Liposome-entrapped plasmid DNA: characterisation studies

Liposome-entrapped plasmid DNA: characterisation studies

Biochimica et Biophysica Acta 1475 (2000) 125^132 www.elsevier.com/locate/bba Liposome-entrapped plasmid DNA: characterisation studies Yvonne Perrie...

551KB Sizes 1 Downloads 65 Views

Biochimica et Biophysica Acta 1475 (2000) 125^132

www.elsevier.com/locate/bba

Liposome-entrapped plasmid DNA: characterisation studies Yvonne Perrie a

a;b

, Gregory Gregoriadis

a;

*

Centre for Drug Delivery Research, School of Pharmacy, University of London, 29^39 Brunswick Square, London WC1N 1AX, UK b Lipoxen Ltd, 29^39 Brunswick Square, London WC1N 1AX, UK Received 18 November 1999 ; received in revised form 25 February 2000 ; accepted 28 March 2000

Abstract Plasmid DNA pRc/CMV HBS (5.6 kb) (100 Wg) encoding the S (small) region of hepatitis B surface antigen was incorporated by the dehydration^rehydration method into liposomes composed of 16 Wmol egg phosphatidylcholine (PC), 8 Wmol dioleoylphosphatidylcholine (DOPE) and 1,2-diodeoyl-3-(trimethylammonium)propane (DOTAP) (cationic liposomes) or phosphatidylglycerol (anionic liposomes) in a variety of molar ratios. The method, entailing mixing of small unilamellar vesicles (SUV) with the DNA, followed by dehydration and rehydration, yielded incorporation values of 95^97 and 48^54% of the DNA used, respectively. Mixing of preformed cationic liposomes with 100 Wg plasmid DNA also led to high complexation values of 73^97%. As expected, the association of DNA with preformed anionic liposomes was low (9%). Further work with cationic PC/DOPE/DOTAP liposomes attempted to establish differences in the nature of DNA association with the vesicles after complexation and the constructs generated by the process of dehydration/rehydration. Several lines of evidence obtained from studies on vesicle size and zeta-potential, fluorescent microscopy and gel electrophoresis in the presence of the anion sodium dodecyl sulphate (SDS) indicate that, under the conditions employed, interaction of DNA with preformed cationic SUV as above, or with cationic SUV made of DOPE and DOTAP (1:1 molar ratio; ESCORT Transfection Reagent), leads to the formation of large complexes with externally bound DNA. For instance, such DNA is accessible to and can be dissociated by competing anionic SDS molecules. However, dehydration of the DNA^SUV complexes and subsequent rehydration, generates submicron size liposomes incorporating most of the DNA in a fashion that prevents DNA displacement through anion competition. It is suggested that, in this case, DNA is entrapped within the aqueous compartments, in between bilayers, presumably bound to the cationic charges. ß 2000 Elsevier Science B.V. All rights reserved. Keywords : Liposome ; DNA vaccine; Plasmid DNA; Hepatitis B surface antigen; Microelectrophoresis; Gel electrophoresis

1. Introduction Numerous studies [1^3] have shown that the immunisation of experimental animals with naked plasmid DNA encoding antigens from a variety of bacteria, viruses, protozoa and cancers, induces humoral and cell-mediated protective immunity. This novel approach to immunisation, i.e. de novo production of vaccines by the host's cells in vivo, promises to revolutionise vaccinology and a number of clinical trials are already in progress [1]. It is widely accepted [1^3] that injected naked DNA, usually by the intramuscular route, is taken up by myocytes which it transfects episomally. Produced antigen is then subjected to pathways that are similar to those undergone by the antigens of internalised viruses, leading to such types of immunity as cytotoxic T-lymphocyte response. However,

* Corresponding author. Fax: +44-171-753-5820; E-mail : [email protected]

uptake of DNA by myocytes is minimal [2] and involves only a minor fraction of the cells. It is also likely that some of the injected DNA su¡ers deoxyribonuclease attack. Moreover, although myocytes carry MHC class I molecules, they are not professional antigen-presenting cells (APC) as they lack vital costimulatory molecules [1,2]. It is thought that responses to genetic vaccines are, at least in part, the result of transfer of antigenic material from myocytes to professional APC. Work from this laboratory [3^6] has recently shown that APC might be a preferred alternative to myocytes as targets for the uptake and expression of DNA vaccines and to that end, liposomes would be a suitable means to deliver these to such cells: locally injected liposomes are endocytosed rapidly by APC in¢ltrating the site of injection or in the draining lymph nodes [7,8] and are also expected [9] to protect their DNA content from deoxyribonuclease. Indeed, in work [3,4] with pRc/CMV HBS (encoding the S (small) region of the hepatitis B surface antigen ; HBsAg), intramuscular or subcutaneous injection of the plasmid

0304-4165 / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 4 1 6 5 ( 0 0 ) 0 0 0 5 5 - 6

BBAGEN 25017 18-5-00

Cyaan Magenta Geel Zwart

126

Y. Perrie, G. Gregoriadis / Biochimica et Biophysica Acta 1475 (2000) 125^132

entrapped in cationic liposomes by the dehydration/rehydration method (whereby unilamellar vesicles mixed with DNA are dehydrated and then rehydrated to generate multilamellar vesicles) elicited much greater humoral (IgG subclasses and splenic interleukin-4) and cell-mediated (splenic interferon-Q) immune responses in mice than naked DNA or DNA complexed to similar, preformed liposomes. Here we present data on the physicochemical characteristics of DNA-containing liposomes previously used successfully in DNA immunisation. It appears from studies on vesicle size and zeta-potential, £uorescent microscopy and anion competition that, in contrast to DNA^vesicle complexes where DNA is external, DNA is rendered internal in the multilamellar liposomes generated by dehydration and rehydration of such complexes.

non-complexed material and resuspended in 0.01 M sodium phosphate containing 0.15 M NaCl, pH 7.4 (phosphate-bu¡ered saline, PBS) to the required volume. DNA entrapment into liposomes (DRV(DNA)) or complexation with their surface (DRV^DNA) was estimated on the basis of 35 S radioactivity recovered in the suspended pellets. In some experiments DRV(DNA) or DRV^DNA were made as above in the presence of £uorescent pRc/CMV HBS. In others, cationic SUV (made as above from 0.21 Wmol DOPE and 0.21 Wmol DOTAP) of similar size (100^ 150 nm) and identical composition to those of ESCORT Transfection Reagent (DOPE and DOTAP, 1:1 molar ratio; Sigma) were incubated at 20³C for 30 min with 200 Wg of DNA and used as such (SUV^DNA) or freeze-dried to prepare DRV(DNA) as above. 2.2. Measurement of vesicle size and zeta-potential

2. Materials and methods Egg phosphatidylcholine (PC) was purchased from Lipid Products, Nut¢eld, Surrey, UK. Cholesterol (CHOL), dioleoyl phosphatidylethanolamine (DOPE) and phosphatidylglycerol (PG) were from Sigma, Poole, Dorset, UK and 1,2-dioleoyl-3-(trimethylammonium) propane (DOTAP) from Avanti Polar Lipids, AL, USA. Polylysine (average molecular weight 22 000) was purchased from Sigma, Poole, Dorset, UK. Plasmid pRc/CMV HBS (5.6 kb) expressing the sequence coding for the S region of HBsAg (subtype ayw) was cloned by Dr. R. Whalen using pRc/CMV (Invitrogen) as a vector backbone. The pRc/ CMV HBS was radiolabelled with 35 S-dATP by the method of Wheeler and Coutelle [10] and in some experiments, coupled by the same method [10] to the £uorescent marker FluorX dATP from Amersham Life Science, Amersham Place, Buckinghamshire, UK. All other reagents were of analytical grade. 2.1. Preparation of plasmid DNA-containing liposomes The dehydration^rehydration procedure [9,11,12] was used for the incorporation of pRc/CMV HBS into liposomes. In brief, 2 ml of small unilamellar vesicles (SUV) prepared by sonication and composed of 16 Wmol PC, 8 Wmol DOPE or CHOL (molar ratio 1:0.5) and 0^16 Wmol of the cationic DOTAP or the anionic PG were mixed with 100 Wg of plasmid DNA (and 35 S-labelled tracer of the same plasmid) or with 100 Wg polylysine (anionic SUV only), frozen at 320³C and freeze-dried overnight. Controlled [9,11,12] rehydration of the dry powders led to the formation of dehydration^rehydration multilamellar [9] vesicles (DRV liposomes). For the preparation of DRV^DNA complexes, DRV prepared as above in the absence of DNA, were incubated at 20³C for 30 min with 100 Wg plasmid DNA mixed with radiolabelled plasmid. DRV preparations were then centrifuged twice at 25 000Ug for 40 min to remove non-entrapped or

BBAGEN 25017 18-5-00

The z-average diameter of liposomes was measured on an Autosizer 2c by photon correlation spectroscopy (PCS) (DRV or DRV(DNA)) or on a Malvern Mastersizer (DRV^DNA) at 20³C by diluting 20 Wl of the dispersion to the appropriate volume with doubly-¢ltered (0.22 Wm pore size) distilled water. The zeta-potential, which is an indirect measurement of the vesicle surface charge, was measured in 0.001 M PBS at 25³C on a zetasizer 3000. 2.3. Fluorescence microscopy Fluorescence photomicrographs of liposome-entrapped or liposome surface-complexed £uorescent DNA were recorded using a Nikon Microphot FXA on a 340^380 nm wavelength band. 2.4. Agarose-gel electrophoresis Samples of liposome-entrapped or liposome surfacecomplexed DNA were subjected to agarose gel (1.0%) electrophoresis [9] to determine the retention of DNA by the liposomes. In brief, 8 Wl (1.6^2.0 Wg, DNA) of DRV or SUV suspension were mixed with 4 Wl gel loading bu¡er (bromophenol blue 0.05% w/v; sodium dodecyl sulphate, 0.05% w/v; sucrose, 40% w/v; EDTA, 0.1 M; pH 8) and subjected to agarose gel electrophoresis in the presence of ethidium bromide (0.5 Wg/ml) for 1 h at 90 V. In some experiments both bromophenol blue and sodium dodecyl sulphate were omitted from the medium. DNA visualisation of the gels was carried out using a UV lamp. 3. Results and discussion 3.1. Incorporation of DNA into liposomes pRc/CMV HBS values of complexation (% of total used) with preformed cationic DRV PC/DOPE liposomes

Cyaan Magenta Geel Zwart

Y. Perrie, G. Gregoriadis / Biochimica et Biophysica Acta 1475 (2000) 125^132

127

Table 1 Incorporation of DNA into cationic liposomes DOTAP content (Wmol)

DNA incorporation (% of amount used)

1 2 4 8 12 16

DRV^DNA complexes

DRV with entrapped DNA

73.0 þ 14.0 95.0 þ 4.6 96.9 þ 3.7 92.7 þ 6.3 96.0 þ 3.5 97.5 þ 1.2

95.9 þ 2.5 95.7 þ 2.8 94.0 þ 2.8 96.3 þ 2.0 96.6 þ 3.9 95.5 þ 4.9

35

S-labelled DNA (pRc/CMV HBS; 100 Wg) was either complexed (DRV^DNA) with or, entrapped (DRV(DNA)) in liposomes prepared by the dehydration^rehydration procedure and composed of 16 Wmol PC, 8 Wmol DOPE and 1^16 Wmol DOTAP. Values denote mean þ S.D. from ¢ve experiments.

incorporating increasing amounts of DOTAP were 73% (of the total used) for 1 Wmol DOTAP but increased to 95^97.5% (P 6 0.05) for greater amounts of DOTAP (2^16 Wmol DOTAP) (Table 1), presumably because DOTAP concentration on the DRV surface was su¤ciently high to complex most of the DNA present. In contrast, entrapment of DNA in DRV liposomes was quantitative (94^ 97%) for all DOTAP contents (Table 1). This could be attributed to the accessibility of much more cationic lipid during the process of dehydration of the cationic SUV^ DNA mixture to form, on rehydration, multilamellar vesicles with the DNA mostly distributed in the inner bilayers. It was previously shown [9] that entrapment of the same plasmid DNA (100 Wg) in neutral (DOTAP-free) liposomes was signi¢cant (48.3%) when the dehydration^ rehydration method was employed. On the other hand, when neutral preformed PC/DOPE DRV were incubated with DNA, association with the vesicles (presumably adsorbed to their surface) was low (9.8%) [9]. Similar observations were made with anionic PC/DOPE liposomes (Table 2) incorporating increasing amounts of PG (0.5^16 Wmol), with DNA entrapment values ranging from 48 to 57%. As expected [9], incubation of preformed anionic DRV (0.5^16 Wmol PG) with DNA led to low values (9.2%) of DNA adsorption to the liposomal surface (results not shown).

DRV prepared in the absence or presence of 100 Wg DNA is shown in Fig. 1. Results indicate that, at a low DOTAP content (1 Wmol), the size of DRV devoid of DNA although considerable (1200 nm), is far lower than that of the same DRV incorporating DNA (DRV(DNA)) where the size is well outside the range of measurement by PCS. However, as the DOTAP content increases, vesicle size is reduced for both preparations, with values becoming similar (580^700 nm) at 4 Wmol DOTAP. The same trend of size reduction for both the empty DRV and DRV(DNA) with increasing DOTAP content is also observed (Table 3) when DOPE is replaced by CHOL although vesicle sizes are greater at all DOTAP contents studied (compare Table 3 and Fig. 1). Again, as with the DOPE DRV(DNA) (Fig. 1), CHOL DRV(DNA) vesicles are signi¢cantly larger than empty DRV of the same composition (Table 3). The e¡ect of vesicle charge on reducing the size of DRV formed by the method of dehydration^

3.2. Vesicle size : the e¡ect of lipid components The e¡ect of DOTAP content on the size (diameter) of Table 2 Incorporation of DNA into anionic liposomes PG content (Wmol)

DNA incorporation (% of amount used) PC:DOPE:PG

PC:CHOL:PG

0.5 1 4 16

55.0 þ 5.2 53.5 þ 2.7 54.3 þ 9.7 56.8 þ 3.9

52.5 þ 3.2 48.3 þ 1.4 47.8 þ 4.4 54.1 þ 5.0

35

S-labelled DNA (pRc/CMV HBS; 100 Wg) was entrapped in DRV liposomes composed of 16 Wmol PC, 8 Wmol DOPE or CHOL and 0.5^ 16 Wmol PG. Values denote mean þ S.D. from at least three experiments.

BBAGEN 25017 18-5-00

Fig. 1. z-Average diameter of cationic PC/DOPE `empty' DRV and DRV(DNA) prepared in the presence of 100 Wg DNA. Vesicle z-average diameter was determined in an Autosizer 2c at 20³C. Values denote mean þ S.D. (n = 4). The size of DRV(DNA) at 1 Wmol DOTAP was outside the range of measurement. Di¡erences between DRV(DNA) and empty DRV values were signi¢cant for 2 and 3 Wmol DOTAP (P 6 0.004). The sizes of DRV^DNA complexes were measured in a Malvern Mastersizer.

Cyaan Magenta Geel Zwart

128

Y. Perrie, G. Gregoriadis / Biochimica et Biophysica Acta 1475 (2000) 125^132

Fig. 3. Zeta-potential of DRV liposomes. `Empty' DRV, DRV^DNA and DRV(DNA) composed of 16 Wmol PC, 8 Wmol DOPE and 1^16 Wmol DOTAP were prepared in the presence of 100 Wg DNA where appropriate and subjected to microelectrophoresis. Values denote mean þ S.D. from ¢ve measurements. *Indicates statistically signi¢cant (P 6 0.03^0.0001) di¡erences in values between empty DRV and DRV(DNA) or DRV^DNA. Di¡erences between DRV(DNA) and DRV^DNA were signi¢cant as follows for the Wmol of DOTAP shown in parentheses: P 6 0.001 (1), P 6 0.01 (2), P 6 0.035 (4) and P 6 0.032 (8).

Fig. 2. z-Average diameter of anionic `empty' DRV, DRV(DNA) prepared in the presence of 100 Wg DNA, and DRV(polylysine) prepared in the presence of 100 Wg polylysine. DRV were composed of 16 Wmol PC, 8 Wmol DOPE and 0.5^16 Wmol PG. Vesicle z-average diameter was determined by PCS at 20³C. Results denote mean þ S.D. (n = 4).

rehydration, could be related to the fusion processes present during the procedure. It is conceivable that charged SUV surfaces repel each other su¤ciently so as to interfere with e¡ective membrane fusion, thus leading to smaller DRV. We are not able at present to explain the substantial increase in DRV size when DOPE in the precursor anionic or cationic SUV is replaced by CHOL (Table 3). In contrast, the sizes of the complexes formed on mixing preformed cationic DOPE DRV (Fig. 1) or CHOL DRV (legend to Table 3) and 100 Wg DNA remained outside the range of measurement by PCS (5^20 Wm as measured in the Malvern Mastersizer) regardless of the DOTAP content. Similarly to the cationic DRV (Fig. 1) and probably for the same reasons outlined above for cationic DRV, the size of anionic (PG) DNA-free PC/DOPE DRV (Fig. 2) or PC/CHOL DRV (Table 3) is also reduced as the PG content increases. However, in contrast to the cationic DRV, the presence of entrapped DNA does not in£uence vesicle size (Table 3 and Fig. 2). Moreover, as with cationic DRV, substitution of DOPE with CHOL in anionic DRV results in greater sizes at all PG contents (compare Table 3 and Fig. 2). Interestingly, when the cationic polylysine is entrapped in anionic DRV, the pattern of vesicle size changes with increasing PG content (Fig. 2) is similar to that observed with the cationic DRV entrap-

ping the anionic DNA (Fig. 1): whereas at 1 Wmol PG the size of DRV(polylysine) is outside the range of measurement by PCS, size is gradually reduced to become similar to that of empty (anionic) DRV at a PG content above 3 Wmol. Taken together, these results indicate that, within the range of cationic lipid (DOTAP) content (4^16 Wmol for 16 Wmol PC and 8 Wmol DOPE or CHOL) used here, incorporation of 100 Wg pRc/CMV HBS into DRV by the dehydration^rehydration procedure leads to the formation of DRV(DNA) constructs that are much smaller (micron or submicron size) than the DRV^DNA complexes seen on mixing the same amount of DNA with preformed DRV of identical lipid composition and concentration. Additional work (results not shown) has revealed that the use of higher amounts of pRC/CMV HBS (e.g. up to 800 Wg DNA) for entrapment into PC/ DOPE cationic DRV with a low (e.g. 2 Wmol) DOTAP content leads to substantial increases in the size (up to 30 Wm) of the DRV(DNA) constructs formed. However, this can be avoided by employing a higher DOTAP (e.g. 8 Wmol) content whereupon, submicron DRV(DNA) constructs are obtained with a high ( s 90% of starting material) plasmid content. Assuming an average DNA base

Table 3 z-Average diameter (nm) of CHOL-containing DRV DOTAP or PG content (Wmol) 2 4 8

PC:CHOL :DOTAP

PC:CHOL :PG

DRV

DRV(DNA)

DRV

DRV(DNA)

1070 þ 122 922 þ 90a 732 þ 123b

outside range 1290 þ 125c 1154 þ 147d

1181 þ 287 821 þ 136 705 þ 40

1040 þ 277 913 þ 70 693 þ 28

Cationic and anionic `empty' DRV and DRV(DNA) were prepared from 16 Wmol PC, 8 Wmol CHOL and 2, 4 or 8 Wmol of either DOTAP (cationic DRV) or PG (anionic DRV). z-Average diameters were measured using an Autosizer 2c where possible. Results denote mean þ S.D. (n = 4). Entrapment of DNA, based on 35 S assay of radiolabelled DNA, was 90^95% (cationic) and 50^60% (anionic DRV) of total (100 Wg) DNA used. The size of complexes of preformed cationic DRV with DNA as measured in a Malvern Mastersizer were 5^10 Wm (results not shown). c vs a, P 6 0.003 (n = 4); d vs b, P 6 0.005 (n = 4). There was no signi¢cant di¡erence between sizes of anionic DRV and DRV(DNA) at all PG contents.

BBAGEN 25017 18-5-00

Cyaan Magenta Geel Zwart

Y. Perrie, G. Gregoriadis / Biochimica et Biophysica Acta 1475 (2000) 125^132

129

residue weight of 320 (and 320 nmol phosphate group charges per 100 Wg DNA), it is of interest that the ratios of positive to negative (+/3) charges used to obtain in the present work the submicron size DRV(DNA) constructs range from 6.25 to 50.00 (for 2 to 16 Wmol DOTAP). These +/3 values are much higher than those (e.g. 0.08^ 2.6) used by others [13^16] in studies with lipoplexes and may explain as to why the latter are often di¤cult to control in terms of size. 3.3. Zeta-potential studies Values of the zeta-potential of liposomes indirectly re£ect vesicle surface net charge and can therefore be used to evaluate the extent of interaction of the liposomal surface cationic charges with the anionic charges of DNA. On this basis, the question of entrapped versus complexed DNA in the PC/DOPE DRV(DNA) cationic liposomes was investigated using DRV with entrapped DNA and preformed DRV before and after complexing with DNA. Results in Fig. 3 show trends of increasing zetapotential values with increasing DOTAP content (1^16 Wmol) for all preparations, with absolute values (for up to 8 Wmol DOTAP) being in the order of `empty' DRV s DRV(DNA) s DRV^DNA. As similar amounts of DNA were present in DRV(DNA) (94^96 Wg) and DRV^DNA (73^96 Wg), results suggest that in the latter case there is more DNA on the vesicle surface (thus neutralising more of the cationic charges) and that, with DRV(DNA), some of the DNA is located within the liposomes, presumably bound to cationic charges hidden in the inner bilayers. However, with DOTAP content at 12 Wmol or more (Fig. 3), the amount of DNA present in DRV(DNA) and DRV^DNA is probably too low (relative to DOTAP) to cause measurable di¡erences in their zeta-potential values. The same order of zeta-potential values (i.e. empty DRV s DRV(DNA) s DRV^DNA) was observed when DOPE in PC/DOPE DRV was replaced with CHOL, although values in this case were much greater (Fig. 4;

Fig. 4. Zeta-potential of DRV liposomes. `Empty' DRV, DRV^DNA and DRV(DNA) composed of 16 Wmol PC, 8 Wmol CHOL and 0.5^2 Wmol DOTAP were prepared in the presence of 100 Wg DNA where appropriate and subjected to microelectrophoresis. Results denote mean þ S.D. from ¢ve measurements.

BBAGEN 25017 18-5-00

Fig. 5. Zeta-potential of empty DRV liposomes composed of 0^4 Wmol DOTAP and 24 Wmol PC (PC:DOTAP), 16 Wmol PC and 8 Wmol DOPE (PC:DOPE:DOTAP), 16 Wmol PC and 8 Wmol CHOL (PC:CHOL :DOTAP) or 16 Wmol PC, 4 Wmol DOPE and 4 Wmol CHOL (PC:DOPE:CHOL :DOTAP). All four preparations exhibited similar zeta-potential values (33.5 to 36.3 mV) in the absence of DOTAP. Results denote mean þ S.D. from ¢ve measurements.

0.5^2 Wmol DOTAP) than those seen when DOPE was present (Fig. 3; 1^2 Wmol DOTAP). This could be attributed either to the zwitterionic nature of DOPE (at pH 7.4) which, when present, would lead to the masking of some of the cationic charges of DOTAP or to CHOL in some way rendering such charges more available for measurement, or both. As with the PC/DOPE DRV, there was no di¡erence in the zeta-potential values for the CHOL-containing PC DRV(DNA) and DRV^DNA when the DOTAP content increased above a certain level (e.g. 51.6 þ 1.0 mV for 4 Wmol DOTAP content with both formulations ; results not shown). The role of the zwitterionic DOPE and of CHOL on the zeta-potential values was further clari¢ed in experiments (Fig. 5) where DNA-free cationic PC DRV as such (PC/DOTAP) or supplemented with DOPE, CHOL or both lipids, were measured for zeta-potential values. Results clearly show that for all DOTAP contents (1^4 Wmol) studied, values are decreased or increased by the presence of DOPE and CHOL, respectively, whereas in the presence of both lipids values remain unaltered (i.e.

Fig. 6. Zeta-potential of `empty' DRV liposomes composed of 0^2 Wmol PG and 24 Wmol PC (PC:PG), 16 Wmol PC and 8 Wmol DOPE (PC:DOPE :PG), 16 Wmol PC and 8 Wmol CHOL (PC:CHOL :PG) or 16 Wmol PC, 4 Wmol DOPE and 4 Wmol CHOL (PC:DOPE : CHOL :PG). All four preparations exhibited similar zeta-potential values (33.5 to 36.3 mV) in the absence of PG. Results denote mean þ S.D. from ¢ve measurements.

Cyaan Magenta Geel Zwart

130

Y. Perrie, G. Gregoriadis / Biochimica et Biophysica Acta 1475 (2000) 125^132

Fig. 7. Fluorescence microscopy studies. (A) DNA (pRc/CMV HBS) entrapped in cationic DRV composed of 16 Wmol PC, 8 Wmol DOPE and 4 Wmol DOTAP. (B) DNA complexed with preformed cationic DRV of the same lipid composition as in (A). (C) and (D) DNA complexed with cationic SUV of the same lipid composition as in (A). Bar represents 2 Wm.

similar to those of PC/DOTAP DRV), probably because the two lipids cancel each other (Fig. 5). The e¡ect of CHOL on the net charge of the liposomal surface was also observed with DNA-free DRV preparations in which DOTAP was substituted with the anionic PG (Fig. 6): addition of CHOL to DRV led to greater availability of the ionic charges and reduction of the zeta-potential values for all PG contents studied. Again, there was no change when both CHOL and DOPE were present. 3.4. Fluorescence microscopy studies The results on DNA incorporation with the neutral (see above) and anionic (Table 2) PC/DOPE DRV strongly suggest that most of the pRc/CMV HBS plasmid is entrapped within the aqueous spaces of the multilamellar [11] vesicles rather than being absorbed to their surface. In the case of cationic PC/DOPE DRV, however, because of the similarity of DNA association values for the DRV(DNA) and the DRV^DNA preparation (Table 1), actual DNA entrapment within the vesicles (as opposed to

BBAGEN 25017 18-5-00

surface complexation) cannot be ascertained. On the other hand, the observed submicron vesicle size of DRV(DNA) with increased DOTAP content ( s 4 Wmol) as opposed to the much larger size of DRV^DNA complexes under the same conditions (Fig. 1 and legend), as well as di¡erences in zeta-potential values (Figs. 3 and 4) are indicative of di¡erences in the mode of DNA interaction with the cationic charges of the two constructs. This was further substantiated when liposomes with `entrapped' or complexed £uorescent DNA were examined under £uorescence microscopy (Fig. 7): in contrast to the relatively small £uorescent particles observed in the case of DRV(DNA) (Fig. 7a), DRV^DNA complexes were seen as large aggregates of various sizes (Fig. 7b). Similarly large aggregates (Fig. 7d) mixed with smaller aggregates (Fig. 7c) were also observed with the SUV^DNA complexes used for the preparation of DRV(DNA), strongly suggesting a re-organisation of these complexes during the dehydration^ rehydration procedure to produce discrete DRV(DNA) vesicles constructs as seen in Fig. 7a rather than the DRV^DNA complexes of Fig. 7b.

Cyaan Magenta Geel Zwart

Y. Perrie, G. Gregoriadis / Biochimica et Biophysica Acta 1475 (2000) 125^132

131

Fig. 8. (a) Gel electrophoresis of DRV(DNA) (lanes 1 and 3) and DRV^DNA (lanes 2 and 4) composed of 16 Wmol PC, 8 Wmol DOPE and 1 (lanes 1 and 2) or 2 Wmol (lanes 3 and 4) DOTAP. Lane C, naked DNA. (b) As in (a) but in the presence of the anionic SDS. (c) Gel electrophoresis of DRV(DNA) (lanes 1 and 3) and SUV^DNA (lanes 2 and 4) both composed of 16 Wmol PC, 8 Wmol DOPE and 1 (lanes 1 and 2) or 2 Wmol (lanes 3 and 4) DOTAP. Lane C, naked DNA. (d) As in (c) but in the presence of the anionic SDS. (e) Gel electrophoresis of DRV(DNA) (lane 1) composed of 0.21 Wmol DOPE and 0.21 Wmol DOTAP and the precursor SUV^DNA (lanes 2 and 4) (used to prepare the DRV(DNA) of lane 1; see Section 2). Lanes 1 and 2: SDS present ; lanes 3 and 4: no SDS. For other details see Section 2.

3.5. Gel electrophoresis studies The spatial localisation of DNA within the cationic DRV(DNA) and DRV^DNA preparations was also investigated by subjecting these to gel electrophoresis in the presence of sodium dodecyl sulphate (SDS) at a concentration (0.05%) below the critical micelle concentration of the surfactant. It was anticipated that the anionic SDS, although unable to solubilise the liposomal lipids, would be able to partition into the outer monolayer of the constructs, compete with DNA bound to the cationic surface charges and release it into the medium. Released DNA would then be expected to migrate towards the cathode, with DNA unavailable (presumably entrapped) to SDS remaining at the site of application. Fig. 8a shows that, on gel electrophoresis of PC/DOPE DRV(DNA) and DRV^DNA in the absence of anionic molecules, DNA remains at the site of application, bound to the cationic charges of the preparations. In contrast, following electrophoresis in the presence of SDS (Fig. 8b) displaced DNA is seen to migrate towards the cathode. As expected, much more DNA is displaced in the DRV^DNA complexes (Fig. 8b, lanes 2 and 4) than the DRV with the entrapped DNA (lanes 1 and 3). DNA displacement in the presence of SDS was also shown to occur in DNA complexes with the cationic PC/DOPE SUV (Fig. 8d, lanes 2 and 4) used as precursors for the

BBAGEN 25017 18-5-00

generation of the multilamellar cationic DRV as well as DNA complexes with the ESCORT Transfection Reagent type DOPE/DOTAP SUV (Fig. 8e, lane 2) but not in its absence (Fig. 8c, lanes 2 and 4, and e, lane 4, respectively). Again, there was much less DNA displacement with DRV made from such (precursor) vesicles (Fig. 8d, lanes 1 and 3, and e, lane 1, respectively). It has been suggested [17^20] that cationic DOPE SUV (such as the ESCORT Transfection Reagent vesicles used here) incubated with appropriate amounts of plasmid DNA to give cationic lipid/ DNA charge ratios that are similar to those in the present work, form larger complexes (lipoplexes) in which DNA is localized within bilayers, bound electrostatically to the cationic charges. However, the observation (Fig. 8d, lanes 2 and 4, and e, lane 2) that such lipoplexes lose most of their DNA content on electrophoresis in the presence of SDS, indicates that the structural characteristics of lipoplexes are such that SDS is allowed to access DNA and displace most of it. As this does not occur with ESCORT vesicle-derived DOPE DRV(DNA) (Fig. 8e, lane 1) or with PC/DOPE SUV-derived DRV(DNA) (Fig. 8d, lanes 1 and 3), such DRV are likely to incorporate most of their DNA within closed bilayers, probably bound to the inner cationic charges. In conclusion, several lines of evidence including vesicle size and vesicle surface zeta-potential measurements, morphological observations (£uorescent microscopy) and

Cyaan Magenta Geel Zwart

132

Y. Perrie, G. Gregoriadis / Biochimica et Biophysica Acta 1475 (2000) 125^132

anion competition experiments (gel electrophoresis), strongly suggest that the localisation of plasmid DNA in the cationic multilamellar DRV liposomes produced by the dehydration^rehydration procedure is di¡erent not only from that obtained on DNA complexation with preformed cationic DRV but also from SUV^DNA complexes that are used as precursors for the production of cationic DRV. It appears that freeze-drying of these large (Fig. 7c,d) SUV^DNA complexes (or lipoplexes) and subsequent rehydration results in smaller structures (Fig. 7a) where DNA is predominantly entrapped within the bilayers, presumably bound to the inner cationic charges. Some of the DNA also interacts with surface charges and it is probably this minor portion of DNA that is displaced by SDS on gel electrophoresis (e.g. Fig. 8e, lane 1). In contrast, a much greater quantity of DNA is displaced in the DRV^DNA (Fig. 8b, lanes 2 and 4) or SUV^DNA (e.g. Fig. 8d, lanes 2 and 4, and e, lane 2) complexes, an event that indicates predominant DNA binding to external surfaces. Acknowledgements This work was supported by a European Community grant BI04 CT972191.

[2] H.L. Davis, R.G. Whalen, B.A. Demeneix, Hum. Gene Ther. 4 (1993) 151^156. [3] G. Gregoriadis, Pharm. Res. 15 (1998) 661^670. [4] G. Gregoriadis, R. Sa¤e, B. deSouza, FEBS Lett. 402 (1997) 107^ 110. [5] Y. Perrie, G. Gregoriadis, J. Liposome Res. 8 (1998) 95^96. [6] B. McCormack, G. Gregoriadis, J. Liposome Res. 8 (1998) 48^49. [7] A. Tumer, C. Kirby, J. Senior, G. Gregoriadis, Biochim. Biophys. Acta 760 (1983) 119^125. [8] M. Velinova, N. Read, C. Kirby, G. Gregoriadis, Biochim. Biophys. Acta 1299 (1996) 207^215. [9] G. Gregoriadis, R. Sa¤e, S.L. Hart, J. Drug Target. 3 (1996) 469^ 475. [10] V.C. Wheeler, C. Coutelle, Anal. Biochem. 225 (1995) 374^376. [11] G. Gregoriadis, B. McCormack, Y. Morrison-Perrie, R. Sa¤e, B. Zadi, in: J. Celis (Ed.), Cell Biology: a Laboratory Handbook, Academic Press, Orlando, CA, 1998, pp. 131^139. [12] C. Kirby, G. Gregoriadis, Biotechnology 2 (1984) 979^984. [13] B. Sternberg, F.L. Sorgi, L. Huang, FEBS Lett. 356 (1994) 361^366. [14] J.O. Ra«dler, I. Koltover, T. Salditt, C.R. Sa¢nya, Science 275 (1997) 810^814. [15] H. Gershon, R. Ghirlando, S.B. Guttman, A. Minsky, Biochemistry 32 (1993) 7143^7151. [16] S. May, A. Ben-Shaul, Biophys. J. 73 (1997) 2427^2440. [17] N. Zhu, D. Liggitt, Y. Liu, R. Debs, Science 261 (1993) 209^211. [18] P.L. Felgner, T.R. Gadek, M. Holm, R. Roman, H.W. Chan, M. Wenz, J.P. Northrop, G.M. Ringold, M. Danielsen, Proc. Natl. Acad. Sci. USA 84 (1987) 7413^7417. [19] M.J. McCluskie, Y. Chu, J.-L. Xia, J. Jessee, G. Gebyehu, H.L. Davis, Antisense Nucleic Acid Drug Dev. 8 (1998) 401^414. [20] A.R. Thierry, Y. Lunardi-Iskandar, J.L. Bryant, P. Rabinovich, R.C. Gallo, L.C. Mahan, Proc. Natl. Acad. Sci. USA 92 (1995) 9742^9746.

References [1] M. Chattergoon, J. Boyer, D.B. Weiner, FASEB 11 (1997) 754^763.

BBAGEN 25017 18-5-00

Cyaan Magenta Geel Zwart