Bioconjugate Chem. 1995, 6, 283-291
283
Targeted Gene Delivery with a Low Molecular Weight Glycopeptide Carrier Manpreet S. Wadhwa,+ Daren L. Knoell, Anthony P. Young, a n d Kevin G. Rice",' College of Pharmacy, The Ohio State University, Columbus, Ohio 43210. Received November 28, 1994@
A low molecular weight glycopeptide carrier was prepared by coupling a tyrosinamide-triantennary oligosaccharide to dp19 poly-L-lysine resulting in a 1:1conjugate. The glycopeptide carrier complexed with plasmid DNA as evidenced by displacement of intercalated dye, light scattering by condensed DNA, and immobility of complexed DNA upon agarose gel electrophoresis. DNA-carrier complexes were endocytosed into HepG2 cells via the asialoglycoprotein receptor due to recognition of terminal galactose residues on the oligosaccharide. The resulting luciferase reporter gene expression was dramatically influenced by the solubility of complexes, the extent of complexation, and the presence of the lysosomotropic agent chloroquine. The results suggest that low molecular weight glycopeptides may be suitable for further development as well-defined DNA carriers for receptor-mediated gene delivery in vivo.
INTRODUCTION
Genes are attractive candidates for use in a variety of disease states due to the ability to produce therapeutic biomolecules using the biosynthetic machinery provided by host cells (1-4). Established protocols for transfecting genes into cells include calcium phosphate or DEAEl dextran coprecipitation, electroporation, particle bombardment, scrape loading, sonication, liposomal delivery, gene transfer by viral vectors, and receptor-mediated gene delivery (5-8). Although all of these methods can be applied to mammalian cells in culture, transfection of cells i n vivo for gene therapy or for probing biological function poses special problems, thus restricting the use of most of these methods to ex vivo protocols. Viral vectors, cationic liposomes, receptor-mediated gene delivery, and direct injection of DNA have emerged as promising noninvasive approaches for the introduction of DNA into cells in vivo (5-8). Of these, receptormediated gene delivery has the greatest potential for targeting specific tissue or cell types based on the specific recognition of ligands by unique receptors on these cells. DNA carriers have been designed to transfect hepatocytes via the asialoglycoprotein receptor (ASGP-R) (9-15) or the insulin receptor (16,171. Mouse lung endothelial cells have been targeted for gene delivery via thrombomodulin (18)and leukaemic T cells targeted by means of a mucin antigen (19). Alternatively, diverse cell types derived
* To whom correspondence should be addressed.
' Present Address: College of Pharmacy, University of Michigan, 428 Church St., Ann Arbor, MI 48109. Tel: 313-763-1032. Fax: 313-763-2022. Abstract published in Advance ACS Abstracts, April 1,1995. Abbreviations: ASGP-R, asialoglycoprotein receptor; AUFS, absorbance units full scale; Boc, tert-butyloxycarbonyl; DEAE dextran, [(diethylamino)ethyl]dextran;DMF, dimethylformamide; DTT, dithiothreitol; EDC, l-ethyl-3-[3-(dimethylamino)propyllcarbodiimide methiodide; dp, degree of polymerization; EDTA, ethylenediaminetetraacetic acid; FAB-MS, fast atom bombardment-mass spectroscopy; FCS, fetal calf serum; HBS, hepes-buffered saline; HPAEC, high pH anion exchange chromatography; MEM, minimum essential media; PL, poly-L-lysine (of average dp 19);RLU, relative light units; RP-HPLC, reversed phase HPLC; TFA, triflouroacetic acid; Tri, galactose-terminated triantennary oligosaccharide; TriPL, Tri-polylysine (average dp19) conjugate; Agal-TriPL, TriPL devoid of terminal galactose residues. @
1043-1802/95/2906-0283$09.00/0
from both normal and tumor tissues have been transfected by means of the transferrin receptor (20-22). Carriers for receptor-mediated gene delivery typically employ a receptor ligand covalently attached to a polycationic anchor that binds DNA by ionic interaction (9). Selective transfection of hepatocytes via the ASGP-R has been accomplished with ligands possessing terminal galactose residues such as asialoorosomucoid (9,10,12), galactosylated proteins or polymers (13, 14, 161, and galactosylated synthetic ligands (11, 15). The anchor utilized most often is poly-L-lysine in the molecular weight range of 20-60 kDa. An important property of polylysine is its ability to condense DNA into compact structures which may be small enough for internalization into cells by endocytosis (23, 24). Carriers may also include effector molecules like fusogenic peptides (25)to allow efficient lysosomal escape of internalized DNA (13, 26-28). A further consideration is the use of well-characterized components to produce carriers of defined structure. This may be essential to carefully control the preparation of DNA-carrier complexes in order to achieve highly reliable gene delivery. We have taken a step toward this objective by selecting a triantennary oligosaccharide as a low molecular weight, high affinity ligand for the ASGP-R (29, 30), which was coupled to 2.5 kDa poly-Llysine to produce a low molecular weight glycopeptide carrier. The carrier allowed formulation of soluble DNA complexes which were monitored and optimized by spectroscopic methods. Carrier-complexed plasmid was endocytosed into hepatoma cells possessing the ASGP-R and resulted in reporter gene expression, indicating the potential of low molecular weight carriers for efficient DNA delivery i n vivo. EXPERIMENTAL PROCEDURES
Materials. Poly-L-lysine hydrobromide, of average dp 19 (PL), succinic anhydride, l-ethyl-3-[3-(dimethylamino)propyl]carbodiimide methiodide (EDC), and chloroquine were obtained from Sigma Chemical Co., St. Louis, MO. D-Luciferin, luciferase from Photinus pyralis (EC 1.13.12.7), and P-galactosidase (EC 3.2.1.23) from bovine testes were obtained from Boehringer Mannheim, Indianapolis, IN. HepG2 cells were from American Type Culture Collection, Rockville, MD. Bradford reagent was 0 1995 American Chemical Society
284 Bioconjugafe Chem., Vol. 6,No. 3, 1995
purchased from BioRad, Hercules, CA, and thiazole orange was a gift from Beckton Dickinson Immunocytometry Systems, San Jose, CA. Gel filtration HPLC column (G2000 SWXL) was purchased from Tosohaas, Montgomeryville, PA; C8 reversed phase columns (MV Microsorb, with 5 pm packing) were from Rainin, Emeryville, CA, and analytical and semipreparative reversed phase polymer columns (PRP-1, with 10 pm packing) were from Hamilton Co., Reno, NV. HPLC was performed using equipment from ISCO (Lincoln, NE), consisting of computer-interfaced pumps, variable wavelength UV detector, and automated fraction collector. High pH anion exchange chromatography (HPAEC) was performed on a carbohydrate analyzer from Dionex Corp., Sunnyvale, CA, with a CarboPac PA1 column. Fluorescence and light scattering measurements were performed using a computer-interfaced fluorimeter (LS50B) from Perkin-Elmer, U.K. UV spectroscopy was conducted on a Beckman DU640 spectrophotometer, and luciferase light units were recorded on a luminometer (Lumat LB 9501) from Berthold Systems, Pittsburgh, PA. Preparation of Oligosaccharide-Polylysine ('I'riPL)Conjugate. Succinylation of Triantennary Oligosaccharide. Boc-protected tyrosinamide triantennary oligosaccharide (Boc-Tri)was prepared from bovine fetuin as previously described (29).Boc-Tri (1pmol) was freeze dried and reacted with 200 pL of TFA for 10 min a t room temperature. The Boc-deprotected oligosaccharide (Tri) was then freeze dried three times to obtain a neutral pH. Tri (500 nmol, prepared in 450 pL of 0.2 M sodium bicarbonate buffer, pH 8.0) was reacted with succinic anhydride (12.5 mg in 50 pL DMF) for 15 min a t room temperature, along with the addition of 150 pL of 1 M sodium hydroxide to maintain the pH between 7.5-8. The reaction was terminated by adjusting the pH to 12 by the addition of 150 pL of 1 M sodium hydroxide and incubated a t 37 "C for 10 min, followed by acidification to pH 3 by adding 100 pL of 4 M TFA prior to HPLC purification. Succinyl-Tri was purified from a semipreparative polymeric RP-HPLC column (305 x 7 mm) equilibrated a t 3 m u m i n with 0.1%TFA and 4% acetonitrile. Following injection of 500 nmol of succinyl-Tri (900 pL) into a 1 mL loop, an acetonitrile gradient of 4%to 15% was developed over 13 min while A b s ~ 7 40.2 ~ ~AUFS was monitored. The peak eluting a t 9 min was collected and freeze dried and the yield estimated by A b s ~ 7 (4E ~=~1450 M-l cm-'1. Boc-Tri, Tri, and succinyl-Tri were prepared for proton NMR spectroscopy by freeze drying 1 pmol twice in 99.98%deuterium oxide containing 0.01%acetone as an internal standard and analyzed on a Bruker 500 MHz NMR spectrometer operating a t 23 "C. The acquired spectra were processed utilizing resolution enhancement parameters supplied with Felix software (Hare Research, Eugene, OR). Samples were prepared for FAB-MS by dissolving 20 nmol in 10 pL of water and 1 pL of a-monothioglycerol. The water was removed by speed vacuum, and the 1 pL sample was applied to the probe of a Finnigan Matt 900 FAB-MS operated in the positive ion mode. Conjugation of Succinyl-Tri with Polylysine. SUCcinyl-Tri (400 nmol in 400 pL water) was added to PL (2.5 pmol in 400 pL of water), and the coupling reaction was initiated by adding 800 pL of EDC (500 mM in 20 mM boraxhydrochloride buffer, pH 7.5). After incubation a t room temperature for 2 h, the reaction was quenched by the addition of 16 pL of 2 M hydrochloric acid.
Wadhwa et al.
The glycopeptide conjugate (TriPLj was purified in 200 nmol(800 pL) portions on an analytical (250 x 4.1 mm) polymeric RP-HPLC column (50 "C) equilibrated at 1 mL! min with 0.1%TFA and 1%acetonitrile. The acetonitrile concentration was held a t 1% for 20 min followed by a step to 15%acetonitrile over 1 min and elution continued ~ ~AUFS. The for 10 min while monitoring A b s ~ 7 40.1 peak eluting a t 23 min was collected and freeze dried and the yield determined by Abs27dnm. Preparation of Agalatco-Rantennary -Polylysine Conjugate. Agal-TriPL was prepared by incubating 100 nmol of TriPL (in 200 pL of 50 mM sodium phosphate citrate buffer, pH 4.3) with 40 mU of @-galactosidasefor 24 h a t 37 "C. The product was purified on a n analytical polymer reversed phase column as described for TriPL and characterized by monosaccharide compositional analysis as described below. Compositional Analysis of TriPL. The amine content of TriPL (determined by Abs274) was obtained by fluorescamine assay as described (31j, using a PL standard, with the fluorescence being measured a t excitation and emission wavelengths of 390 and 475 nm (5 nm slit widths). Amino acid analysis of the conjugates was performed by Picotag analysis (321,and monosaccharide composition analysis was performed by HPAEC following TFA and hydrochloric acid hydrolysis (33). Gene Delivery and Expression. Complexation of Plasmid D N A with Carrier. A 5.6 kbp plasmid pCMVL encoding the gene for luciferase (34)under the control of cytomegalovirus promoter was prepared by the alkaline lysis method and purified on a cesium chloride gradient to obtain the supercoiled form (35). Quantitation was based on absorbance of DNA a t 260 nm (1 pg = 0.02 0.d. units). TriPL-pCMVL complexes were prepared with DNA concentrations ranging from 0.2 to 40 pg/mL and TriPLpCMVL ratios (nmol carriedpg DNA) varying from 0.1 to 1.2. The optimized complex was prepared a t a DNA concentration of 20 pg/mL and a carrier-DNA ratio of 0.8 by adding TriPL (16 nmol, in 500 pL of solvent) to pCMVL (20 pg, in 500pL of solvent) while vortexing and allowing the mixture to incubate a t room temperature for 30 min. The amounts and volumes of carrier and DNA were linearly scaled for preparing different amounts of complex. Solvents used were 0.15 M sodium chloride (saline), 20 mM Hepes (pH 7.4) with 0.15 M sodium chloride (HBS), and 0.72 M mannitol. For control experiments, TriPL was substituted with PL or Agal-TriPL. Assays for Monitoring DNA-Carrier Complexes. TriPL-pCMVL solubility was determined by analyzing an aliquot (1 pg of DNA) of the complex before and after centrifugation a t 13000g for 4 min a t room temperature. The aliquots were diluted to 1 mL in the appropriate solvent (saline, HBS, or mannitol), and the DNA remaining in solution was measured by Abs260. Complexation was monitored by a fluorescence assay based on displacement of intercalating dye from DNA by TriPL or PL. An aliquot of TriPL-pCMVL (1 pg of DNA, in 25-500 pL) was diluted to 0.5 mL in solvent and then diluted with 3 mL of solvent containing 0.117 pM thiazole orange (from a 0.1 mg/mL stock in 1% methanol, 6476 = 30 000 M-' cm-l). Fluorescence of the intercalated dye was measured using excitation a t 500 nm and emission a t 530 nm, with the slits set a t 15 and 20 nm, respectively, to maximize sensitivity. Complexes were also measured by light scattering in order to monitor DNA condensation. Traces of dust were removed from sample tubes by means of 0.1 pm filtered pressurized air, and solvents were filtered through 0.2 pm surfactant free cellulose acetate filters. TriPL-
Targeted Gene Delivery with a Glycopeptide Carrier
A
Bioconjugate Chem., Vol. 6,No. 3, 1995 285
Boc-Tri
Galp 1-4GlcNAcp1-2Mana 1, y
2
6Manpl-4GlcNAcp1-4GkNAcp1-NH-CO-CH-NH-CO-O-C-(CH&
/3
Galp1-4GlcNAcp1-2Mana 1
I
4
Galp14GlcNAcpi
B
Tri I
Galp l-4GlcNAcp1-2Manal\
CH2
6Manp1-4G1cNAc~1-4G1cN~l-NH~~H-NH2 /3
I
Galp l-4GlcNAcp1-2Mana 1 4
Gal pl4GlcNAcpi
C
Succinyl-Tri
Succinic Anhydride
6
Galp 1-4GlcNAcpl-2Mana 1\ Galpl-4GlcNAcp 1-2Manal
CH2 6Manpl-4GlcNAcp1-4GlcNAcp1-NHCO-CH-NH-CO-(C~$XOH .3
4
Galpl-4GlcNAcpi
Polylysine
D
TriPL
Galpl-4GlcNAcp1-2Manal~
6
CH2 Manpl4GlcNAcpl-4GlcNAc~l-NH-C&H-NH-CO-(Ct-l&CO-NHI
Gal4.f 14GlcNAcp1-2Manal l 3 4 Galpl4GlcNAcpi
C b y
2
NH2 y 2 NH2 NH2 NH2 I I 1 I CCM4 (CW4 (CH34 CCM4 ( y J 4 (7M4 CH2 H-[NH-CH-CO-NH-CH-CO-NH-CH-CO-NH~HC~NH-CHCO-NHCH-CO]gNHCH-WOH NH2 I
y
2
Figure 1. Reaction scheme for conjugation of triantennary oligosaccharide with polylysine. Boc-protected tyrosinamide triantennary oligosaccharide (Boc-Tri, A) was converted into Tri by treatment with TFA which exposed the N-terminus of tyrosine by removal of Boc (B). The amine was reacted with succinic anhydride to introduce a carboxylic group in succinyl-Tri (C). The carboxylic group TriPL was a prepared as 1:l conjugate was activated by EDC and coupled with an amine on polylysine to produce TriPL conjugate (D). by controlling stoichiometry of reactants; however, the structure shown is only meant for illustration since the oligosaccharide was randomly coupled to one of the amines on polylysine of average dp 19.
pCMVL (1 pg of DNA) was diluted to 3.5 mL in solvent, and scattered light intensity a t 90" was measured by keeping both monochromaters a t 350 nm (2.5 nm slits). Light scattering synchronous scans were obtained by simultaneous scanning of emission and excitation monochromaters paired a t the same wavelength. Additionally, band retardation assay was used to monitor complexation (9). TriPL-pCMVL (200 ng of DNA) was mixed with gel loading buffer and electrophoresed on a 1%agarose gel a t 70 V, and DNA was visualized postrun by ethidium bromide staining and U V detection. Transfection and Gene Expression. HepG2 cells were plated on 6 x 35 mm wells ((0.5-1) x lo6 cells per well) and grown to 40-70% confluency in minimum essential media (MEM) with 10% fetal calf serum (FCS) supplemented with penicillin and streptomycin. Transfections were performed in MEM (2 mL per 35 mm well) with 2% FCS, with or without 100 pM chloroquine. pCMVLTriPL (0.1-20 pg of DNA, in 0.5 mL) was added dropwise
to triplicate wells. After 5 h incubation a t 37 "C, the media was replaced with MEM supplemented with 10% FCS. Luciferase expression was determined a t 24 h (19 h post transfection). Cells were washed twice with ice-cold phosphate-buffered saline (calcium, magnesium free) and then treated with 0.5 mL of ice-cold lysis buffer (25 mM tris chloride pH 7.8, 1 mM EDTA, 8 mM magnesium chloride, 1%Triton X-100, with 1mM DTT added fresh) for 10 min. The cell lysate mixture was scraped, transferred to 1.5 mL microcentrifuge tubes, and centrifuged for 7 min a t 13000g at 4 "C t o pellet debris. Lysis buffer (350 pL), sodium-ATP (4 pL of a 180 mM solution, pH 7, 4 "C), and cell lysate (100 pL, 4 "C)were combined in a test tube, briefly mixed, and immediately placed in the luminometer. Luciferase light units were recorded with 10 s integration after automatic injection of 100 pL of 0.5 mM D-lUCiferin (prepared fresh in lysis buffer without DTT). Relative light units (RLU) of luciferase activity were converted to femtomoles of luciferase (after subtraction of background counts, typically
Wadhwa et al.
286 Bioconjugate Chem., Vol. 6, No. 3, 1995
150-200 RLU) based on a standard curve obtained with known triplicate concentrations of firefly luciferase in lysis buffer in the presence of untransfected HepG2 cell lysate. The standard curve was linear between 200 and lo6 RLU, and each femtomole was equivalent to 740 000 RLU. Protein concentration in the cell lysate was measured by Bradford assay (36)using bovine serum albumin as a standard (the sample size of the cell lysate was 50 pL or less, and no interference from triton X-100 was observed). The femtomoles of luciferase in each sample were normalized to mg of protein and the mean and standard deviation obtained from each triplicate.
0.01
0.oc
r A
BOC-Tri
B
Tri
0.01
E * b
;
~~
0.00
RESULTS
Preparation of Oligosaccharide-Polylysine (TriPL) Conjugate. A hepatocyte-targeted gene delivery carrier was prepared by conjugating a natural triantennary oligosaccharide to an amine side chain of dp 19 polylysine. The conjugation results in a n amide linkage between the oligosaccharide and polylysine such that the three galactose-terminated antennae are accessible for binding with the ASGP-R. The reaction scheme is outlined in Figure 1. Triantennary oligosaccharide was purified from fetuin after reducing end modification as previously described (29) and obtained as a Boc-protected tyrosinamide oligosaccharide (Boc-Tri, Figure 1A). Treatment of Boc-Tri with TFA removed the Boc group and exposed the N-terminal amine on tyrosine for further modification (Figure 1B). Deprotection was carried out for 10 min in the absence of water and the product freeze dried immediately to avoid hydrolysis of glycosidic linkages of the oligosaccharide. Under these conditions, the conversion to deprotected oligosaccharide was quantitative, as evidenced by elution on RP-HPLC. A carboxylic group was introduced into the oligosaccharide by succinylation of the exposed amine (Figure 10. The reaction proceeded quickly when the pH was maintained between 7.5 and 8 in order to deprotonate the amine terminus on Tri. In addition to succinylation of the amine, esterification of hydroxyl groups of the oligosaccharide also occurred but was reversed a t an elevated pH while the succinimide linkage was maintained. Succinyl-Tri was isolated from a polymeric reversed phase column a t '90% yield. The elution of Boc-Tri, Tri, and succinyl-Tri on RPHPLC is shown in Figure 2A-C. Tri was much less hydrophobic than Boc-Tri and eluted earlier, while succinyl-Tri had intermediate hydrophobicity under identical conditions. These differences in properties on RP-HPLC allowed the monitoring of Boc-Tri, Tri, and succinyl-Tri for optimizing reaction conditions and evaluating purity. Proton NMR analysis of the purified Boc-Tri, Tri, and succinyl-Tri showed characteristic signals from anomeric protons on the oligosaccharide (37, 29). In addition, proton signals a t 1.35 ppm from methyl groups on BocTri were absent from Tri and succinyl-Tri, and the latter showed the presence of methylene protons on the succinyl group between 2.30 and 2.45 ppm. FAB-MS analysis provided molecular ions corresponding to M Na for BocTri and Tri (29) and a M 2Na ion for succinyl-Tri (2313.3) which was within 0.5 amu of the calculated molecular weight. Succinyl-Tri was converted to TriPL by coupling its carboxylic group to a n amine side chain of PL after activation with EDC (Figure 1D). The reaction was monitored by gel filtration HPLC. The appearance of a n early eluting product peak having absorbance a t 274 nm was dependent upon the pH, time of reaction, and the
+
+
86
e
C
Succinyl-Tri
D
TriPL
0.01
P
0.w 0.01
J L 0.00
7
5
10
~
T
15
.-
20
5
Elution Time (min)
Figure 2. Analytical RP-HPLC characterization of oligosaccharide conjugates. Purified Boc-Tri (A), Tri (B), and succinylTri (C) (4nmol each) were injected into a C8 silica column (50 "C) equilibrated at 1m u m i n with 0.1% TFA and 1%acetonitrile. Acetonitrile concentration was held at 1%for 5 min, ramped to 20% over 5 min, and kept constant a t 20% for 15 min while Abs274 was monitored at AUFS 0.02. TriPL (D) was chromatographed on the same column using similar conditions, however, with a step gradient: acetonitrile concentration was held constant at 1%for 9 min, stepped upto 25% over 1 min, and kept constant at 25%for another 15 min. Analytical RP-HPLC allowed monitoring of reactions shown in Figure 1 and established the purity of each intermediate.
molar concentrations of succinyl-Tri, PL, and EDC. The reaction was optimized and a succinyl-Tri to PL stoichiometry of 1:6 chosen to allow the reproducible isolation of a 1:l conjugate as determined by amino acid analysis. Increasing the molar ratio (succinyl-Tri to PL) progressively to 4:3 resulted in higher molecular weight conjugates which were resolved on gel filtration HPLC. Amino acid analysis of the isolated products indicated that these contained multiple oligosaccharide units conjugated to each polylysine. Alternatively, decreasing the molar ratio to 1:12 inhibited the reaction. TriPL was isolated with an overall 60% yield (starting from Boc-Tri) on a polymer RP-HPLC column. The purification removed free polylysine and excess EDC which eluted earlier. RP-HPLC of TriPL on a C-8 column is shown in Figure 2D, while gel filtration chromatography is shown in Figure 3. TriPL eluted earlier compared to succinylated oligosaccharide on gel-filtration HPLC due to its increased molecular weight. The monosaccharide composition of TriPL was identical to that of BocTri, and fluorescamine analysis of TriPL using a PL standard indicated 1.03 nmol of polylysine dp 19 for every nmol of tyrosine absorbance. Furthermore, amino acid analysis resulted in a lysine to tyrosine ratio of 20:1, establishing an approximate 1:1conjugate between polylysine and oligosaccharide. A control carrier molecule was prepared by trimming terminal galactose residues from TriPL with P-galactosidase. Agal-TriPL was purified in a manner similar to
Bioconjugate Chem., Vol. 6,No. 3, 1995 287
Targeted Gene Delivery with a Glycopeptide Carrier
I
A Succinyl-Trl
100
.-alC nE.
c)
GP
8o a? 4
10
W
6oB 3' tn
al
t
al
5J .c
1
0
40
E
20
-80
E
f? 3
2
2 0.1 0.05 6
7
8
9
10
11
12
13
I
14
Elution Time (min)
Figure 3. Analytical gel filtration HPLC characterization of TriPL. Succinyl-Tri (A) and TriPL (B) were chromatographed on a gel filtration HPLC column eluted at 1 m u m i n with 50 mM sodium phosphate (pH 4.5) and 300 mM sodium chloride. TriPL (MW: -4700) eluted 40 s earlier than succinyl-Tri (MW 2269) due to its higher molecular weight. Analytical gel filtration HPLC was used to monitor reaction progress of TriPL formation.
TriPL and was found to be devoid of galactose residues upon monosaccharide analysis. DNA Complexation and Transfection. TriPLpCMVL complexes were defined as soluble if centrifugation under the conditions selected failed to remove DNA from the supernatant. This assay established that the solvent used to prepare complexes dramatically influenced solubility. Complexation in 'saline or HBS a t a final DNA concentration of 20 pg/mL led to the formation of a fine precipitate which sedimented upon centrifugation. Alternatively, DNA-carrier complexes prepared in a mannitol solution were more soluble. This effect was unrelated to viscosity. Subsequently, it was found that complexes prepared in low or nonionic conditions did not sediment upon centrifugation, and under these conditions, the presence of mannitol was not necessary for solubility, but its presence avoided hypotonicity. The formation of TriPL-pCMVL precipitates correlated with greatly diminished transfection efficiencies. Complexes prepared in saline or HBS were 5-12% soluble and resulted in gene expression levels 2 orders of magnitude lower than complexes prepared in mannitol which were typically 90% soluble (Figure 4). The complexation of DNA with TriPL was initially evaluated by a band retardation assay (9). A fxed amount of pCMVL was titrated with increasing amounts of TriPL resulting in complexes which showed complete retardation on a 1%agarose gel when the carrier to DNA ratio was 0.2 nmol/pg, which approximates the calculated ratio for neutralization of assumed unit charges on DNA by unit charges on TriPL (Figure 5, inset). Similar titrations were monitored by fluorescence and light scattering assays. The addition of PL or TriPL was found to displace intercalated probe from DNA resulting in a decrease of fluorescence (Figure 5). This assay indicated 45% fluorescence quench a t a carrier-DNA ratio of 0.2 and a 95% quench a t a ratio of 0.8. Alternatively, complexes were monitored by light scattering at 90". Synchronous scans showed that the light scattering spectra were dependent upon the energy
0
Saline
HBS
Mannitol
Figure 4. Effect of solvent on solubility and gene transfection. TriPL-pCMVL complexes were prepared (pCMVL 20 pglmL and TriPL 16 nmoUmL) in saline, HBS, or mannitol solution and were analyzed for solubility and transfection competency in HepG2 cells in the presence of chloroquine as described in the Experimental Procedures. The shaded bars represent percent of DNA remaining in solution after centrifugation, while the solid bars represent luciferase expression obtained a t 24 h.
profile of the source lamp which provided the highest intensity of scattered light a t 350 nm. Complexes prepared with increasing ratios of TriPL to pCMVL resulted in increased light scattering a t 350 nm until an asymptote was reached. Light scattering was nearly a t solvent background a t a carrier-DNA ratio of 0.2 and a t 95% of the asymptote a t a ratio of 0.8 (Figure 5). Light scattering and fluorescence intensity could be measured for the same sample by adjusting slit widths and monochromators since the presence of thiazole orange did not produce any significant difference in the scattered light intensity. HepG2 cells were transfected with complexes containing different ratios of TriPL to pCMVL or PL to pCMVL in either the presence or absence of chloroquine. The level of reporter gene expression was amplified by more than 2 orders of magnitude when the carrier-DNA ratio was increased from 0.2 to 0.8 nmol/pg. Further increase in the carrier-DNA ratio to 1.2 caused only a modest 2-fold increase in reporter gene expression (Figure 6 ) . Gene expression was also observed when TriPL was substituted with PL; however, this was 2 orders of magnitude lower than that observed with TriPL. In either case, a n enhancement of gene expression (1-2 orders of magnitude) was obtained when the transfection was carried out in the presence of chloroquine (Figure 7). On the basis of these results, an optimized ratio of 0.8 nmol of TriPL carrier per pg of DNA (approximate molar ratio of 3000; approximate positive:negative charge ratio of 5) was selected for further experiments. To confirm that the enhancement of gene expression was due to specific recognition of terminal galactose residues on TriPL by the ASGP-R, TriPL was substituted with Agal-TriPL. The resulting gene expression was approximately equivalent to that seen with PL (Figure 7). In addition, TriPL-pCMVL complexes provided the same background gene expression as PL-pCMVL when incubated with HeLa cells, a human cell line lacking the ASGP-R (data not shown). The influence of fetal calf serum (FCS) concentration on transfection in the presence of chloroquine was investigated. Transfections in the presence of serum free
Wadhwa et al.
288 Bioconjugafe Chem., Vol. 6,No. 3, 1995 160
I20
t E
-
1" 9
G 40 /I
-
\
+' - * i _
\
0 I
0.2
0.4
0.6
1.0
0.8
nMol of TriPL per 4 pcMvl
1.2
Figure 5. Assays for monitoring DNA complexation. TriPL-pCMVL complexation in mannitol solution was evaluated at different ratios of carrier to DNA by a band retardation assay on 1% agarose gel (inset), by fluorescence quench of intercalated dye (solid line, circles), and by light scattering of complexes (dashed line, diamonds). The band retardation assay showed complete complexation at a TriPL-oCMVL ratio (nmoVue) of 0.2 or more. However. fluorescence quench and light scattering reached 95% of their asymptotic values at-a ratio of 0.8 or higher.
100 3 c
e 3 al
100 -3
1
3 Y
2 n
8
10
8 a a2 a
$ I4 Q4
0
20 2
0.1
-I-
T
Is 0.01
g
0.005 0.1 0.2 0.4 0.6 0.8 1 1.2 nMol of TriPL per pg pCMVL Figure 6. Influence of carrier-DNA ratios upon gene transfection. Complexes were prepared in mannitol solution at increasing ratios of TriPL to pCMVL (10 pg of DNA), and their transfection competency was analyzed in HepG2 cells in the presence of chloroquine. The bars show luciferase expression at 24 h. A good correlation was observed between complexation level indicated by spectroscopic assays and transfection competency.
media or in 10% FCS resulted in similar luciferase levels (not shown). However incubation in the presence of 2% FCS was chosen since gene expression was 2-fold higher compared to transfection in 10% FCS or serum free media. Although peak level of expression was a t 3 days (not shown), luciferase activity was routinely assayed a t 24 h for rapidity. Dose-response experiments established a nonlinear increase in reporter gene expression with increasing dose
PL
Agd-TriPL
TriPL
Figure 7. Influence of carrier type on receptor-mediated gene delivery to hepatoma cells. Carrier-pCMVL complexes were prepared with 10 pg of pCMVL and 8 nmol of PL, Agal-TriPL, or TriPL in mannitol solution and transfected into HepG2 cells in the absence (shaded bars) and presence (solid bars) of 80 pM chloroquine. Complexes prepared with TriPL showed luciferase expression 2 orders of magnitude above those made with PL (polylysine dp19). However, this enhancement was dependent only on the presence of terminal galactose residues on TriPL which were recognized by the asialoglycoprotein receptor on hepatocytes since Agal-TriPL-pCMVL showed expression similar to PL-pCMVL.
of TriPL-pCMVL up to the maximum of 20 pg of DNA dose tested (Figure 8). In the presence of chloroquine, luciferase expression was observed even with 0.1 pg of pCMVL. Increasing the dose from 1to 10 pg resulted in a 1000-fold increase in luciferase expression, while further increasing the dose to 20 pg led to a 2.5-fold increase in luciferase levels. The highest level of expres-
Targeted Gene Delivety with a Glycopeptide Carrier
U
&
0.01
0.005 1 5 10 20 Dose of p C W (pg) per well
0.1
Figure 8. Effect of T r i P L - p C W dose on reporter gene expression. TriPL-pCMVL (0.1-20 pg of DNA) complexes prepared in mannitol solution were transfected into HepG2 cells at a carrier-DNA ratio 0.8 nmol/pg in the presence of chloroquine and the transfection competency analyzed a t 24 h. Increasing levels of expression were obtained with increasing dose of DNA.
sion seen with 20 pg of DNA in the presence of chloroquine corresponds to a n average of 7 x lo6 light units from the entire 35 mm well (37 x lo6 light units per mg protein). DISCUSSION
One of the major limiting factors for the widespread application of gene therapy is the development of safe, efficient, and well characterized DNA delivery methods for routine transfection in vivo (4-7). Receptor-mediated gene delivery may potentially provide such a method. We have constructed a well-characterized, low molecular weight carrier for receptor-mediated gene delivery. The carrier (TriPL) consists of a galactose terminated triantennary oligosaccharide covalently coupled to low molecular weight polylysine (dp 19). This oligosaccharide was chosen since it can be prepared from bovine fetuin in high yield (291, and it targets hepatocytes in vivo (30) due to its high affinity (Kd= 4 nM) for the ASGP-R. The use of a low molecular weight ligand and anchor creates a distinct advantage in preparation and characterization of carrier conjugates. Purification procedures employed RP-HPLC separations, which gave reproducible and scalable isolation of desired products. In contrast, carriers made with proteinaceous ligands and high molecular weight polylysine pose many difficulties in characterization and purification (22,381. This is in part due to the presence of several reactive groups on both ligand and anchor which makes the conjugation chemistry difficult to control and results in considerable heterogeneity in products. In the case of TriPL, the triantennary oligosaccharide has only one reactive carboxylic group, but it is attached randomly to one of the amines of polylysine. An additional source of heterogeneity is the polydispersity of commercially available polylysine, and the potential to form crosslinks, which is compounded in the case of higher molecular weight polylysines . Complexation between DNA and carrier is a critical parameter for receptor-mediated gene delivery. Unfortunately, complexation also frequently results in precipi-
Bioconjugafe Chem., Vol. 6,No. 3, 1995 289
tation of DNA, limiting the concentration of DNA that can be used (21, 38). Complexation of pCMVL with TriPL in saline or HBS resulted in precipitates a t 20 pg/ mL of DNA, even though TriPL is nearly 50% carbohydrate by weight. These precipitates resulted in transfection levels that were 2 orders of magnitude lower than those observed when soluble complexes were prepared in a solution of mannitol, a nontoxic carbohydrate which can be used in vivo (Figure 4). Further experimentation has demonstrated that DNA-carrier complexes remain soluble in a variety of solutions of very low ionic strength, making possible the use of isotonic vehicles prepared with nonionic excipients like mannitol (unpublished data). The process of complexation between DNA and carrier is usually monitored by a n electrophoretic band retardation assay on a n agarose gel. In order to further characterize DNA complexes, we have used two spectroscopic assays (Figure 5). Dye displacement assays utilize dyes that show increased quantum yield of fluorescence when intercalated within nucleic acids and are usually utilized to monitor binding of intercalating molecules. However, our results demonstrate that this type of assay may also be used to monitor ionic binding of polycations to DNA, even though the modes of binding of dye and polycation are different. Thiazole orange (39) was selected due to its almost nonfluorescent nature when unbound and moderate affinity for DNA which allowed facile displacement of the dye. However, other dyes fulfilling these criteria may also be used, especially in situations where buffer components interfere with the wavelength being monitored. An important consequence of complexation between polycations and DNA is condensation of DNA into compact structures (23, 24). Band retardation assay cannot be used to evaluate DNA condensation. However, total intensity laser light scattering may be used to evaluate extent of condensation while quasielastic laser light scattering may be used for size analysis of condensates (40-41). At the DNA concentrations used in our study, light scattering may result from DNA condensation as well as aggregation. Nevertheless, the results with light scattering intensity measurements complement those obtained with the dye displacement assay. Band retardation on a n agarose gel indicated complete complexation a t a TriPL-pCMVL ratio (nmoL'pg) of 0.2 or more. Fluorescence quench and light scattering assays indicated 95% complexation or condensation at a carrierDNA ratio of 0.8. At the same time, luciferase expression levels were 2 orders of magnitude higher a t a carrierDNA ratio of 0.8 compared to 0.2 (Figure 6). Thus, spectroscopic assays were more predictive of transfection competency compared to band retardation assay. Further, these can be performed in the same buffer or solvent that the complex is prepared in and may allow complexation to be studied under varying conditions. We investigated the specificity of gene delivery based on the ligand recognition characteristics of the ASGP-R (42, 43). The results established that the transfection seen was dependent upon the presence of terminal galactose residues on TriPL (Figure 7). Millimolar concentrations of galactose can inhibit binding of galactose terminated triantennary to the ASGP-R (42);however, expression levels were not inhibited by transfection in the presence of 100 mM free galactose (not shown). This may be due to the very high affinity of complexes for the ASGP-R when multiple oligosaccharide residues are clustered on condensed DNA. Reporter gene expression was enhanced 1-2 orders of magnitude when transfection was conducted in the presence of the lyso-
290 Bioconjugate Chem., Vol. 6,No. 3, 1995
somotropic agent chloroquine, consistent with a receptormediated endocytosis process. From a standpoint of in vivo gene delivery, it is significant that transfection levels with 10% FCS were the same as with serum free media. At the same time, dose response results emphasize the importance of achieving high solubility of plasmid complexes in order to maximize expression levels. The triantennary galactose-terminated oligosaccharide used in this study is a good ligand for the ASGP-R. However, it is possible to create even higher affinity oligosaccharide ligands by enzymatic remodeling (30) which may lead to more efficient gene delivery, especially in a n in vivo situation. There is also potential for substitution with oligosaccharides that are ligands for other cell surface receptors. We have used dp19 (2.5 kDa) polylysine as the DNA anchor for reasons discussed earlier. Most previous studies on receptor mediated gene delivery have used high molecular weight polylysine, and one report cites low transfections levels with transferrin conjugates when polylysines of 30 kDa or lower were used (22). This may be due to the large size of ligand used, requiring comparable size polylysine for stable binding to DNA. Although strict comparisons cannot be made, levels of luciferase expression obtained in this study are similar to those observed in previous studies where high molecular weight carriers were used for chloroquine-enhanced, ASGP-R-mediated gene delivery to HepG2 cells (11,131. The results obtained from this study lead us to conclude that it is possible to achieve reliable receptormediated gene delivery with low molecular weight DNA carriers. Oligosaccharide ligands or synthetic analogs designed to bind with high affinity to target receptors may be linked to polycationic oligopeptides of known sequence to obtain completely defined carrier structures. The use of highly characterized carriers will lead to greater reproducibility in gene delivery and expression. Furthermore, it will allow rational manipulation of carrier design for stability and solubility, modulation of DNA and receptor binding characteristics, and incorporation of additional effector molecules like fusogenic and nuclear targeting peptides. ACKNOWLEDGMENT
We acknowledge the technical assistance of Dr. Charles Cottrel and Dr. David Chang for performing NMR and FAELMS services and Dr. John Lowbridge for amino acid analyses. We thank Dr. Robert Smith-McCollum a t Beckton Dickinson Immunocytometry Systems for providing us with thiazole orange and Dr. M. A. Hickman a t University of California, Davis, for the gift of pCMVL. This work was supported by the American College of Clinical Pharmacy, Amgen Biotechnology Research Award to D.L.K.; a grant from the Developmental Therapeutics Program a t The Ohio State University; and NIH Grants DK45742 and GM48048 to K.G.R. LITERATURE CITED (1) Morgan, R. A., and Anderson, W. F. (1993) Human gene therapy. Annu. Rev. Biochem. 62, 191-217. (2) Roemer, K., and Friedmann, T. (1992) Concepts and strategies for human gene therapy. Eur. J. Biochem. 208, 211225. (3) Miller, A. D. (1992) Human gene therapy comes of age. Nature 357, 455-460. (4) Mulligan, R. C. (1993) The basic science of gene therapy. Science 260, 926-932. (5) Schreier, H. (1994) The new frontier: gene and oligonucleotide therapy. Pharm. Acta Helv. 68, 145-159.
Wadhwa et al. (6) Lyerly, H. K., and DiMaio, J. M. (1993) Gene delivery systems in surgery. Arch. Surg. 128, 1197-1206. (7) Ledley, F. D. (1993) Hepatic gene therapy: present and future. Hepatology 18, 1263-1273. (8) Chang, A. G. Y., and Wu, G. Y. (1994) Gene therapy: Applications to the treatment of gastrointestinal and liver diseases. Gastroenterology 106, 1076-1084. (9) Wu, G. Y., and Wu, C. H. (1988) Evidence for targeted gene delivery to HepG2 hepatoma cells in vitro. Biochemistry 27, 887-892. (10) Wu, G. Y., and Wu, C. H. (1988) Receptor-mediated gene delivery and expression in vivo. J . Biol. Chem. 263, 1462114624. (11) Plank, C., Zatloukal, K., Cotten, M., Mechtler, K., and Wagner, E. (1992) Gene transfer into hepatocytes using asialoglycoprotein receptor mediated endocytosis of DNA complexed with an artificial tetra-antennary galactose ligand. Bioconjugate Chem. 3, 533-539. (12) Cristiano, R. J., Smith, L. C., and Woo, S. L. C. (1993) Hepatic gene therapy: Adenovirus enhancement of receptormediated gene delivery and expression in primary hepatocytes. Proc. Natl. Acad. Sci. U.S.A. 90, 2122-2126. (13) Midoux, P., Mendes, C., Legrand, A., Raimond, J.,Mayer, R., Monsigny, M., and Roche, A. C. (1993) Specific gene transfer mediated by lactosylated poly-L-lysine into hepatoma cells. Nucleic Acids Res. 21, 871-878. (14) Ferkol, T., Lindberg, G. L., Chen, J., Perales, J. C., Crawford, D. R., Ratnoff, 0. D., and Hanson, R. W. (1993) Regulation of the phosphoenolpyruvate carboxykinasehuman factor M gene introduced into the livers of adult rats by receptor-mediated gene transfer. Faseb J . 7, 1081-1091. (15) Haensler, J., and Szoka, F. C., Jr. (1993) Synthesis and characterization of a trigalactosylated bisacridine compound to target DNA to hepatocytes. Bioconjugate Chem. 4,85-93. (16) Huckett, B., Ariatti, M., and Hawtrey, A. 0. (1990) Evidence for targeted gene transfer by receptor-mediated endocytosis. Stable expression following insulin-directed entry of neo into HepG2 cells. Biochem. Pharmacol. 40, 253-263. (17) Rosenkranz, A. A., Yachmenev, S. V., Jans, D. A., Serebryakova, N. V., Murav’ev, V. I., Peters, R., and Sobolev, A. S. (1992) Receptor-mediated endocytosis and nuclear transport of a transfecting DNA construct. Exp. Cell Res. 199,323329. (18) Trubetskoy, V. S., Torchilin, V. P., Kennel, S. J., and Huang, L. (1992) Use of N-terminal modified poly(L4ysine)antibody conjugate as a carrier for targeted gene delivery in mouse lung endothelial cells. Bioconjugate Chem. 3,323-327. (19) Thurnher, M., Wagner, E., Clausen, H., Mechtler, K., Rusconi, S., Dinter, A., Birnstiel, M. L., Berger, E. G., and Cotten, M. (1994) Carbohydrate receptor-mediated gene transfer t o human T leukaemic cells. Glycobiology 4, 429435. (20) Wagner, E., Zenke, M., Cotten, M., Beug, H., and Birnstiel, M. L. (1990) Transferrin-polycation conjugates as carriers for DNA uptake into cells. Proc. Natl. Acad. Sci. 87, 34103414. (21) Cotten, M., Wagner, W., and Birnstiel, M. L. (1993) Receptor-mediated transport of DNA into eukaryotic cells. Meth. Enzymol. 217, 618-644. (22) Taxman, D. J., Lee, E. S., and Wojchowski, D. M. (1993) Receptor-targeted transfection using stable maleimido-transferridthio-poly-L-lysine conjugates. Anal. Biochem. 213,97103. (23) Laemmli, U. K. (1975) Characterization of DNA condensates induced by poly(ethy1ene oxide) and polylysine. Proc. Natl. Acad. Sci. U.S.A. 72, 4288-4292. (24) Wagner, E., Cotten, M., Foisner, R., and Birnstiel, M. L. (1991) Transferrin-polycation-DNA complexes: The effect of polycations on the structure of the complex and DNA delivery to cells. Proc. Natl. Acad. Sci. U.S.A. 88, 4255-4259. (25) Lear, J. D., and DeGrado, W. F. (1987) Membrane binding and conformational properties of peptides representing the NH2 terminus of influenza HA-2. J.Biol. Chem. 262, 65006505. (26) Wagner, E., Plank, C., Zatloukal, K., Cotten, M., and Birnstiel, M. L. (1992) Influenza virus hemagglutinin HA-2 terminal fusogenic peptides augment gene transfer by trans-
Bioconjugafe Chem., Vol. 6,No. 3, 1995 291
Targeted Gene Delivery with a Glycopeptide Carrier ferrin-polylysine-DNA complexes: Toward a synthetic viruslike gene-transfer vehicle. Proc. Natl. Acad. Sei. U.S.A. 89, 7934-7938. (27) Plank, C., Oberhauser, B., Mechtler, K., Koch, C., and Wagner, E. (1994) The influence of endosome-disruptive peptides on gene transfer using synthetic virus-like gene transfer systems. J . Biol. Chem. 269, 12918-12924. (28) Haensler, J., and Szoka, F. C., Jr. (1993) Polyamidoamine cascade polymers mediate efficient transfection of cells in culture. Bioconjugate Chem. 4 , 372-379. (29) Tamura, T., Wadhwa, M. S., and Rice, K. G. (1994) Reducing-end modification of N-linked oligosaccharides with tyrosine. Anal. Biochem. 216, 335-344. (30) Chiu, M. H., Tamura, T., Wadhwa, M. S., and Rice, K. G. (1994) In vivo targeting function of N-linked oligosaccharides with terminating galactose and N-acetylgalactosamine residues. J . Biol. Chem. 269, 16195-16202. (31) Naoi, M., and Lee, Y. C. (1974) A fluorometric measurement of ligands incorporated into BrCn-activated polysaccharides. Anal. Biochem. 57, 640-644. (32) Bidlingmeyer, B. A., Cohen, S. A., and Tarvin, T. L. (1984) Rapid analysis of amino acids using pre-column derivatization. J. Chromatogr. 336, 93-104. (33) Hardy, M. R. (1989) Monosaccharide analysis of glycoconjugates by high-performance anion-exchange chromatography with pulsed amperometric detection. Methods Enzymol. 179, 76-82. (34) de Wet, J. R., Wood, K. V., DeLuca, M., Helsinki, D. R., and Subramani, S. (1987) Firefly luciferase gene: Structure and expression in mammalian cells. Mol. Cell. Biol. 7, 725737. (35) Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual; Cold Spring Harbor Laboratory Press: Plainview, NY.
(36) Bradford, M. M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72,248254. (37) Vliegenthart, J. F. G., Lambertus, D., and van Halbeek, H. (1983) High-resolution, lH-nuclear magnetic resonance spectroscopy as a tool in the structural analysis of carbohydrates related to glycoproteins. Adu. Carbohydr. Chem. Biochem. 41, 209-374. (38) McKee, T. D., DeRome, M. E., Wu, G. Y., and Findeis, M. A. (1994) Preparation of asialoorosomucoid-polylysine conjugates. Bioconjugate Chem. 5, 306-311. (39) Lee, L. G., Chen, C. H., and Chiu, L. A. (1986) Thiazole orange: A new dye for reticulocyte analysis. Cytometry 7, 508-517. (40) Wilson, R. W., and Bloomfield, V. A. (1979) Counterioninduced condensation of deoxyribonucleic acid. A light scattering study. Biochemistry 18, 2192-2196. (41) Bloomfield, V. A., He, S., Li, A. Z., and Arscott, P. B. (1991) Light scattering studies on DNA condensation. Biochem. SOC. Trans. 19, 496. (42) Lee, Y. C., Townsend, R. R., Hardy, M. R., Lonngren, J., Arnarp, J., Haraldsson, M., and Lonn, H. (1983) Binding of synthetic oligosaccharides to the hepatic gaVgalNAc lectin. Dependence on fine structural features. J . Biol. Chem. 258, 199-202. (43) Rice, K. G., Weisz, 0. A,, Barthel, T., Lee, R. T., and Lee, Y. C. (1990) Defined geometry of binding between triantennary glycopeptide and the asialoglycoprotein receptor on rat hepatocytes. J . Biol. Chem. 265, 18429-18434. BC9500103
