Dendrimeric alkylated polyethylenimine nano-carriers with acid-cleavable outer cationic shells mediate improved transfection efficiency without increasing toxicity

#

Pharmaceutical Research, Vol. 27, No. 4, April 2010 ( 2010)

DOI: 10.1007/s11095-010-0058-1

Research Paper

Dendrimeric Alkylated Polyethylenimine Nano-carriers with Acid-Cleavable

Outer Cationic Shells Mediate Improved Transfection Efficiency

Without Increasing Toxicity

Terry W. J. Steele^1,2and Wayne Thomas Shier^1,3

Received September 10, 2009; accepted January 5, 2010; published online February 17, 2010

Purpose. Improved polycation-based non-viral DNA vectors were sought by preparing dendrimers with

polyethylenimine cores surrounded by various shells incorporating structural features intended to

facilitate steps in transfection mechanisms. Dendrimeric vectors were designed with (a) an outer

oligocation shell, intended to facilitate DNA release inside cells, (b) a hydrophobic C-16 alkyl shell, and

(c) a polycationic core, the latter two intended to combine lipid-depletion and osmotic burst endosome

escape mechanisms, respectively, and were (d) attached through an a acid-cleavable linker reported to

hydrolyze at endosomal pH values.

Methods. Vectors and DNA complexes were characterized by dynamic and static light scattering. Flow

cytometry was used to quantitate transfection activity and cytotoxicity in CHO–K1 cells.

Results. About 5-fold increased transfection activity was obtained for a vector constructed with an outer

shell of oligocations attached through acid-cleavable linkers, relative to a control dendrimer with an acid-

stable linker. The most effective oligocation component of outer shells tested was spermine. Neither

modiﬁcation was associated with increased cytotoxicity. This vector design did not permit an evaluation

of the beneﬁt of combining endosome release mechanisms.

Conclusion. Using acid-cleavable linkers to attach an outer shell of short, highly-charged oligocations to

a PEI-based dendrimeric vector substantially increased transfection efﬁciency without increasing

cytotoxicity.

KEY WORDS: acid-cleavable; dendrimers; gene therapy; polyethylenimine; transfection.

INTRODUCTION

A wide variety of strategies have been investigated to

increase the level and duration of gene expression and

Gene therapy, the use of therapeutic DNA for the treat- decrease the toxicity obtained with non-viral DNA vectors.

ment of disease, has not achieved its anticipated therapeutic The strategies include novel approaches, such as Sleeping

potential due to severe drug delivery problems resulting from Beauty transposons (7), and simpler structural variations on

the chemical and physical properties of DNA. DNA is too big, the standard polycation vector, polyethylenimine (PEI) (8,9)

too polar and too charged (1) to allow effective delivery to its aimed at increasing transfection efﬁciency and/or decreasing

intended site of action inside the cell without the use of a suitable cytotxicity (10). Simply adding ligands to PEI has yet to

vector. Viral vectors have had problems with immunogenicity achieve major increases in transfection efﬁciency. Slow

(2,3) and carcinogenicity resulting from insertional mutagenicity progress is at least partly due to incomplete understanding

(4,5). Non-viral vectors are being investigated (6) as alternatives of the mechanisms by which PEI-DNA complexes are taken

with potential advantages of being non-immunogenic, simple to into cells, or by which the DNA is released from the complex

produce, stable, non-carcinogenic and able to give transient in a functional form. When hydrophobic moieties are

gene expression with little or no insertion in host DNA.

attached to PEI, an appropriate balance between the lip-

ophilicity and hydrophilicity appears to be necessary for

efﬁcient transfection of DNA in polyplexes, presumably

because lipophilicity of a PEI-based vector may inﬂuence

both its interactions with DNA and interactions of the DNA-

PEI complexes with cell membranes (11–14). An additional

concern is that many of the types of reactions used to attach

modifying groups to branched PEI involve preferential

coupling to and loss of primary amines. The branched form

of PEI contains primary, secondary, and tertiary amines, all of

which typically become protonated in the pH range 9–11

when they are in an isolated form (15). However, it is

Adapted from the PhD thesis of T.W.J. Steele, University of

Minnesota, 2006.

¹Department of Medicinal Chemistry, College of Pharmacy, University

of Minnesota—Twin Cities, 308 Harvard St., SE, Minneapolis,

Minnesota 55455, USA.

²Present Address: School of Materials Science and Engineering,

Nanyang Technological University, 50 Nanyang Avenue, Singapore,

639798, Singapore.

³To whom correspondence should be addressed. (e-mail: shier001@

umn.edu)

#

683

0724-8741/10/0400-0683/0

2010 Springer Science+Business Media, LLC

684

Steele and Shier

buffering capacity in the pH range through which Na⁺,K⁺- mine (manufacturer’s designated molecular weights of 600,

ATPases lower endosomal pH (pH 5.2–7.0) that is the basis of 1,800 and 25,000 Da) was purchased from Polysciences Inc.

the osmotic burst mechanism for escaping endosomes before (Warrington, PA) and used without further puriﬁcation.

they fuse with lysosomes (9). The 25 kDa form of bPEI Polyethyleneimine with the manufacturer’s designated molec-

exhibits substantial buffering capacity in this critical pH ular weight of 25,000 Da was determined by size-exclusion

range, possibly associated with internal amines affected by chromatography to have an average molecular weight of

undetermined environmental effects.

32,000 Da. The experimentally determined value has been

Recent efforts have aimed to reduce cytotoxicity through used throughout this presentation. 1,4-Bis-maleimidobutane

use of biodegradable cationic polymers. Polymers were synthe- was purchased from Pierce Chemical (Rockford, IL). 3,9-

sized incorporating chemistries known to be labile within the Diethylidene-2,4,8,10-tetraoxaspiro[5.5]undecane was the gift

physiological milieu. These include disulﬁde (16,17), diacrylate of Dr. Jorge Heller, Advanced Polymer Systems, Inc., Red-

(18), and acid-labile poly-orthoester (19) and ketal (20,21) wood City CA. CHO–K1 cells were purchased from Amer-

bonds. Another commonly used strategy has been to attach ican Type Culture Collection (Manassas, VA). Bovine calf

different types of ligands to branched or linear PEI, including serum was purchased from Hyclone Laboratories, Inc.

attaching various hydrophobic moieties (11,22,23), the hydro- (Logan, UT). pEGFP plasmid was purchased from Clontech

philic polymer poly(ethyleneglycol) (24,25), carbohydrates such Laboratories, Inc. (Mountain View, CA). Membrane ﬁlters

as dextran (26), and ligands which bind to cell surfaces with a (0.02, 0.2 and 0.45 μm) were obtained from Fisher Scientiﬁc

greater degree of speciﬁcity than polycations do (27,28).

(Pittsburgh, PA). Deuterated solvents (CDCl₃, D₂O, and

One potentially useful strategy for making polycation DMF-D₇) were obtained from Cambridge Isotope Laborato-

derivatives with improved transfection activity is to alter them ries (Andover, MA).

in ways that should facilitate known transfection mechanisms of

cationic liposomes and cationic polymers (29–31). In this study,

we report the preparation, partial characterization and measure-

Instrumentation

ment of DNA transfection and cytotoxic activities of dendrimeric

¹H NMR spectra were obtained on an Oxford 300 MHZ-

vectors designed to improve non-viral gene delivery by incorpo-

Mercury spectrometer (Varian, Inc., Palo Alto, CA) in 5 mm

rating into various shells structures intended to facilitate steps in

o.d. NMR tubes in deuterated solvents. ¹³C NMR spectra

were obtained on an Oxford 600 MHZ–Inova spectrometer

the transfection mechanism. As presented in Scheme 1, these

vectors consist of a polycationic core, a hydrophobic shell, a

(Varian, Inc., Palo Alto, CA) in 8 mm o.d. NMR tubes using

linker shell expected to hydrolyze at pH values near the low end

D₂O solvent. Fluorescence was measured on a Varian Cary

of the endosomal pH range, and an outer oligocation shell. These

vectors were designed to test two mechanism-based hypotheses.

The ﬁrst hypothesis is that transfection efﬁciency can be

increased by incorporating into the vector a mechanism for

Eclipse ﬂuorescence spectrophotometer (Varian, Inc., Palo

Alto, CA). Dynamic laser light scattering was carried out on a

custom-assembled modular instrument consisting of a Lexel

Model 95 Ion 488 nm adjustable intensity laser (Lexel Laser,

releasing the DNA from the vector when inside the cell. This

Inc., Fremont, CA) directed into a Malvern PCS100

hypothesis has been investigated by preparing a series of

goniometer and photomultiplier tube (Malvern Instruments,

dendrimeric vectors in which the outer shell is composed of

Inc., Southborough, MA). Photomultipler signals were

oligocations held together in place by acid-labile linkers reported

analyzed using a Brookhaven Instruments BI-9000AT

to be stable at pH values found outside cells, but to hydrolyze

autocorrelation card and processed with the Windows

rapidly at pH values found in endosomes. Linker hydrolysis

version of Brookhaven Instruments DLS software version

should release small oligocations, intended to bind DNA weakly

3.24 (Brookhaven Instruments Corp., Holtsville, NY).

enough to not interfere with its expression inside the cell.

Fluorescence polarization measurements were carried out

The second hypothesis is that transfection efﬁciency can

be increased by incorporating into the same dendrimeric

vector two endosome release mechanisms. Some studies on

the transfection mechanisms of polycation vectors have

suggested that endosome release is a rate-limiting factor in

transfection efﬁciency (31,32). This hypothesis has been

on ﬂuorescein isothiocyanate derivatives using a 96-well

spectroﬂuorometer (Tecan Corp. model Genios Pro,

Durham, NC) with excitation at 480 nm and a 530 nm ﬁlter

block.

The primary method for analyzing dendrimeric polyca-

tion vectors and synthetic intermediates was size-exclusion

investigated by preparing dendrimeric vectors containing (a)

chromatography with multi-angle laser light scattering detec-

a branched polyethyleimine (bPEI) core, which along with

tion on a custom-assembled modular HPLC instrument with

the outer oligocationic shell provides buffering capacity for a

‘proton sponge’ or osmotic release mechanism, and (b) a

hydrophobic shell of C-16 alkyl chains designed to provide

a “lipid sink” to absorb structural lipids from endosome

multiple detectors in series. The instrument included a

Beckman (Beckman Coulter, Inc., Fullerton, CA) HPLC

system with a Rheodyne (PerkinElmer Life and Analytical

Sciences, Inc., Waltham, MA) manual injector ﬁtted with a

membranes causing lysis by membrane depletion.

100 μL sample loop and Beckman 125 pumps controlled by

Beckman 32 karat software, and equipped with PSS Novema

MATERIALS AND METHODS

Materials

8×300 mm 300 Å and 1,000 Å columns (Polymer Standards

Service-USA, Inc., Warwick, RI) in series. The primary

detection method for molecular weight analysis was static

light scattering performed on a Wyatt Dawn EOS multi-angle

Unless otherwise indicated, all biochemicals were pur- laser light scattering detector (Wyatt Technology Corporation

chased from Sigma–Aldrich (St. Louis, MO). Polyethylenei- Santa Barbara, CA) controlled by Wyatt Astra software

Dendrimeric Alkylated Polyethylenimine Nano-carriers

685

2

2 15

Scheme 1. Synthesis of dendrimeric alkylated polyethylenimine nano-carriers with acid-

cleavable outer cationic shells. Top line: The central core consisting of 32,000 Da branched

polyethylenimine (bPEI) was alkylated with 16-iodo-hexadecanoic acid homocysteine

thiolactone amide to add a hydrophobic shell, and the thiolactone ring opened under

alkaline conditions. Second line: The addition of an acid-cleavable linker, or a non-

cleavable control linker by a Michael addition reaction with one end of a dimaleimido

derivative of the linker present in excess. Third line: The addition of the outer oligocation

shell by reaction of thiolated spermine (S) with terminal unreacted maleimide residues to

attach the outer oligocation shell and complete the assembly (bottom). Abbreviation: DMF

N,N′-dimethylformamide.

version 5.1 (available at ftp.wyatt.com). Additional detectors Synthesis of Dendrimer Shell-Forming Reagents

in series included an Altex 156 Refractive Index detector

(Beckman Coulter, Inc., Fullerton, CA) calibrated with Wyatt Synthesis of the Hydrophobic Shell-Forming Reagent:

RICAL software (available at www.wyatt.com), a Beckman 16-iodo-hexadecanoyl-homocysteine Thiolactone Amide (IV)

166 UV detector, and a JASCO FP-1520 programmable (Scheme 2)

ﬂuorometer (JASCO Inc., Easton, MD). Flow cytometry

was carried out on a Becton–Dickinson FACScaliber with cell

A solution of 16-hexadecanolide (24.5 g, 96 mmol) in 2 L

counts analyzed with CellQuest Pro software (BD Bioscien- ethanol was mixed with 40 mL of 200 mM aqueous sodium

ces, San Jose, CA). Zeta-potentials were calculated using a hydroxide and heated under reﬂux for 1 h. The solution was

ZM-80 zeta–meter (Zeta–Meter Inc., Staunton, VA) with a acidiﬁed with concentrated hydrochloric acid and crystals

type GT-2 electrophoresis cell using a molybdenum cylinder allowed to form overnight at 4°C. The crystals were collected

anode and a platinum rod cathode.

by ﬁltration and the ﬁltrate concentrated on a rotary

686

Steele and Shier

was cooled to room temperature yielding 20 g of 16-

hydroxy-hexadecanoic acid, II, crystals, which were col-

lected by ﬁltration and dried in vacuum (76% yield). ¹H

NMR (300 MHZ, CD₃OD, δ) 1.29 (br s, 22H, CH₂), 1.55 (m,

4H, CH₂), 2.65 (t, 2H, CH₂), 3.53 (t, 2H, CH₂), ¹³C NMR

(CD₃OD, δ) 25.07 (s, 1C), 25.93 (s, 1C), 29.22 (s, 1C), 29.5 (br s,

11C), 32.5 (s, 1C), 33.92 (s, 1C), 61.91 (s, 1C), 176.42 (s, 1C).

16-Hydroxy-hexadecanoic acid, II (1.36 g, 5 mmol), was

placed into a 500 mL one-necked ﬂame-dried round-bottom

ﬂask along with homocysteine thiolactone (785 mg, 5.1 mmol),

hydroxybenzotriazole (690 mg, 5.1 mmol), 1-(3-dimethlyamino-

propyl)-3-ethylcarbodiimide (978 mg, 5.1 mmol), and 4 Å heat-

activated molecular sieves (3 g), mixed with 150 mL of dry

dimethylformamide in a ﬂame-dried ﬂask and stirred for 72 h at

room temperature. The reaction mixture was ﬁltered to remove

molecular sieves. The ﬁltrate was poured into a separatory

funnel containing 500 mL distilled water and extracted three

times with 100 mL aliquots of ethyl acetate. The combined

organic phases were washed with water, 1 N hydrochloric acid,

5% aqueous sodium hydroxide, and saturated sodium chloride

solution. The organic phase was dried over anhydrous magne-

sium sulfate, and evaporated on a rotary evaporator to yield a

white powder. The white powder was crystallized from acetone

to yield 1.45 g (78%) of 16-hydroxy-hexadecanoyl-homocys-

1

teine thiolactone amide, III. H NMR (300 MHZ, CDCl₃, δ)

1.25 (br s, J=2.97, OCCH₂CH₂CH₂(CH₂)₉CH₂CH₂CH₂OH,

18H), 1.61 (m, J=1.89, 6.07, 7.3, OCCH₂CH₂CH₂(CH₂)₉

CH₂CH₂CH₂OH, 4H), 2.25 (m, J=1.89, 2.96, OCCH₂

CH₂CH₂(CH₂)₉CH₂CH₂CH₂OH, 4H), 2.45 (br s, OH, 1H),

2.94 (m, J=1.22, 4,71 COSCH₂CH₂CH, 2H), 3.36 (m, J=1.22,

COSCH₂CH₂CH, 2H), 3.65 (t, J=6.07, OCCH₂CH₂CH₂

(CH₂)₉CH₂CH₂CH₂OH, 2H), 4.05 (t, J=7.3, OCCH₂

CH₂CH₂(CH₂)₉CH₂CH₂CH₂OH, 2H), 4.50 (m, J=4.58, 4.71,

CH₂CH₂CH(NH)(CO), 1H), 6.05 (br s, J=4.58, CHNHCO,

1H), ¹³C (CDCl₃, δ) 25.9–36.8 (s, 13C), 59.8 (s, 1C), 63.38 (s, 1C),

173.84 (s, 1C), 205.64 (s, 1C).

16-Hydroxy-hexadecanoyl-homocysteine thiolactone

amide (372 mg, 1 mmol) and sodium iodide (600 mg,

4.0 mmol) were dissolved in 10 mL dry acetonitrile under

an argon atmosphere. The mixture was heated to 70°C,

chloro-trimethylsilane (326 mg, 3.0 mmol) was added by

syringe, and the yellow solution was allowed to react for

90 min. The reaction mixture was cooled to room temper-

ature and poured into 100 mL of 10% aqueous sodium

thiosulfate. The mixture was extracted twice with 50 mL

aliquots of ethyl acetate, the extracts combined, washed with

10% aqueous sodium thiosulfate and with saturated sodium

chloride solution and dried over anhydrous magnesium

sulfate. The solution was evaporated on a rotary evaporator,

and the tan residue was crystallized from methanol to yield

16-iodo-hexadecanoyl-homocysteine thiolactone amide, IV,

(400 mg, 83% yield): ¹H NMR (300 MHZ, CDCl₃, δ) 1.27

(br s, J=3.01, OCCH₂CH₂CH₂(CH₂)₉CH₂CH₂CH₂I, 18H),

1.64 (m, J=1.87, 4.65, 7.22, OCCH₂CH₂CH₂(CH₂)₉CH₂

CH₂CH₂I, 4H), 2.25 (m, J=1.87, 3.01, OCCH₂CH₂CH₂

(CH₂)₉CH₂CH₂CH₂I, 4H), 2.94 (m, J=1.27, 4,77 COSCH₂

CH₂CH, 2H), 3.20 (t, J=4.65, OCCH₂CH₂CH₂(CH₂)₉

CH₂CH₂CH₂I, 2H), 3.36 (m, J=1.27, COSCH₂CH₂CH, 2H),

Scheme 2. Synthesis of reagents used to add shells to branched

polyethylenimine cores to construct dendrimeric alkylated polyethy-

lenimine nano-carriers with acid-cleavable outer cationic shells. The

alkylating reagent used for the addition of the hydrophobic shell was

16-iodo-hexadecanoic acid homocysteine thiolactone amide, IV.

Maleimide coupling units were added to each end of acid-cleavable

linkers. Outer shell-forming oligocation reagents were prepared by

thiolation of spermine and low molecular weight (600 Da and

1,800 Da) polyethylenimines with Traut’s reagent, 2-iminothiolane.

Abbreviations: ETOH ethanol; eq. equivalent, DMF N,N′-dimethyl-

formamide; EDCI 1-[3-(dimethylamino)propyl]-3-ethylcarbodiimide

hydrochloride; HOBt 1-hydroxybenzotriazole; Mol. molecular; RT

room temperature; TMSCl chlorotrimethylsilane; deg degrees; ACN

acetonitrile; NaOAc sodium acetate; pTSA p-toluenesulfonic acid; RT

room temperature.

evaporator to ∼800 mL, and a second crop of crystals was

obtained at 4°C. The two crops of crystals were combined,

dissolved in benzene and heated under reﬂux to remove 4.05 (t, J=7.22, OCCH₂CH₂CH₂(CH₂)₉CH₂CH₂CH₂I, 2H),

water azeotropically using a Dean–Stark trap. The solution

4.50 (m, J=4.77, 4.97, CH₂CH₂CH(NH)(CO), 1H), 6.05 (br s,

Dendrimeric Alkylated Polyethylenimine Nano-carriers

687

J=4.97, CHNHCO, 1H), ¹³C (CDCl₃, δ) 7.86 (s, 1C), 25.8–

36.8 (s, 13C), 59.8 (s, 1C), 63.38 (s, 1C), 173.97 (s, 1C), 205.76

(s, 1C).

2-Hydroxyethylmaleimide (VI, 262 mg, 2 mmol) and 3,9-

diethylidene-2,4,8,10-tetraoxaspiro[5.5]undecane (290 mg,

1.04 mmol) were added anaerobically to 50 mL anhydrous

ether in a septum-sealed round-bottom ﬂask under argon,

and stirred at room temperature for 2 h until the solution

turned turbid. The mixture was ﬁltered and the ﬁltrate

concentrated on a rotary evaporator to a clear oil that

crystallized on standing at −20°C to yield 3,9-bis(maleimdioe-

thoxy)-3,9-ethane-2,4,8,10-tetraoxaspiro[5.5]undecane (33)

Synthesis of Acid-Cleavable Linker-Forming Reagents

(Scheme 2). Synthesis of 2-Hydroxyethyl-maleimide (VI)

Maleic anhydride (V, 15.4 g, 157 mmol) and freshly-

distilled ethanolamine (9.59 g, 157 mmol) were each dis-

solved in 50 mL acetone and slowly added to 50 mL acetone

in a round-bottom ﬂask in an ice bath. After addition was

complete, the ice bath was removed and the mixture allowed

to react overnight at room temperature. Approximately

75% of the acetone was removed on a rotary evaporator and

the solution placed at −20°C until crystallization was

complete, yielding 25.3 g (81% yield) 2-[(3-carboxy-1-oxo-

2-propenyl)amino]-ethanol (33): mp=73°C, ¹H NMR

(300 MHZ, CD₃OD, δ) 3.07 (m, 1H, NH) 3.42 (t, 2H,

CH₂), 3.66 (t, 2H, CH₂), 5.56 (br s, 1H, OH), 6.26 (d, 1H,

CH), 6.46 (d, 1H, CH), ¹³C NMR (CD₃OD, δ) 42.48 (s, 1C),

59.94 (s, 1C), 132.38 (s, 1C), 134.06 (s, 1C), 166.81 (s, 1C),

167.44 (s, 1C)

2-[(3-Carboxy-1-oxo-2-propenyl)amino]-ethanol (1.59 g,

10 mmol) was dissolved in 10 mL acetic anhydride. Freshly

fused and crushed sodium acetate (0.82 g, 10 mmol) was

added and the solution heated at 70°C overnight. The

resulting black solution was poured onto 50 mL crushed ice

plus 50 mL 1 N hydrochloric acid in a separatory funnel and

mixed until the ice melted. The mixture was extracted with

50 mL ethyl acetate and the aqueous layer discarded. The

organic layer was washed twice with water, once with

saturated sodium chloride solution, dried over anhydrous

magnesium sulfate and concentrated on a rotary evaporator

to few mL of liquid. The dark brown liquid was fractionated

using 1:1 ethyl acetate:hexane on a silica gel column to yield

0.5 g of 2-hydroxyethylmaleimide, VI (35% yield): mp=70–

72°C, ¹H NMR (300 MHZ, CDCl₃, δ) 3.8 (t, 2H, CH₂), 4.2 (t,

2H, CH₂), 6.74 (s, 2H, CH), ¹³C NMR (CDCl₃, δ) 36.61 (s,

1C), 61.12 (s, 1C), 134.48 (s, 2C), 205.5 (s, 2C).

1

(VIII, 450 mg, 90% yield): H NMR (300 MHZ, CDCl₃, δ)

0.83 (m, CH₂, 6H), 1.10 (m, CH₂, 4H), 1.66 (m, CH₂, 4H).

3.53 (m, CH, 2H), 3.69 (m, CH, 2H), 3.93 (m, CH₂, 4H), 6.82

(s, CH, 4H) ¹³C NMR (CDCl₃, δ) 7.20 (s, 2C), 28.38 (s, 2C),

30.43 (s, 2C), 37.66 (s, 2C), 59.85 (s, 2C), 62.65 (s, 4C), 117.59

(s, 1C), 134.64 (s, 4H), 171.07 (s, 4C).

Synthesis of Outer Shell-Forming Oligocation Reagents

(Scheme 2)

Three oligocations that span a range of relatively low

molecular weights were used: spermine and two low molec-

ular weight PEIs (600 and 1,800 Da). An oligocation (100 mg,

containing 1 mmol of primary amine) was dissolved in 50 mL

of 50 mM sodium borate, 1 mM EDTA, pH 8.0 buffer. 2-

Iminothiolane (Traut’s reagent, 68 mg, 0.5 mmol) was added

as a 427.4 mM stock solution in small aliquots to a vigorously

stirred solution of the oligocation until a 1:1 molar ratio was

reached. The reaction was allowed to proceed to completion

for 30 min at room temperature. To remove large aggregates,

thiol-derivatized oligoamines were loaded into Amicon-15

ultraﬁltration centrifuge tubes, 10,000 MWCO, and centri-

fuged for 30 min at 2000×g. Comparison of the free thiol

concentration before and after ultraﬁltration using the

ﬂuorogenic thiol quantitation assay described below indicated

a 6.5, 9.2, and 15.9% free thiol concentration decrease for

spermine, 600 Da PEI, and 1,800 Da PEI, respectively. No

further puriﬁcation was employed.

Dendrimeric Nanoparticle Assembly (Scheme 1)

Synthesis of Acid-Cleavable Linker

2,2-bis(2-maleimidoethoxy)-propane (VII)

The dendrimeric nanoparticle was assembled using a

divergent strategy in the initial steps, beginning with a core

consisting of commercially available bPEI (32,000 Da meas-

2-Hydroxyethylmaleimide (VI, 140 mg, 1 mmol), 2,2- ured molecular weight). The hydrophobic second shell was

dimethoxypropane (75 mg, 0.07 mmol), p-toluenesulfonic added to the core by diluting 32,000 Da bPEI (1.0 g, about

acid (2 mg), and crushed heat-activated 4 Å molecular sieves 7.8 mmol primary amines) in 10 mL of a solution of the

(100 mg) were added to 10 mL dry benzene in a 50 mL hydrophobic shell reagent, 16-iodo-hexadecanoyl-homocys-

round-bottom ﬂask under an argon atmosphere. The mixture teine thiolactone amide in anhydrous DMF. One mL aliquots

was stirred for 2 h and evaporated to dryness on a rotary of the solution (100 mg bPEI) were diluted 1:10, and

evaporator. An additional 10 mL benzene and 50 mg 2,2- sufﬁcient hydrophobic shell reagent was added to give

dimethoxypropane were added, and the solution stirred hydrophobic shell: bPEI molar ratios of 10, 19, 48 and 96

another 2 h, ﬁltered, mixed with two drops of pyridine and per mole of bPEI, which corresponds to alkylation of about

evaporated on a rotary evaporator. The clear oil remaining 5%, 10%, 25% and 50% of primary amines in bPEI,

was crystallized from acetone: hexane at −20°C to yield 2,2- respectively, assuming a 1:2:1 ratio of primary: secondary:

bis(2-maleimidoethoxy)-propane (VII, 50 mg, 15% yield) tertiary amines in bPEI and complete reaction with primary

(34): mp. 123–125°C (lit. rep. 124–126), (300 MHZ, CD₃CN, amines. A brief yellow coloration of the reaction mixture was

δ) 1,21 (s, 6H, CH₃), 3.46 (t, 24, CH₂), 3.56 (t, 4H, CH₂), 6.77 observed initially, which then turned colorless. The solution

(s, 4H, CH), ¹³C NMR (CD₃CN, δ) 24.4 (s, 2C), 37.96 (s, 2C), was vigorously mixed on a Buchler Evapo-mix vortex mixer

58.1 (s, 2C), 117.58 (s, 2C), 134.45 (s, 2C), 171.0 (s, 1C).

Synthesis of acid-cleavable linker 3,9-bis(2-maleimidoe-

at room temperature for 1 h.

The reaction mixtures containing the alkylated bPEI

thoxy)-3,9-ethane-2,4,8,10-tetraoxaspiro[5.5]undecane (VIII), cores were directly submitted to homocysteine thiolactone

688

Steele and Shier

ring opening conditions by adding them drop-wise to a linkers were complexed with DNA and used in transfection

vigorously mixed solution of 45 mL of 100 mM sodium

borate plus 1 mM EDTA pH 9.5 or 100 mM ammonium

hydroxide plus 1 mM EDTA pH 9.5 and allowed to react

overnight at room temperature. Alternatively, the solution of

alkylated bPEI cores was added drop-wise to 45 mL of

vigorously mixed 100 mM ammonium hydroxide plus 1 mM

EDTA at pH 9.5 and heated under reﬂux for 30 min for faster

homocysteine thiolactone ring opening. Ten mL of the

solution of ring-opened alkylated bPEI cores was loaded into

a 10,000 Da molecular weight cutoff ultraﬁltration tube

(Amicon Ultra-15) and centrifuged at 1000×g for 30 min.

The retentate was dissolved in 10 mL of 50 mM sodium

borate, 1 mM EDTA, 1 mM tris-(2-carboxyethyl)-phosphine,

pH 8.0, and centrifuged again. The procedure was repeated.

The number of ring-opened alkyl chains per bPEI core was

determined by measuring the amount of free thiol with a

ﬂuorogenic reagent (see below).

The next shell to be added consisted of the acid-

cleavable linker. Three types of maleimido-containing linker

reagents were used: 2,2-bis(2-maleimidoethoxy)-propane

(VII), 3,9-bis(2-maleimidoethoxy)-3,9-ethane-2,4,8,10-tet-

raoxaspiro[5.5]undecane (VIII) and 1,4-bis-maleimidobutane,

which is a non-cleavable structural analog of VII included in

the study as a negative control. Each maleimido-containing

linker reagent was dissolved in anhydrous DMF at 5 mg/mL.

Ring-opened alkylated bPEI cores (100 mg) were dissolved in

10 mL of 50 mM sodium borate, 1 mM EDTA, pH 8.0, and

added dropwise to sufﬁcient vigorously stirred maleimido-

containing linker reagent solution to give a 5- to 10-fold molar

equivalent excess of maleimido moiety to the amount of free

thiol groups measured on the ring-opened alkylated bPEI

cores. The reaction mixture was stirred an additional 10 min,

loaded into a 10,000 Da molecular weight cutoff ultraﬁltration

centrifuge tube (Amicon Ultra-15), and centrifuged at 2000×g

for 15 min. The retentate was dissolved in a 10 mL aliquot of

buffer (50 mM sodium borate, 1 mM EDTA, 1 mM tris-(2-

carboxyethyl)-phosphine, pH 8.0) and centrifuged again. This

procedure was repeated. The amount of thiol groups remain-

ing unreacted on the dendrimer preparations were measured

with a ﬂuorogenic reagent (see below).

studies without further characterization.

Preparation of pDNA Complexed with Dendimeric

Nanoparticles

Preparation of pEGFP

The pEGFP plasmid was transformed into DH5α com-

petent cells for ampliﬁcation. The DNA was puriﬁed using

both cesium chloride-ethidium bromide gradients (35) and

Qiagen Plasmid Maxi kits (Qiagen Inc. Valencia, CA USA).

Preparation of pDNA-Dendrimer Complexes

Dendrimer samples (10 μg) were dissolved in 1 mL of

deionized water or saline and ﬁltered through a 0.2 μm

membrane ﬁlter. Stock solutions of pEGFP were prepared by

diluting 3.20 μg of pDNA into deionized water or saline and

ﬁltering through a 0.45 μm membrane ﬁlter. The pDNA

solution was pipetted into dendrimer-containing solutions in

an amount calculated to achieve a 10:1 molar ratio of vector

nitrogen to DNA phosphate. The pDNA and dendrimer were

allowed to complex for 30 min at room temperature before

being analyzed by laser light scattering.

Characterization of Constructs

Quantitation of Thiols Using a Fluorogenic Reagent

The ﬂuorogenic reagent solution was 1 mM ammonium

7-ﬂuorobenzo-2-oxa-1,3-diazole-4-sulfonate (36) (SBD-F) in

buffer (100 mM sodium borate, 1 mM EDTA, pH 8.5). Thiol

groups were measured by dissolving a sample or standard in

0.495 mL of the same buffer and mixing with 0.5 mL of the

ﬂuorogenic reagent solution plus 0.005 mL 100 mM tris-(2-

carboxyethyl)-phosphine in a 1.5 mL Eppendorf tube then

heated at 60°C for 1 h in an aluminum heating block. The

reaction was cooled to room temperature, and ﬂuorescence

was immediately read on a ﬂuorescence spectrophotometer in

a 1 mL quartz cuvette at 385 nm excitation wavelength and

524 nm emission wavelength using a 5 nm slit for 100 μs. The

photomultiplier tube voltage was typically set so that the

highest standard was 10–20% below saturation.

Sufﬁcient outer shell-forming oligocation reagent was

dissolved in 3–10 mL of buffer (50 mM sodium borate, 1 mM

EDTA, 1 mM tris-(2-carboxyethyl)-phosphine, pH 8.0) to

provide a 1.1-fold molar ratio of free thiol moieties to the

maleimide moieties on 10 mg of a dendrimer preparation with Spectrophometric Quantitation of bPEI by Copper

Complexation (37)

an acid-cleavable linker in its outer shell. The dendrimer

preparation dissolved in 5 mL of the same buffer was added

drop-wise to the solution of outer cationic shell reagent with

vigorous vortex mixing. The mixture was allowed to react for

10 min, then loaded into a 5,000 or 10,000 Da molecular

weight cutoff ultraﬁltration tube (Amicon Ultra-15) and

centrifuged at 2000×g for 15 min. The retentate was dissolved

in a 4–15 mL aliquot of buffer (50 mM sodium borate, 1 mM

EDTA, 1 mM tris-(2-carboxyethyl)-phosphine, pH 8.0) and

centrifuged again. The procedure was repeated once to give

the completely assembled nanoparticle vector. The nano-

particle vectors assembled with the non-cleavable 1,4-bis-

maleimidobutane linker were characterized by size-exclusion

chromatography with multiangle laser light scattering detec-

tion. The nanoparticle vectors assembled with acid-labile

A 2 mL aliquot of copper reagent solution (0.145 mg/mL

copper sulfate, 100 mM acetic acid, adjusted to pH 5.4 with

sodium hydroxide) was added to 800 μL of standard bPEI

solution or a sample containing 5.0–50.0 μg/mL bPEI. A blue

color developed immediately. Absorbance was measured at

285 nm. Unknown samples were compared to a bPEI standard

curve prepared with bPEI of the same molecular weight.

Size-Exclusion Chromatography with Detection by Static Light

Scattering

The size-exclusion HPLC system with a multiangle laser

light scattering detector described above was used to provide

Dendrimeric Alkylated Polyethylenimine Nano-carriers

689

relatively accurate and precise molecular weight data. To allow the vector-DNA complexes. Dendrimer-DNA complexes

mass measurements, the Altex 156 refractive index detector were tested in triplicate at four bPEI nitrogen to DNA

was calibrated using ﬁve sodium chloride concentrations phosphate ratios (N:P ratio) from 5:1 to 25:1. CHO–K1 cells

known to the fourth signiﬁcant digit (e.g., 0.1847 mg/mL) (100 μL of 500 cells/μL) in 5% (v/v) calf serum in medium

spanning the range of 0.0000–2.000 mg/mL (NaCl speciﬁc were added to each well and cultured for 24 h and 48 h at 37°

refractive index increment, dn/dc=0.172 mL/g) (38). The Wyatt C in a 5% CO₂atmosphere. The medium was drawn off, and

Dawn EOS 18-angle detectors were normalized using spec- 100 μL of 1% (v/v) trypsin in buffered saline pH 7.2 was

trophotometric grade toluene. The UV detector, Wyatt Dawn added and incubated for 1 min. The trypsin solution was

EOS, and refractive index detector were aligned using an drawn off and immediately replaced with 200 μL culture

injection of 0.50 mg bovine serum albumin monomer fractio- medium containing 5% calf serum and 2 μg/mL propidium

nated by a Superdex 200 10/300 GL gel permeation column. iodide. The cells were removed from the substratum by

Alkylated bPEI and 1,4-bis-maleimidobutane-containing den- ﬂushing with a pipet and transferred to pre-chilled test tubes

drimer constructs were analyzed using three PSS Novema kept on crushed ice until the expressed green ﬂuorescent

columns in series: a guard column, a 300 Å 8×300 mm column, protein (GFP) was quantiﬁed by ﬂow cytometry.

and a 1,000 Å 8×300 mm column (38). All samples were

dissolved in 1% aqueous formic acid ﬁltered through a 0.2 μm

pore size ﬁlter and eluted with the same solvent. All samples

were passed through a 0.45 μm syringe ﬁlter directly into the

injector.

Flow Cytometric Analysis of Transfection Efficiency

Plasmid DNA-transfected CHO–K1 cells were measured

for GFP expression and propidium iodide ﬂuorescence 48 h

after transfection using ﬂow cytometry analysis on a FACS-

Calibur analytical ﬂow cytometer (Becton Dickinson, Erem-

bodegem, Belgium). For each sample, 2,000 cells were

quantitated, and the intensity of FL1 (GFP) and FL3

(propidium iodide) was recorded. Flow cytometry parameters

were set using untransfected CHO–K1 cells (FL1 negative,

FL3 negative), freeze-thawed CHO–K1 cells (FL1 negative,

FL3 positive), and ﬂuorescein diacetate (10 μL, 1 mg/mL

acetone solution)-treated CHO–K1 cells (FL1 positive,

FL3 negative). In this context, negative denotes cells with

no measurable ﬂuorescence, and positive with measurable

ﬂuorescence.

Dynamic Light Scattering

Dendrimeric vector/pEGFP complexes were prepared by

mixing vector with 3.2 μg enhanced green ﬂuorescent protein

plasmid DNA in 1 mL of 0.02 μm ﬁltered phosphate buffered

saline at a N:P ratio of 10:1 and allowed to complex for 15–

30 min before readings were taken. Laser power was adjusted

between 200 and 700 kcounts per second for dendrimer/

pEGFP complexes. Maximum laser intensity was used for

dendrimer size measurements without DNA complexation.

Each sample was analyzed for 1 min after the ﬁrst and last

delay times were experimentally determined on the Brook-

haven Instruments DLS software for Windows. Particle size

distributions were derived according to the non-negative least

squares (NNLS) algorithm included in the Brookhaven

Instruments DLS software for Windows.

Statistics

Signiﬁcance of differences between means was calculated

using unpaired Student’s t-test.

Zeta-Potential Analysis

RESULTS

Suspensions of dendrimer/pDNA complexes (10 μg)

prepared as described above for dynamic light scattering

measurements were diluted into 15 mL of ﬁltered (0.2 μm)

deionized water and the speciﬁc conductance measured in a

Zeta-meter ZM-80. Voltage was set according to speciﬁc

conductance measurements and was typically 200 V, as

speciﬁed by the manufacturer. Between six to ten particles

were timed with a stopwatch for the distance of one micro-

meter. Direction was noted for polarity, and the time-

averaged values were compared with an electrophoretic

mobility/zeta potential conversion chart to yield the zeta

potential of each dendrimer/pDNA complex.

Construction of Dendrimeric Vectors

Dendrimeric vectors were constructed using primarily a

divergent strategy (i.e., sequential addition of shells, one at a

time, proceeding outward from a central core), followed by

limited use of a convergent strategy (i.e., addition of multiple

shells in one step as pre-assembled segments) for the ﬁnal

assembly. The central core consisted of commercially avail-

able 32,000 Da bPEI, which was selected for its established

effectiveness in the ‘proton sponge’ endosomal escape

mechanism (6,8,9,29–31,39).

Addition of a Hydrophobic Shell to Polycationic Cores

Measurement of Transfection Efficiency of Dendrimeric

Vectors

Hydrophobic shells were added to bPEI cores by

alkylating outer primary amines at four molar ratios (10, 19,

Transfection experiments were conducted in 96-well 48 and 96) of alkylating agent IV to bPEI. Alkylation of bPEI

plates. Dendrimer stock solutions (1 mg/mL in distilled water) with a number of different hydrophobic moieties has been

were pipetted into 20 μL of Dulbecco-modiﬁed Eagle’s reported as an approach to enhancing catalytic properties

medium. pEFGP (320 ng) in 25 μL of the same medium (40), attaching gold colloids (41), and enhancing bactericidal

was subsequently added. The plates were placed on an orbital activity (42). The most commonly used reaction has been

shaker and shaken for 15 min at room temperature to form nucleophilic substitution with alkyl halides. The alkyl chain

690

Steele and Shier

reagent developed for the present study, 16-iodo-hexadecanoyl- to bind the particle electrostatically to the surface of the

homocysteine thiolactone amide (IV), was designed to react mammalian cell. In this study, outer polycationic shells were

readily with a bPEI amino group at one end, attaching the constructed of oligoamines, which were expected to bind

hydrophobic alkyl domain with a protected thiol at the opposite pDNA effectively when held together as an aggregate in the

end. The C16 chain was selected to provide a shell with outer shell, but which would release bound pDNA more

hydrophobic character to act as a sink for released endosomal effectively than a larger polycation in the cytoplasm after

lipids and to prevent pDNA released from outer oligocation dispersion of the vector-pDNA complex at low pH. Three

shell from binding electrostatically to the bPEI core. The thiol oligoamines were examined in the present study: spermine,

moiety was included to enable subsequent coupling to a which gives a terminal oligocation with a maximum of 4

maleimide moiety attached to an acid-labile linker. It was positive charges; 600 Da PEI, with a maximum of about 14

initially protected as a thiolactone ring to prevent polymer- positive charges; and 1,800 Da PEI, with a maximum of about

ization by reaction of free thiols with iodo-alkane moieties. The 42 positive charges. The range in MWs allowed differences in

thiolactone ring was opened by alkaline hydrolysis using slow cationic surface charge and electrostatic binding ability to be

addition to a basic buffer (Scheme 1) to yield a free thiol (an compared. To facilitate coupling to the unreacted maleimido

efﬁcient nucleophile) and a carboxylate ion, the negative charge moiety of acid-labile linkers, the oligoamines were derivatized

of which was expected to both aid in solubilization during with Traut’s reagent (2-iminothiolane), which reacts with an

construction of the vector and later in preventing electrostatic amino group attaching a free thiol group. The numbers of

collapse of pDNA onto the polycationic core after decom- free thiol groups attached were quantiﬁed using a ﬂuorogenic

position of vector-pDNA complexes at low pH.

reagent, SBD-F. A reducing agent, 1 mM tris-(2-carbox-

yethyl)-phosphine, was included in assay and reaction buffers

to keep the free thiols in reduced form. Thiolated oligoamines

were allowed to react in borate buffer at pH 8.0 with the

maleimido moieties on acid-labile linkers bound to alkylated

bPEI. Coupling reactions with maleimide derivatives were

carried out in the shortest possible time due to the instability

of the linker moiety. Excess unreacted thiolated oligoamines

were removed using repeated cycles of dilution with buffer

and concentrating in ultraﬁltration tubes. Fully assembled

vectors were dissolved in 5% DMF/50 mM sodium borate

buffer, pH 8.0, ﬂash frozen and stored in liquid nitrogen.

Addition of Acid-Labile Linkers

Acid-cleavable linkers were selected for their reported

ability to maintain moderate stability above pH 7.0, but

hydrolyze at pH 5.5, a pH value achieved inside the

endocytotic vesicle (33). The following three types of linker

reagents were prepared for attaching outer shells to inner

shells, and maleimide moieties were added at each end: (a)

1,4-bis maleimidobutane, (b) 2,2-bis(2-maleimidoethoxy)-pro-

pane, VII, and (c) 3,9-bis(2-maleimidoethoxy)-3,9-ethane-

2,4,8,10-tetraoxaspiro[5.5]undecane, VIII (see Scheme 2).

1,4-Bis-maleimidobutane is a non-cleavable linker included Characterization of Assembled Dendrimeric Vectors

as a negative control to assess whether acid-cleavable linkers

would improve transfection efﬁciency, cytotoxicity, or both.

Various approaches were used to characterize the four

The other two linkers, VII and VIII, have reported cleavage intermediates at the two-shell stages and the nine fully

half-times of 42 and 18 min, respectively, at pH 5.5 (33). assembled vectors prepared from each by attaching one of

Maleimide moieties undergo facile, essentially irreversible three different linkers and to each one of three different outer

Michael addition to strong nucleophiles such as thiol groups. shell oligocations, except for alkylation at the 96 molar ration

An acid-cleavable linker was attached to the free thiol groups for which only spermine and 600 Da PEI oligocations were

at the ends of each alkyl chain by reacting with excess bis- used (43). The characterization methods included (a) ratios of

maleimido-linker construct, which attached the linker using alkyl chains to bPEI cores by integrating proton NMR signals

one maleimido moiety, while leaving an unreacted maleimido for methylene groups in core bPEI and hydrophobic shell

moiety at the new terminus of each chain. Maleimide- alkyl chains, (b) quantitation of bPEI by copper complexation

containing dendrimers were handled with extreme care, due (37) and quantitation of thiol groups by ﬂuorescent derivati-

to their inherent lability. Aqueous solutions of VII and VIII zation (36), (c) molecular weight determination by static light

were made on the day they were used, and never stored. VII scattering, and (d) particle size determination by dynamic

and VIII could only be stored neat; degradation was observed light scattering.

during storage in all organic solvents tested. To prevent

aggregation of dendrimers due to crosslinking, the solutions

of alkylated bPEI were added slowly to a solution containing

Characterization Studies on Two-Shell Intermediates

5- to 10-fold molar excess of linker reagent. Reactions were

Characterization studies were carried out on the two-

monitored for completion by determining if free thiols were

shell dendrimeric intermediates consisting of a bPEI core and

present using a ﬂuorogenic reagent (36). Unreacted excess

a hydrophobic shell with terminal free thiol groups by

linker reagents were removed (i.e., reduced to <0.05 equiv-

integration of methylene protons in the NMR spectra in

alents) by repeated cycles of dilution with buffer and

deuterated water. Methylene protons on alkyl chains could be

concentration in ultraﬁltration tubes.

accurately integrated for reaction products with 10 and

19 mol of alkyating agent IV per mole of bPEI core,

Addition of the Outer Oligocationic Shell

indicating the presence of 7.7 and 18.0 alkyl chains per bPEI

core, respectively. The thiol contents of these intermediates

Effective dendrimeric transfection agents must include were consistent with most, if not all, of the thiolactone

an outer polycationic shell to condense the plasmid DNA and moieties having reacted with aqueous base to generate the

Dendrimeric Alkylated Polyethylenimine Nano-carriers

691

free thiols needed for subsequent coupling to an acid-labile acetal moieties for characterization, due to excessive reso-

linker. Molar ratios of alkylating agent IV to bPEI core >20 nance band broadening because the large size and rigidity of

yielded products which gave proton NMR spectra unsuitable the dendrimers prevent molecular tumbling in the NMR time

for quantitating methylene proton due to micelle formation, scale. Fully assembled vectors were analyzed by size exclu-

aggregation, or both (43). The molar ratios obtained by sion chromatography with 1% formic acid as mobile phase

measuring thiols by ﬂuorescence after derivatization with a and multi-angle laser light scattering detection (Table II). A

ﬂuorogenic reagent (36) and bPEI cores by copper complex- wide variety of solvent systems were tested as mobile phases

ation (37) were similar (7.6 and 15.3 alkyl chains per bPEI for elution of PEI derivatives, including 450 mM ammonium

core for alkylation with 10 and 19 mol of alkyating agent IV hydroxide, pH 8.5; 250 mM triethylamine acetate, pH 8.0;

per mole of bPEI core, respectively). Molar ratios of thiol to 500 mM acetate buffer, pH 5.0; with a range of column

bPEI cores measured by chemical analysis indicated 42.7 and settings from room temperature to 60°C, but among them

60.6 alkyl chains per bPEI core for alkylation with 48 and only 1% formic acid successfully eluted the polymers. For size

96 mol of alkyating agent IV per mole of bPEI core, exclusion chromatography of vectors containing the uncleav-

respectively. Solubility of the dendrimeric intermediate with able 1,4-bis-maleimidobutane linker in 1% formic acid as

60.6 alkyl chains per bPEI core was limited, and it could not mobile phase, higher numbers of alkyl chains conjugated to

be subjected to a freeze-thaw cycle without precipitation.

the bPEI core (calculated 48 and 96 alkyl chains per core,

Molecular weights of alkylated bPEI intermediates were representing about 25% and 50% of primary amines alky-

also determined by dynamic and static light scattering (43) lated, respectively) resulted in larger deviations of the

and by size exclusion chromatography with multi-angle laser measured molecular weight from the calculated value. At

light scattering detection (Table I). The observed molecular the highest alkylation level, anomalously high molecular

weights of intermediates prepared with 48:1 and 96:1 molar weight values were measured, consistent with aggregation.

ratios of alkylating agent to bPEI core deviated widely from Attempts to study vectors containing acid-labile linkers (VII

values predicted on the basis of complete reactions. The and VIII) by size exclusion chromatography yielded M_w

deviations may be the result of aggregation of the dendrimers values well below calculated, consistent with cleavage of the

due to high hydrophobicity. Elution of highly alkylated bPEI acid-labile linker in the low pH mobile phase and loss of the

polymers from Novema size exclusion chromatography col- outer cationic shell.

umns required harsh acidic conditions (1% formic acid); none

Given that dendrimers prepared with acid-labile linkers

of a wide variety of neutral solvent mixtures (triethylamine could not be characterized by size exclusion chromatography,

acetate, phosphate, and borate buffers, with or without the dendrimers prepared with the acid-stable control linker,

methanol or acetonitrile) were effective.

1,4-bis maleimidobutane, were extensively characterized

Two-shell intermediates were complexed with pEGFP using the technique, and it was assumed that, because similar

DNA and examined for transfection activity and cytotoxicity bis-maleimide chemistry was used with all the linkers,

(43). Cytotoxicity increased dramatically with N:P ratio, dendrimers prepared with acid-labile VII and VIII would

typically from <5% cytotoxicity at N:P=5:1 to over 50% at have analogous structures. Additional information about the

N:P=50:1. Modest transfection activity was observed only at dendrimers was obtained by preparing model vectors in

the lowest alkylation level (calculated 10 alkyl chains per which the inner polycationic core was ﬂuorescently labeled

bPEI core).

with ﬂuorescamine (FLA) (44) and the outer oligocationic

shells were ﬂuorescently labeled with ﬂuorescein isothiocya-

nate (FITC) (45). Vectors prepared in this manner retained

transfection and cytotoxic activities (43). The spectral proper-

Characterization Studies on Completed Vectors

A variety of strategies were used in the characterization ties of FLA and FITC are sufﬁciently different to allow facile

simultaneous analysis of both. The FITC-labeled outer

of fully assembled dendrimeric vectors. The bPEI content of

each was determined using copper complexation (37). It was cationic shell and the FLA-labeled alkylated bPEI core elute

not possible to use the unique NMR absorption bands of the from size exclusion chromatography (Fig. 1) at the same time

Table I. Estimation of the Average Molecular Weight of Partially Assembled (Two-Domain) Dendrimeric Vector Intermediates

Observed molecular weight^b

Molar ratio of alkylating

agent to bPEI polymer

Calculated molecular weight^a(Da)

Weight average, M_w(Da)

Number average, M_n(Da)

0

3.2×10⁴

3.6×10⁴

3.9×10⁴

5.0×10⁴

6.8×10⁴

3.2×10⁴

5.4×10⁴

6.7×10⁴

1.1×10⁵

1.8×10⁶

1.7×10⁴

3.5×10⁴

3.7×10⁴

8.8×10⁴

1.4×10⁶

10

19

48

96

bPEI branched polyethylenimine

^aAssuming complete reaction of bPEI (32,000 Da measured molecular weight) with 16-iodo-hexadecanoyl-homocysteine thiolactone amide

adding a 371 Da unit

^bSize exclusion chromatography was carried out on custom-assembled modular HPLC instrumentation in 1% formic acid with Novema 8×300

mm 300 Å and 1,000 Å columns in series and multi-angle laser light scattering detection. Dn/dc values used to calculate the mass and

molecular weight were based on the mass ratio of the polyethylenimine core (dn/dc=0.253¹⁶) and n-eicosane (dn/dc=0.169¹⁷).

692

Steele and Shier

Table II. Average Molecular Weights of Selected Fully-Assembled Dendrimeric Vectors Using Static Light Scattering in 1% Formic Acid

Alkylation ratio

Linker

Outer shell oligocation

Calculated MW

Weight average (Mw)^a

Number average (Mn)^a

10

19

48

96

BMB

VIII

BMB

VII

spermine

PEI-600

PEI-1800

spermine

PEI-600

PEI-1800

spermine

PEI-600

PEI-1800

spermine

PEI-600

3.8×10⁴

4.2×10⁴

5.6×10⁴

3.8×10⁴

4.3×10⁴

5.0×10⁴

7.8×10⁴

6.0×10⁴

7.9×10⁴

1.5×10⁵

1.0×10⁵

1.3×10⁵

6.3×10⁴

1.1×10⁵

9.1×10⁴

4.8×10⁴

1.3×10⁵

1.6×10⁵

1.2×10⁵

3.6×10⁵

5.8×10⁵

3.7×10⁴

4.0×10⁵

4.2×10⁶

5.6×10⁶

5.5×10⁴

9.6×10⁴

7.5×10⁴

3.8×10⁴

1.8×10⁵

1.4×10⁵

1.1×10⁵

3.0×10⁵

4.5×10⁵

3.6×10⁴

2.9×10⁵

3.5×10⁶

5.0×10⁶

BMB

MW molecular weight; BMB 1,4-bis-maleimidobutane; VII 2,2-bis(2-maleimidoethoxy)-propane; VIII 3,9-bis(2-maleimidoethoxy)-3,9-ethane-

2,4,8,10-tetraoxaspiro[5.5]undecane, PEI branched polyethylenimine with the manufacturer’s designated molecular weight in Da

^aMass and molecular weight were calculated with a mass ratio for polycationic cores of dn/dc=0.253¹⁶and for n-eicosane of dn/dc=0.169¹⁸. A

5% error was assumed for M_wand M_ndata.

as total sample measured by the 90 degree light scattering Preparation and Characterization Studies on Polyplexes

signal, providing evidence that the two cationic shells are

coupled as one molecule by the 1,4-bis-maleimidobutane

The DNA condensation/complexation capacity of dendri-

linker. FLA and FITC contents were quantitated at their meric vectors with pEGFP DNA was evaluated by dynamic

absorbance maxima (44,45) for vectors prepared with calcu- light scattering. The overall stability of the vector/pEGFP

lated 19 and 48 C16 alkyl chains per bPEI core, the acid- complexes was investigated by determining the zeta-potential,

stable linker, and outer oligocation shells of 600 and 1,800 Da a measure of ﬂocculation potential and colloidal stability.

PEI (Table III). For 19 alkyl chains per bPEI core, the results Table IV presents examples of values obtained for a series of

indicate nearly 100% coverage of alkyl chains with an completely assembled vectors containing 32,000 Da bPEI cores

oligocation, whereas with 48 alkyl chains per bPEI core, alkylated with an amount of alkylating agent calculated to

there was only ∼65% coverage of alkyl chains with an attach 10 alkyl chains per core, and containing one of the three

oligocation. The size of the outer shell oligocation did not outer shell oligocations attached through either an acid-

affect the extent of alkyl chain coverage, because the molar cleavable or the non-cleavable control linker. Fully and

ratios of outer oligocation component to cores were the same partially completed vectors with all levels of alkylation formed

within error despite the size differences between the FITC- complexes with pEGFP plasmid DNA, except the most highly

labeled PEIs with 600 and 1,800 Da molecular weights.

alkylated intermediate (96 molar ratio of alkylating agent to

bPEI core) lacking an outer oligocation shell (43). The simplest

explanation for the failure of the highly alkylated core to

complex with DNA is the presence of a net negative surface

charge due to large numbers of terminal carboxylate moieties.

Table IV presents an example of data for a fully assembled

vector series prepared from 32,000 Da bPEI cores alkylated at

a level calculated to give 10 alkyl chains per bPEI core and

linked to one of three oligocations by an acid-cleavable or non-

cleavable linker. The vectors complexed pEGFP plasmid DNA

tightly with sizes ranging from 74 to 110 nm. Similar sizes were

seen for pEGFP plasmid DNA complexed with the vector

series alkylated at a level calculated to give 19 alkyl chains per

bPEI core and linked to one of three oligocations by an acid-

cleavable or non-cleavable linker. However, particle sizes were

much larger, typically >500 nm, for complexes formed between

DNA and vectors of the corresponding vector series alkylated

Fig. 1. Size-exclusion chromatography of a completed dendrimeric

vector prepared with a ﬂuorescamine-labeled (44) 32,000 Da

branched polyethylenimine core, alkylated with a calculated average at a level calculated to give 48 and 96 alkyl chains per bPEI

of 19 C-16 alkyl chains coupled through the non-cleavable linker 1,4-

bis-maleimidobutane to a ﬂuorescein-labeled (45) 600 Da polyethy-

lenimine oligocation outer shell. Size-exclusion chromatography used

formic acid (1%) as eluent on a custom-assembled modular HPLC

instrument with multiple detectors in series. Matter, measured by

static light scattering (black line), ﬂuorescamine ﬂuorescence (dark

grey line), and ﬂuorescein ﬂuorescence (light grey line) all co-eluted.

Similar results were obtained for alkylation of the polyethyleimine

core with a calculated average of 48 C-16 alkyl chains (43).

core, and the complexes had negative zeta-potentials (43).

Evaluation of the Ability of Acid-Labile Linkers to Increase

Transfection Activity of Dendrimeric Vectors

with Oligocation Outer Shells

Dendrimeric vectors alkylated at a level calculated to

give 10 alkyl chains per bPEI core and linked through a non-

Dendrimeric Alkylated Polyethylenimine Nano-carriers

693

Table III. Composition of Fully-Assembled Dendrimeric Vectors Prepared with Fluorescamine-labeled 32,000 Da Branched Polyethylenimine

Cores, Fluorescein Isothiocyanate-Labeled Outer Oligocationic Shells and the Acid-Stable Linker, 1,4-bis-maleimidobutane

Vector components

Measured^amass/moles of

oligocation outer shells

(µg/nmol)

Ratio of measured moles of

outer of shell oligocations per

mole of polycationic core

Calculated alkylation

ratio

Outer shell oligocation

molecular weight

Measured^amass/moles of

polycation cores (µg/nmol)

19

48

19

48

600 Da

1,800 Da

23/0.34

12/0.11

30/0.45

33/0.31

4.1/6.8

2.1/3.5

14/7.0

20

32

15

31

19/9.6

^aBased on ﬂuorescence measurements following size exclusion chromatography

cleavable linker (1,4-bis-maleimidobutane) to any of the not exhibit higher transfection activity than the unaltered

outer cationic shell components tested exhibited only very bPEI core. Degradation of the acid-cleavable acetal linkers in

low levels (1 and 2%) of transfection efﬁciency. In contrast, dendrimeric vectors was studied by three different methods:

the corresponding vectors prepared with the acid-labile (a) ﬂuorescence polarization, (b) ultraﬁltration, and (c)

linkers, VII and VIII, and with spermine as the cationic shell dynamic light scattering. Initial studies used ﬂuorescence

component exhibited an optimal transfection efﬁciency about polarization, a technique which estimates the molecular

ﬁve times higher (11 1% with 21 5% cytotoxicity at a N:P weight of molecules based on their rotational mobility after

ratio of 10:1, Fig. 2A). This represents a signiﬁcant (Student’s excitation of a ﬂuorescent label with polarized light. In

t-test, P<0.005) increase in transfection efﬁciency relative to dendrimeric vectors prepared with FITC-labeled outer shell

that of the corresponding vector with the non-cleavable oligocations, the ﬂuorescent label experiences a large drop in

linker, 1,4-bis-maleimidobutane. Signiﬁcantly increased trans- molecular weight when linker hydrolysis releases oligocations

fection activity was also observed for vectors with acid- from the alkylated bPEI core. A decrease in ﬂuorescence

cleavable linkers and alkylation at 19 chains per bPEI core polarization at lower pH was expected for vectors prepared

and N:P ratios ≤10, but not at higher N:P or higher alkylation with acid-labile linkers (VII, and VIII), but not expected for

ratios (43). The observed increased transfection activity with vectors synthesized with the 1,4-bis-maleimidobutane linker.

vectors incorporating an acid-cleavable linker provides proof- However, increased ﬂuorescence polarization was observed,

of-concept evidence that a low pH releasable outer oligoca- particularly at low pH with acid-labile linkers. Fig. 3 contains

tion shell can increase transfection efﬁciency. However, the an example obtained with the series of three linkers in vectors

optimal transfection activity observed with acid-cleavable prepared from 32,000 Da bPEI cores alkylated at a level

linkers was still about half that of unaltered 25 kDa bPEI calculated to give 19 alkyl chains per bPEI core and

(Fig. 2A).

containing 1,800 Da FITC-labeled PEI oligocationic outer

Additional characterization of the acid lability of dendri- shell components. The largest change was seen using the VIII

meric vectors prepared with linkers VII and VIII was linker at pH 4.0. The increased ﬂuorescence polarization

undertaken in an attempt to better understand why they did suggests that the predominant reaction occurring under the

Table IV. Plasmid DNA Condensation/Complexation Facility with Branched Polyethylenimine Cores, Partially Assembled (Two-Domain)

Dendrimer Intermediates and Completely Assembled Dendrimeric Vectors

Vector shell components

Size^aof complex

ζ-potential of complex

Dendrimer stage

Alkylation ratio

Linker type

Outer shell oligocation

with pEGFP DNA (nm)

with pEGFP DNA (mV)

Core only

–

430 20

540 30

150 10

160 10

1,600 100

70 10

110 30

80 10

90 10

80 10

110 10

80 10

28

19

20

15

16

22

9

14

13

18

9

“ + Alkyl shell

“ ” + Oligocation

10

19

48

96

10

–

BMB

VII

VIII

PEI-600

PEI-1800

Spermine

PEI-600

PEI-1800

Spermine

PEI-600

PEI-1800

Spermine

12

15

– not present in the complex; BMB 1,4-bis-maleimidobutane; VII 2,2-bis(2-maleimidoethoxy)-propane; VIII 3,9-bis(2-maleimidoethoxy)-3,9-

ethane-2,4,8,10-tetraoxaspiro[5.5]undecane, PEI branched polyethylenimine with the indicated molecular weight in Da

^aHydrodynamic size determined by dynamic light scattering

694

Steele and Shier

hydrophobically. Failure to observe release of ﬂuorescently

labeled oligocations (see below) is consistent with the latter

mechanism.

The second approach to characterizing the acid lability of

dendrimeric vectors prepared with linkers VII and VIII used

ultraﬁltration to monitor release of 1,800 Da FITC-labeled

PEI oligocationic outer shell components from the vector

series prepared from 32,000 Da bPEI cores alkylated at a

level calculated to give 19 alkyl chains per bPEI core. Over

the course of 2 h, no ﬂuorescently labeled outer cationic shell

components were detected crossing the 10,000 Da molecular

weight cutoff ultraﬁltration membrane at pH 4.0, 5.5, or 8.0

(43). It was also observed that dilution by pH 4.0 and pH 5.5

buffers caused the vector solutions to turn opalescent or

cloudy.

Dynamic light scattering was the third method used to

investigate the degradation/aggregation of vectors at lower

pH. Table V presents an example obtained with the series of

three linkers in vectors alkylated at a level calculated to give

ten alkyl chains per bPEI core and containing 1,800 PEI

oligocationic outer shell components. Similar results were

obtained with vectors alkylated at a level calculated to give 19

alkyl chains per bPEI core, whereas vectors alkylated at a

level calculated to give 48 and 96 alkyl chains per bPEI core

exhibited increased aggregation (43). The hydrodynamic radii

for vectors alkylated at a level calculated to give ten alkyl

chains per bPEI core and completed with the three linkers

differed signiﬁcantly only at pH 2.0, at which the vector with

the acid-stable linker was signiﬁcantly larger (Student’s t-test,

P<0.05), consistent with loss of the outer oligocationic shells

from vectors containing the acid-labile linkers VII and VIII

(43). At pH values of 4.0 and 5.5, aggregates of 300–500 nms

were formed, which provides an explanation for the opales-

cence observed during the ultraﬁltration experiments.

Fig. 2. Transfection efﬁciency and cytotoxicity in CHO–K1 cells with

green ﬂuorescent protein plasmid DNA complexed with (A) fully-

assembled dendrimeric vectors prepared with acid-labile and stable

control linkers or (B) incomplete dendrimeric vector intermediates

lacking linkers and an outer oligocation shell. Assays were carried out

in triplicate with the indicated N:P ratio (estimated moles of vector

amines to DNA phosphate) of vectors consisting of branched

polyethylenimine cores alkylated with the indicated calculated

average of C-16 alkyl chains (10, 19, 48 and 96). The positive controls

LF2000 and unaltered 2,500 Da bPEI were tested at an N:P ratio of

Evaluation of Combining Endosomal Escape Mechanisms

as a Strategy for Increasing Transfection Activity

of Dendrimeric Vectors

bPEI cores alkylated at levels calculated to give averages

of 10–96 alkyl chains per bPEI core and lactone rings opened

were complexed with pEGFP plasmid DNA and transfected

into cells to determine the latent transfection efﬁciency and

10:1. In (A) 10 C-16 alkyl chains were coupled through one of three cytotoxicity not associated with the presence of a releasable

linkers (acid-cleavable linkers, VII or VIII, or the non-cleavable

control linker, BMB) to one of three outer shell oligocations (600 Da

branched polyethylenimine, 1,800 Da branched polyethylenimine or

spermine). Transfection and cytotoxicity activities were measured

simultaneously on a ﬂuorescence-activated cell sorter as described in

“Materials and Methods.” Abbreviations: BMB 1,4-bis-maleimidobu-

tane; VII 2,2-bis(2-maleimidoethoxy)-propane; VIII 3,9-bis(2-malei-

midoethoxy)-3,9-ethane-2,4,8,10-tetraoxaspiro[5.5]undecane; 600

600 Da bPEI; 1,800 1,800 Da bPEI; 25,000 25,000 Da (manufacturer’s

speciﬁcation) bPEI; SP spermine; LF2000 Lipofectamine 2000®.

outer cationic shell. bPEI cores alkylated at levels calculated

to give 10 and 19 alkyl chains per bPEI core had ∼8%

transfection efﬁciency and 10–12% cytotoxicity at a 10:1N:P

ratio (see Fig. 2B). This represents about half the transfection

efﬁciency of unalkylated bPEI cores complexed with pEGFP

plasmid DNA under the same transfection conditions. bPEI

cores alkylated at levels calculated to give 48 and 96 alkyl

chains per bPEI core exhibited no observable transfection

efﬁciency or increased cytotoxicity (43).

Increased alkylation of bPEI cores beyond the level

calculated to give ten alkyl chains per bPEI core did not

increase the transfection efﬁciency or lessen the cytotoxicity

in the fully assembled dendrimeric vectors. No dendrimeric

vector prepared with alkylation at levels calculated to give 19

and 48 alkyl chains per bPEI core exhibited transfection

efﬁciencies of over 5% (43) even if constructed with an acid-

cleavable linker and spermine as the outer shell oligocation.

conditions used was aggregation, not hydrolytic release of

oligocation outer shell components. Aggregation could result

from release of outer shell oligocations or from altered

dendrimer conformations which have lost the shell structure

and assumed a form which allows the previously internal

hydrophobic regions of different dendrimers to interact

Dendrimeric Alkylated Polyethylenimine Nano-carriers

695

Fig. 3. Hydrolysis of fully-assembled dendrimeric vectors with acid-

cleavable and stable, control linkers at various pH values as

monitored by ﬂuorescence polarization. Fully-assembled dendrimeric

vectors prepared with 32,000 Da branched polyethylenimine cores

were alkylated with a calculated average of 19 C-16 alkyl chains and

linked to ﬂuorescein-labeled 1,800 Da branched polyethylenimine

outer shell oligocations through one of two acid-cleavable linkers, VII

or VIII, or the non-cleavable control linker, 1,4-bis-maleimidobutane.

Vector solutions (5 mg/mL) were diluted 1:20 in one of the following

three buffers: pH 4.0: 50 mM sodium citrate; pH 5.5: 50 mM sodium

citrate; or pH 8.0: 50 mM sodium borate. Fluorescence polarization

was measured every 10 min for 17 h at room temperature on a 96-well

spectroﬂuorometer. Fluorescence polarization decreases when small

ﬂuorescein-labeled oligocations are released from large dendrimers

by linker hydrolysis, and it increases when intact vectors aggregate.

Abbreviations: BMB 1,4-bis-maleimidobutane; VII 2,2-bis(2-maleimi-

doethoxy)-propane; VIII 3,9-bis(2-maleimidoethoxy)-3,9-ethane-

2,4,8,10-tetraoxaspiro[5.5]undecane.

highest transfection activities at all levels of alkylation

examined except at a calculated average of 96 alkyl chains

per bPEI core (43). Cytotoxicity was heavily dependent on N:

P ratios with little effect of alkylation level even at a

calculated average of 96 alkyl chains per bPEI core

(Fig. 2A, B and reference 43). Thus, optimal transfection to

cytotoxicity ratios were observed at low alkylation levels with

spermine as the outer cation.

DISCUSSION

DNA Release with Acid-Labile Linkers as a Strategy

for Increased Transfection Efficiency

Studies were carried out to investigate the hypothesis

that transfection efﬁciency can be increased by incorporating

into the structure of a dendrimeric vector a mechanism that

facilitates release of the DNA from the vector after the

complex has been taken into the target cell. The approach

used was to prepare a dendrimeric vector in which the outer

shell (which binds the DNA being transfected) consists of

oligocations held in place by linkers reported to be stable at

neutral pH, but to hydrolyze rapidly at pH values found

inside endosomes (pH 4.5–5.5). During the initial electrostatic

binding of DNA to the fully-assembled vector, and the

binding of the vector-DNA complex to the negatively

charged outer surface of the cell at neutral pH, the acid-

labile linkers should remain intact. However, following

endocytosis Na, K-ATPases pump protons into endocytic

vesicles, lowering the pH in preparation for fusion with a

primary lysosome (30). When exposed to lower pH inside

endosomes, acid-labile linkers should hydrolyze, releasing

the outer shell and its bound DNA. Small oligocations

should bind electrostatically to DNA weakly enough not to

interfere with its expression inside the cell. Consistent with

this hypothesis, the transfection efﬁciency of vectors con-

taining acid-cleavable linkers was observed to be up to ﬁve-

fold greater than for vectors containing the non-cleavable

No completed dendrimeric vector prepared with alkylation at

levels calculated to give 96 alkyl chains per bPEI core

exhibited transfection efﬁciencies of over 1% (43). The

underivatized bPEI core was a more efﬁcient vector for

transfection with generally comparable or lower cytotoxicity

than any of the multi-shell dendrimers (Fig. 2A, B and

reference 43).

Other Structural Features of Dendrimeric Vectors Associated control linker, 1,4-bis-maleimidobutane (Fig. 2A). However,

with Increased Transfection Activity

optimal transfection efﬁciency with vectors containing acid-

labile vectors was less than that observed in control experi-

Of the three outer shell oligocations tested (spermine, ments with the unaltered bPEI core. There are various

600 Da PEI and 1,800 Da PEI), spermine generally gave the possible explanations for these observations including (a)

696

Steele and Shier

Table V. Hydrodynamic Radius (R_H) of Dendrimeric Vectors Following Exposure to a Range of pH Values

Hydrodynamic radius^a(R_H, nm) at indicated buffer pH

Linker

pH 2.0^b

pH 4.0^b

pH 5.5^b

pH 8.0^c

1,4-bis-maleimidobutane (BMB)

2,2-bis(2-maleimidoethoxy)-propane (VII)

3,9-bis(2-maleimidoethoxy)-3,9-ethane-2,4,8,10-tetraoxaspiro-[5.5]undecane (VIII)

16

12

2

300 10

270 10

300 10

490 20

480 20

530 30

10

11

9

2

^aParticle sizes were continuously measured by dynamic light scattering for 60 min and the ﬁnal value noted, except for the pH 8.0 value, which

was read after 1 min.

^bMeasurements at pH 2.0, 4.0, and 5.5 used a 50 mM sodium borate citrate buffer

^cMeasurements at pH 8.0 was 50 mM sodium borate, 1 mM EDTA buffer

other structural features built into the dendrimeric vectors reductions in transfection efﬁciency. Similar reductions in

reduced transfection efﬁciency more than the improvement transfection efﬁciency with increased alkylation were

provided by acid-labile linkers, (b) very little acid hydrolysis observed in completed dendrimeric vectors containing an

of linkers occurred under the conditions encountered, and acid-releasable outer oligocationic shell (see Fig. 2A). One

(c) DNA release from the vector is not a rate-limiting step in explanation for reduced transfection efﬁciency resulting from

transfection.

addition of an alkyl shell is provided by the observations of

Additional studies conducted on degradation of pure Suh et al. (46), who showed that alkylating bPEI shifted pK_a

vectors in low-pH buffers indicated that hydrolysis and values two pH units lower so that buffering capacity in the pH

release of outer oligocations did not occur as expected. range found in endosomes (pH 5 and 6), which is believed

Hydrolysis of acid-labile linkers with release of the outer responsible for the osmotic burst/proton sponge mechanism

cationic shell was not observed at pH 4.0 and 5.5 using several proposed by Behr (30), was shifted to lower pH values (pH 3

experimental approaches (Table V), whereas hydrolysis did and 4). Alkylation of the less toxic 10 kDa bPEI resulted in

occur at pH 2.0 in the buffer used for size exclusion improved transfection efﬁciency when the primary amine

chromatography (Fig. 1). Spontaneous aggregation of vectors content was maintained (47), whereas alkylcarboxylate graft-

was also observed at pH 4.0 and 5.5 by both ﬂuorescence ing to 10 kDa bPEI resulted in lowered cytotoxicity, but

polarization and dynamic light scattering measurements increased transfection efﬁciency only when shorter alkyl

(Fig. 3 and Table V), but not at pH 2.0, when hydrolysis did chains (C-6, C-10) were used at low (<20%) percent

occur. The observations are consistent with stabilization of alkylation of the calculated amount of primary amines (48).

acid-labile bonds in the environment found in the completely Thus, the dendrimeric vectors prepared in this study did not

assembled dendrimer, slowing or preventing release of outer provide an adequate experimental system to fully test the

shell oligocations, and with vector aggregation that occurs at advantage of combining osmotic burst and lipid depletion

least until hydrolysis has taken place. Suh et al. (46) showed mechanisms, because the structural alterations used to add

that for bPEI alkylated at a 10:1 molar ratio, less than 2% of the lipid depletion mechanism reduced the osmotic burst

the amino groups are protonated at pH 8, 19% at pH 5, 34% mechanism too much. Studies on alkylcarboxylate PEI

at pH 4 and 53% at pH 3. The low percent protonation at derivatives (47,48) suggest that greater transfection efﬁciency,

pH 4.0 and 5.5 may permit aggregation, which may result in and presumably better lipid depletion, is achieved when

reduced transfection efﬁciency. However, the relevance of hydrophobic shells are formed with shorter alkyl chains (C-6

studies on linker stability in buffer to stability in endosomes to C-10) than used in this study (C-16).

has not been established. The presence of cellular proteins in

the endosome may prevent aggregation of alkylated bPEI

cores, and they may also prevent re-binding of the pEGFP to

the aggregated alkylated bPEI cores.

Other Structural Features Associated with Increased

Transfection Efficiency

The transfection efﬁciency was greatest when the outer

shell oligocation was spermine. One explanation may be that

spermine derivatives, which have only four charges with high

charge density, may be more easily displaced from DNA by

salts and proteins in the cytoplasm or nucleus than outer shell

components prepared with more heavily charged 600 or

1,800 Da bPEI. Additional studies will be needed to

determine if more efﬁcient displacement of spermine moieties

from pEGFP DNA is involved in the improved transfection

efﬁciencies with spermine-containing dendrimers.

Combining Endosome Disruption Mechanisms as a Strategy

for Increasing Transfection Efficiency

In the second hypothesis being tested in these studies,

structural components required for two endosome release

mechanisms were incorporated into the same dendrimeric

vector to determine if they would be additive or synergistic in

facilitating DNA transfection. The hypothesis was tested with

two sets of dendrimeric vectors, the alkylated bPEI cores and

the completed dendrimers including acid-releasable outer

oligocationic shells. The reduction in transfection efﬁciency CONCLUSIONS

of bPEI was associated with alkylation by 16-iodo-hexadeca-

noyl-homocysteine thiolactone amide (see Fig. 2B), and more

A series of dendrimeric non-viral vectors were prepared

extensive alkylation of the bPEI core resulted in further from commercial 25 kDa branched PEI as an approach to

Dendrimeric Alkylated Polyethylenimine Nano-carriers

697

identifying structural features associated with increased trans- 15. Hall HK. Correlation of the base strengths of amines. J Amer

Chem Soc. 1957;79:5441–4.

16. Gosselin MA, Guo W, Lee RJ. Efﬁcient gene transfer using

fection activity without increased cytotoxicity. When a vector

was constructed with an outer shell of oligocations attached

through acid-cleavable linkers designed to be cleaved inside

reversibly cross-linked low molecular weight polyethylenimine.

Bioconjug Chem. 2001;12:989–94.

target cells when the endosomal pH is lowered, about ﬁve-

fold increased transfection activity was obtained relative to an

analogous control dendrimer with an acid-stable linker. Of

the oligocations examined as outer shell components of the

dendrimeric vectors prepared in the study, spermine, with

four possible positive charges, exhibited the highest trans-

fection efﬁciencies. The highest transfection efﬁciencies were

observed at the lowest levels of alkylation of bPEI cores

tested. The dendrimer design used in this study did not permit

an adequate evaluation of the beneﬁt of combining two

endosome release mechanisms, osmotic burst and lipid

depletion, into a single dendrimeric non-viral vector.

17. Breunig M, Hozsa C, Lungwitz U, Watanabe K, Umeda I, Kato

H, et al. Mechanistic investigation of poly(ethylene imine)-based

siRNA delivery: disulﬁde bonds boost intracellular release of the

cargo. J Control Release. 2008;130:57–63.

18. Kloeckner J, Bruzzano S, Ogris M, Wagner E. Gene carriers

based on hexanediol diacrylate linked oligoethylenimine: effect

of chemical structure of polymer on biological properties.

Bioconjug Chem. 2006;17:1339–45.

19. Zhong Z, Feijen J, Lok MC, Hennink WE, Christensen LV,

Yockman JW, et al. Low molecular weight linear polyethyleni-

mine-b-poly(ethylene glycol)-b-polyethylenimine triblock

copolymers: synthesis, characterization, and in vitro gene transfer

properties. Biomacromolecules. 2005;6:3440–8.

20. Knorr V, Ogris M, Wagner E. An acid sensitive ketal-based

polyethylene glycol-oligoethylenimine copolymer mediates

improved transfection efﬁciency at reduced toxicity. Pharm Res.

2008;25:2937–45.

ACKNOWLEDGEMENTS

21. Shim MS, Kwon YJ. Acid-responsive linear polyethylenimine for

efﬁcient, speciﬁc, and biocompatible siRNA delivery. Bioconjug

Chem. 2009;20:488–99.

22. Kim S, Choi HS, Jang HS, Suh H, Park J. Hydrophobic

modiﬁcation of polyethylenimine for gene transfectants. Bull

Korean Chem Soc. 2001;22:1069–75.

The authors thank Dr. Robert Vince and University of

Minnesota Department of Medicinal Chemistry Developmen-

tal Grant in Drug Design #882-1010 for support of this

research, and Dr. Mohammad Ramezani for useful discussions.

23. Brownlie A, Uchegbu IF, Schatzlein AG. PEI-based vesicle-

polymer hybrid gene delivery system with improved biocompat-

ibility. Int J Pharm. 2004;274:41–52.

24. Kursa M, Walker GF, Roessler V, Ogris M, Roedl W, Kircheis R,

et al. Novel shielded transferrin-polyethylene glycol-polyethyle-

nimine/DNA complexes for systemic tumor-targeted gene trans-

fer. Bioconjug Chem. 2003;14:222–31.

25. Kircheis R, Schuller S, Brunner S, Ogris M, Heider KH, Zauner

W, et al. Polycation-based DNA complexes for tumor-targeted

gene delivery in vivo. J Gene Med. 1999;1:111–20.

REFERENCES

1. Lipinski CA. Drug-like properties and the causes of poor

solubility and poor permeability. J Pharmacol Toxicol Methods.

2000;44:235–49.

2. Lehrman S. Virus treatment questioned after gene therapy

death. Nature. 1999;401:517–8.

26. Tseng WC, Jong CM. Improved stability of polycationic vector

by dextran-grafted branched polyethylenimine. Biomacromole-

cules. 2003;4:1277–84.

27. Kircheis R, Wightman L, Schreiber A, Robitza B, Rossler V,

Kursa M, et al. Polyethylenimine/DNA complexes shielded by

transferrin target gene expression to tumors after systemic

application. Gene Ther. 2001;8:28–40.

3. Volpers C, Kochanek S. Adenoviral vectors for gene transfer and

therapy. J Gene Med. 2004;6 Suppl 1:S164–71.

4. Check E. Regulators split on gene therapy as patient shows signs

of cancer. Nature. 2002;419:545–6.

5. Li Z, Dullmann J, Schiedlmeier B, Schmidt M, von Kalle C,

Meyer J, et al. Murine leukemia induced by retroviral gene

marking. Science. 2002;296:497.

28. Ogris M, Walker G, Blessing T, Kircheis R, Wolschek M, Wagner

E. Tumor-targeted gene therapy: strategies for the preparation of

ligand-polyethylene glycol-polyethylenimine/DNA complexes. J

Control Release. 2003;91:173–81.

29. Zhang S, Xu Y, Wang B, Qiao W, Liu D, Li Z. Cationic

compounds used in lipoplexes and polyplexes for gene delivery. J

Control Release. 2004;100:165–80.

30. Behr J. The proton sponge: a trick to enter cells viruses did not

exploit. Chimia. 1997;51:34–6.

31. Godbey WT, Wu KK, Mikos AG. Tracking the intracellular path

of poly(ethylenimine)/DNA complexes for gene delivery. Proc

Natl Acad Sci U S A. 1999;96:5177–81.

32. Remy-Kristensen A, Clamme JP, Vuilleumier C, Kuhry JG, Mely

Y. Role of endocytosis in the transfection of L929 ﬁbroblasts by

polyethylenimine/DNA complexes. Biochim Biophys Acta.

2001;1514:21–32.

33. Srinivasachar K, Neville DM. New protein cross-linking reagents

that are cleaved by mild acid. Biochemistry. 1989;28:2501–9.

34. Ambekar S, Gowda DC. Synthesis of heterobifunctional cross-

linking reagents: w-(N-maleimido)alkanoic acid hydrazides.

Indian J Chem. 1996;35B:184–6.

6. Merdan T, Kopecek J, Kissel T. Prospects for cationic polymers

in gene and oligonucleotide therapy against cancer. Adv Drug

Deliv Rev. 2002;54:715–58.

7. Izsvak Z, Ivics Z, Plasterk RH. Sleeping Beauty, a wide host-

range transposon vector for genetic transformation in verte-

brates. J Mol Biol. 2000;302:93–102.

8. Kunath K, von Harpe A, Fischer D, Petersen H, Bickel U, Voigt K,

et al. Low-molecular-weight polyethylenimine as a non-viral vector

for DNA delivery: comparison of physicochemical properties,

transfection efﬁciency and in vivo distribution with high-molec-

ular-weight polyethylenimine. J Control Release. 2003;89:113–25.

9. Boussif O, Lezoualc’h F, Zanta MA, Mergny MD, Scherman D,

Demeneix B, et al. A versatile vector for gene and oligonucleo-

tide transfer into cells in culture and in vivo: polyethylenimine.

Proc Natl Acad Sci U S A. 1995;92:7297–301.

10. Neu M, Fischer D, Kissel T. Recent advances in rational gene

transfer vector design based on poly(ethylene imine) and its

derivatives. J Gene Med. 2005;7:992–1009.

11. Gabrielson NP, Pack DW. Acetylation of polyethylenimine

enhances gene delivery via weakened polymer/DNA interac-

tions. Biomacromolecules. 2006;7:2427–35.

35. Garger SJ, Grifﬁth OM, Grill LK. Rapid puriﬁcation of plasmid

DNA by a single centrifugation in a two-step cesium chloride-

ethidium bromide gradient. Biochem Biophys Res Commun.

1983;117:835–42.

36. Imai K, Toyo’oka T, Watanabe Y. A novel ﬂuorogenic reagent

for thiols: ammonium 7-ﬂuorobenzo-2-oxa-1, 3-diazole-4-sulfo-

nate. Anal Biochem. 1983;128:471–3.

12. Nimesh S, Aggarwal A, Kumar P, Singh Y, Gupta KC, Chandra

R. Inﬂuence of acyl chain length on transfection mediated by

acylated PEI nanoparticles. Int J Pharm. 2007;337:265–74.

13. Forrest ML, Meister GE, Koerber JT, Pack DW. Partial

acetylation of polyethylenimine enhances in vitro gene delivery.

Pharm Res. 2004;21:365–71.

14. Thomas M, Klibanov AM. Enhancing polyethylenimine’s delivery

of plasmid DNA into mammalian cells. Proc Natl Acad Sci U S A.

2002;99:14640–5.

37. Ungaro F, De Rosa G, Miro A, Quaglia F. Spectrophotometric

determination of polyethylenimine in the presence of an

698

Steele and Shier

oligonucleotide for the characterization of controlled release 44. Udenfriend S, Stein S, Bohlen P, Dairman W, Leimgruber W,

formulations. J Pharm Biomed Anal. 2003;31:143–9.

Weigele M. Fluorescamine: a reagent for assay of amino acids,

peptides, proteins, and primary amines in the picomole range.

Science. 1972;178:871–2.

38. Wagner HL, Hoeve CAJ. Effect of molecular weight on the

refractive increment of polyethylene and n-alkanes. J Polymer

Sci A-2. 1971;9:1763–76.

39. Merdan T. Polyethylenimine and its derivatives: investigation of

in vivo fate, subcellular trafﬁcking and development of novel

vector systems. Marburg: Philips-Universität Marburg; 2003.

40. Johnson TW, Klotz IM. Preparation and characterization of some

derivatives of poly(ethylenimine). Macromolecules. 1974;7:149–53.

45. Weigele M, DeBernardo S, Leimgruber W. Fluorometric assay of

secondary amino acids. Biochem Biophys Res Commun.

1973;50:352–6.

46. Suh J, Paik HJ, Hwang BK. Ionization of polyethylenimine and

polyallylamine at various pH’s. Bioorgan Chem. 1994;22:318–

27.

41. Noeding G, Heitz W. Amphiphilic poly(ethyleneimine) based on 47. Dehshahri A, Oskuee RK, Shier WT, Hateﬁ A, Ramezani M.

long-chain alkyl bromides. Macro Chem Phys. 1998;199:1637–44.

42. Lin J, Qiu S, Lewis K, Klibanov AM. Bactericidal properties of

ﬂat surfaces and nanoparticles derivatized with alkylated poly-

ethylenimines. Biotechnol Prog. 2002;18:1082–6.

43. Steele TWJ. Oligo-l-lysine-, dextran-, and alkyl-derivatives of

polyethylenimine for the development of novel gene transfection

vectors. Minneapolis: University of Minnesota; 2006.

Hydrophobized PEI coupled to various oligoamines results in

efﬁcient nanocarriers for plasmid DNA transfer. Biomaterials.

2009;30:4187–94.

48. Oskuee RK, Dehshahri A, Shier WT, Ramezani M. Alkylcar-

boxylate grafting to polyethylenimine: a simple approach to

producing a DNA nano-carrier with low toxicity. J Gene Med.

2009;11:921–32.

Article Doi

Dendrimeric alkylated polyethylenimine nano-carriers with acid-cleavable outer cationic shells mediate improved transfection efficiency without increasing toxicity

DOI: 10.1007/s11095-010-0058-1

Source and publish data:

Authors:

Article abstract of DOI:10.1007/s11095-010-0058-1

Full text of DOI:10.1007/s11095-010-0058-1

Products guided by the article

R&D Labs maybe for 118377-62-1

Relevant to this article

Hot Product