts per murine mammary gland and sheep mammary gland, respectively. Lactation of ewes was induced according to a 2-week regime of hormonal induction (39Fowler P. Knight C. Foster M. J. Dairy Res. 1991; 58: 151-157Crossref PubMed Scopus (15) Google Scholar). Luciferase activity in cultured cells was assayed in the presence of CoA (40Simpson W.J. Hammond R.M. J. Biolumin. Chemilumin. 1991; 6: 97-106Crossref PubMed Scopus (32) Google Scholar). Cells were washed with 100 mmpotassium phosphate buffer, pH 7.8, containing 1 mmdithiothreitol and 2 mm EDTA, disrupted using 1% Triton X-100 (Serva) in the same buffer, and centrifuged. 50 μl of the resultant supernatant were mixed with 350 μl of 25 mmglycylglycine buffer, pH 7.8, containing 15 mmMgSO4, 2 mm ATP, 10 mmdithiothreitol, then 20 μl of 1 mm CoA and 100 μl of 1 mm luciferin were added, and the luminescence was measured using an LKB 1250 luminometer. A standard of 1 ng of luciferase (Sigma) corresponds to 1250 AU. Mammary glands were washed with Hanks' solution, homogenized in 100 mm potassium phosphate buffer, pH 7.8, with 1 mm EGTA, 3 mmMg(CH3COO)2, 0.1% bovine serum albumin, 1 mm dithiothreitol, 1% Triton X-100, 20 μg/ml aprotinin (Sigma), and centrifuged. The luciferase activity was measured in the obtained supernatant as above. Biopsy material was frozen and kept in a liquid nitrogen until use. Milk proteins (10 μl of milk/lane) were resolved by sodium dodecyl sulfate polyacrylamide gel (12.5%) electrophoresis and transferred to nitrocellulose membranes. Membranes were incubated at 25 °C for 60 min in Tris-buffered saline containing 0.05% Tween 20 and 3% bovine serum albumin and then for 60 min with a primary antibody. We used affinity-purified rabbit anti-luc obtained from Promega diluted 1:5000. After being washed with Tris-buffered Tween saline, membranes were incubated for 60 min with anti-rabbit IgG alkaline phosphatase conjugate (Sigma), and the immunoblots were finally developed using 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium (Sigma) as substrate. We had previously used insulin-containing conjugates to target foreign genetic materials (29Sobolev, A. S. (1988) International Conference on Medical Biochemistry, October 17–21, 1988, USSR Ministry of Health, Moscow.Google Scholar, 30Rosenkranz A.A. Yachmenev S.V. Sobolev A.S. Dokl. Akad. Nauk. SSSR (Russia). 1990; 312: 493-494PubMed Google Scholar, 31Rosenkranz A.A. Yachmenev S.V. Jans D.A. Serebryakova N.V. Murav'ev V.I. Peters R. Sobolev A.S. Exp. Cell Res. 1992; 199: 323-329Crossref PubMed Scopus (98) Google Scholar, 32Sobolev, A. S., Rosenkranz, A. A., and Nikitin, V. A. (1994) Russian Patent 2025487.Google Scholar)2 and photosensitizers (41Akhlynina T.V. Rosenkranz A.A. Jans D.A. Gulak P.V. Serebryakova N.V. Sobolev A.S. Photochem. Photobiol. 1993; 58 (45–48): 1Crossref PubMed Scopus (32) Google Scholar, 42Akhlynina T.V. Rosenkranz A.A. Jans D.A. Sobolev A.S. Cancer Res. 1995; 55: 1014-1019PubMed Google Scholar, 43Akhlynina T.V. Jans D.A. Rosenkranz A.A. Statsyuk N.V. Balashova I.Y. Toth G. Pavo I. Rubin A.B. Sobolev A.S. J. Biol. Chem. 1997; 272: 20328-20331Abstract Full Text Full Text PDF PubMed Scopus (120) Google Scholar) into hepatoma cells and, using confocal laser scanning microscopy and video-intensified microscopy, had been able to directly visualize that the constructs could be endocytosed. Murine mammary epithelial HC-11 cells possess insulin receptors (data not shown) and are thus able to bind and internalize insulin molecules. The fluorescein isothiocyanate-labeled Ins-pLys plasmid construct was incubated with HC-11 cells for 3 h, and free excess insulin inhibited binding and internalization of the construct (Fig.1 a–d). Video-intensified microscopy of a single cell with higher magnification reveals that the construct was localized in the cytoplasm near the cell nuclei (Fig.1 e). The image was processed using a gradient Sobel image filter, which resulted in the direction-independent enhancement of brightness gradients in the image (44Pratt W.K. Digital Image Processing. John Wiley & Sons, Inc., Chichester, UK1978: 492-535Google Scholar). This processing revealed that the construct appeared to be localized in vesicular and/or tubular bodies, probably endocytotic compartments (Fig. 1 f). The (Ins-pLys)-pRSVL construct (12.5 nmplasmid DNA and a lysine/nucleotide ratio of 0.4) was used to transfect HC-11 cells, with luciferase activity readily detectable in cell lysates (4.9 ± 0.7 AU/106 cells). Excess free insulin almost completely inhibited the transfection (less than 0.5 AU). Endocytosed DNA is known to face many obstacles before it reaches the nucleus, one of which is degradation in endosomes/lysosomes. Adenoviruses are able to exit from endosomes along with other components of these vesicles, and this property of adenoviruses has been employed successfully to enhance the transformation efficiency of the transfecting constructs (25Curiel D.T. Agarwal S. Wagner E. Cotten M. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 8850-8854Crossref PubMed Scopus (413) Google Scholar, 26Curiel D.T. Wagner E. Cotten M. Birnstiel M.L. Agarwal S. Li C.M. Loechel S. Hu P.C. Hum. Gene Ther. 1992; 3: 147-154Crossref PubMed Scopus (215) Google Scholar). We decided to test whether the replication-deficient human adenovirus, serotype 5, strain dl312 (Ad5) adenoviruses could increase transfection of mammary gland epithelial cells using our insulin-containing constructs. A significant increase in cellular luciferase activity was obtained when the virus was coupled directly to pLys through a streptavidin-biotin linkage; the activity was dependent on the Str-pLys/virion ratio in the construct (Fig.2 A). Free Ad5 was less efficient when co-incubated with (Ins-pLys)-pRSVL construct; for example, the addition of 3.3 × 1010 virions/6-cm dish 60-fold increased transfection of HC-11 cells by 5.1 nmIns-pLys construct. Greater addition of Ad5 led to cytotoxicity. Excess free insulin added together with the transfecting construct significantly decreased reporter gene expression, again indicating that transfection in this system was mainly an insulin-receptor-mediated process (Fig. 2 B). Avian adenoviruses (chicken embryo lethal orphan virus or EDS-76) lacking determinants for mammalian cells were also tested and found to be able to enhance HC-11 cell transfection provided they were linked to the transfecting constructs. Free avian viruses, in contrast to human Ad5, were unable to enhance transfection (data not shown). The (Ins-pLys)-pRSVL construct was infused into the milk ducts of mice. On days 1, 3, 6, and 9 the mice were sacrificed, their transfected and intact mammary glands and livers were homogenized, and luciferase activity was measured. Luciferase activity could be detected in extracts prepared from the transfected mammary glands over all 9 days, whereas no activity could be detected in the intact glands (Fig.3). We also tested mammary gland transfection using adenovirus-containing constructs. The (biotinylated Ad5-Str-pLys)-pRSVL-(Ins-pLys) (2.7 nm) construct was injected into the mouse milk ducts, and luciferase activity was measured in gland homogenates 2 days later, the activity being about 10 times higher than that after injection of the construct in the absence of viruses (179 ± 92 AU,n = 10 as compared with 18.6 ± 10.2 AUn = 9 of luciferase activity/g of tissue without adenoviruses). Increasing the concentration of DNA-delivering complex up to 30 nm led to a more than 10-fold increase in luciferase activity in tissue homogenates for (Ins-pLys)-DNA construct. In this case adenoviruses did not enhance transfection above that in their absence (Fig. 4), implying that increasing the amount of transfecting construct can enable a significant amount of construct to enter the endosomolytic pathway. Increasing the lysine/nucleotide ratio up to about 2 in the DNA-transfecting constructs led to a significant increase in luciferase activity in the transfected HC-11 cells; further increase reduced the activity (Fig. 5). In contrast to these in vitro results, we could not detect luciferase activity in homogenates of murine mammary glands after infusion into the ducts of the construct with a lysine/nucleotide ratio of 2. Sheep mammary glands were transformed with the (biotinylated EDS-76-Str-pLys)-pRSVL-(Ins-pLys) construct using the same approach. On days