Gene transfer into hepatoma cell lines via the serpin enzyme complex receptor

Abstract

The serpin enzyme complex receptor (SECR) expressed on hepatocytes binds to a conserved sequence in α₁-antitrypsin (α₁-AT) and other serpins. A molecular conjugate consisting of a synthetic peptide (C1315) based on the SECR binding motif of human α₁-AT covalently coupled to poly-L-lysine was used to introduce reporter genes into hepatoma cell lines in culture. This conjugate condensed DNA into spheroidal particles 18-25 nm in diameter. When transfected with the SECR-directed complex containing pGL3, Hep G2 cells that express the receptor, but not Rep G2 cells that do not, expressed a peak luciferase activity of 538,731 ± 144,346 integrated light units/mg protein 4 days after transfection. Free peptide inhibited uptake and expression in a dose-dependent manner. Complexes of DNA condensed with polyly-sine or LC-sulfo-N-succinimidyl-3-(2-pyridyldithio )propionate-substituted polylysine were ineffective. Transfection with a plasmid encoding human factor IX produced expression in Hep G2 (high) and HuH7 cells that express SECR but not Hep G2 (low) cells that lack the receptor. Fluorescein-labeled C1315 peptide labeled 9-31% of Hep G2 (high), 10–14% of HuH7, and 0.6-3.4% ofHep G2 (low) cells, and when the lac Z gene was transfected, only these cells expressed β-galactosidase. SECR-mediated gene transfer gives efficient, specific uptake and high-level expression of three reporter genes, and the system merits further study for gene therapy.

A number of gene transfer systems have been developed in the search for efficient, yet safe, vectors for gene therapy (23). Replication-deficient recombinant viral vectors, as well as liposomes, have been used to deliver genes to a variety of cells (7, 10, 11, 29). However, despite their ability to accomplish gene delivery, these vectors are limited by practical and theoretical constraints (5, 7, 10, 23, 24). A less cytotoxic, more specific alternative to these gene transfer systems is gene delivery via the receptor-mediated endocytosis pathway (25, 30–32). For this approach, a carrier consisting of poly-L-lysine chemically conjugated to a receptor-specific ligand is noncovalently bound to expression plasmids. With careful manipulation of the ionic strength, DNA is condensed into complexes ranging in size from 10 to 25 nm in diameter (8, 9). Cotten et al. (6) reported delivery and expression of genes contained in 48-kilobase (kb) plasmids, demonstrating efficiency of this method in the transfer of large DNA (6). Thus the system affords tissue target specificity and is noninfectious. Furthermore, because receptor-mediated internalization is a ubiquitous physiological process, this form of gene therapy could be applied to a variety of cell types.

Since the initial reports by Wu and Wu (32) of receptor-mediated gene delivery to cells (32), a number of groups have demonstrated efficient gene transfer through targeting the transferrin (30, 31), asialoglyco-protein (25), polymeric immunoglobulin (8), and man-nose (9) receptors in vitro and in vivo. These receptors have in common the ability to internalize large molecules in large quantities and, other than the binding motif of the ligand, have relaxed structural requirements for the remainder of the molecule.

The serine protease inhibitor (serpin) enzyme complex receptor (SECR) was originally described as a cell surface binding site on human hepatoma Hep G2 cells and blood monocytes. It recognizes a sequence on the carboxy-terminal tail of α₁-antitrypsin (α₁-AT) (27). A synthetic peptide (peptide 105Y) based in sequence on amino acids 359–37 4 of α₁-AT bound to Hep G2 cells in a manner both specific and saturable (13, 16, 18, 27, 28). Scatchard analysis estimated a dissociation constant (Kd) of ~40 nM and ~405,000 plasma membrane receptors per cell. The binding site defined in these early studies has been shown to bind α₁-AT itself but only when it is complexed with a serine protease, such as neutrophil elastase, or modified by either metalloelastase or by the collaborative action of active oxygen intermediates and neutrophil elastase (16, 27). Subsequent studies have shown that SECR mediates internalization of its ligands and delivers them to an acidic compartment, either late endosome or lysosome, for degradation (2, 28). It is now known to be expressed on a number of cell types, including mononuclear phagocytes, neutrophils, myeloid cell lines U937 and HL60, human intestinal epithelial cell line Caco-2, mouse fibroblast L cells, rat neuronal cell line PC12, and human glial cell line U373MG (2, 14, 26). On hepatoma cells SECR mediates feedback upregulation of endogenous α₁-AT synthesis (27). On neutrophils (14) it mediates the known chemotactic effect of α₁-AT-elastase complexes (1).

The SECR binding domain in the carboxy-terminal tail of α₁-AT is highly conserved among the serpin family, and, indeed, a number of SECs compete for binding to SECR (16, 27). Compatible sequences are also found in some tachykinins, including substance P and substance K, amyloid-13 peptide, and bombesin (14, 15). Thus, similar to other receptors favorable for gene transfer, it is adapted for binding and internalizing large molecular complexes with relatively low selectivity as long as a pentapeptide binding domain FV(F/Y)L(IIM) is present (13, 15, 16, 18, 26, 27).

Exogenous DNA complexes bearing the pentapeptide binding motif could be targeted to and internalized by the SECR. Its abundance and bulk flow characteristics coupled to the prospect of targeting hepatocytes (the site of many recessive inherited disorders such as α₁AT deficiency) as well as cells primarily affected by Alzheimer's disease (reviewed in Ref. 26), make this receptor system an attractive candidate for receptor-mediated gene delivery. Furthermore, its presence in the brain may provide the potential to transfer therapeutic genes across the blood-brain barrier. In this study we tested the hypothesis that foreign DNA condensed by poly-L-lysine coupled to a synthetic peptide ligand for the SECR (13, 18) can be targeted to and expressed in cells bearing the receptor in vitro.

MATERIALS AND METHODS

Materials

DNA-modifying enzymes, nucleotides, and 5-bromo-4 chloro-3-indolyl-(β-D-galactopyranoside (X-gal) were purchased from Boehringer Mannheim (Indianapolis, IN). Poly-L-lysine was obtained from Sigma Chemical (St. Louis, MO) and lc-sulfo-N-succinimidyl-3-(2-pyridyldithio)propio-nate (LC-sulfo-SPDP) was purchased from Pierce Chemical (Rockford, IL). Luciferase activity was measured with the use of Promega (Madison, WI) assay reagents. Sephadex G₁₀ columns and Bradford protein assay reagents were obtained from Bio-Rad (Richmond, CA). ¹²⁵I used for iodination was purchased from DuPont-New England Nuclear (Boston, MA). Peptide C105Y (CSIPPeVKFNKPFVYLI), and C1315 (CFLEAIPMSIPPEVKFNKPFVFLIIHRD) were synthesized by solid-phase method, purified, and subjected to amino acid composition and sequence analysis as described previously (13).

Cell culture

Two populations of human hepatoma Hep G2 cells were maintained as previously described (28). Hep G2 (high) cells (passage 2) were obtained from ATCC (Rockville, MD). Hep G2 (low) cells (passage 300) were kindly provided by Dr. Lucyndia Marino (Cleveland, OH). These cells were designated high or low on the basis of their ability to bind SECR ligands C105Y and C1315 (see below). HuH7 cells were cultured in RPMI medium. Fresh medium was added every second day. All transfection experiments were performed in medium containing 10% fetal calf serum (FCS). Serum had no effect on complex stability.

Determination of cell surface SECR binding

Peptide C105Y was radioiodinated by a modification of the chloramine T method (12) and purified on a Sephadex G₁₀ column. Specific radioactivity of ¹²⁵I-labeled peptide C105Y ranged between 3,500 and 11,700 disintegrations· min ^–1. ng^–1. Binding studies were conducted on HuH7 cells and two populations ofHep G2 cells as previously described (15). Binding parameters were determined by Scatchard analysis. Assays were performed on all three cell lines [HuH7, Hep G2 (high), and Hep G2 (low)J simultaneously with the same batch of iodinated C105Y peptide so that comparison between cell lines could be made.

Production of cDNA complexes that target the SECR

Two features of this system are critical for successful introduction of genes into the cells: 1) efficient coupling of the DNA-condensing agent, poly-l-lysine, to the peptide used to target the SECR and 2) proper condensation of the DNA by the peptide-based carrier into highly compact complexes suitable for efficient internalization via an endocytic pathway. Peptides covalently linked to poly-L-lysine (average relative mol mass = 22.5 kDa) with the use of the heterobifunctional crosslinking reagent LC-sulfo-SPDP, as previously described (17). Briefly, 77 μl of 20 mM LC-sulfo-SPDP in water were incubated with 3 mg poly-L-lysine (10-fold molar excess of LC-sulfo-SPDP to polylysine) in 0.1 M phosphate-buffered saline (PBS), pH 7.4, at room temperature ( ~22°C) for 30 min. The reaction mixture was then dialyzed exhaustively to remove unreacted LC-sulfo-SPDP and low molecular weight reaction products. A threefold molar excess of modified poly-L-lysine was then added to peptide C1315, and the reaction was allowed to proceed at 22° for 24 h. The conjugate was dialyzed to remove unreacted peptide and low molecular weight reaction products.

Nuclear magnetic resonance spectra

An aliquot (5–10 mg) of the conjugate was exhaustively dialyzed against water, lyophilized from water, and subsequently from D₂O, then resuspended in 0.75 ml of 99.99% D₂O (Aldrich), or 90% dimethyl sulfoxide d6 (DMSOd₆ )-10% D₂O. Proton nuclear magnetic resonance (NMR) spectra were obtained at 600 MHz on a Varian Unity Plus 600 NMR spectrometer using standard proton parameters. Spectra acquisitions typically required between 0.5 and 16 h. Chemical shifts were referenced to the residual water HDO resonance at ~4.8 parts/ million (ppm) or the DMSO_d6 multiplet at 2.5 ppm. Spectra of C1315 peptide and C1315-poly-L-lysine conjugate were obtained in 90% DMSO_d6. Aliquots of dialysis bath water as well as dialysis bag wash were also lyophilized and examined by NMR to verify the absence of contaminants.

Reporter genes and plasmid preparation

Three plasmids coding for three different reporter genes were used. The expression plasmid pGL3 control (Promega) encodes the Photinus pyralis luciferase gene and was inserted into the Escherichia coli pUC19 vector. The plasmids pCMV lac Z II (20) and pFIX (originally obtained from Dr. Earl Davie, University of Washington, Seattle, WA) contained the cytomegalovirus (CMV) promoter ligated to the E. coli (β-galactosidase (lac Z) and the human factor IX (hFIX) genes, respectively. Plasmids were grown and purified as previously described (21). The sizes of the plasmids were as follows: pGL3, 5.6 kb; pCMV lac Z, 10.8 kb; pCMV FIX, 5.4 kb.

Formation of the C1315 peptide-based DNA complexes

The eDNA complexes were formed with the use of general techniques previously described for the galactosylated polylysine ligand (25). The final volume of the solutions was typically 500 μl (0.5–1 μg plasmid DNA/5 μ, containing a mixture of 1:0.45 (wtlwt) DNA-to-peptide-poly-L-lysine conjugate ratio in 0.8–1 M NaCl. Different final concentrations of sodium chloride were due to minor differences among preparations in DNA and poly-l-lysine size and physical state (25). An aliquot of the reaction mixture was examined under the electron microscope (EM) to assess condensation.

Electron microscopy of the condensed DNA complexes

Micrograph grids were prepared as previously described (25). Briefly, immediately after formation of DNA complexes, a drop of a solution (1:10 dilution of complex mixture in water) was added to a 1,000-mesh EM carbon grid, blotted, and stained with 0.04% uranyl acetate. The samples were then shadowed with the use of rotary shadowing and examined using a JEOL-100C EM.

Transfection of hepatoma cells in culture

For pGL3 control, pCMV lac Z II, and pFIX transfection, the HuH7 or Rep G2 cells were washed twice with PBS, pH 7.4, trypsinized with 0.05% trypsin in Dulbecco's minimal medium (DMEM), and plated in six-well plates in 10% serum DMEM with glutamine 2 days before transfection. The cells were allowed to adhere to the plate and become 30% confluent. Cell density was typically 5 × 105 cells per plate at the time of transfection. On the day of transfection, growth medium was changed and the cells were washed with Ca²⁺/Mg²⁺ PBS. Aliquots containing C1315 peptide-poly-l-lysine-DNA complex [0.83, 1.11, or 1.34 pmol pGL3 control, pFIX (0.83 or 1.11 pmol), or pCMV lac Z II (0.83, 1.11, or 1.34 pmol) DNA condensed with 62 (122 for lac Z II), 80 (160), or 97 (194) pmol C1315-polylysine conjugate, respectively] were added to 2 ml of serum-containing medium in individual wells. Controls included: 1) HuH7 or Hep G2 (high) (as determined by radioligand binding assay) cells transfectcd with 1.11 pmol pGL3 control, pFIX, or pCMV lac Z II DNA alone, condensed with 80 (160 for lac Z II) pmol unconjugated, or condensed with LC-sulfo-SPDP-modified polylysine in the presence of80 (160) pmol C1315 peptide and 200 (400) pmol LC-sulfo-SPDP linker; 2) Rep G2 (low) cells transfected with 1.11 pmol pGL3 control or pCMV lac Z II DNA condensed with 80 or 160 pmol C1315-polylysine conjugate, respectively; 3) Hep G2 (high), Hep G2 (low), or HuH7 cells transfected with 1.0 pmol of pGL3 control, pFIX, or pCMV lac Z II DNA by lipofectin; 4) Rep G2 (high), Rep G2 (low), or HuH7 cells transfected with 1.11 pmol of polylysine-condensed DNA by lipofectin. Controls 1 and 2 were designed to test for nonspecific uptake; controls 3 and 4 were designed to confirm that target cells could express the transgene if delivered. After addition of the complex and/or lipofectin, all cells were incubated at 37° for 6 h. Cells were then rinsed with Ca²⁺/Mg²⁺ PBS, and fresh growth medium was added and incubated at 37°C(with a change of medium every 2 days) until the functional assay was performed. Competition experiments were conducted, transfecting Hep G2 (high) cells with 1.11 pmol C1315 carrier-condensed DNA in the presence and absence of free C1315 peptide (fresh medium was added after 2 h of incubation). All transfections were done in duplicate. No cell death was observed in any of the wells transfected with the DNA-conjugated polylysine complexes throughout the incubation. Luciferase expression was assessed at 2, 4, 6, 8, 10, and 12 days after transfection. Staining for [β-galactosidase was done 36 h after transfection. Medium from cells transfected with pFIX was assayed for hFIX 4 days after transfection.

Assay for luciferase expression

Duplicate samples of cells were harvested and assayed for luciferase activity as previously described (3). Protein was determined by the Bradford method (Bio-Rad kit). Results were expressed as integrated light units (ILU) per milligram protein. All measurements were done in duplicate and averaged.

Assay for β-galactosidase activity

Individual RuR7 and Rep G2 cells expressing [β-galactosidase were identified as previously described (19). Briefly, cells transfected with the pCMV lac Z II plasmid were thoroughly washed with PBS, fixed (in the 6-well plates) with a solution of 0.5% glutaralde-hyde in PBS for 10 min, washed again, then incubated in a solution containing 0.5% X-gal for 4-–5 h at 37°C. Cells were then lightly counterstained with nuclear fast red. Blue-colored cells were identified and photographed through a phase-contrast inverted-light microscope. Efficiency was calculated by the number of clearly blue cells in 100 cells counted.

Assay for hFIX production

hFIX was assayed by enzyme-linked immunosorbent assay (ELISA). Standards, ranging in concentration from 0.2 to 1 ng/ml, were prepared using purified hFIX (American Diagnostics, Greenwich, CT). ELISA plates were coated with the capturing monoclonal mouse immunoglobulin G (IgG)-derived anti-hFIX (Hematological Technology, Essex, VT), incubated at 4°C overnight, thoroughly rinsed 0.1% Tween 20 PBS, and blocked with 100 μ of 10% FCS in RPMI medium (GIBCO). Standards and aliquots of media collected from transfected HuH7 and Hep G2 cells were then added and incubated at room temperature for 2 h. After stringent wash, 50 μl of primary antibody (rabbit IgG-derived polyclonal anti-hFIX; California Biotechnology) diluted in 10% FCS RPMI (1:1,000) was added to the wells and incubated at room temperature for 1 h. After a stringent wash, 50 μl of a 1:2,000 dilution in 10% FCS RPMI of goat anti-rabbit IgG conjugated to horseradish peroxidase (Boeh-ringer Mannheim) was added. After the final wash, horseradish peroxidase activity in each sample was assessed by optical density measurement of the samples after incubation for 1 h with tetramethyl benzidine dihydrochloride. All assays were done in duplicate, and the results were expressed as nano-grams hFIX per milliliter medium per million cells (ng·ml^–1·10⁶ cells^–l).

β-Galactosidase-SECR cytochemical staining colocalization

HuH7, Hep G2 (high), and Hcp G2 (low) cells were plated in six-well plates and transfected as described above. Fluorescein labeling was carried out with fluorescein isothiocyanate, as described previously (22). Two days after transfection, cells were washed with Ca²⁺ /Mg²⁺ PBS, incubated with 100 nM fluorescein-labeled C1315 peptide, and diluted in binding buffer (DMEM containing 50 mM N-2-hydroxyethylpiperazine-N’ -2-ethanesulfonic acid, 0.1 mg/ml cytochrome c, 0.01% Tween 80, 2 mg/ml bovine serum albumin) at 4°C. Individual cells were imaged on a Zeiss axiovert 35 microscope at an excitation wavelength of 493.5 nm and a measurement wavelength of 530 nm. Digital images were captured by a cooled CCD camera model CH250 (Photometries, Tucson, AZ) and quantified by a Nu 2000 camera controller board (Photometries) with a Macintosh Quadra 900 configuration. Data were processed with Oncor image software (Oncor Imaging, Rockville, MD). After measurement, the cells were rinsed repeatedly and the plates were marked for future reference of orientation. The cells were then assayed for β-galactosidase activity, imaged on a phase-contrast light microscope in the exact orientation used during fluorescein measurements. Photographs were taken to scale to compare cell binding and uptake of fluorescein-labeled C1315 with cellular expression of β-galactosidase.

Statistical analysis

Data are expressed as means ± SE. Statistical comparisons of treatment groups were evaluated with the use of a parametric analysis of variance using the Student-Newman-Keuls test.

RESULTS

Structure of the gene transfer complex

Construction of the protein conjugate of poly-L-lysine to C1315 peptide was monitored by proton NMR, both at the step of LC-sulfo-SPDP modification of poly lysine and at the step of conjugation of the C1315 peptide to modified poly lysine. Key to the analysis is the presence of unique aromatic proton resonances that have chemical shifts distinct from peaks produced by polylysine for both the LC-sulfo-SPDP conjugated to polylysine and for the C1315 peptide conjugated to SPDP-polylysine. Integration of these resonances relative to the C α-protons of polylysine reveal that 1 in 14 lysine side chains were reacted with the LC-sulfo-SPDP reagent, and 1 in 345–385lysines were modified by C1315.

Because tightly condensed particles apparently increase the efficiency of internalization, we examined our complexes by electron microscopy. TYpically, C1315-polylysine-pGL3 control DNA (5.5 kb) complex mixtures contained complexes between 17 and 23 nm in diameter (Fig. 1C). Solutions used to make the complexes (high-salt conditions) were also examined and contained no visible structures. The lighter zones bordering the complexes, produced by rotary shadowing, indicate that the height is comparable to the diameter, i.e., the complexes are roughly spherical. Figure 1A shows aggregated complexes present in solution before the addition of 5 M NaCl (MATERIALS AND METHODS). Final complex mixtures contained ≤0.5% of these aggregates. Mixtures that contained ≥50% of the aggregated form failed to transfect HuH7, Hep G2 (high), or Hep G2 (low) cells (data not shown). In addition, some complexes appear to unravel at lower salt concentrations into more relaxed crescentlike structures. Mixtures containing these complexes achieved a very low transfection efficiency (data not shown). These data are consistent with previous reports showing that only tightly formed complexes transfect cells efficiently (9). Because plasmid size might affect the size of these particles, the pCMV lac Z II DNA (10.8 kb) and C1315-polylysine-pCMV lac Z II complexes were compared (Fig. 1B). These complexes also ranged in size from 20 to 25 nm in diameter. Complexes with pFIX (5.4 kb) were identical to pGL3 control complexes in size.

Fig. 1.

Electron micrographs ofC1315-polylysine condensed plasmid DNA. Complexes were formed under high-salt conditions (~ 1 M), then diluted 10 times and immediately pi petted onto a 1,000-mesh electron microscope carbon grid, fixed, blotted, and stained with 0.04% uranyl acetate. A: aggregated complexes [formed under low-salt (~ 0.3 M) conditionsJ. B: spheroid particles observed when cytomegalovirus (CMV) lac Z II plasmid l10.8 kilobase (kb)] was properly condensed (arrowheads). C: spheroid particles observed when SV40 pGL3 control plasmid (5.25 kb) was properly condensed (arrowheads). Bar, 100 nm.

Determination of surface receptor binding

HuH7, Hep G2 (high), and Hep G2 (low) cells were incubated with different concentrations of ¹²⁵I-labeled C105Y peptide in the presence and absence of 200-fold molar excess of unlabeled peptide. Both HuH7 and Hep G2 (high) cells exhibited specific and saturable binding, as shown in Fig. 2. Scatchard analysis of Hep G2 (high) binding revealed a Kd of 48.7 ± 2.4 nM, consistent with previous reports (13, 16, 18, 27, 28). HuH7 cells bound more C105Y [1.5-fold more than Hep G2 (high)] with a Kd of76.3 ± 7.3 nM. Hep G2 (low) cells exhibited 10-fold less specific binding of iodinated ligand than HuH7 cells and 7 .5-fold less than Hep G2 (high) cells (Fig. 2). Hep G2 (low) cells bound iodinated C105Y with a Kd of 124.2 ± 34.2 nM. These binding trends were consistent in seven experiments that compared binding in HuH7, Hep G2 (high), and Hep G2 (low) cells.

Fig. 2.

Specific binding of iodinated Cl05Y to hepatoma cell lines HuH7 (O), Hep G2 (high) (□), and Hep G2 (low) cells (○). Specific binding represents difference between total binding and binding in the presence of 200-fold excess of unlabeled peptide (nonspecific binding). A rept·esentative example from 6 experiments is shown. cpm, Counts/min.

Transfection of Hep G2 cells with the pGL3 control luciferase expression plasmid

Various concentrations ofC1315-polylysine-pGL3 control DNA complexes were applied to Hep G2 cells. Transfection and expression were assessed by luciferase enzyme activity in cell extracts. Positive controls, described in MATERIALS AND METHODS, established the capability of both Hep G2 (high) and Hep G2 (low) cells to express the pGL3 luciferase gene product. Luciferase activity transfected by receptor-mediated means peaked at 538,731 ± 144,346 ILU/mg protein between days 2 and 4. Luciferase activity declined to background 10 days after transfection. Figure 3A demonstrates the dose dependence and time course of the transfection with the peptide-based complex. Gene transfer was greatest with DNA contentofl.ll pmol/5 × 10⁵ cells in a 10-mm well. Complex concentrations either below or above the optimum concentration achieved less efficient transfer and expression (8, 25, 32). Hep G2 Gow) cells were transfected by the complex with a much lower efficiency (Fig. 3A). Hep G2 (high) cells exposed to unconjugated poly-L-lysine-condensed DNA (1.11 pmol), as well as lc-sulfo-SPDP modified poly-l-lysine-condensed DNA(l.11 pmol) in the presence of corresponding concentrations of free C1315, were not transfected (Fig. 3B).

Fig. 3.

Luciferase expression: dose dependence and time course of expression. A: luciferase activity assayed 2, 4, 6, and 8 days after transfection with 0.83 pmol (hatched bars), 1.11 pmol (crosshatched bars), or 1.34 pmol (shaded bars) of pGL3 plasmid DNA. Protein extracts obtained from cells treated with 1.34 pmol ofpGL3 plasmid DNA condensed with unmodified poly-L-lysine in the presence of free peptide were used as negative controls (open bars). In the 6 experiments performed, all DNA c-oncentrations produced significant (P < 0.05) transgene expression at days 2 and 4. Hep G2 (high) expression was statistically significant from Hep G2 (low) (black bars) expression at days 2, 4, and 6. None of the transfected Hep G2 (low) samples were significantly different from negative controls. Lipofectin transfection yielded similar transfection in both cell types: Hep G2 (high), 1,023,000; and Hep G2 (low), 934,000 integrated light units (ILU)/mg protein. B: luciferase expression in cells transfected with nonfunctional conjugates. In the 6 experiments conducted, 1.11 pmol of DNA properly condensed with either poly-L-Iysine (open bars) or poly-l-lysine modified with LC-sulfo-N -succinimidyl-3-(2-pyridyldithio)propionate (SPDP; gray bars) in the presence offree C1315 peptide failed to result in luciferase activity significantly different from cells incubated with naked DNA (black bars). All controls were significantly less effective than active complex. Note different scale on y-axis. C: dose-dependent competition of transgene expression with free ligand. Hep G2 (high) cells were transfected with optimal DNA concentration (1.11 pmol) in the presence and absence of a 1-, 10-, 50-, 100-, and 200-fold molar excess of free C1315 peptide. Fold molar excess competitor refers to free C1315 added in excess to peptide conjugated to the transfection complexes. Inhibited expression levels differed from uninhibited transfection (P < 0.05).

Addition of a 1-, 10-, 50-, 100-, and 200-fold molar excess offree peptide at the time oftransfection blocked uptake and expression in a dose-dependent fashion, as shown in Fig. 3C. High excess of peptide completely blocked uptake and transfection. Excess free peptide had no effect on cell viability.

Transfer of hFIX to Hep G2 cells

We tested the ability of our system to deliver a clinically relevant gene as well. Hep G2 and HuH7 cells do not express endogenous human coagulation factor IX. Thus we transduced cells with a plasmid coding for the hFIX gene and measured product secreted into growth medium 4 days later. The medium did not interfere with ELISA for factor IX. Figure 4 depicts the results. When transfected with 1.11 pmol carrier-condensed DNA, HuH7 cells produced 7.01 ± 1.49 nglml, whereas Hep G2 (high) cells produced 5.07 ± 1.60 ng/ml hFIX. As with other reporter genes, Hep G2 (low) cells expressed minimal factor IX, peaking at 0.86 ± 0.44 ng/ml, with 1.11 pmol carrier-condensed DNA. Again, unconjugated poly-L-lysine-condensed DNA failed to transduce any of the cell types.

Fig. 4.

Human factor IX expression in transfected cells. HuH7, Hep G2 (high), and Hep G2 (low) cells were transfected with control complexes (open bars), 0.83 pmol (crosshatched bars), and 1.11 pmol of CMV pFIX plasmid coding for human coagulation factor IX (black bars). Human factor IX secreted into the media was assayed at 4 days posttransfection by enzyme-linked immunosorbent assay for human factor IX. In the 5 experiments performed, HuH7 and Hep G2 (high) cells expressed levels of human factor IX that were significantly higher (P < 0.05) than Hep G2 (low). HuH7 and Hep G2 (high) cell expression was also significantly different from control transfections (P < 0.05), whereas Hep G2 (low) cell expression was not. Lipofectin transfection yielded similar results in all3 cell types: HuH7, 22.35; Hep G2 (high), 25.45; and Hep G2 (low)26.74 ng·^–·million cells^–1.

Transfection with pCMV lac Z II β-galactosidase expression plasmid

Both HuH7 and Hep G2 cell lines were transfected with the pCMV lac Z II plasmid coding for the β-galactosidase protein as described above. As shown in Table 1, HuH7 and Hep G2 (high) cells, but not Hep G2 (low), displayed substantial β-galactosidase staining. The pattern of staining varied with DNA concentration, similar to luciferase expression. For both HuH7 and Hep G2 (high), 1.11 pmol DNA/well produced the highest percentage of positive cells (Table 1). Nonspecific lipofectin transfection of cells yielded, on average, twice the proportion of positive cells seen with our complex. DNA condensed with unconjugated C1315-poly-L-lysine failed to transfect any of the cell types. Positive cells were intensely stained, and no background β-galactosidase activity was detected.

Table 1.

Transgene and SECR expression in hepatocytes

Cell Type	Average %Blue Cells	Average %Fluorescent Cells
Hep G2 (high)	11.3 ± 2.3	20.4 ± 4.6
HuH7	5.0 ± 1.5	14.3 ± 1.6
Hep G2 (low)	1.0 ± 0.4	1.8 ± 0.6

Data are means ± SE; n = 6 experiments for each cell type. HuH7, Hep G2 (high), and Hep G2 (low) cells were treated with 1.11 pmol of CMV lac Z II plasmid DNA condensed with peptide carrier. Cells were assayed for β-galactosidase activity 3 days after transfection. Percentages represent number of clearly blue or fluorescent cells per 100 cells counted. SECR, serpin enzyme complex receptor.

Colocalization of SECR and β-galactosidase expression

We designed a set of experiments to colocalize transgene expression with expression of SECR. Cyto-chemical staining for the receptor with fluorescein-labeled C1315 peptide revealed that only some cells in cultures of HuH7, Hep G2 (high), and Hep G2 (low) cells bind detectable amounts of the ligand (shown in Table 1). Only those Hep G2 (high) and HuH7 cells that bound the fluorescein-labeled peptide took up the complex, expressed the transgene, and stained positive for β-galactosidase expression. Hep G2 (low) cells exhibiting minimal fluorescence did not stain positive for β-galactosidase. Furthermore, HuH7 cells bound the fluorescein-labeled peptide with less frequency than Hep G2 (high) cells (Table 1). However, HuH7 binding, as well as β-galactosidase expression, was more intense (Fig. 5, A and B, respectively). Cells treated with unlabeled peptide or free fluorescein had no detectable autofluorescence.

Fig. 5.

Comparison of cellular or histochemical staining to identify β-galactosidase expression (A, C, and E; arrowheads indicate blue-stained cells) and immunofluorescent identification of cells that express high levels of serpin enzyme complex receptor (SECR; B, D, and F ). A and B: HuH7 cells. C and D: Hep G2 (high) cells. E and F: Hep G2 (low) cells. Same field was photographed for each pair. No example was found of a cell positive for β-galactosidase expression that was not also positive for SECR expression, but not all SECR-positive cells stained intensely for fβ-galactosidase activity. Lipofectin transfection ofHep G2 (high) cells yielded 34.8 ± 6.7; RuH7 cells, 14.7 ± 3.6; and Rep G2 (low), 9. 7 ± 1.6 blue cells in 100 cells counted. Magnification ×20.

DISCUSSION

We have shown that expression plasmids tightly condensed (18–25 nm in diameter) with polylysine conjugated to the C1315 peptide, a ligand previously shown to bind to the SECR (13, 16, 26–28), can be targeted to cells bearing the receptor in vitro. The size of the peptide ligand, as well as the repetitive nature of poly-L-lysine, allows us to assess the coupling of the C1315 peptide to the poly-L-lysine by NMR, which demonstrates that one receptor ligand for every three poly-L-lysine molecules (or 35 and 72 ligands bound to each of the small and large plasmid DNA molecules, respectively) is sufficient to direct receptor-mediated gene transfer. At the structural level, we are able, using electron microscopy, to verify that compact condensation occurs. Our data confirm that tightly condensed complexes are far more efficient for transfection than the aggregates that form at lower N aCl concentrations or relaxed complexes that form at higher NaCl concentrations.

Several lines of evidence indicate that gene transfer is mediated by the SECR and does not occur by nonspecific mechanisms. Hep G2 (low) cells, which express few SECRs, take up and express minimal levels of DNA (although they are capable of expressing the identical plasmid when it is delivered by lipofectin), whereas HuH7 and Hep G2 (high) cells, which express abundant SECR, can be transduced with a 10-fold higher efficiency. This is true for all genes tested. In addition, free ligand added at the time of transfection inhibited gene transfer in a dose-dependent manner, so receptor ligands apparently compete with the complex for receptor binding and uptake. HuH7 and Hep G2 (high) cells transfected with DNA condensed with unmodified poly lysine in the presence or absence of free peptide did not exhibit gene expression, so uptake is not due to nonspecific pinocytosis of condensed DNA particles. Only cells shown by fluorescent ligand binding and uptake to bind the C1315 peptide exhibit intense β-galactosidase activity, whereas cells that bind no fluorescent C1315 do not express the lac Z gene. Taken together, these data demonstrate that uptake and expression of exogenous genes are mediated through the SECR.

The specificity, as well as success, of gene transfer in vitro in cells that bear the SEC receptor allows us to speculate that this receptor system could promote gene transfer in vivo. High-level expression was achieved for all three reporter genes. In addition, substantial gene transfer occurred even in the presence of a 10-fold molar excess of competitive ligand in vitro, so the presence of the natural ligand in vivo will probably not prohibit gene transfer. SECR has been found in lung, liver, and brain (reviewed in Ref. 26), all of which might be potential target tissues for therapeutic gene transfer in common inherited (e.g., α₁-AT deficiency) or acquired (e.g., Alzheimer's disease) disorders.

Thus the tissue distribution of the SECR coupled to the specificity and in vitro success of SECR-directed gene transfer give this system promise for gene therapy. However, more detailed knowledge of receptor-mediated gene transfer and the elements that affect its efficiency is crucial to optimizing the system for gene therapy. The SECR-directed system offers special advantages in this regard, because NMR can be used to assess the stoichiometry and internal structure of the protein portion of the gene transfer complex and the receptor ligand can be labeled with fluorescent tags for tracking its fate inside the cell. In addition, examination of the in vivo response to targeting this system (for example, immune responses or receptor-mediated signal transduction) is essential to evaluate the potential of this system for therapeutic use.