ArticlePDF Available

Lipid Nanoparticle Technologies for Nucleic Acid Delivery: A Nanoarchitectonics Perspective

Advanced Functional Materials

June 2022
32(37)

DOI:10.1002/adfm.202203669

Authors:

Abdul Rahim Ferhan

Nanyang Technological University

Soohyun Park

BioMed X Institute

Hyeonjin Park

Sungkyunkwan University

Hyunhyuk Tae

Nanyang Technological University

Show all 6 authorsHide

Lipid‐based nanoparticles have emerged as a clinically viable platform technology to deliver nucleic acids for a wide range of healthcare applications. Within this scope, one of the most exciting areas of recent progress and future innovation potential lies in the material science of lipid‐based nanoparticles, both to refine existing nanoparticle strategies and to develop new ones. Herein, the latest efforts to develop next‐generation lipid‐based nanoparticles are covered by taking a nanoarchitectonics perspective and the design, nucleic acid encapsulation methods, scalable production, and application prospects are critically analyzed for three classes of lipid‐based nanoparticles: 1) traditional lipid nanoparticles (LNPs); 2) lipoplexes; and 3) bicelles. Particular focus is placed on rationalizing how molecular self‐assembly principles enable advanced functionalities along with comparing and contrasting the different nanoarchitectures. The current development status of each class of lipid‐based nanoparticle is also evaluated and possible future directions in terms of overcoming clinical translation challenges and realizing new application opportunities are suggested.

Illustration of the pH‐dependent behavior of lipid nanoparticles due to ionizable lipids. A) Principle of the mechanism of ionizable lipids. During formulation, the ionizable lipids present a positive charge at low pH (below its pKa) to promote favorable interactions with negatively charged nucleic acids and facilitate nucleic acid encapsulation. During storage and in circulation under physiological pH (above its pKa), the ionizable lipids will switch to being neutral and remain stable. Once the LNP reaches the endosome, which typically presents an acidic environment (below its pKa), the ionizable lipids will become protonated again and resume a positive charge. B) Schematic representation of LNP interaction with the endosome, leading to a change in lipid geometry due to the formation of ion pairs. Protonated ionizable lipids interact with anionic lipids adopting an inverted cone geometry, which promotes the formation of the inverted hexagonal (HII) phase associated with membrane fusion.

…

Structural classes of ionizable lipids. Three major classes of ionizable lipids depending on the tail structure, namely two‐tail, multi‐tail, and branched‐tail.

…

General overview of the approach to obtain LNPs incorporating XMaN lipidoids to achieve the delivery of different types of nucleic acids. A) Four lipid excipients, including XMaN lipidoids, are first mixed in ethanol, before they are combined, using a microfluidic platform, with B) a wide range of nucleic acid payloads that are dissolved in an aqueous buffer. C) A representation of a possible architectural configuration of nucleic acid‐encapsulating XMaN LNPs. D) Cryo‐TEM images of mRNA‐loaded XMaN LNPs, their size distribution and evaluation of their functional stability after prolonged storage (i.e., up to five weeks) at 4 °C. Adapted with permission.[⁶³] Copyright 2021, Wiley‐VCH.

…

Self‐assembly process of LNPs within microfluidic systems. A) The self‐assembly process of nucleic acid–LNPs in a microfluidic mixer begins when lipid excipients (i.e., ionizable lipid, PEG lipid, helper lipid, and cholesterol) dissolved in ethanol are rapidly mixed with the nucleic acid payload (e.g., siRNA, mRNA, etc.) dissolved in an aqueous buffer. This initially occurs at a low buffer pH whereby the ionizable lipid will assume a protonated form and bind, via electrostatic interactions, to the negatively charged backbone of the nucleic acid. Concurrently, the increasingly polar environment drives the formation of vesicles and the encapsulation of the nucleic acid. As the mixing progresses, and with increasing buffer pH, the ionizable lipid becomes increasingly neutral, leading to the fusion of adjacent vesicles to form the interior core of the LNP. The extent of vesicle fusion is influenced by the amount of PEG lipid added due to its hydrophilicity as well as steric effects. Adapted with permission.[⁸¹] Copyright 2021, MDPI Publishing. B) Major microfluidic mixing architectures for LNP formulation (top to bottom): T‐junction mixer, staggered herringbone micromixer, bifurcating mixer, and baffle mixer. C) The effect of varying flow rate on the mixing profile of ethanol (red) and water (blue) within a baffle mixer, visualized through a computational fluid dynamics simulation. Adapted with permission.[⁸⁹] Copyright 2018, American Chemical Society. D) A diagrammatic representation of an example of microfluidics parallelization, showing (from top to bottom) the arrangement of mixing channels, each mixing unit, and each mixing cycle within a parallelized microfluidics device, which adopts the staggered herringbone architecture. Evaluation of the in vivo delivery performance of E) siRNA–LNPs and F) mRNA–LNPs obtained using the parallelized device compared to bulk mixing. The evaluation was performed by means of administering to mice. The mRNA encodes for luciferase and the in vivo delivery performance was evaluated by directly visualizing the luminescence from selected regions of the mice. The results clearly show that the in vivo delivery performance of nucleic acid–LNPs obtained via microfluidic mixing is superior to those obtained via bulk mixing. D–F) Adapted with permission.[⁸⁵] Copyright 2021, American Chemical Society.

…

Lipoplex components and nanoarchitecture organization. Schematic illustration of the lipoplex and molecular structures of the cationic lipids and neutral lipids.

…

Figures - available from: Advanced Functional Materials

This content is subject to copyright. Terms and conditions apply.

Content uploaded by Hyeonjin Park

Content may be subject to copyright.

www.afm-journal.de

2203669 (1 of 38)

Review

Lipid Nanoparticle Technologies for Nucleic Acid Delivery:

A Nanoarchitectonics Perspective

Abdul Rahim Ferhan, Soohyun Park, Hyeonjin Park, Hyunhyuk Tae, Joshua A. Jackman,*

and Nam-Joon Cho*

Lipid-based nanoparticles have emerged as a clinically viable platform

technology to deliver nucleic acids for a wide range of healthcare applica-

tions. Within this scope, one of the most exciting areas of recent progress

and future innovation potential lies in the material science of lipid-based

nanoparticles, both to reﬁne existing nanoparticle strategies and to develop

new ones. Herein, the latest eorts to develop next-generation lipid-based

nanoparticles are covered by taking a nanoarchitectonics perspective and the

design, nucleic acid encapsulation methods, scalable production, and applica-

tion prospects are critically analyzed for three classes of lipid-based nanopar-

ticles: 1) traditional lipid nanoparticles (LNPs); 2) lipoplexes; and 3) bicelles.

Particular focus is placed on rationalizing how molecular self-assembly prin-

ciples enable advanced functionalities along with comparing and contrasting

the dierent nanoarchitectures. The current development status of each class

of lipid-based nanoparticle is also evaluated and possible future directions in

terms of overcoming clinical translation challenges and realizing new applica-

tion opportunities are suggested.

DOI: 10.1002/adfm.202203669

A. R. Ferhan, S. Park, H. Park, H. Tae, N.-J. Cho

School of Materials Science and Engineering

Nanyang Technological University

Singapore 639798, Singapore

E-mail: njcho@ntu.edu.sg

H. Park, J. A. Jackman

School of Chemical Engineering and Translational Nanobioscience

Research Center

Sungkyunkwan University

Suwon 16419, Republic of Korea

E-mail: jjackman@skku.edu

possibilities have enabled rapid develop-

ment and deployment of mRNA vaccines

for coronavirus disease- (COVID-)

prevention, for example.[] Ongoing

advances in the ﬁeld related to issues like

nucleic acid engineering, nanoparticle sta-

bility improvement, and immune safety

considerations along with application

trends have been extensively covered in

the past two years.[] However, the archi-

tectural design principles of lipid-based

nanoparticles have been relatively less

covered, especially from a material science

perspective that takes into account the

latest concepts and trends.

One emerging concept to rationally con-

trol the structure and function of nanoscale

objects is called nanoarchitectonics and has

been used to design advanced nanomate-

rials for biomedical, energy, and environ-

mental applications.[] Nanoarchitectonics

refers to the fusion of nanotechnology with organic chemistry,

supramolecular chemistry, and biology and involves combining

molecular-level manipulation with self-assembly and nanofabri-

cation.[] It is thus ideally suited to serve as a conceptual frame-

work to evaluate design trends and to suggest future directions

for the development of lipid-based nanoparticles for nucleic acid

delivery applications, especially as the ﬁeld grows to include not

only traditional lipid nanoparticles (LNPs) but also lipoplexes and

bicelles among various possibilities (Scheme1). The molecular

design and self-assembly properties of natural and synthetic

lipids underlies the core nanoparticle technology and under-

standing how to modulate the coassembly of lipids by themselves

and in combination with nucleic acids can lead to new types of

lipid nanoarchitectures and support future innovation.

In this review, we cover the latest eorts to develop next-

generation lipid-based nanoparticles for nucleic acid delivery

applications by taking a nanoarchitectonics perspective and criti-

cally analyze the design, production, and application prospects for

three classes of lipid-based nanoparticles as follows: ) traditional

LNPs; ) lipoplexes; and ) bicelles. The nucleic acid encapsula-

tion strategies for each nanoparticle type are introduced and help

to rationalize how dierent lipid nanoarchitectures can be useful

depending on the delivery objective and application scope. While

the main focus is on distilling molecular self-assembly principles

to build advanced lipid-based nanoparticle systems, we also dis-

cuss pertinent issues related to overcoming clinical translation

challenges and enabling scaled-up manufacture that can help to

realize new application opportunities.

The ORCID identiﬁcation number(s) for the author(s) of this article

can be found under https://doi.org/10.1002/adfm.202203669.

1. Introduction

Lipid-based nanoparticles are enabling new possibilities for

nucleic acid medicine across various dimensions such as infec-

tious disease vaccines and gene editing.[] These capabilities are

broadly applicable to various types of nucleic acids, including

deoxyribonucleic acid (DNA)[] and ribonucleic acid (RNA),[]

and some of the most promising types are messenger RNA

(mRNA),[] small interfering RNA (siRNA),[] transfer RNA

(tRNA),[] and self-amplifying RNA (saRNA).[] One of the

greatest beneﬁts of lipid-based nanoparticle technology is the

platform aspect whereby one nucleic acid molecule can be

swapped out with another one, while utilizing the same nan-

oparticle design principles and manufacturing route.[] Such

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

2. Lipid Nanoparticles

Lipid nanoparticles (LNPs) represent the most clinically

advanced non-viral vector for the delivery of ribonucleic acids

(RNAs).[] LNPs have been widely used for encapsulating and

delivering small interfering RNA (siRNA) in RNA interference

(RNAi) therapy (see ref. []) and, more recently, have gained

widespread attention by enabling the rapid development of

messenger RNA (mRNA)-based vaccines to prevent COVID-

infection.[] The remarkable success of LNPs in the clinical

setting can be attributed to a few major technological mile-

stones with respect to lipid design and scalable methods of

production. In particular, the utilization of ionizable lipids has

led to tremendous improvements in LNP functional perfor-

mance, especially transfection eciency.[]

At the same time, the shift from bulk preparation methods

such as batch extrusion to microﬂuidic-based production strat-

egies has allowed more precise control over the physicochem-

ical properties and overall quality of LNPs along with scalable

manufacturing capabilities.[] This enabled systematic studies

to be conducted to understand the relationship between the

physicochemical properties of LNPs, which depend on the

nanoarchitectonic design strategy and production methods,

and in vivo delivery performance. Finally, advances in microﬂu-

idic mixing technologies have facilitated scale-up through par-

allelization, which paved the way for large-scale manufacturing

Scheme 1. Applying the nanoarchitectonics concept to support innovation of lipid-based nanoparticle technology for nucleic acid delivery. A) Timeline

of lipid-based nanoparticle technology progress and conceptual advances in the nanoarchitectonics ﬁeld. B) Schematic representations of dierent

lipid-based nanoparticle types, including (1) traditional lipid nanoparticles (LNPs), (2) lipoplexes, and (3) bicelles. The nanoparticles are enabled by

the combination of controlling molecular self-assembly with various-shape lipids and using dierent nanofabriation strategies, highlighting how a

nanoarchitectonics perspective can be useful to support future innovation. FDA, United States Food and Drug Administration; EMA, European Medi-

cines Agency; DOPE, 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine; DOPC, 1,2-dioleoyl-sn-glycero-3-phosphocholine; DSPC, 1,2-distearoyl-sn-glycero-

3-phosphocholine; DHPC, 1,2-dihexanoyl-sn-glycero-3-phosphocholine; GML, glycerol monolaurate.

Adv. Funct. Mater. 2022, 32, 2203669

16163028, 2022, 37, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202203669 by Sungkyunkwan University, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

www.afm-journal.dewww.advancedsciencenews.com

of high-quality RNA-encapsulated LNPs.[] This ultimately

accelerated the translation of RNA-based LNP therapeutic

solutions from the benchtop to the clinical setting. Neverthe-

less, eorts to continually improve the performance of LNPs

in delivering dierent types of nucleic acid payloads are still

ongoing. Along this line, it is crucial to understand the inﬂu-

ence of dierent physicochemical parameters on the perfor-

mance of LNPs, and how speciﬁc design characteristics can

be incorporated to overcome limitations particularly associ-

ated with the transport of nucleic acids across biological bar-

riers. In this section, we will review the evolution of LNPs as

delivery systems for nucleic acids, especially RNAs, based on

their design principles and production platforms and shed

light on how nanoarchitectonic concepts are relevant to the

design of LNPs.

2.1. General Design Rules from Biological Inspiration

Within the context of nucleic acid delivery using lipid-based sys-

tems, a series of works have taken inspiration from the delivery

of native nucleic acids by lipoproteins.[] In biological systems,

native nucleic acids such as microRNAs are delivered to target

cells by lipoproteins such as high-density lipoproteins (HDLs).[]

This occurs via binding of apolipoprotein A-I (ApoA-I),

which is a major structural protein of HDLs, to speciﬁc scav-

enger receptors, such as SR-B that are present on target cell

membranes.[a] While HDLs are most popularly known for

their role in reverse cholesterol transport (RCT)[] in which

case they remove excess cholesterol from peripheral tissues and

transport it to the liver for biliary excretion, they have also been

found to play critical roles in the transport of other biomolecules

(e.g., signaling lipids, proteins, and nucleic acids).[] Endoge-

nous HDL particles are relatively small (≈–nm in diameter)

and are highly dynamic in terms of physicochemical properties.

Notably, they transition from a discoidal shape (i.e.,in their nas-

cent form) to a spherical shape (i.e., in their mature form) and

undergo changes in lipid composition during RCT (i.e.,due to

the adsorption, accumulation, and subsequent release of cho-

lesterol). Such morphological transitions oer insights into

key design criteria related to multiple aspects of nucleic acid

delivery, namely loading eciency, physiological stability, and

cell targeting and uptake eciency.[] Consequently, HDL par-

ticles represent excellent natural model systems for the formu-

lation of lipid-based nanoparticle drug delivery vehicles[] and

signiﬁcant work has been performed to develop synthetic lipid

and lipoprotein nucleic acid delivery vehicles that are inspired

by these systems and follow a few general design principles as

described below.[,]

With regard to nucleic acid loading, the shape and size

transition of HDL particles upon cholesterol adsorption

during RCT demonstrate how shape and size constraints

may limit the number of nucleic acid molecules that can

be loaded and transported per particle, with spherical parti-

cles oering higher loading capacities than discoidal particles

of the same diameter. As such, the diameter of the particles

needs to be carefully considered to maximize loading without

compromising particle distribution, particle stability, or cell

uptake eciency. In addition, the amount of loading is also

inﬂuenced by the loading method. While passive loading is

generally more convenient, the loaded molecules can diuse

or leak out during circulation even before the delivery vehicle

reaches target cells. On the other hand, loaded molecules

that are covalently bound to the delivery particle are retained

until reaching target cells. However, an active release mecha-

nism needs to be implemented to ensure that the molecules

in that case can be eectively unloaded once the delivery

vehicle reaches a target cell. Practically, this issue highlights

how careful consideration of the noncovalent intermolecular

forces in the delivery system is warranted in order to ensure

that the delivery is optimally suited to enable both nucleic acid

encapsulation and eventual release. In some cases, stimuli-

responsive systems that are sensitive to environmental condi-

tions such as solution pH have proven useful to achieve these

performance objectives.

To further enable systemic circulation and attain a long

circulation half-life, the particles should first be physically

robust under flow conditions at physiological tempera-

ture (i.e., maintain structural integrity and not degrade or

destabilize under shear stress).[] In addition, the particles

should avoid binding with blood components, which could

lead to interparticle aggregation and uptake by the mononu-

clear phagocyte system (MPS). MPS evasion is traditionally

achieved through the introduction of hydrophilic moieties or

the attachment of non-fouling polymers (e.g., poly ethylene

glycol or PEG) to the outer surface of the particles.[] In

terms of size, the particles should not be too small (i.e.,

not less than ≈ nm) to avoid kidney clearance.[] Along

this line, lipid conjugation strategies have been explored to

obtain lipidated therapeutic nucleic acids that boast higher

circulation times and are less prone to degradation.[] Ulti-

mately, such protection improves the overall biodistribution

of the nucleic acids and ensures that the nucleic acids effec-

tively reach target cells.[] Besides providing protection from

nuclease degradation, the conjugation of lipids to nucleic

acids has also led to new functionalities. For example, lipi-

dated nucleic acids can be designed to assemble into pre-

programmed shapes with precise control over geometrical

dimensions.[] This offers greater flexibility in terms of

architectural design and more faithfully mimics natural

delivery systems.

Finally, cell targeting and nucleic acid uptake in the cyto-

plasm represent two closely related critical considerations in

the development of targeted delivery vehicles. Inspired by the

ability of HDL particles to achieve targeted delivery via binding

of ApoA-I to speciﬁc scavenger receptors, cell targeting in syn-

thetic delivery vehicles can likewise be accomplished either

through opsonization by apolipoprotein A or E (ApoA or ApoE)

when the delivery vehicles are injected into the systemic circula-

tion or by functionalizing the delivery vehicles with natural tar-

geting molecules (e.g., ApoA or ApoE) or synthetic ligands.[]

Once the delivery vehicle reaches the target cell, the uptake

mechanism is also inﬂuenced by the targeting strategy. While

most lipid-based delivery systems rely on endocytosis, delivery

vehicles that undergo opsonization are internalized by cells

through phagocytosis.[] In addition, the mechanism of uptake

also depends on the surface charge, shape, size, and surface

chemistry of the delivery vehicle.

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

2.2. LNP Design Principles

In addition to the general design rules inspired by biological

nanoparticle examples such as HDL particles described above,

the design principles behind the formulation of LNPs also

expand on the early designs of synthetic liposome-based car-

riers, which consist of a neutral unilamellar lipid bilayer with

an aqueous core, encapsulating short-stranded nucleic acid

molecules (i.e., oligonucleotides).[] While liposomes can serve

as carriers for a variety of therapeutic molecules, their simple

structure and the use of largely neutral lipids for the delivery of

oligonucleotides posed several challenges.[] Firstly, weak inter-

actions between the lipids and negatively charged nucleic acids

hampered ecient entrapment during production. Secondly,

when introduced into biological systems, liposomes are prone

to clearance by the reticuloendothelial system (RES), opsoniza-

tion and destabilization, leading to reduced circulation time.

Over the years, the design has signiﬁcantly evolved to achieve

lipid-based delivery systems that overcome these limitations

through employing a combination of dierent types of phos-

pholipids. Within the context of nucleic acid delivery, it is vital

to ) ensure that the phospholipid carrier can eciently interact

with, condense, and entrap the nucleic acids, in a process col-

lectively known as encapsulation and ) eectively deliver the

encapsulated nucleic acids to the cells with minimal biological

interference during circulation. Encapsulation serves not only

to spatially entrap the nucleic acids within the delivery vehicle,

but to also isolate them from the external environment while

protecting their structure and function. Initially, permanently

cationic lipids were largely used to improve the interaction

between positively charged lipid membrane interfaces and neg-

atively charged nucleic acids.

While it was successful in increasing the encapsulation e-

ciency, the use of permanently cationic lipids attracted nonspe-

ciﬁc biological interactions and suers from problems associ-

ated with triggering the innate immune response (e.g., arising

from the activation of the complement system) and cytotoxicity

(i.e., arising from the activation of proapoptotic and proinﬂam-

matory cascades), thus failing to provide a sustainable solution

to issues related to reduced systemic circulation times. Early

eorts to mitigate these issues focused on utilizing cationic

lipids with delocalized charge while the issues were eventually

tackled through the design of a unique class of lipids with

switchable charge known as ionizable lipids, which achieved

signiﬁcant improvements in encapsulation eciency by being

positively charged at low pH during production and avoided

nonspeciﬁc biological interactions by remaining neutral at

physiological pH during circulation. A careful balance of head-

group and tail properties further enabled control over mem-

brane biophysics to guide the development of tailored LNPs

with heightened functionalities.

When combined with PEG lipid, helper lipid, and choles-

terol, the delivery system becomes distinct from a traditional

liposomal carrier in terms of structure, morphology, and lipid

composition, and is therefore referred to as an LNP. Archi-

tecturally, LNPs are characterized by a phospholipid bilayer

surrounding an electron-dense core in which ionizable lipid-

encapsulated nucleic acid payloads are densely packed in the

interior[] (Scheme2). Compared to liposomes, which consist

of phospholipid bilayers and cholesterol with distinct phase

transitions from ordered gel phase to liquid crystalline phase,

Larson et al. reported that cationic mRNA–LNPs transition

from an inverse hexagonal phase at pH values below the pKa

of the cationic lipid, to a lamellar phase above the pKa.[] Like-

wise, such phase changes also occur when the temperature is

increased, indicating that the inverted hexagonal phase is more

thermodynamically favorable. Based on these developments, it

becomes clear that the most critical aspects of LNP formulation

are nanoarchitecture design and lipid composition. By varying

the lipid composition, the interaction with the nucleic acid pay-

load can be carefully tuned. This directly impacts the encapsu-

lation eciency and inﬂuences the particle morphology and

overall charge, which taken together will determine the overall

performance of the delivery system. Conversely, the choice of

lipids also depends on the nature of the nucleic acid payload. In

the following subsections, we elaborate on the rationale behind

the choice of lipid compositions, as well as the considerations

for dierent payloads, in designing RNA–LNP delivery systems.

2.2.1. Lipid Composition

The formulation of LNPs is uniquely characterized by the uti-

lization of four dierent lipid excipients namely: ) ionizable

Scheme 2. Lipid nanoparticle architecture. Structural conﬁguration and key components of a lipid nanoparticle with nucleic acid payload, highlighting

how each component contributes to the overall structure and function.

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

lipids, ) helper phospholipids, ) cholesterol, and ) PEG lipid.

Each one of these excipients plays speciﬁc roles in ensuring the

overall eectiveness of LNP as a nucleic acid delivery vehicle.

Recently, Siegwart Daniel and co-workers developed selec-

tive organ targeting (SORT) nanoparticles, which enable the

delivery of mRNA and gene editing systems to non-liver tis-

sues.[] SORT nanoparticles include a supplemental molecule

whereby the chemical structure determines the LNP’s tissue-

speciﬁc activity. Their work demonstrates how a “ﬁfth excip-

ient” can further enhance the physicochemical properties and

targeted delivery performance of LNPs. In all cases, the key to

designing LNPs with high delivery ecacy lies in achieving an

optimal balance between particle stability, sustained systemic

circulation, and the ability to release the payload in a controlled

manner inside the target cell, all of which rely on the proper

selection of the dierent lipid excipients as described below.

Ionizable Lipids: Ionizable lipids are positively charged at

acidic pH and neutral at physiological pH. During LNP produc-

tion, they speciﬁcally serve to condense and entrap negatively

charged nucleic acids (e.g., siRNA, mRNA, pDNA) through elec-

trostatic interactions under acidic conditions. However, when

the LNPs are introduced to the physiological environment, the

ionizable lipids will switch to being neutral[b] (Figure1A).

Compared to permanently charged cationic lipids, the use of

ionizable lipids with neutral surface charge in the physiological

system is especially beneﬁcial in preventing the adsorption of

negatively charged biomolecules onto the LNP and preventing

rapid sequestration by immune cells during systemic circulation.

Once the LNP reaches the endosome, which typically presents

an acidic environment, the ionizable lipids will revert to being

positively charged. This leads to electrostatic interactions with

the anionic endosomal membrane, resulting in the formation of

a nonbilayer hexagonal phase that momentarily destabilizes the

endosomal membrane and thereby facilitates the release of the

nucleic acid to the cytosol[] (FigureB).

There are three major classes of ionizable lipids depending

on the tail structure of lipids, namely conventional two-

tail, multi-tail, and branched-tail ones (Figure2). While the

number/structure of the tail region largely aects the overall

geometry of the lipid self-assembly (i.e., a bulkier tail group

promotes a more cone-shaped structure), another important

factor to consider is tail saturation. Unsaturated ionizable lipids

generally increase ﬂuidity and a tendency to form a nonbilayer

phase, which facilitates membrane disruption and payload

release.[] One of the most widely used two-tail, unsaturated

ionizable lipids is DLin-MC-DMA, which is incorporated in

the FDA-approved siRNA–LNP drug (ONPATTRO). Multi-tail

ionizable lipids have three or more tails, which can result in

a more cone-shaped geometry to enhance endosome disrup-

tion.[] Along this line, branched-tail ionizable lipids have also

been developed, which may present methacrylate, acrylate, or

ester chains that extend either as a C branch near the head or

a C branch at the end. This can promote endosomal escape

arising from stronger protonation of spaced ionizable lipids at

endosomal pH and the increased cross-section of lipid tails,

allowing for the formation of a more cone-shaped geometry.[]

Figure 1. Illustration of the pH-dependent behavior of lipid nanoparticles due to ionizable lipids. A) Principle of the mechanism of ionizable lipids.

During formulation, the ionizable lipids present a positive charge at low pH (below its pKa) to promote favorable interactions with negatively charged

nucleic acids and facilitate nucleic acid encapsulation. During storage and in circulation under physiological pH (above its pKa), the ionizable lipids

will switch to being neutral and remain stable. Once the LNP reaches the endosome, which typically presents an acidic environment (below its pKa),

the ionizable lipids will become protonated again and resume a positive charge. B) Schematic representation of LNP interaction with the endosome,

leading to a change in lipid geometry due to the formation of ion pairs. Protonated ionizable lipids interact with anionic lipids adopting an inverted

cone geometry, which promotes the formation of the inverted hexagonal (HII) phase associated with membrane fusion.

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

Biodegradability can be improved by incorporating ester bonds

to promote degradation into nontoxic metabolites after intra-

cellular delivery.[] This is critical to reduce systemic accu-

mulation and eliminate potential side eects. Finally, since

chain branching increases the cross-section of phospholipid

tails (i.e., reduces chain packing) and results in the formation

of phospholipids with cone-shaped geometry that promote

spontaneous negative curvature to the membrane (i.e., due to

the inversion of cone-shaped structures), it generally leads to

better fusogenicity.

As a critical factor in the nucleic acid delivery process, it

is important to understand the structure-function relation-

ship between the molecular structure of ionizable lipids to the

transfection eciency of the LNPs, and how this in turn will

determine the therapeutic eciencies of LNPs.[] For example,

it has been found that lipid tail saturation, as well as the type

of linker between the lipid tail and the headgroup can signiﬁ-

cantly inﬂuence transfection eciency.[c,] More interestingly,

small structural variations such as those arising from dier-

ences in the double bond positions of unsaturated lipid tails

Figure 2. Structural classes of ionizable lipids. Three major classes of ionizable lipids depending on the tail structure, namely two-tail, multi-tail, and

branched-tail.

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

or branching positions, are sucient to produce signiﬁcant

variations in delivery performance.[,] Following this line, it

is anticipated that ionizable lipids with novel structures will

continue to emerge, which will facilitate the rational optimiza-

tion of LNP lipid composition. Such developments may also be

guided by the need for improvements in biodegradability, which

is a key feature for clinical translation, as well as the promising

opportunity of ﬁne tuning the lipid geometry through adjust-

ments in tail branching (i.e., in terms of extent and position),

tail length, and number of tails.

Helper Phospholipids: The roles of helper phospholipids are

various, ranging from facilitating membrane fusion, which

contributes to higher transfection eciencies, and increasing

bilayer stability.[] Helper phospholipids can have dierent

molecular structures and geometry. For example, ,-dio-

leoyl-sn-glycero--phosphoethanolamine (DOPE), which is

widely regarded as a fusogenic lipid, has a small phosphoe-

thanolamine head (i.e., comprising of a primary amine and

a phosphoric acid moiety), and two bulky and unsaturated

oleoyl chains, resulting in a cone-shaped geometry. It is sug-

gested that such geometry has the capability to stabilize the

non-bilayer inverted hexagonal (HII) phase, which transiently

forms during membrane fusion. Hence, the addition of DOPE

as helper lipids enhances fusogenicity, eectively increasing the

amount of LNP accumulation in target cells leading to higher

transfection eciencies. On the other hand, phosphotidylcho-

lines (PCs), such as ,-distearoyl-sn-glycero--phosphocholine

(DPSC) and ,-dioleoyl-sn-glycero--phosphocholine (DOPC)

form a cylindrical geometry and provide greater bilayer stability

relative to DOPE. In particular, the high melting temperature

(Tm) of DSPC, which is a saturated PC, leads to highly stable

bilayers at physiological temperatures. However, such high

stabilities will hamper endosomal release and reduce the e-

ciency of payload delivery. In comparison, DOPC, which is an

unsaturated PC, has a lower Tm and exists in the ﬂuid phase

at physiological temperatures. As a result, they contribute to

lower bilayer stability while allowing better eciencies in pay-

load delivery compared to DSPC. Taken together, DSPC oers

the highest bilayer stability, followed by DOPC and DOPE. In

other words, DOPC represents a good intermediate choice as a

helper lipid and the balance between bilayer stability and trans-

fection eciency is often achieved through the introduction of

cholesterol.

Cholesterol: The introduction of cholesterol can aect the

degree of lipid packing, which inﬂuences membrane ﬂuidity

and permeability of the bilayer of the LNP.[] Structurally, cho-

lesterol comprises of three domains namely the head, body, and

tail.[] The head is a hydroxyl group positioned in a tilted con-

ﬁguration above one of the cyclohexane or cyclohexene rings

of the body. The body comprises of a fused four-ring system

containing three cyclohexane or cyclohexene rings and one

cyclopentane ring, while the tail is a saturated alkyl side chain

of the cyclopentane ring. It is suggested that the ﬂexibility of

cholesterol arises from the lack of functional groups in the tail

and saturated nature of the alkyl chain. As a result, cholesterol

molecules can ﬁll gaps between other phospholipid molecules,

through the interaction of the hydroxyl head groups with the

aqueous phase of the phospholipid membrane (i.e., via polar

interactions and hydrogen bonding). Together, the body and

tail of the cholesterol molecule promote orderly arrangement

of the hydrophobic portion of the bilayer, leading to tighter lipid

packing of the membrane. Consequently, this reduces mem-

brane permeability and increases the overall structural integrity

of the LNPs. In addition, cholesterol can traverse lipid bilayers

and oer the possibility of equilibration across a concentration

gradient should the need arise. During circulation, cholesterol

also helps to prevent the diusion of the other lipid excipients

of the LNPs into high-density lipoproteins (HDL)-endogenous

lipids thereby improving the LNP stability in vivo.

Poly(ethylene glycol) PEG Lipid: The introduction of PEG lipid

serves to protect the LNP surface from opsonization, reticu-

loendothelial clearance, and destabilization during systemic

circulation.[a,] This is achieved through a combination of

steric repulsion and the formation of a hydration layer arising

from the extension of hydrophilic PEG chains from the LNP

surface. Based on the same mechanism, PEG-lipids also pre-

vent the aggregation of LNPs during production, storage, and

in circulation. Taken together, the structure of PEG extending

from the LNP surface increases the overall stability of LNPs,

which is largely advantageous. However, the prevention of non-

speciﬁc molecular adsorption on the LNP surface oered by

PEG lipids comes with a caveat as it also implies a signiﬁcant

reduction in the adsorption of apolipoprotein E (ApoE), ApoE

is an extracellular protein that plays a key role in inﬂuencing

the systemic circulation time of LNPs. The binding of ApoE has

been found to trigger a redistribution of lipids at the shell and

the core and plays a critical role in fusogenicity. In particular, it

facilitates the cellular uptake of LNPs through receptor-medi-

ated cellular entry.[] Hence, it is vital to strike an optimal bal-

ance when deciding on the appropriate amount of PEG to be

introduced to LNPs to ensure good protection against opsoni-

zation and reticuloendothelial clearance without compromising

fusogenicity and cellular uptake. The adjustment of PEG den-

sity also allows the tuning of LNP pharmacokinetics for dif-

ferent applications.

While there are obvious beneﬁts of including PEG lipids in

LNPs, it is worthy to note that PEG is widely known to trigger

immune responses and hypersensitivity reactions by activating

the complement system. In addition, it can also suer from

accelerated blood clearance (ABC) (i.e., following the binding

of anti-PEG antibodies). This has been intensively discussed,

most recently during the development, and ongoing evalu-

ation of COVID- vaccines, in which the potential of PEG

lipids to cause anaphylaxis especially among patients with

PEG allergy represents a signiﬁcant concern.[] In fact, there

is evidence that highlights the importance of the form of PEG;

speciﬁcally suggesting that PEG conjugated with lipids, not

PEG alone, is more likely to cause allergic reactions.[] Along

this line, it is also important to note that PEG-lipids have the

tendency to desorb and therefore escape the LNP architec-

ture to circulate freely, increasing the likelihood of triggering

immune responses. Considering these issues, several natural

(e.g., polyaminoacids, glucosaminoglycans) and synthetic (e.g.,

polyacrylamide, polysulfobetaine methacrylates) polymers have

been investigated as alternatives to PEG.

Cell-Speciﬁc Targeting: In addition to their primary roles in

encapsulation, maintaining particle stability, and supporting

systemic circulation, the lipid excipients can also play a critical

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

role in achieving cell-speciﬁc targeting as well as in modulating

biodistribution and uptake mechanisms. For example, it has

been established that ApoE adsorption is largely responsible for

LNP accumulation in the liver since lipid-associated ApoE binds

to low-density lipoprotein (LDL) receptors in the liver.[] Along

this line, several works have explored the possibility of varying

the helper lipid composition to modulate the LNP interaction

with ApoE.[,c,] It has been observed that LNPs containing

DOPE helper lipid interact more favorably with ApoE, therefore

resulting in increased levels of accumulation in the liver.[c]

On the other hand, LNPs containing DSPC helper lipid exhib-

ited weaker interactions with ApoE, leading to accumulation

in the spleen. This was attributed to the unsaturated nature of

DOPE, which leads to higher membrane ﬂuidity; greater mem-

brane ﬂuidity has been previously observed to increase protein

adsorption.[]

Based on this rationale, lipid compositions can therefore

be tuned to modulate the uptake mechanism for LNPs in line

with controlling protein adsorption. In the most common sce-

nario in which the cell uptake of LNPs occurs via the endocytic

pathway,[] the collective role of helper lipids and PEG lipids is to

reduce protein adsorption and prevent corona formation. In the

case of receptor-mediated endocytosis of ligand-targeting LNPs,

lipids can be conjugated with antibodies or aptamers. It is worthy

to note that for uptake via endocytosis, the ionizability of the

lipids represents another crucial factor that enables the LNPs to

withstand pH changes throughout the process, achieve endolyso-

somal escape, and eventual release of the nucleic acid cargo.

Alternatively, the cell uptake of LNPs can also occur via phago-

cytosis.[] In this case, lipid compositions should preferably be

tuned to induce protein adsorption and trigger opsonization

involving apolipoproteins (e.g., ApoA-I and ApoE), which would

lead to phagocytosis. Furthermore, the concept of decorating

pre-synthesized LNPs with native or synthetic biomolecules for

the purpose of modulating biodistribution and cell uptake can be

extended to improve cell targeting. Recently, SORT nanoparticles

developed by Siegwart Daniel and co-workers demonstrated the

ability to deliver mRNA with increased cell-targeting capabilities

(i.e., to non-liver tissues) by endowing LNPs with an additional

molecule that exhibits tissue-speciﬁc activity.[]

2.2.2. Payload Considerations

LNPs were initially developed and optimized for the delivery of

siRNAs. Although similar lipid compositions can theoretically

encapsulate other types of nucleic acids, it may be challenging

to apply the same formulation across the board considering the

dierences in molecular size, net charge, and conformational

structure.[] For example, small activating RNAs (saRNAs), and

microRNAs (miRNAs), both of which are also involved in the

RNA interference pathway like siRNAs, are double-stranded

and with lengths between  and  base pairs. On the other

hand, single-stranded mRNAs can and double-stranded plasmid

DNAs (pDNAs), can vary from several hundred to several hun-

dred thousand nucleotides. Since the negative charge of nucleic

acids arises from the phosphate backbone, their net charge is

directly correlated to the number of nucleotides and whether

they are single- or double stranded. In addition, nucleic acids

such as pDNAs and transfer RNAs (tRNAs) have well-deﬁned

conformational structures; pDNAs are circular and tRNAs have

a distinctive three-leafed clover structure comprising of three

hairpin loops. As such, adjustments to the lipid composition

are often required to accommodate dierent nucleic acids. Spe-

ciﬁcally, the ratio between the four lipid excipients, as well as

the ratio of ionizable lipid to RNA has been found to signiﬁ-

cantly aect the delivery ecacy both in vitro and in vivo.[]

siRNA versus mRNA: In order to exemplify the importance

of considering the dierent characteristics of nucleic acid pay-

loads on the ecacy of LNP delivery systems, we examine two

types of RNAs, namely siRNA and mRNA, that are most com-

monly encapsulated by LNPs. It is well known that siRNA and

mRNA structurally dier based on size and eective charge.

While siRNA has a short and well-deﬁned structure (i.e.,

double-stranded with a typical length of between – base

pairs[]), mRNA has a relatively less deﬁned structure with

variable lengths (i.e., threadlike, single-stranded with lengths

in the range of several hundred to a thousand bases[])—as

a reference, the SARS-CoV- single-stranded genomic RNA

is approximately  kilobases in length.[] These dierences

potentially lead to variations in lipid packing and the overall

LNP conﬁguration and size. For example, it was found that

while siRNA molecules in siRNA–LNPs are largely conﬁned

within multilamellar nanostructures (i.e., consisting of siRNA

molecules sandwiched between tightly packed concentric lipid

bilayers of ionizable lipids with PEG lipids on the outermost

layer), mRNAs, due to their larger size, can rearrange the

compartmental organization within the internal core of the

mRNA–LNPs and form inverted hexagonal nanostructures

instead.[] In this conﬁguration, mRNA molecules are encapsu-

lated within an aqueous environment surrounded by a mono-

layer of ionizable lipids with their headgroups facing inwards.

These individual compartments are in turn densely packed and

contained within a nonpolar environment bounded by a phos-

pholipid monolayer containing a mixture of ionizable lipids,

helper lipids, and PEG lipids with their headgroups facing the

external aqueous environment. Based on these arguments,

it becomes clear that the chemical, as well as architectural,

requirements of the lipid formulation would greatly depend

on size and structure of the intended encapsulated material.

For example, a smaller payload (e.g., siRNA) with well-deﬁned

structure would result in a denser, well-packed arrangement

of encapsulated nucleic acids in the interior of the LNPs and

potentially a smaller overall LNP. Conversely, a larger payload

(e.g., mRNA) with a less ordered structure would result in a

less dense arrangement within the LNPs and possibly a larger

overall LNP. Indeed, previous works have shown that siRNA

and mRNA are most eectively delivered in dierentiated LNP

formulations.[a]

Besides adjusting the ratio between the four lipid excipients,

various proprietary ionizable lipids with new structures and

functionalities have emerged driven by the need to enhance

the ecacy of mRNA–LNPs. In a noteworthy development,

Ball et al. explored the possibility of codelivery of siRNA and

mRNA using a single LNP formulation.[] To achieve this, they

employed a formulation comprising of an ionizable and biode-

gradable amine-containing lipidoid, cholesterol, DSPC, DOPE,

and PEG lipid. They began with two previously reported LNP

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

formulations that were found to be potent for siRNA and mRNA

delivery and systematically varied the lipidoid-to-RNA ratio, as

well as the percentage of lipidoid in relation to the helper lipid,

cholesterol, and PEG lipid. They also gradually changed the

composition of helper lipids from a blend containing majority

DSPC to DOPE, and introduced a negatively charged helper

polymer in the form of poly(sodium -styrenesulfonate) (PSS) to

enhance the ecacy. Surprisingly, they found that coformulation

of siRNA and mRNA in the same LNP was not only possible but

enhanced the ecacy of both drugs in vitro and in vivo. It was

suggested that since DSPC has two saturated aliphatic tails while

DOPE has a cis-double bond in each of its two aliphatic tails, the

packing of nucleic acids within the internal core of the LNP can

be carefully optimized by varying the ratio of DSPC to DOPE

to accommodate the signiﬁcant structural dierences of both

siRNA and mRNA. Furthermore, the additional negative charge

of PSS promotes electrostatic attraction in the particle and

increased overall particle stability. Separately, Hejdankovaet al.

recently developed a novel structural class of ionizable adaman-

tane-based lipidoids named XMaNs, which overcomes issues

related to tedious optimization of LNP components for a wide

range of nucleic acids.[] Brieﬂy, XMaN LNPs were formu-

lated with DMG-PEG, cholesterol, the helper lipid DOPE,

XMaN lipidoids (Figure3A), and combined with diverse types of

nucleic acids (mRNA, siRNA, pDNA, ′,′-cGAMP) (FigureB)

in a microﬂuidic device. Nucleic acid-loaded XMaN LNPs have

an intricate architecture (Figure C) and the microﬂuidic pro-

duction has allowed the fabrication of XMaN LNPs with good

sizes and excellent size distribution (FigureD). Using the best

performing member of the XMaN family, they were successful

in entrapping and delivering siRNA, mRNA, plasmid DNA as

well as a cyclic dinucleotide. Taken together, these works high-

light that the delivery of dierent nucleic acid payloads, whether

isolated or in combination, is highly achievable and can be sig-

niﬁcantly enhanced through a series of systematic and ecient

formulation optimization.

Figure 3. General overview of the approach to obtain LNPs incorporating XMaN lipidoids to achieve the delivery of dierent types of nucleic acids.

A) Four lipid excipients, including XMaN lipidoids, are ﬁrst mixed in ethanol, before they are combined, using a microﬂuidic platform, with B) a wide

range of nucleic acid payloads that are dissolved in an aqueous buer. C) A representation of a possible architectural conﬁguration of nucleic acid-

encapsulating XMaN LNPs. D) Cryo-TEM images of mRNA-loaded XMaN LNPs, their size distribution and evaluation of their functional stability after

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

Formulation Optimization: The evolution of LNPs from

encapsulating siRNA to mRNA highlights the need to optimize

the lipid formulation for each type of nucleic acid payload since

the formulations are not interchangeable. To achieve the most

ideal LNP formulation for the delivery of dierent nucleic acid

payloads, most works rely either on the one-factor-at-a-time

(OFAT) approach or the design-of-experiment (DoE) approach.

In the OFAT approach, one lipid is varied systematically while

keeping all other factors constant. While this method is advan-

tageous to reveal the direct correlation between the varied

parameter to the overall physicochemical property of the LNP,

it does not consider higher order interactions between the dif-

ferent lipid excipients. Consequently, a possible optimal formu-

lation arising from second or third order interactions between

the lipids could be missed. The method is also time consuming

and could lead to wastage of precious reagents, especially when

it involves optimizing lipid parameters for dierent nucleic

acids. The DoE approach, on the other hand, gains maximum

insights about the eect of varying several parameters to the

overall property of the LNPs through minimal number of

experiments.

Generally, it involves a screening design paired with

response surface methodology (RSM). Screening design ﬁrst

identiﬁes potential signiﬁcant variables (e.g., choice of ioniz-

able lipid, helper lipid, amount of cholesterol, ratio of lipid to

nucleic acid payload, etc.) while RSM determines the critical

degree of adjustment that should be applied to the most impor-

tant variable.[] The DoE approach provides information on

the interaction between several parameters, leading to more

accurate conclusions compared to the OFAT approach. More

importantly, it streamlines the formulation optimization pro-

cess. This was clearly demonstrated by Love et al., where the

formulation of LNPs containing erythropoietin (EPO)-mRNA

and C- phospholipid was optimized from a previous for-

mulation designed for siRNA–LNP.[] The study identiﬁed

the choice of helper phospholipid as the most important para-

meter for eective delivery EPO–mRNA; the use of DOPE was

favorable over DSPC, which was initially used in the formula-

tion of siRNA–LNP. The weight ratio of C- phospholipid

to mRNA was also identiﬁed as another important parameter.

Notably, second-order interactions (e.g., between C- mol%

and C--to-mRNA weight ratio) were detected, highlighting

the advantage of the DoE approach over the OFAT approach.

Such capability to detect higher-order interactions is especially

vital when performing formulation optimization involving ion-

izable lipids with novel functionalities, of which multiple layers

of interdependencies with other parameters could be dicult

to resolve.

In a separate demonstration, Dahlman and co-workers

described a cluster-based screening approach, which was

inspired by statistical DoE methodologies, to optimize the for-

mulation of mRNA–LNPs for the delivery of nebulized thera-

peutic mRNA.[b] The cluster-based approach involves, ﬁrstly,

the identiﬁcation of LNP chemical traits that inﬂuence the in

vivo delivery of nebulized LNPs (e.g., the amount of PEG added

to the LNP, the structure of the lipid-PEG, the charge of the

phospholipid and the presence or absence of cholesterol). A

hypothesis relating to the eect of varying each of these traits

is assigned as an axis in an N-dimensional chemical space and

asmall group of LNPs (e.g., around  to ) are then formulated

at the extreme of each axis. Within these “extreme groups,” a

selection process is further applied to determine whether )the

LNPs meet size and stability requirements, ) the stability

remains unchanged when LNPs are pooled together, and

) the pooled LNPs functionally deliver mRNA following nebu-

lization. Promising groups that pass this selection process are

then expanded and combined by formulating LNPs near the

intersection of these groups, with newly formulated LNPs also

subject to the same selection process. All these data, including

from groups that initially failed, were considered to design

subsequent LNPs. Through this regimented process, they

were able to eciently scan a diverse LNP chemical space, and

manipulate several pertinent parameters that inﬂuence in vivo

delivery, iteratively.

2.2.3. Physicochemical Parameters

The size and surface charge of the resultant LNPs have been

identiﬁed as critical parameters that aect not only the amount

of nucleic acid payload that can be packaged within the internal

core but also the in vivo behavior of LNPs.[] These parameters

inﬂuence the pharmacokinetic proﬁle of LNPs. For example,

smaller particles are observed to exhibit longer systemic circula-

tion times and slower clearance from the blood stream. More

importantly, it has been observed that particle size aects cell

uptake since the endocytic pathway varies with particle size,

which has implications on the intracellular processing and

overall ecacy of the LNPs.[] For instance, it was found that

the endocytosis of nm diameter LNPs is based on a dynamin-

dependent pathway while the endocytosis of LNPs larger than

nm in diameter is based on a clathrin-dependent pathway.[]

More recently, Hassetet al. evaluated the eect of LNP size on

mRNA vaccine immunogenicity on mice. They systematically

changed the LNP diameter at a ﬁxed lipid composition and

found that LNPs with smaller diameter did not lead to a signiﬁ-

cant immune response in mice. However, signiﬁcant immune

responses were observed in non-human primates regardless of

the LNP diameter.[] In a separate study, Chenetal. also found

that the hepatic gene silencing of FVII in mice was signiﬁcantly

inﬂuenced by the size of the siRNA–LNPs.[] Speciﬁcally, gene

silencing was far more ecient using LNPs with an optimal

size of –nm in diameter compared to smaller (i.e.,  nm

in diameter) and larger (i.e., nm in diameter) particles. The

observation was attributed to the inability of large LNPs to inﬁl-

trate the fenestrations in the liver vasculature and the instability

of smaller LNPs in serum. In addition, the ionizable lipids also

tend to dissociate more rapidly from smaller LNPs leading to

lower transfection eciencies. Aside from demonstrating the

eect of size on the ecacy of LNPs, these studies also high-

light the importance of producing LNPs with high monodis-

persity without overlapping size distributions. Likewise, the

phospholipid charge not only aects the interaction between

the lipid excipients and the nucleic acid payload, which deter-

mines the encapsulation eciency but also determines the

overall surface charge and zeta potential of the LNPs, which

strongly inﬂuences interactions with biological components in

vivo. While the uptake pathway is inﬂuenced by particle size,

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

the degree of initial adsorption of the LNP on the cell mem-

brane is governed by its zeta potential. It is therefore essential

to achieve precise control over these parameters.

2.3. Platform Development

As evidenced by the previous section, the key to the success

of LNPs as nucleic acid delivery systems lies in a solid under-

standing of the structure–function relationship. Clearly, the

molecular structures deﬁne the resultant geometry of the lipid

components, aect their interaction with the nucleic acids, and

inﬂuence the overall morphology and physicochemical proper-

ties of the LNP (i.e., the concept of nanoarchitectonics[]), all

of which in turn determine the behavior of LNPs in vivo, and

ultimately the performance of the LNPs within the context of a

nucleic acid delivery system.[]

While the design principles are built upon these correla-

tions, they would not have been realized without proper and

consistent control over the LNP fabrication process, which has

allowed systematic studies to be conducted in a reproducible

manner. At the same time, reproducibility and scalability rep-

resent key prerequisites for the clinical development of LNPs.

Along this line, eorts have been continually made to improve

the production methods of LNPs. The development of produc-

tion platforms serves two overarching objectives. Firstly, it aims

to achieve controlled and reproducible production of LNPs so

that fundamental investigations looking into the structure-func-

tion relationship can be performed reliably thereby facilitating

formulation optimization of LNPs. Second, it promotes scalable

manufacturing, which paves the way for the entry of LNPs to

the clinic. In the following subsections, we will elaborate on

how the production platforms have evolved over recent years,

particularly through the advent of microﬂuidics technologies.

2.3.1. Bulk Mixing Methods

Early production of liposome- and LNP-based delivery systems

relies on bulk mixing methods. Liposome-based carriers, par-

ticularly those containing siRNA, have been commonly pre-

pared through lipid ﬁlm hydration followed by extrusion, or

using the sonication and homogenization methods. Extrusion

involves passing the solution containing the lipid mixtures

through a series of ﬁlters to attain liposomes with the desired

size. The intended payload is encapsulated through the sonica-

tion or homogenization steps.

Besides extrusion, liposomes can be fabricated following the

ethanol injection method. In this method, lipid mixtures in

ethanolic solution are injected into a large volume of aqueous

solution under continuous stirring. The ethanol is then

removed via evaporation leading to the spontaneous formation

of vesicles. This method has several advantages over the ﬁlm

hydration method. Firstly, the method is relatively straightfor-

ward as it does not involve any intermediate processing steps.

Secondly, the method does not involve any sonication steps,

avoiding unnecessary damage to the phospholipid molecules.

Most importantly, it produces unilamellar vesicles with smaller

diameters and better monodispersity. While the ﬁlm hydration

and ethanol injection methods can be easily adopted and per-

formed at the laboratory scale, they are labor-intensive and lack

scalability. The crossﬂow injection method was then developed

by Wagneretal. to streamline the process and execution of the

ethanol injection method.[] In the crossﬂow injection method,

the setup comprises of two stainless steel tubes welded per-

pendicular to one another with a small injection hole at the

intersection between the tubes. Through this injection hole,

lipid mixtures in ethanol are injected into a stream of aqueous

buer leading to the spontaneous self-assembly of lipids to

form liposomes. It was found that the introduction of high

lipid concentrations under high injection pressures leads to

liposomes with a narrow size distribution. It was suggested that

unlike the lipid ﬁlm hydration and ethanol injection methods,

the crossﬂow injection method was suitable for continuous and

scalable manufacturing of liposomes. Overall, these bulk prep-

aration methods establish the guiding principles behind LNP

production and provide the basis for the development of more

advanced production platforms that are aimed to improve the

physicochemical properties of the LNPs.

2.3.2. Advent of Microﬂuidics

Microﬂuidics refers to the control and manipulation of ﬂuids

within channels, cavities, and structural features with dimen-

sions in the micrometer range.[] Microﬂuidics have found rel-

evance in a wide variety of biomedical applications, including in

the preparation of gene and drug delivery vehicles.[] It allows

greater consistency in the preparation process since the ﬂow

is well controlled by means of optimizing and ﬁxing the ﬂow

channel dimensions and ﬂow rates of the introduced ﬂuids,

thus leading to improved reproducibility. Besides, microﬂuidics

also increases throughput and opens the possibility of integra-

tion with digital automation.[] Within the context of LNP pro-

duction, microﬂuidics overcome challenges related to manual

experimental execution and batch processes. For example,

microﬂuidics allows controlled mixing of lipids in organic sol-

vent and nucleic acids in aqueous buer under continuous

ﬂow, with minimal operator intervention. This contrasts with

the lipid injection method, which, although principally similar

in terms of solvent mixing, is a batch process that requires con-

tinuous stirring and subsequent removal of the ethanol via an

evaporation step.

Motivated by the growing promise of nanoparticles as an

alternative to viral vectors for the delivery of nucleic acids into

speciﬁc types of cells for gene therapy, Wang et al. designed

a rapid developmental pathway for generating nanoparticle-

based vectors for highly ecient delivery of a variety of nucleic

acid payloads for gene therapy applications.[] It relied on a

combinatorial synthetic approach based on supramolecular

assembly executed within a custom-designed digital microre-

actor to achieve a convenient, ﬂexible, and modular method

for generating DNA-encapsulated supramolecular nanoparti-

cles. The employment of controlled microﬂuidic formulation

for the rapid discovery of potent siRNA-containing LNPs was

later reported by Anderson and co-workers.[] The method

involved the production of large amounts of siRNA–LNPs on

a microliter scale through stepwise ethanol dilution. Brieﬂy,

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

an ethanolic solution containing the lipid excipients was ﬁrst

rapidly mixed with an equal volume of aqueous solution con-

taining siRNA. This leads to lipid self-assembly into LNPs while

encapsulating the siRNA through electrostatic interactions. The

siRNA–LNPs were then further diluted with aqueous buer to

prevent LNP aggregation. These steps were performed within a

polydimethylsiloxane (PDMS) microﬂuidic channel, with peri-

odic trenches and ridges on the ﬂoor of the channel to promote

mixing. Around the same time, the group of Langer developed

a pattern-tunable microvortex platform for the mass produc-

tion and size control of lipid-polymer hybrid nanoparticles.[]

The platform involves the formation of symmetric microvor-

tices at the intersection of three inlets (i.e., two ﬂanking inlets

introducing lipid and lipid-PEG dissolved in % ethanol/water,

and one central inlet introducing PLGA dissolved in acetoni-

trile). These microvortices promote the rapid mixing of the

lipid excipients with PLGA, followed by nanoprecipitation and

the formation of the nanoparticles. The microvortex formation

was regulated through careful prediction and manipulation of

the D ﬂuid ﬂow patterns. Using this approach, they reported

that the resultant nanoparticle size can be controlled with high

productivity and low polydispersity by adjusting the ﬂow rates,

particularly by varying the Reynolds number between  and

. In a subsequent development, Valenciaet al. developed a

fully integrated microﬂuidic platform with a multi-inlet micro-

mixer for programmable and systematic mixing of large quanti-

ties of precursors to make nanoparticles.[] The platform was

successful in synthesizing nanoparticles with a wide range of

properties by combining  dierent precursors in dierent

ratios, achieving  dierent formulations with dierent sizes

and surface compositions. Taken together, the obvious advan-

tages of ﬂow-based platforms oered by microﬂuidics within

the context of preparing nanoparticle-based delivery systems

have led to the realization of the importance of rapid mixing,

and further development of microﬂuidic mixing technologies

thus followed suit to expand its capabilities.[]

2.3.3. Microﬂuidic Mixing Architectures

Due to their nonpolar tails, lipids readily dissolve in ethanol

but not in aqueous solvent. The addition of lipids dissolved in

ethanol to an aqueous buer containing nucleic acids results

in the formation lipid bilayer structures upon hydration, which

subsequently interact with the nucleic acids to form small mul-

tilamellar liposomes with nucleic acids sandwiched between

the bilayers. Based on this principle, LNPs are formed when

there is a rapid increase in polarity within the bulk environ-

ment when two miscible phases combine (e.g., when lipids

dissolved in an organic solvent are rapidly introduced to an

aqueous buer). This occurs through the supersaturation of

lipid molecules, which leads to their self-assembly to form

LNPs[] (Figure4A). Along this line, it is important to note

that in order for the multilamellar bilayers to eciently trap

the nucleic acids, the ethanol concentration must be near the

region of breakdown of the liposomal structure.[c] In other

words, while increasing the concentration of ethanol leads to

higher nucleic acid entrapment, it also leads to a higher degree

of membrane destabilization. Hence it has been found that

dilution of the ethanol/buer mixture to around % ethanol is

typically ideal for the formation of LNPs with good encapsula-

tion eciency and sucient stability.

In microﬂuidic channels, the Reynolds (Re) number is

usually smaller than , which means that the concentration

of lipids in the organic solvent is dominated by molecular

diusion. Based on this principle, microﬂuidic mixing archi-

tectures have been developed with the aim of enhancing the

mixing eciency and attain a uniform population of LNPs

while achieving excellent entrapment of the nucleic acid

payload.[c,] Microﬂuidic mixing can be broadly categorized

into passive and active mixing; passive mixers solely rely on

the geometry of the microﬂuidic channels while active mixers

control the ﬂow of sample solutions by electrohydrodynamic

disturbances. Within the context of LNP fabrication, most

micromixers can be categorized as passive mixers. Some

examples of advanced passive micromixers used for LNP

production include the T-junction mixer, staggered herring-

bone micromixer (SHM), hydrodynamic ﬂow focuser (HFF),

bifurcating and bae mixers (FigureB). These micromixers

dier in their D structure and induce rapid mixing between

the organic and aqueous phases in a controlled environment

through dierent mechanisms.

T-Junction Mixer: T-junction mixing is a rapid mixing

method where input streams are introduced directly opposite

to each other at very high ﬂow rates (i.e., – mL min−),

with a perpendicular output.[] When the aqueous phase con-

taining nucleic acids and organic phase containing lipids are

introduced as inputs, the collision of the input ﬂows results in

turbulent mixing, resulting in the formation of nucleic acid–

LNPs. It has been found that the particle size can be reduced

by increasing the ﬂow rates. Likewise, increasing the ﬂow rates

also resulted in signiﬁcant improvements in the polydispersity

index (PDI), which is an indication of the size distribution.[a]

However, as T-junction mixing is operated at high ﬂow rates, it

is not a preferred method especially in the laboratory where the

reagents for high throughput screening need to be conserved.

Hence, T-junction mixing is favored for large-scale production

of siRNA–LNPs.

Staggered Herringbone Micromixer (SHM): The staggered

herringbone micromixer architecture was introduced by

Stroocketal. as a passive mixing method for steady pressure-

driven ﬂows in microchannels at low Reynolds number.[] It

comprises a series of asymmetric protrusions, which mix sol-

vents based on chaotic advection. Within the context of LNP

production, the characteristic diusion length between the

organic lipid and nucleic acid–aqueous streams is signiﬁcantly

reduced. This allows controllable mixing to occur faster (i.e.,

in the millisecond scale) than the characteristic timescale for

lipids to aggregate, hence producing LNPs with uniform sizes.

Unlike T-junction mixing, microﬂuidic devices with SHM are

suitable for mixing small quantities of input solutions, hence

making them ideal for screening applications in the laboratory.

Nevertheless, scale-up can be achieved through parallelization

of microﬂuidic chips.[] For example, high production rates of

LNPs up to mL min− have been achieved by incorporating

six sets of SHM in a single microﬂuidic device.[] Most notably,

SHM devices were commercialized by Precision NanoSystems

in their ﬁrst-generation NanoAssemblr platform.

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

Hydrodynamic Flow Focuser (HFF): Hydrodynamic ﬂow

focusing (HFF) was ﬁrst reported by Jahn et al. and involves

the continuous-ﬂow hydrodynamic focusing of the organic

phase (i.e., lipids dissolved in organic solvent) by the aqueous

phase.[] Typically, a central stream introduces the organic

phase, which is pinched by two aqueous streams. These ﬂows

are laminar, which results in a well-deﬁned interface between

the two phases where interfacial forces can be carefully tuned

by adjusting the operating parameters. Two main operating

parameters are the total ﬂow rate (TFR) of the two phases and

Figure 4. Self-assembly process of LNPs within microﬂuidic systems. A) The self-assembly process of nucleic acid–LNPs in a microﬂuidic mixer begins

when lipid excipients (i.e., ionizable lipid, PEG lipid, helper lipid, and cholesterol) dissolved in ethanol are rapidly mixed with the nucleic acid payload

(e.g., siRNA, mRNA, etc.) dissolved in an aqueous buer. This initially occurs at a low buer pH whereby the ionizable lipid will assume a protonated

form and bind, via electrostatic interactions, to the negatively charged backbone of the nucleic acid. Concurrently, the increasingly polar environment

drives the formation of vesicles and the encapsulation of the nucleic acid. As the mixing progresses, and with increasing buer pH, the ionizable lipid

becomes increasingly neutral, leading to the fusion of adjacent vesicles to form the interior core of the LNP. The extent of vesicle fusion is inﬂuenced

B)Major microﬂuidic mixing architectures for LNP formulation (top to bottom): T-junction mixer, staggered herringbone micromixer, bifurcating mixer,

and bae mixer. C) The eect of varying ﬂow rate on the mixing proﬁle of ethanol (red) and water (blue) within a bae mixer, visualized through a

of an example of microﬂuidics parallelization, showing (from top to bottom) the arrangement of mixing channels, each mixing unit, and each mixing

cycle within a parallelized microﬂuidics device, which adopts the staggered herringbone architecture. Evaluation of the in vivo delivery performance

of E) siRNA–LNPs and F) mRNA–LNPs obtained using the parallelized device compared to bulk mixing. The evaluation was performed by means of

administering to mice. The mRNA encodes for luciferase and the in vivo delivery performance was evaluated by directly visualizing the luminescence

from selected regions of the mice. The results clearly show that the in vivo delivery performance of nucleic acid–LNPs obtained via microﬂuidic mixing

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

the ﬂow rate ratio (FRR) between the two phases. Both param-

eters inﬂuence the width of the central stream, which is repre-

sentative of the degree of hydrodynamic focusing. Since LNP

formation is based on a reduction in lipid solubility at the inter-

face between the two phases, the size and size distribution of

the resultant LNP are dependent on diusion, which is inﬂu-

enced by the degree of hydrodynamic focusing. Speciﬁcally, it

was found that increasing the FRR can reduce the LNP size

while attaining narrower size distribution, and higher TFRs

resulted in larger LNPs only at low FRRs.

Bifurcating and Bae Mixers: Motivated by the commercial

success of SHM devices for the production of LNPs, Precision

NanoSystems later introduced the NxGen, which is a micro-

mixing architecture consisting of a sequence of bifurcating

mixers for scalable, nonturbulent mixing.[] Brieﬂy, the organic

and aqueous phases containing lipids and nucleic acids, respec-

tively, are introduced from two pronged inlets that merge into

a single ﬂow. Downstream, a series of toroidal mixers promote

chaotic advection to the merged stream. The stream initially

splits upon entry into the ﬁrst toroidal mixer and travels dif-

ferent path lengths before merging again. This process, which

induces rapid mixing in a single layer by centrifugal force, is

repeated a few more times leading to the formation of nucleic

acid–LNPs with higher encapsulation eciency and reproduci-

bility compared to SHM. More importantly, it allows production

rates to be increased by up to -fold. A similar architecture

that operates based on a similar principle is the bae mixer.

In bae mixers, the organic and aqueous phases are also intro-

duced from two pronged inlets that merge into a single ﬂow.

However, instead of having a series circular, toroidal mixtures,

it comprises sharp perpendicular turns to induce chaotic advec-

tion. Bae mixers can produce LNPs with controllable sizes in

the range of –nm, by varying the TFR, FRR, and device

dimensions (FigureC).[] Taken together, both bifurcating and

bae mixers are single-layer mixing architectures, which serve

as excellent alternatives to SHM and HFF for LNP production.

Each of the abovementioned approaches has their advan-

tages and limitations. For example, while the conventional

method of lipid ﬁlm hydration followed by extrusion is easy to

perform, it is a time-consuming multistep production process

with relatively low encapsulation eciencies (≈–%). The

crossﬂow injection method is highly suitable and is already in

use for large-scale production in the industry but is less suited

for laboratory-scale research. On the other hand, microﬂuidic-

based approaches allow controlled mixing and generally result

in relatively high encapsulation eciencies. For example, the

T-junction and HFF methods can produce uniform LNPs,

has broad solvent compatibility and can achieve high encap-

sulation eciencies (≈%). The SHM method can produce

highly uniformed particles with a polydispersity index of less

than .. In addition, the encapsulation eciency of the SHM

method is very high (>%). However, due to the architecture

of the mixture, SHM is prone to clogging within the channel.

All microﬂuidic approaches require parallelization for scale-up

(FigureD). Nevertheless, parallelization has shown to improve

in vivo delivery for siRNA (FigureE) and mRNA (FigureF)

In conclusion, compared to conventional preparation methods,

the advent of microﬂuidics especially with the development

of microﬂuidic mixing architectures has oered better control

over the production process of LNPs leading to LNPs with

greatly improved physicochemical properties and encapsula-

tion eciencies. More importantly, microﬂuidics has paved the

way for more rapid screening of LNP formulations and scale-

up manufacturing, as well as the prospect of automated LNP

synthesis.[]

2.4. Applications

A better understanding of the structure–function relationship,

as well as improvements in the design principles and platform

development has enabled the exploration of LNPs for use in dif-

ferent clinical applications. Conversely, the promise of LNPs as

delivery systems in certain applications has driven the reﬁne-

ment of their design principles and platform development. In

the following subsections, we will review how two main appli-

cations, namely RNAi therapy and mRNA-based vaccines for

COVID-, have led to the progress of LNPs.

2.4.1. RNAi Therapy

The development of LNPs was originally motivated by RNA

interference (RNAi) therapy.[] RNAi refers to a fundamental

cellular mechanism for silencing gene expression, which can

be harnessed for reducing the expression of pathological pro-

teins and therefore exploited as a novel therapeutic approach.[]

In brief, it relies on the speciﬁc binding and silencing of thera-

peutic targets by means of introducing siRNA molecules.[]

While the therapeutic mechanism is very promising since the

approach can reversibly silence any gene, the eective delivery

of siRNA proved to be challenging.[b,]

LNPs were developed as an alternative to viral vectors

and represent the most advanced non-viral vector for gene

delivery.[c,] Nevertheless, formulation optimization had to

be conducted rigorously to achieve maximum siRNA encap-

sulation and transfection eciencies.[b,] These studies had

been largely facilitated by the advent of microﬂuidics.[b,,,]

For example, Chenetal. demonstrated the utilization of con-

trolled microﬂuidic formulation leading to the rapid discovery

of potent siRNA–LNPs.[] In the industry, the success of RNAi

can be attributed to T-junction mixing, which enabled the pro-

duction of stable siRNA–LNPs that are capable of potent knock-

down of Apolipoprotein B (ApoB) in non-human primates.[b]

This is a prominent study as it represents one of the earliest

reports highlighting the eectiveness of RNAi in a large animal

study, revealing the signiﬁcant reduction of ApoB mRNA and

protein even beyond  days. Separately, Merck & Co. also

adopted a similar approach by employing T-junction mixing

to produce siRNA–LNPs targeting the murine gene Ssb.[a]

Likewise, they were successful in reducing the target mRNA

levels in vivo by over %. Undoubtedly, microﬂuidic devices

have proven beneﬁcial for formulation optimization within the

context of siRNA–LNPs, as they allow mixing of small amounts

of reagents, which is ideal for screening lipid structures with

minimal wastage. In addition, LNPs produced using micro-

ﬂuidic devices are typically smaller, and more uniform, than

those produced via bulk preparation techniques. This has led

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

to the discovery of novel lipid structures for potent hepatic

gene silencing in vivo. In addition, with the ability to carefully

tune the mixing rates between the organic phase containing

the lipids and the aqueous phase containing the nucleic acids,

the size of the LNPs can be better controlled to understand the

eect of LNP size to gene silencing performance. For example,

the use of high ﬂow rates to produce small LNPs corresponded

to better gene silencing capability compared to the use of low

ﬂow rates, which produced large LNPs. All these advances are

critical in the eort to overcome outstanding challenges in the

delivery of LNPs. For example, while an siRNA–LNP formula-

tion has been established and developed into an FDA-approved

prescription medicine (i.e., known as patisiran and marketed as

ONPATTRO, NCT) to treat polyneuropathy in people

with hereditary transthyretin-mediated amyloidosis, clinical

trials led by Roche/Ionis Pharmaceuticals (NCT) and

WAVE Life Sciences/Takeda (NCT and NCT)

to explore similar approaches with LNP candidates for treating

Huntington’s disease are ongoing.

2.4.2. mRNA Vaccine for COVID-19

The translation of mRNA–LNPs to the clinic had been recently

realized in the form of mRNA vaccines against COVID-

.[a,a,] This development not only highlights the beneﬁts of

mRNA-based vaccines but also establishes the biocompatibility

of LNPs as a delivery vehicle.[] While the success accelerated

by the COVID- pandemic was unprecedented considering

there had been only one FDA-approved LNP-based siRNA drug

(i.e., ONPATTRO) prior to that,[] it was not entirely surprising

considering an extensive amount of eort has been put in to

understand the structure-function relationship and improve

the performance of LNPs over the last decade. Such collective

eorts have facilitated the quick adoption of LNPs, which was

initially largely developed for siRNA, to incorporate mRNA pay-

loads instead. As discussed in the previous sections, the evolu-

tion of design principles and platform development play critical

roles in incrementally improving the design and manufacturing

of potent LNPs, culminating in the maturity of the LNP-based

delivery systems that meet the requirements of good manufac-

turing practice and safety. Key milestones that contributed to

the realization of LNP-based mRNA vaccine include the intro-

duction of ionizable lipids as well as the development of micro-

ﬂuidics and advanced micromixing architectures. In particular,

microﬂuidic platforms that oer high degrees of parallelization

for large-scale manufacturing (e.g., Precision NanoSystems’

NanoAssemblr platform) have played a pivotal role in ensuring

the clinical success of mRNA–LNPs. In addition to the avail-

able COVID- vaccines (NCT, NCT, etc.),

other LNP–mRNA vaccines against inﬂuenza (NCT

and NCT), Zika virus (NCT), rabies virus

(NCT), and multiple other viruses are currently under-

going human clinical trials.

With the establishment of such platforms, LNPs containing

mRNA as well as other types of nucleic acids continue to be

explored for other therapeutic applications and following

other routes of administration.[] For example, Dahlman and

co-workers recently optimized the formulation of LNPs so that

mRNA can be delivered to the lungs in the form of nebulized

aerosols.[b] Based on all the data gathered via a cluster-based

screening approach, they found that PEG lipids are critical in

attaining stable LNPs. Furthermore, nebulized mRNA delivery

can be improved with the use of cationic helper lipids and a

higher molar fraction of PEG, and more PEG is needed for for-

mulations with cationic helper lipids than those with neutral

lipids. The resultant LNPs optimized following their approach

exhibited greater structural stability after nebulization, as

opposed to LNPs designed for systemic delivery. In essence,

their work demonstrated the possibility of adopting nucleic

acid-based LNP delivery systems for developing therapeutic

application following dierent routes of administration, at the

same time revealed that formulations developed for systemic

delivery cannot be simply adopted for inhalation, which under-

scores that the importance of formulation optimization.

3. Lipoplex

While LNPs are fabricated by direct coassembly of ionizable

lipids, helper lipids, PEGylated lipids, sterols, and nucleic acids

during solvent mixing, another intriguing lipid nanoarchi-

tecture for delivery applications involves the aqueous-phase

mixing of preformed liposomes with nucleic acids, termed

“lipoplexes.” The lipoplex strategy evolved from early eorts[]

to encapsulate nucleic acids into liposomes composed of ani-

onic and zwitterionic lipids through passive loading, which

faced challenges such as low eciency in terms of nucleic acid

entrapment and delivery.[a,] Although negatively charged

liposomes may reduce the leakage of contents with increased

stability and protection, inecient entrapment (ranging

%–% with the highest eciency obtained by the freeze/

thawing procedure[a]) due to the repulsion between the mem-

brane and nucleic acid was the major concern of encapsula-

tion method.[] To enhance interactions between liposomes

and nucleic acids, cationic lipids have been utilized in order to

facilitate electrostatic attraction with negatively charged nucleic

acid molecules.[] As a result, passive loading of nucleic acids

into liposomal interiors is not required, and instead electro-

static interactions between liposome surfaces and nucleic acids

spontaneously drive lipoplex self-assembly. The ﬁrst lipoplex

demonstrations utilized liposomes composed of cationic and

zwitterionic lipids that were mixed with pDNA[] or mRNA[]

and exhibited superior transfection eciency (especially with

RNA transfection yielding - to -fold more) with less

cytotoxicity compared to a polymer-based delivery vehicle, a

diethylaminoethyl (DEAE)-dextran complex, that was tested in

parallel.

Motivated by these pioneering studies, lipoplexes have

emerged as a powerful platform technology that can potentially

take a single liposome formulation and mix it with dierent,

customizable nucleic acids to create distinct lipoplex options for

various applications. Indeed, while LNPs incorporating ioniz-

able lipids have demonstrated high practical utility for mRNA

vaccine applications, lipoplexes have reached human clinical

trials for applications such as cancer therapy.[] Even so, there

continue to be key issues related to the cationic lipid design

along with paired neutral lipid choice, processing parameters,

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

and the corresponding inﬂuence on the nanoparticle’s architec-

ture and delivery functionality. In this section, we cover these

design points and also discuss selected cases that directly com-

pare the performance of LNPs and lipoplexes.

3.1. Nanoarchitecture Design Principles

3.1.1. Lipid Composition

The composition of a lipoplex is deﬁned by two lipid types:

) cationic lipid and ) neutral (helper) lipid. As with LNPs, in

lipoplexes, cationic lipids mainly interact with nucleic acids while

neutral lipids inﬂuence self-assembly properties and assist with

conferring particle stability. A noteworthy distinction is that the

majority of cationic lipids used in lipoplex formation are perma-

nently charged, meaning that pKa is larger than  (or without

pKa), while a few cases feature slightly ionizable lipid-like charac-

teristics (e.g., DC-Chol, DODMA, multivalent cationic lipids, etc.

that are discussed below). Hence, the stability and eciency of

the complex are highly dependent on the neutral lipid choice and

molar charge ratio (N/P, cationic lipid-to-nucleic acid ratio). It

is also worthy to note that the transfection eciency might vary

depending on the cell types, and there are several comprehensive

studies comparing commercially available transfection reagents

such as Lipofectamine, RNAiMAX, and Lipofectin.[]

Cationic Lipid: Cationic lipids incorporated in lipoplexes

should be able to capture the nucleic acid, maintain stability

before entry, facilitate cellular entry (via fusion or destabili-

zation), and subsequently release nucleic acids in the target

cell.[] Lipid structures can be divided into three functional

units consisting of a hydrophilic headgroup, hydrophobic tail,

and linker, connecting the former two moieties. Modifying

each element can aect the stability (compaction with nucleic

acid), disassembly, and phase behavior of the lipoplex, which

have been extensively discussed.[] Depending on the tail

group, cationic lipids can be categorized into aliphatic chains or

cholesterol-based derivatives while the majority of headgroups

are composed of permanently positively charged, quaternary

ammonium or amines and their derivatives (Figure5).

The ﬁrst cationic lipid used to form a DNA lipoplex

was ,-di-O-octadecenyl--trimethylammonium propane

(DOTMA), which is a double-chain quaternary ammonium

with ether bonds. One of the major commercial transfection

reagents, Lipofectin, is based on DOTMA/DOPE (/ mol%)

composition and is recommended to deliver nucleic acids to

endothelial cells.[] Since then, several other related com-

pounds with aliphatic chains were developed to improve trans-

fection eciency and reduce cytotoxicity. ,-dioleoyl--trimeth-

ylammonium-propane (DOTAP) with ester bonds,[] dimethyl-

dioctadecyl ammonium bromide (DDAB), and single-chain

lipid, cetyltrimethyl ammonium bromide (CTAB), were subse-

quently utilized to fabricate lipoplexes while the most widely

used cationic lipids have remained DOTMA and DOTAP.

Based on these lipids, various derivatives were also developed

by replacing the methyl group with the hydroxyethyl group,

leading to lipids such as ,-dioleoyl--dimethyl-hydroxyethyl

ammonium bromide (DORIE) and ,-dioleoyloxypropyl--di-

methyl-hydroxyethyl ammonium chloride (DORI) that demon-

strated enhanced compaction of nucleic acids with lipids.[b,b]

Aside from dioleoyl derivatives, dimyristoyl compounds were

recognized later to display higher transfection eciency, leading

to the development of another commonly used cationic lipid,

,-dimyristyloxypropyl--dimethyl-hydroxyethyl ammonium bro-

mide (DMRIE), which is currently used in clinical-stage lipoplex

formulations together with DOTAP and DOTMA.[]

Figure 5. Lipoplex components and nanoarchitecture organization. Schematic illustration of the lipoplex and molecular structures of the cationic lipids

and neutral lipids.

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

Furthermore, multivalent cationic lipids have gained atten-

tion by showing higher transfection eciencies due to their

ability to condense nucleic acids more tightly and to par-

tially ionize to act as a buer to better protect the nucleic

acid from degradation compared to their monovalent coun-

terparts.[] For example, another widely used transfection

agent, Lipofectamine, is based on polycationic ,-dioleyloxy-

N-[(sperminecarboxamido)ethyl]-N,N-dimethyl-l-propana-

minium triﬂuoroacetate (DOSPA) and DOPE at a ≈.: molar

ratio and is often considered a benchmark composition for

evaluating transfection eciency.[] Another advantage mul-

tivalent lipids might confer is their high eciency to capture

nucleic acid with a lower amount of lipid (lower N/P) compared

to the monovalent ones.[]

On the other hand, cholesterol derivatives such as

β-[N-(N′,N′-dimethylaminoethane)carbamoyl]-cholesterol

hydrochloride (DC-Chol) are another popular class of cationic

lipids used in lipoplexes.[] As cholesterol itself is commonly

used as a neutral, helper lipid in lipoplex formation to increase

membrane packing by embedding its bulky steroid moiety

between hydrophobic tails of neighboring lipids, the addi-

tion of a cationic moiety in cholesterol derivatives was found

to exhibit superior performance, enabling delivery of nucleic

acids into various cell types in vitro and in vivo.[b,] Also,

DC-Chol contains a carbamate linker and a tertiary amine with

a relatively low pKa value (% charged in buer[]) compared

to other permanently charged lipids, providing higher biodegra-

dability and aiding the ecient release of nucleic acid.[] Due

to these advantages, a DC-Chol and pDNA lipoplex formulation

has been shown to exhibit low toxicity upon local or systemic

administration and has entered clinical trials for the treatment

of genetic disorders and cancer along with its multivalent deriv-

ative with spermine, N-cholesteryl-Spermine (Genzyme lipid

, GL).[b,]

Typically, ionizable lipids used in lipid nanoparticle formula-

tion possess apparent pKa less than  in order to exhibit posi-

tive surface charges at acidic (endosomal) pH while neutral

charges at physiological pH. Abovementioned DC-Chol and

-dioleyloxy-N,N-dimethyl--aminopropane (DODMA) possess

pKa value around , oering pH-dependent nucleic acid and

lipid interactions.[,] For example, DODMA molecules rear-

range toward the hydrophobic region of the bilayer at high pH,

oering an increased ﬂip-ﬂop rate that might aect the fusion

process.[] As ionizable lipids may oer more ecient endo-

somal escape by membrane rearrangement upon contact with

anionic lipid membrane, there were several eorts to discover

ecient ionizable lipids that can be used without a neutral lipid,

leading to a development of a series of aminoglycoside deriva-

tives that exhibit higher stability and smaller mean diameters

compared to cationic lipid-based lipoplexes.[] Other eorts

include an ionizable lipid based on vitamin E-scaolds[] and

α-amino-lipophosphonates to target dendritic cells.[]

Neutral Lipid: Neutral lipids have many useful functions as

part of lipoplex compositions, including assisting with forma-

tion-related phase changes and reducing interparticle aggrega-

tion. While certain cationic lipids forming lipoplexes without

neutral lipids can be functional to deliver nucleic acids, appro-

priate inclusion of neutral lipids such as DOPE and/or choles-

terol can signiﬁcantly improve the stability and the transfection

activity of lipoplex.[] Thus, neutral lipids in lipoplex are also

called “co-lipids,” functioning similarly to “helper lipids” in

LNP excipients. Interestingly, the choice of neutral lipids might

signiﬁcantly enhance or hamper the transfection eciency.

For example, DOTMA,[,b] DOTAP,[] and DC-Chol[]

have shown that the incorporation of DOPE largely augmented

transfection eciency whereas DOPC hampered the nucleic

acid delivery both in vitro and in vivo.[] These propensities are

closely related to the geometry of lipids and in turn, the archi-

tectural conﬁguration of lipoplex that will be more discussed in

the next Section...

Cholesterol is another widely used neutral lipid[] aiding col-

loidal stability of lipoplex and participating in membrane fusion

by promoting contact sites for lipid mixing and expanding the

pore induced by fusion.[] Also, its inclusion in lipoplex has

shown higher resistance to serum-induced aggregation and

protein binding, which may explain enhanced transfection.[]

Indeed, a quantitative analysis performed by replacing DOTAP

lipids with cholesterol (or DOPE) showed reduced anity

of negatively charged plasma proteins that might be respon-

sible for protein corona formation.[] Interestingly, the choice

between cholesterol and DOPE may vary depending on the

tail group of cationic lipid.[] When various chain lengths of

lipids were tested with an equimolar ratio of either cholesterol

or DOPE inclusion, the highest eects were observed with sat-

urated  carbon or  carbon chains, respectively, while both

neutral lipids boosted the transfection eciency compared to

cationic lipid alone.[] Therefore, it should be noted that the

selection and optimum mole fraction of cationic and neutral

lipid depend on the lipid types, possibly N/P charge ratio, types

of nucleic acid, and cells.

3.1.2. Architectural Conﬁguration

The architectural conﬁguration of lipoplex is closely related to

the phase behavior governed by several factors such as types of

lipids, the composition of cationic lipids, and N/P charge ratio.

Despite recent clinical achievements, the structure–activity

relationship of lipoplex continues to be unraveled and func-

tional optimization is often based on empirical approaches.

To improve clinical performances, it is vital to understand the

physicochemical properties aected by dierent parameters.

The supramolecular structure of lipoplexes depends on the

participating components’ intrinsic properties such as their

geometry and other environmental (extrinsic) factors as well as

the balance between electrostatic interaction (headgroup) and

repulsive force driven by hydrophobic domain upon the addi-

tion of nucleic acid.[]

The most common structures are lamellar phase (

) and

inverted hexagonal phase ( II

), which can also be referred to

external and internal model, respectively, depending on the

relative orientations of nucleic acid (Figure6A).[] Saﬁnya pio-

neered early studies (and continue to explore more) of supra-

molecular structures of lipoplexes.[] Initiated by applying

the small-angle X-ray scattering (SAXS) technique to a mix-

ture of two cylindrical-shaped lipids, DOTAP/DOPC lipoplex,

a lamellar phase composed of multiple sheets of bilayers was

discovered (Figure B).[] SAXS measurements show that

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

the DNA monolayers are sandwiched between cationic mem-

branes with an interlayer spacing of .nm (≈nm of bilayer

and ≈. nm of water spacing with intercalated DNA) in an

equimolar fraction case, and the spacing varied with the molar

fraction of lipid molecules as the bilayer thickness changed

due to the dierent length between DOPC and DOTAP.[]

Interestingly, the spacing between DNA was relevant to the

membrane charge density as indicated by the arrows: the

spacing increased as the membrane charge density decreased

(increasing DOPC fraction). A cryo-EM image in the absence

of DNA shows enclosed liposomes with an equimolar composi-

tion of DOTAP/DOPC (Figure C).[] When the nucleic acid

is added, these liposomes transform into a layered structure to

accommodate nucleic acid molecules.[]

Soon after verifying lamellar phase, the same group also

identiﬁed inverted hexagonal phase (

) when the inverted

cone-shaped DOPE was included (Figure D).[] With

increasing DOPE mole fraction in DOTAP/DOPE lipoplex from

. to ., SAXS scans showed that the internal structure of

the DOTAP/DOPE lipoplex changes from

to something else,

which was deﬁned as II

with a diameter of ≈.nm embed-

ding DNA within a lipid monolayer arranged in a hexagonal lat-

tice (FigureE).[] Of note, when much smaller-sized siRNA is

entrapped instead of DNA, the peaks in SAXS may appear less

obvious (meaning less orientation) because of the higher disor-

dered phase.[] Cryo-EM analysis also showed that the distinct

feature of equimolar DOTAP/DOPE with the coexistence of

and

(FigureF).[]

Although it is well known that particular

complexes

(DOTAP/DOPE = .) successfully delivered DNA, whereas

similarly high fraction of DOPC-based lipoplex failed to deliver

nucleic acids,[] it should be noted that the architectural con-

ﬁguration of lipoplex and biological functionality cannot be

directly correlated.[] In other words, lipoplexes with

Figure 6. Architectural conﬁguration of lipoplexes. A) Schematic of the lamellar phase (

) of DOTAP/DOPC lipoplex. B) SAXS scans of lipoplex at

constant N/P charge ratio of 2.2 with increasing DOPC ratio. The arrows indicate the DNA peaks, whereby the spacing between DNA increase as the

membrane charge density decreased. C) Cryo-EM images of DOTAP/DOPC lipoplex without/with nucleic acids. D) Schematic of the inverted hexagonal

phase ( II

) of DOTAP/DOPE lipoplex. E) SAXS patterns of the C

and II

of lipoplexes with increasing DOPE ratio. F) Cryo-EM images of DOTAP/DOPE

lipoplex without/with nucleic acids. G) Coarse-grained molecular dynamics simulations to visualize the fusion mechanism between lipoplexes and

endosomal membrane: i) free lipoplex complexed with dsDNAs approaches the endosomal bilayer; ii) stalk formation upon initial contact; iii) hemifu-

sion diaphragm; iv) membrane fusion with successful delivery of nucleic acids. H) Schematic of the hexagonal phase ( I

) of lipoplex. B,E) Adapted with

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

not always show higher transfection eciency compared to the

ones with

, thus the phase is not a universal parameter to

predict transfection eciencies.[] Also, the existence of serum

can aect the outcome, in which membrane fusion is not a

dominant factor, and lipoplexes with

phase outperformed the

others.[] Some studies have reported that the morphological

changes or phase transitions from

to II

upon lipoplex–cell

membrane contact trigger better transfection eciency, and a

series of studies have attempted to optimize the cationic lipids

by tuning their molecular structures.[]

More recently, the interaction of lipoplex and negatively

charged endosomal membrane has been computationally simu-

lated, interrogating key parameters such as lipid types (dierent

headgroups and the extent of chain length and saturation),

endosomal membrane composition, and sizes of lipoplex.[]

By using this approach, two alternative fusion pathways were

suggested (Figure G, perpendicular and parallel) with a pos-

sible endosomal membrane composition dependency. Also, an

unsaturated chain in a cationic lipid was found to be necessary

for eective membrane fusion. It was also simulated that larger

lipoplexes (≈ nm) confer more thermodynamically stable self-

assembled structures than smaller ones (≈ nm). For example,

the highest amount of nucleic acid delivery was achieved for

the unsaturated DOTAP/DOPE lipoplex compared to its satu-

rated counterparts, which can be explained by the decreased

bending modulus of unsaturated lipids to favor the fusion.

About a decade after the

report, continuous eorts

revealed another representative phase called hexagonal ( I

)

structure based on dendritic cationic lipid/DOPC lipoplex that

was highly ecient in in vitro transfection (FigureH).[] Due

to highly charged (+) bulky headgroups of dendritic lipid,

the complex forms long micelle-like structure to surround the

nucleic acid. The improved eciency was ascribed to the e-

cient packing of nucleic acid within lipoplex that can facilitate

easier release of nucleic acid in cells. More recently, a concept

of cuboplex has emerged, utilizing bicontinuous cubic phases

(cubosome) made of glycerol monooleate (GMO), DOTAP, and

GMO-PEG (// mol%).[] As marginal amount of cationic

lipid inclusion suggests, the endosomal membrane interac-

tions of cuboplex is not governed by electrostatics but by elas-

ticity energetics, and possess intrinsic fusogenic properties to

aid endosomal escape. Although cuboplex is largely dierent

from conventional lipoplex composition, the development and

structural characterization of this novel vehicle provide valuable

insights how nanoarchitectonic approach can contribute to the

advancement of lipid-based gene delivery technologies.

3.2. Platform Development

3.2.1. Lipoplex Manufacturing Considerations

One of the biggest advantages of lipoplex comes from the

simplicity and ﬂexibility of its fabrication, which takes two

steps: ) formation of cationic liposomes and ) complexation

with nucleic acid (Figure7A). Liposomes can be prepared via

traditionally known liposome production methods, ranging

from simple rehydration of lipids, sonication, or ethanol injec-

tion; sophisticated monodisperse ones can be formulated by

extrusion and microﬂuidics. In this section, a few considera-

tions while fabricating lipoplex (cationic liposome fabrication

followed by nucleic acid complexation), their eects on func-

tional activities, and the importance of N/P charge ratio are

critically discussed.

3.2.2. Liposome Fabrication

Liposomes are sphere-shaped lipid vesicles consisting of one

(unilamellar) or more (multilamellar) lipid bilayers. When

amphipathic lipids are hydrated, hydrophilic headgroups face

aqueous phase whereas hydrophobic tail groups orient or aggre-

gate towards each other to form thermodynamically favored

self-assembly structure. Therefore, the simplest liposomes can

be fabricated by mechanically dispersing the lipid suspension in

aqueous medium, which can be done by shaking or vortexing.

Largely, the preparation of methods can be divided into

mechanical and organic solvent-assisted methods. Most of

mechanical preparations start from lipid thin-ﬁlm hydration

as mentioned above, then subjected to further processing

including sonication, freeze-thawing, and/or extrusion. Among

them, sonication is the most used method to prepare liposomes

for lipoplex formation even though its size distribution can be

relatively larger than other processing methods.[] From the

ﬁrst reported lipoplex fabrication to recent optimization studies,

sonication has been widely used and proven eective even com-

pared to other methods.[,b,] For more homogenous and

smaller unilamellar liposome production, freeze-thaw and/or

extrusion can be applied after lipid hydration or after mild soni-

cation. Freeze-thawing is adopted commonly to improve the

encapsulation eciency of molecules by expanding the inner

water phase by ice crystal formation, which result in high pop-

ulation of unilamellar vesicles while physically disrupting the

lamellar structure.[] Extrusion can be accompanied to yield

more homogenous particles, often using nm ﬁlter to pro-

duce cationic liposomes. Ethanol injection is another widely

used method to fabricate liposomes, which can be versatile and

easy to scale up. In this method, the lipid–ethanol solution is

injected into the aqueous phase (or vice versa), then ethanol

is removed by rotary evaporation or dialysis to produce con-

centrated liposome suspensions.[] The microﬂuidic method

is a sophisticated improvement of ethanol injection that the

mixing of organic and aqueous phases can be controlled by the

interfacial diusion of narrow microchannel, which has been

described in the previous section. In lipoplex formation, micro-

ﬂuidics can be used in either fabricating initial liposomes or

mixing nucleic acid with liposomes, which will be discussed in

the next section.

Multiple studies have compared the transfection eciency

of lipoplex formulated from dierent methods, whereby the

reason of one’s performance is debatable whether it is due

to the particle size and/or lamellarity, or other factors. Early

studies have reported that the transfection eciency is higher

in large multilamellar liposomes (– nm) formed by

vortexing rather than in relatively smaller sized liposomes

(< nm) formed by sonication (FigureB).[b,,] Similar

conclusion was drawn when the transfection eciency of lipo-

plex was tested with or without serum as well.[] On the other

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

Figure 7. Platform development of lipoplex. A) Schematic illustrations of lipoplex formulation from cationic liposome formation to addition of nucleic

acids. B) Comparison of transfection eciency depending on the liposome preparation method. Vortexing and sonication were used to produce dominantly

large multilamellar (red) and smaller (blue) liposomes, respectively. C) One-step formation of lipoplex assisted by a microﬂuidic device. The graph (right)

shows the hydrodynamic diameter and polydispersity index of microﬂuidic-obtained and bulk mixing-prepared lipoplexes. D) Schematic representation to

show that the N/P charge ratio largely aects the colloidal stability of lipoplexes, which can be presumed by their particle size (green line) and net surface

charge (orange line). E) Lipoplexes formulated at various N/P ratios display a distinct instable range around N/P = 1, whereby the zeta potential is close

to neutral. F) Bioluminescence imaging of BALB/c mice after intravenous administration of Luc-RNA-lipoplexes from panel E) at various N/P charge

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

hand, ethanol injection method was reported to be equally or

more eective than the dry-ﬁlm method[] and its eciency

proven in human clinical studies.[d,]

3.2.3. Nucleic Acid Complexation

During the formation of lipid-nucleic acid complex, the

nucleic acid electrostatically binds to the cationic lipid by fast

exothermic process, then subsequently rearrange/fuse into

lipoplex by slower endothermic reaction.[] While the ﬁrst fast

regime occurs instantly when the lipid and nucleic acid con-

tact, the kinetics of slow rearrangement and fusion-induced

growth of lipoplexes largely depend on the mixing protocol.

The mode of mixing can vary by the sequence of adding com-

ponents and whether the addition is bulk (one-step mixing by

pipetting or vortexing) or titration (more than two steps). In

permanently charged cationic lipid, the complexation solution

conditions are generally maintained at physiological pH or in

deionized water to facilitate strong electrostatic interactions

between nucleic acid and lipids. Whereas excess electrolytes

can shield and decrease the driving force of self-assembly of

lipoplex when there was signiﬁcant dierences (.  and

. ),[] a direct link between transfection eciency and

solution conditions during the formation of lipoplex (mild salt

concentrations or deionized water) remains controversial.[b]

However, in the cases of pH-responsive lipids such as DC-Chol

and DODMA, slightly acidic or neutral buer solution is pre-

ferred over deionized water to increase the payload and trans-

fection eciency.[]

One of the ﬁrst studies comparing the order of addition of

nucleic acid and lipids was reported by Zelphatiet al. that the

small, stable lipoplexes could be formed at high N/P charge

ratio (which will be explained in the next section) by bulk-

adding the nucleic acids into the excess of liposomes while the

opposite order produces large aggregates.[] At low N/P charge

ratio, the trend was reversed such that the addition of liposomes

to the excess of nucleic acid could produce stable lipoplexes.

This observation could be explained by titration approach.[]

During the nucleic acid addition into liposomes (in excess),

nucleic acids come into contact with multiple liposomes, asso-

ciating at the interface of two (or more) liposomes at once. This

interaction may generate asymmetrical stresses, leading to rup-

ture and aggregation of the lipoplex. In the opposite case when

the liposome is added, the excess nucleic acids rapidly bind and

wrap the surface of the cationic liposomes, forming a relatively

stable complex until a critical point where liposomes bind to

other ones that are already coated with nucleic acid. Then, the

aggregates start to form, and the generated stress drives the

rupture and growth of the particles. The former case might pro-

duce too large particle that can limit the transfection eciency,

but the latter case could inhibit the intimate lipid mixing and

necessary structural rearrangements.[a,] Likewise, small

variations in each step can result in largely dierent self-assem-

bled structure, thus care should be taken during formulation

and optimization of lipoplex preparation. Of note, an optimal

mixing method might vary depending on the nucleic acid types

and concentration as well. For example, vortex-mixing resulted

in better transfection eciency than spontaneous mixing

when the siRNA concentration was higher than a certain point

whereas the opposite was true when N/P charge ratio was

high.[]

After presenting that the order of mixing can be a critical

parameter inﬂuencing the complex behavior, Zelphatietal. from

the same study proposed a controlled mixing method to mix

both components simultaneously at a ﬁxed ratio/speed with the

aid of a syringe pump.[] The optimal mixing rate was higher

for the smaller liposomes, whereby the results were reproduc-

ible and more homogenous lipoplexes could be formed. This

controlled mixing was further developed recently via micro-

ﬂuidics to formulate the liposome as well as to complex with

nucleic acids.[] This method has demonstrated potential for

controlling physical properties of liposomes including size dis-

tribution and lamellarity in precisely conﬁned microenviron-

ment. Compared to the bulk mixing method, microﬂuidics

generated lipoplexes with more homogenous and smaller sizes

(Figure C).[a,] This feature might explain higher transfec-

tion eciency in some cases, whereby about half number of

stacked bilayers were needed (reduced amount of cationic lipid)

to incorporate the same amount of nucleic acid.[b]

Lastly, it is worthy to note that generalizing the eect of lipo-

plex size by comparing dierent literature should be avoided

because multiple parameters involve in the ﬁnal resulting

aggregates such as lipid compositions, N/P charge ratio,

nucleic acid, and reported transfection eciency might vary a

lot depending on the cell lines as mentioned above.[c] Nev-

ertheless, large lipoplexes (up to a certain extent) feature wide

range of contact area with cells and tend to form large intra-

cellular vesicles after endocytosis, which are more susceptible

to release the nucleic acid.[,] Consequently, the transfection

eciency also depends on whether a particular cell dominantly

uptake particles by endocytosis.[] While most of the men-

tioned studies were conducted in vitro, the physicochemical

properties of lipoplex might be more pronounced in the in vivo

experiments because cellular uptake of nanoparticles is strongly

dependent on their size and surface charge.[] Generally, the

mean diameter of lipoplex particles that have entered clinical

trials ranged –nm.[b]

3.2.4. Charge Ratio (N/P)

The charge ratio (N/P) denotes the molar ratio between cati-

onic moieties, such as protonatable nitrogen (N) atoms of

the lipid headgroup, and the anionic charges brought by

the phosphate (P) groups of the nucleic acids. Theoretically,

N/P = / will result in the net charge of  but usually this point

can be achieved empirically by a slight excess of either negative

(nucleic acid) or positive charges (cationic lipid) due to steric

hindrance and/or other geometrical constraints.[] As N/P is

the most easily tunable and dominant factor that directly aects

the colloidal stability of the complex, examining its eect on

colloidal stability is a prerequisite to optimize lipoplex formula.

From a functional perspective, too low N/P increases uncom-

plexed nucleic acid,[] but too high N/P leads to increased cyto-

toxicity due to excess cationic lipids as mentioned above. With

respect to stability, there are three main regimes depending on

the N/P (Figure D,E). Starting from higher N/P (low amount

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

of nucleic acid), most particles are positively charged and repel

each other because the population is dominantly cationic with

a small part of lipoplexes. Upon adding more nucleic acid, lipo-

plexes become near to neutral surface charge (see zeta poten-

tial measurement) such that they start to aggregate because of

colloidal instability.[] With increasing nucleic acid, the lipo-

plex reaches again colloidally stable structure that is negatively

charged with some uncomplexed or not suciently condensed

nucleic acid molecules surrounding the particle.

The optimal N/P charge ratio might vary depending on the

lipid choices and respective applications, ranging from >

for monovalent cationic lipids such as DOTMA to less than 

for multivalent GL or several DNA-complexed cases.[b,]

Although negatively charged regime is often regarded as less

eective considering the anionic cell membrane,[a] in vivo

studies have reported that Luc expression was predominant

in the lungs of mice when the particle is positively charged

whereas the gradual increase of nucleic acid can shift the tar-

geted organ to the spleens (FigureF).[,] Interestingly, not

only the lead composition of DOTMA/DOPE, but also other

combinations of lipids including cholesterol and DOTAP have

led to exclusive targeting of spleen when the lipoplex was nega-

tively charged, prompting the selection of N/P = . condition

for clinical testing.

To conclude the manufacturing considerations with per-

spective of scalable production, lipoplex-based gene delivery

features a simple preparation step for liposome and a ﬂexible

choice of nucleic acid that can be complexed without aids of

sophisticated devices or speciﬁc skills. All liposome produc-

tion methods have been established in depth, and fabrication

methods except for microﬂuidics are quite straightforward and

widely available, allowing us to scale-up the liposome manu-

facturing and to easily combine with commercially established

protocols for nucleic acid production. Therefore, the ﬂexibility

regarding nucleic acid choice comes as a major advantage in

utilizing lipoplex in “on demand” situations such as imminent

viral outbreaks or personalized production of gene delivery

vehicle.[a] Furthermore, the qualities of liposome and nucleic

acid can be separately screened prior to complex formation,

which can minimize the contamination and enhance the batch-

to-batch reproducibility with lower cost.[]

3.3. Current Applications and Future Possibilities

3.3.1. Cancer Vaccines

There has been intense interest in developing lipoplex as a vac-

cine that stimulates targeted immune response to treat cancer.

Toward this goal, antigen-presenting cells such as dendritic

cells (DCs) are ideal targets of lipoplexes to boost the immune

response. Among the pioneering works, intravenously admin-

istered nonfunctionalized DOTAP/DOPE lipoplex has been

developed for systematic targeting of DCs.[a,] The admin-

istration of lipoplex induced both adaptive and innate immune

responses via antigen-encoding RNA.[] Importantly, this

study highlights that surface charge-dependent organ-targeting

is achievable by simply tuning the N/P charge ratio. The site-

speciﬁc RNA expression from lungs to spleen was achieved

when the N/P was – and .–., respectively (FigureF),[]

which is distinguishable from targeting function of neutral

LNP mainly to hepatocytes.[]

Additionally, lipoplex has been used as a vaccine speciﬁcally

for treating melanoma with various nucleic acids such as pDNA

or RNA.[d,b,] For example, multiple nonmutated, tumor-

associated antigens were targeted by intravenously administered

melanoma FixVac (BNT-mRNA)-DOTMA/DOPE lipoplex,

leading to the elevated cytokines such as IFN-α, γ, IL- with

ex vivo CD+ T-cell responses across dierent cancer types.[]

On the other hand, OVA-encoding mRNA–DOTAP/DOPE

lipoplex induced a potent type I IFN response upon subcu-

taneous, intradermal, and intranodal injection, interfering

with the generation of potent cytolytic T-cell responses.[]

When the antitumor eect of mRNA–DOTAP/DOPE lipo-

plex was compared with pDNA–lipoplex, mRNA–lipoplex

showed higher lipofection eciency than pDNA one.[]

Thus, mRNA–DOTAP/DOPE lipoplexes demonstrate great

potential to treat cancer including melanoma in the context

of cancer immunotherapy, gene, and vaccine therapy. Indeed,

several human clinical studies are ongoing with lipoplex vac-

cines to treat melanoma (NCT, NCT, and

NCT), breast cancer (NCT), advanced esopha-

geal cancer and (non-small cell) lung cancer (NCT),

or multiple tumors including pancreatic, colorectal and lung

cancers (NCT), whereby the lipoplex carriers include

DOTAP/DOPE or DOTMA/DOPE compositions or a proprietary

RNAiMAX composition. Both intravenous and subcutaneous

administration routes have been utilized for lipoplex admin-

istration while emerging evidence shows that the administra-

tion route can critically determine whether mRNA–lipoplexes

stimulate or inhibit type I interferon (IFN) immune responses

based on a recent in vivo mouse study.[] Promoting strong

immune responses should also be balanced with safety con-

siderations, highlighting the potential of lipoplexes along with

important translational needs for reﬁning vaccination regimens

and advancing mechanistic understanding of how administra-

tion route and dosing relate to immune responses.

3.3.2. Infectious Disease Vaccines

Likewise, lipoplexes can be utilized as a vaccine for infectious

disease such as immunodeﬁciency syndrome caused by human

immunodeﬁciency virus (HIV). Similar to the immunotherapy

for cancer, DOTAP/DOPE lipoplex carrying mRNA encoding

Gag protein modulated DCs to stimulate HIV-speciﬁc immune

responses.[] The same lipoplex system with HIV- antigen

Gag-encoding mRNA produced antigen-speciﬁc functional

T cells in spleens and lymph nodes, together with the immune-

activation characterized by the release of type I IFN.[] Another

inspiring example includes the parallel evaluation of lipoplex

and LNP entrapping HIV- Env-encoding saRNA to compare

the transfection eciency in vitro (Figure8A) and in vivo.[]

Strikingly, immunogenicity comparison of both formulations

resulted in a distinctive trend that the cationic lipid-based lipo-

plexes (DDAB rather than DOTAP) achieved the maximum

response of HIV IgG titers after a single dose whereas all LNPs

showed the enhanced antibody response only after the second

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

Figure 8. Possibilities of tuning the lipid compositions in conventional lipoplex or LNP compositions. A) Transfection eciency of Luc-saRNA delivered

by LNP/lipoplex under standard conditions without FCS ((−) FCS), with FCS ((+) 50% FCS), or with RNAse ((+) RNAse). B) Immunogenicity measured

by antibody titers after IM injection of HIV-1 Env saRNA complexed to the LNP/lipoplex. C) Zeta potential measurements of mRNA lipoplexes with

varying complexing lipids. D) Total protein expression mediated by C) lipoplexes within dierent cell subtype. E) Schematic representation showing

selective organ targeting (SORT) strategy for tissue-speciﬁc nucleic acid delivery by adding dierent net charged lipid molecules to aim speciﬁc biodis-

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

injection (FigureB). This proof-of-concept study highlights the

potential usage of cationic lipids with adjuvant properties such

as enhancing humoral immune response and immunostimula-

tory eects in vivo.[]

Recently, DNA-lipoplex vaccine was also developed for the

treatment of SARS-CoV- utilizing saturated neutral lipid

(DOTAP/DOPE/DPPC = // mol%) and low N/P charge

ratio of . with highly negative net charge.[a] The study

selected intramuscular (IM) injection method and has proven

its eective spike protein expression, neutralizing antibody

responses, and T-cell activation comparable to the parallelly

tested gold standard, intramuscular electroporation method.

Similar trend was reported in the eorts to develop malaria

surface protein-encoded pDNA-lipoplex vaccine (equimolar

DOTAP/cholesterol), whereby the N/P of . (anionic) lipoplex

performed higher immunogenicity in mice compared to the

N/P of . (cationic) one when it was administered IM.[b] In

both cases with DNA-lipoplex, higher N/P triggered increased

cytotoxicity while lower N/P displayed higher transfection, pre-

sumably because of the more ecient release of DNA from

negatively charged lipoplexes (lower N/P) compared to the

higher ones. Considering that DNA-based drugs can be stored

at – °C for a few years and room temperature for  year,[]

ecient entrapment of DNA using lipoplex can improve the

availability of the vaccine, avoiding costly frozen-state supply

chain for existing mRNA vaccines.[]

3.3.3. Lipid Nanoparticles versus Lipoplexes: Which One Is Better?

As the formulation and concept of LNP and lipoplex are dis-

tinctive, it is an immediate point of inquiry to consider which

one works better and under what conditions. However, there

have been a few direct comparison studies to address this

point. In one example, Kubota et al. evaluated cationic lipid-

based LNP and lipoplex in conventional LNP composition

(DODMA/DSPC/cholesterol/PEG-lipid = //./.), using

ethanol mixing and extrusion ( nm-ﬁltered), respectively, to

compare their physicochemical properties as well as biological

functionalities including particle characterization, transfection

eciency, and inﬂammatory cytokine response.[] While both

particles showed the maximum size at N/P = , lipoplexes were

only slightly larger than LNP with ≈ nm margin ( and

 nm, respectively), which was probably attributed to the

addition of PEG lipid and DODMA with low pKa that could be

veriﬁed again by unilamellar structure of lipoplexes observed by

cryo-EM images. While cellular uptake eciency and inﬂam-

matory cytokine responses indicated LNP is more compatible

with the given cell line, it should be noted that the selected

composition was ideal for LNP formulation and not for lipo-

plexes as the entrapment mechanism of the latter one is mainly

governed by electrostatic attraction on surface.

On the other hand, Blakney et al. have fabricated both

ionizable and cationic lipid-based LNP and lipoplex with

a conventional lipoplex formula (complexing/neutral

lipid =  / mol%) to characterize physicochemical proper-

ties and biological activities.[] Although both liposomes and

LNPs were produced by microﬂuidics, all resulted particle sizes

ranged –nm, which are quite larger than typical LNPs,

postulating that the self-assembled structure of LNPs formed

here is dierent from the one from classical composition with

PEG lipids. Only ionizable lipid-based lipoplex showed negative

surface charge, reﬂecting incompletely coupled saRNA due to

the ionization of the lipid at physiological pH on surface, which

later signiﬁcantly aected the susceptibility to degradation, and

lower transfection in vitro as well as in vivo (FigureA). Conclu-

sively, the overall functionality was the highest from ionizable

lipid (C-) and cationic lipid (DDAB rather than DOTAP

in this case) in LNP and lipoplex, respectively, conﬁrming that

the optimization strategy of entrapping nucleic acid should be

distinguished between LNP and lipoplex platform.

For further study, the same group has formulated lipo-

plexes by varying complexing lipids from cationic to zwitteri-

onic or ionizable lipids (complexing/neutral lipid = / mol%)

to intradermally deliver mRNA in human skin explants and

evaluated the eects of complexing lipids on cellular uptake

and expression.[] Intriguingly, the combinations with zwit-

terionic (cephalin) and ionizable lipids (C- and MC)

could also successfully entrap nucleic acid (complexation e-

ciency > %) with varied extent of negative zeta potentials

(FigureC). Surprisingly, the dierences in the surface charge

after forming lipoplex were strongly correlated with the pat-

tern of nucleic acid uptake and expression among cell types

(Figure D), not the inherent lipid types classiﬁed (i.e., cati-

onic/zwitterionic/ionizable lipids). For instance, cellular uptake

depending on the cell types shared similarities among moder-

ately charged cephalin and DDAB lipoplexes, whereas highly

charged DOTAP and MC lipoplexes were expressed in dif-

ferent distributions of cells within skin explants.

Another example includes recent eorts on developing selec-

tive organ targeting (SORT) lipid nanoparticles designed to

deliver nucleic acid to speciﬁc organs such as lungs, spleens,

and livers by tuning the membrane charge of the composi-

tion.[,] Based on conventional LNP composition (ionizable

lipid/DOPE/cholesterol/PEG-lipid = /// mol%), ioniz-

able (cationic) lipids, anionic lipids, and permanently charged

cationic lipids were supplemented, and the result clearly indi-

cated that the respective inclusion of particles delivered mRNA

to the liver, spleen, and lungs after IV injection (FigureE).[]

This distinctive tissue-targeting property was derived from the

recognition, anity, and adsorption of speciﬁc serum proteins

toward particle’s surface charge/composition.[] These ﬁndings

strongly support the nanoarchitectonic design strategy of lipid

self-assemblies that can be adopted to engineer optimized lipid-

based nucleic acid delivery vehicles.

4. Lipid Bicelles

While lipid bicelles were originally invented as suitable mem-

brane-mimetic environments to reconstitute membrane pro-

teins for structural biology studies,[] they have also emerged

as promising tools for biotechnology applications such as drug

delivery,[] antibacterial medicine,[] and lipid bilayer coat-

ings.[] To date, LNPs and lipoplexes are the most advanced

types of lipid-based nanoparticles for nucleic acid delivery appli-

cations, but there is growing attention to lipid bicelles. Com-

pared to the typical spherical geometry of LNPs and lipoplexes,

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

bicelles have a wider range of possible geometries, including

a disk-like shape in some cases that can be advantageous for

tissue penetration and cellular uptake. As such, lipid bicelles

represent a promising frontier for lipid-based nanoparticles for

nucleic acid delivery applications and are covered in this part.

4.1. Nanoarchitecture Design Principles

Bicelle morphology depends on a wide range of factors and

controlling pertinent parameters such as lipid concentration,

molar ratios, and temperature can enable rational control over

nanoarchitecture features. In this section, we begin by intro-

ducing the self-assembly principles of bicelles and relevant con-

siderations related to morphology and lipid geometry, before

expanding our coverage to more advanced functionalization

concepts that enhance bicelle stability.

4.1.1. Self-Assembly Principles

It is widely appreciated that long-chain phospholipids self-

assemble to form liposomes in aqueous solution whereas

detergents self-assemble into micelles.[] Interestingly, new

morphological phases emerge when long-chain phospholipids

and detergents are mixed together and this original discovery

gave rise to the “bicelle” concept.[] Practical utilization of

bicelles was further advanced by recognizing that short-

chain phospholipids can behave as detergent-like molecules

and hence it is possible to form bicelles using phospholipids

only.[] The molar ratio of long-chain phospholipids to short-

chain phospholipids is deﬁned as the q-ratio and strongly inﬂu-

ences bicellar phase properties.[] Depending on the lipid

composition and q-ratio, bicelles can form a wide range of mor-

phologies, including multilamellar vesicles, toroidal pore com-

plexes, extended lamellae, chiral nematic ribbons, disk-shaped

nanostructures, and spherical micelles[] (Figure9A).

While the phase properties of bicelles are complex and

depend on various parameters such as lipid concentration,

ionic strength, and temperature, it is possible to control them,

and we turn our attention to the case of disk-like bicelles that

form in certain conditions. The self-assembly of bicelle disks

can be rationalized by considering the geometrical properties

of the dierent lipid components, as depicted in FigureB,C.

In general, lipid molecules have a hydrophilic headgroup and

one or more hydrophobic alkyl chains. For short-chain phos-

pholipids, the headgroup is relatively large compared to the two

short chains and hence the geometry is cone-like. A similar con-

ical geometry is also found for typical detergent molecules. As

such, short-chain phospholipid and detergent molecules tend

to pack in high-curvature regions. On the other hand, long-

chain phospholipids usually have a more cylindrical appear-

ance because the headgroup and two long chains are more

similar shape wise. It is therefore more energetically favorable

for long-chain phospholipids to reside in planar regions. For

these reasons, upon mixing of lipid molecules in bicelle com-

positions, the overall self-assembly process is driven by hydro-

phobic interactions and the dierent types of lipids are prefer-

entially organized into distinct bicelle regions. Indeed, due to

the curvature-related energetic preferences, short-chain phos-

pholipids and detergents have low miscibility with long-chain

phospholipids[] and the cone-like and cylinder-like molecules

tend to self-assemble into the disk edges and central planar

region, respectively.[] An interesting case concerns the use

of headgroup-functionalized, long-chain phospholipids such as

PEGylated lipids. Due to the increased size of the headgroup

Figure 9. Lipid bicelle concept and molecular building blocks. A) Schematic illustrations of dierent bicelle morphologies that can form depending on

molecular components and experimental conditions as follows: (i) multilamellar vesicles; (ii) toroidal pore complexes; (iii) extended lamellae; (iv) chiral

nematic ribbons; (v) disk-shaped nanostructures; and (vi) spherical micelles. B) Cross-sectional view of a bicelle disk that is composed of long-chain

and short-chain phospholipids. C) Molecular geometry of dierent molecular building blocks, including (i) detergents and short-chain phospholipids

with cone-like geometries, (ii) long-chain phospholipids with cylinder-like geometries, and (iii) PEGylated phospholipids with cone-like geometries.

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

in such cases, PEGylated lipids can assume a more cone-like

appearance than typical long-chain phospholipids and may

locate in the edge and/or planar regions of a bicelle disk.[]

There are also a few practical points to highlight: ) the

q-ratio must be within a certain range to form bicelle disks

while other morphologies can form in dierent q-ratio

ranges;[] ) the planar region of bicelle disks can present a

membrane-mimicking environment, especially in cases where

biologically relevant phospholipids are used. In certain cases,

phospholipids and cholesterol can both be incorporated in

the planar region and can form phase-separated domains in

some cases;[] ) bicelle preparation should take place at a suf-

ﬁciently high temperature so that all lipid components, espe-

cially the long-chain phospholipids, in the mixture are in the

ﬂuid phase.[] The fabricated sample can then be cooled down

afterward; and ) the speciﬁc morphology of bicelles is highly

sensitive to the speciﬁc phospholipids used in a mixture and

the corresponding phase diagram must be carefully considered

in each particular case. For example, including saturated or

unsaturated lipids in bicelle compositions can cause wide vari-

ance in the phase diagram and possible morphologies[] (for

some compositions, the disk morphology is not possible), and

similar attention must be placed on charged lipids which are

sometimes added to enhance bicelle stability but can also aect

possible morphologies.[]

4.1.2. Advanced Functionalization Strategies

There have been various functionalization strategies introduced

into bicelle nanoarchitecture designs in order to improve perfor-

mance and enable new application opportunities. One direction

has focused on improving bicelle stability to prevent coalescence

and to inhibit the structural transformation of bicellar disks into

vesicles over time, especially at higher temperatures. Charged

lipids such as negatively charged phosphatidylglycerols (PG)

have often been incorporated into bicelles and prevent bicelle

coalescence due to charge repulsion.[] Inspired by other

classes of lipid-based nanomedicines, researchers have also

added PEGylated lipids such as distearoyl phosphoethanolamine-

[methoxy (polyethyleneglycol)-] (DSPE-PEG) to further

improve bicelle stability and reduce uptake by the reticuloen-

dothelial system, potentially enabling increased circulation time

in vivo. For example, it is possible to form disk-like bicelles in

the presence of  mol% PEG-DSPE, which can enhance

bicelle stability due to steric hindrance of the PEG chains.[]

While PEGylated lipids have a mainly cone-like geometry, more

detailed structural studies have shown that a neutral PEGylated

lipid (C-PEG-ceramide) can distribute across both the

planar region and rims of bicellar disk. Of note, there is indeed

an approximately, two-fold greater density of PEGylated lipids in

the rims,[] however, it should also be noted that the broader

distribution of the PEGylated lipid is advantageous for poten-

tial in vivo applications. In molecular dynamics simulations, it

has also been shown that bicelles containing ligand-modiﬁed

PEGylated lipids are more easily internalized by cells than simi-

larly functionalized liposomes due to more favorable membrane

energetics[] (see also other examples of bicelles with improved

cellular uptake properties over liposomes[]).

Other eorts have explored how to create inorganic–organic

hybrid bicelles based on mixing short-chain phospholipids

and long-chain organoalkoxysilane lipids (termed cerasome-

forming lipids, CFL).[] After lipid hydration, the bicelles were

incubated overnight at room temperature, during which time

a sol-gel reaction occurs between the alkoxysilyl headgroups

of the long-chain lipid molecules and results in a crosslinked

siloxane network (a silica-like coating). The hybrid bicelles had

typical diameters and heights of around  and nm, respec-

tively, and retained their structural morphology in the air envi-

ronment upon drying. Whereas conventional bicelles trans-

formed into vesicles at high temperature, the hybrid bicelles

also maintained their disk-like nanostructures at high tem-

perature.[] Based on these advantageous properties, there

have been eorts to develop partially silica-coated bicelles by

mixing short-chain phospholipids with CFL and PEGylated

lipids (DSPE-PEG). With increasing DSPE-PEG lipid

fraction, there was improved release of a small-molecule drug

and the PEGylated bicelles also had improved biocompatibility

and cellular uptake properties in vitro.[] As such, there is

increasing control over the functional properties of bicelles and

these eorts have coincided with attempts to utilize bicelles for

nucleic acid delivery.

4.2. Nucleic Acid Complexation

Currently, there are two main strategies to encapsulate nucleic

acids within bicelles. The ﬁrst strategy involves fabricating

bicelles in the presence of nucleic acids while the second

strategy focuses on bicelle fabrication initially, followed by

mixing bicelles and nucleic acids in a lipoplex-type format. The

merits and latest progress of both strategies, including struc-

tural and functional insights, are critically discussed below.

4.2.1. Coassembly of Lipids and Nucleic Acids

In the earliest example, the stacking of disk-like, cationic bicelles

and negatively charged DNA was explored in order to form a new

lipid nanoarchitecture for potential nucleic acid delivery applica-

tions.[] The bicelles were composed of long-chain and short-

chain lipids in a : molar ratio; the short-chain phospholipid

was ,-diheptanoyl-sn-glycero--phosphocholine (diCPC) while

the long-chain phospholipid consisted of a mixture of ,-dihexa-

decanoyl-sn-glycero--phosphocholine (DPPC) lipid and a cationic

lipid termed β-[N-(N′,N′-dimethylaminoethane)-carbamoyl]

cholesterol hydrochloride (DC-Chol). The DNA strands had

 or  base pairs and the lipid-to-DNA molar ratio was

from  to  (– mol% DC-Chol in the long-chain lipid

population). An interesting point in the study was the simplicity

of the fabrication approach. The dry lipid ﬁlm was directly

hydrated in an aqueous solution containing DNA molecules, fol-

lowed by bath sonication at  °C. As such, the bicelles were

directly formed in the presence of DNA, which is distinct from

the lipoplex approach in which case liposomes are ﬁrst formed

before mixing with nucleic acids. It was determined that the

DNA strands intercalate between the bicellar disk layers, and the

number of layers increased at greater cationic lipid fractions in

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

the bicelles (Figure10A). Accordingly, the size of the lipid-DNA

nanostructures could be varied between ≈– nm and the

ability to fabricate sub- nm nanostructures was highlighted

as a potential merit compared to lipoplexes, which typically have

> nm diameter. It was further noted that the DNA–bicelle

interactions were stable in acidic pH conditions and the fabri-

cation approach was also applicable to other commonly used

cationic lipids and surfactants such as ,-dioleoyl--trimethyl-

ammonium-propane (DOTAP) and hexadecyl-trimethyl-ammo-

nium bromide (CTAB).

The bicelle–DNA stacking concept was further extended

to anionic lipid bicelles by using the same basic design prin-

ciples and fabrication approach as in the previous example

and replacing the cationic lipid with an anionic lipid termed

,-dipalmitoyl-sn-glycero--phospho-(’-rac-glycerol) (sodium

salt) (DPPG) ( mol% in the long-chain lipid population).[]

While cationic lipids can readily form complexes with nucleic

acids and are eective transfection agents, they are cytotoxic,

which prompted the exploration of anionic lipids as an alter-

native option. In order to promote bicelle–DNA interactions,

divalent Ca+ ions were incorporated into the system and a min-

imum Ca+ concentration was needed to form stacked bicelle–

DNA complexes (Figure B). Complex formation did not

occur at  × −  Ca+ concentration, whereas bicelle–DNA

complexes did form at ≥ × −  Ca+ concentration and

the interlayer spacing within the complexes did not depend on

the Ca+ concentration. At ≥× −  Ca+ concentration, an

additional lamellar phase of ion-lipid complexes also formed in

cases of low DNA concentration, indicating that all three com-

ponents—lipids, DNA, and ions—play an important role in

driving the self-assembly process with controlled stoichiometry.

The resulting bicelle–DNA complexes had typical dimensions

around –nm diameter and –nm length.

In addition to DNA, there has also been exploration of uti-

lizing bicelles to encapsulate peptide nucleic acids (PNAs)

that target microRNA for potential gene editing and thera-

peutic purposes.[] The bicelles had a lipid composition of

around  mol% DPPC,  mol% DHPC,  mol% charged

lipid (anionic DPPG or cationic DOTAP), and  mol% DSPE-

PEG, and were prepared using a combination of vortexing,

temperature cycling (between °C and °C), and subsequent

room-temperature dilution—all steps were conducted in the

presence of PNAs at a PNA:lipid molar ratio between :

and :. The control bicelles without PNAs had an ≈nm

diameter radius, while PNA-loaded anionic bicelles had an

≈ min diameter and disk-like appearance. In marked con-

trast, PNA-loaded cationic bicelles underwent extensive aggre-

gation due to electrostatic attraction, resulting in large vesicles.

By optimizing the PNA:lipid molar ratio, encapsulation e-

ciencies as high as % could be achieved with anionic bicelles,

which also demonstrated superior cellular uptake properties in

vitro compared to cationic bicelles and a polymeric nanoparticle

control. The PNA-loaded anionic bicelles also had negligible

cell cytotoxicity in vitro. Mechanistic studies further revealed

that the PNA-loaded anionic bicelles were internalized by HeLa

cells by an endocytic mechanism and the delivered PNAs were

functionally active to inhibit a target microRNA.

4.2.2. Mixing Prefabricated Bicelles with Nucleic Acids

Anecdotally, there have also been reports of bicelle–nucleic acid

complexes that were developed using other fabrication proto-

cols and were evaluated in terms of structural properties and

in vitro transfection eciency. For example, lipid mixtures

consisting of gel-phase DPPC helper lipid and one of two cati-

onic lipids (malonic acid diamides) were prepared using thin-

ﬁlm hydration followed by shaker incubation at °C and bath

sonication, and were noted to form heterogenous, disk-like

bicelle aggregates with ≈ nm height and ≈ nm diameter

(: molar ratio of cationic lipid to helper lipid).[] The cationic

lipid bicelles were then mixed with negatively charged plasmid

DNA (pDNA: ≈. to  N/P ratios) using a lipoplex-type protocol

at room temperature before attempting to transfect human

lung carcinoma (A) cells. It was noted that the bicelle–pDNA

complexes had low transfection eciency, which was attributed

to poor mixing and phase separation, whereas liposome–pDNA

complexes formed using other helper lipids such as ﬂuid-phase

DOPC had eective lipoplex-type transfection. A follow-up

study characterized the interactions between DPPC/cationic

lipid mixtures (: molar ratio) and mainly double-stranded

DNA (≈ N/P ratio), and noted that the bicelle–DNA complexes

Figure 10. Nucleic acid interactions with lipid bicelles. A) Schematic illustration of a cationic bicelle–DNA complex. The interaction between cationic

lipid bicelles and DNA molecules induced self-assembly of a multilamellar complex consisting of alternating layers of bicelles and DNA molecules. The

complex formation was mediated by attractive electrostatic interactions between cationic lipids and negatively charged DNA. B) Schematic illustration

of an anionic bicelle–DNA complex. The interaction between anionic lipid bicelles and DNA molecules is mediated by divalent Ca2+ ions and induces

self-assembly of a multilamellar complex that consists of alternating layers of bicelles and DNA molecules.

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

retained disk-like nanostructures with a condensed lamellar

phase.[] In addition, using a similar preparation protocol,

lipid mixtures consisting of DMPC helper lipid and a three-

chain cationic lipid (DiTT; another malonic acid diamide)

were reported to form disk-like bicelles when the DiTT molar

fraction was around – mol% and had ≈–nm diame-

ters.[] In that case, a lipoplex-type protocol was again followed

to mix the DiTT/DMPC bicelles with pDNA (≈. to  N/P

ratio) and the resulting bicelle–pDNA complexes demonstrated

ecient transfection of A cells, including in the presence

of % serum. There is also evidence that the DiTT lipid by

itself can self-assemble into disk-like nanostructures that can be

useful for transfection.[]

In another study, bicelles composed of CFL, DOTAP, and

DHPC along with ligand-modiﬁed DSPE-PEG or DHPE-

PEG were prepared for gene delivery applications.[] The

base bicelle composition was CFL, DOTAP, and DHPC in an

approximately :: molar ratio and, in some cases, ligand-

modiﬁed DSPE-PEG or DHPE-PEG was included too.

The conceptual idea was that ligand-modiﬁed DSPE-PEG

would present the ligand (i.e., cRGD peptide that binds to inte-

grin receptors) on the main plane of the bicellar disk, whereas

the ligand-modiﬁed DHPE-PEG would present cRGD pep-

tide on the bicellar edge instead. In general, the bicelles were

initially prepared by hydrating a dry lipid ﬁlm at °C, followed

by ultrasonication for min and then overnight incubation at

room temperature to induce sol–gel formation on the bicellar

surface due to the presence of the CFL lipid. Afterward, using

a lipoplex-type protocol, the bicelles were incubated for min

at °C with small interfering RNA (siRNA) at a bicelle:siRNA

molar ratio of ≈:. The resulting bicelles had typical height

and diameter values of ≈ and ≈ nm, respectively, while it

was noted that the siRNA-loaded bicelles tended to be slightly

larger. Notably, when bicelles with edge-located cRGD had

superior siRNA loading capacity compared to bicelles with

plane-located cRGD, in which case it was suggested that the

ligand location diminishes bicelle–siRNA interactions. More-

over, the siRNA-loaded bicelles with edge-located cRGD showed

greater performance to inhibit a cellular signaling pathway in

vivo along with improved antitumor activities. The performance

dierences were attributed to the tailored, orthogonal function-

alities aorded by the bicelle nanoarchitecture, whereby the

targeting ligands were located on the bicelle edges while the

cationic lipids (DOTAP) were located in the plane region to

facilitate bicelle–siRNA interactions.

Together, these ﬁndings highlight how multiple strategies

are being implemented to encapsulate and deliver nucleic acids

using bicelles while the ﬁeld evolves from fundamental struc-

tural studies to early application contexts where feasibility and

utility are being demonstrated. Concurrently, eorts are being

made to explore new ways of producing bicelles that can sup-

port the translation of these innovations from laboratory con-

cepts to real-life applications.

4.3. Bicelle Manufacturing Considerations

While lipid bicelles are typically made by the thin-ﬁlm hydra-

tion method as described above, the method has practical

limitations to advance beyond the laboratory scale. Here, we

begin by brieﬂy introducing the method in more detail before

critically discussing the latest progress in developing more

advanced bicelle production strategies.

For the thin-ﬁlm hydration method, lipids in organic sol-

vent are ﬁrst mixed together and the solvent is evaporated in

order to form a dry lipid ﬁlm.[] The ﬁlm is then hydrated and

the aqueous lipid suspension is subjected to multiple rounds

of freeze-thaw-vortex cycling. The freezing step often involves

immersing the sample in liquid nitrogen (°C or below) and

the heating step is usually done in a water bath with ≈–°C

temperature. Other variations with dierent temperature cycles

are also possible to some extent. However, in all cases, the

method is limited to batch production and the harsh processing

conditions can potentially denature biological components such

as proteins and nucleic acids. As such, to enhance translational

prospects, there have been ongoing eorts to develop more suit-

able bicelle fabrication methods that are ecient and scalable.

4.3.1. Semispontaneous Method

One pioneering example is adoption of the semispontaneous

method to easily prepare bicelles from aordable lipids and sur-

factants.[] The method was ﬁrst demonstrated using soybean

lecithin and poly(oxyethylene) sorbitan monooleate (Tween ),

which were initially dissolved in a water-soluble organic sol-

vent, ,-butanediol in this case (and dipropyleneglycol in some

other cases), followed by water dilution; the temperature during

the processing steps was ﬁxed at °C (other studies have used

up to ≈°C). The lipid/surfactant mixture was then subjected

to tip ultrasonication as the ﬁnal processing step. Dynamic

light scattering (DLS) and transmission electron microscopy

(TEM) characterization of the aqueous suspension identiﬁed

the formation of disk-like bicelles at a : molar ratio of lecithin

to Tween . The typical diameter and thickness values of the

bicelles were around  and nm, respectively, and remained

physically stable for at least  d. One recent mechanistic

study discussed how two-tail lecithin molecules constitute the

disk body while one-tail Tween  molecules form the disk

rim.[] The semi-spontaneous method has also been utilized to

fabricate disk-like bicelles from other compositions, including

using a polyglycerol-type nonionic surfactant with two asym-

metrical tails[] and mixtures of lecithin and poly(oxyethylene)

cholesteryl ethers by themselves[] or in combination with up

to  mol% cholesterol in the liquid-ordered phase.[]

4.3.2. Microﬂuidic Production

In addition to the semispontaneous method that focuses on

batch processing, there have been recent eorts to develop a

continuous bicelle production method by utilizing microﬂu-

idic-based hydrodynamic ﬂow focusing, which additionally

does not require heating or cooling.[] A schematic out-

line of the microﬂuidic chip is presented in Figure11A and

involves mixing tangential streams of long-chain phospho-

lipids dispersed in an organic solvent such as methanol or

ethanol with short-chain phospholipids dispersed in aqueous

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

buer (pH .). When the two streams merged, the long-chain

and short-chain phospholipids mixed and self-assembled to

form bicelles in a continuous manner while it was important

to evaporate residual solvent in a low vacuum environment

after sample collection from the microﬂuidic channel. Using

,-dimyristoyl-sn-glycero--phosphocholine (DMPC) and

,-dihexanoyl-sn-glycero--phosphocholine (DHPC) as model

long-chain and short-chain phospholipids, respectively, it was

found that the molar ratio (q-ratio) of DMPC to DHPC was an

important determinant of the resulting lipid nanoparticle prop-

erties. When the DHPC concentration was ﬁxed at  × −

, it was identiﬁed that heterogenous vesicles formed at low

DMPC concentrations (<× − ; equivalent to q-ratio of less

than .) (Figure B). Conversely, bicelles formed at higher

DMPC concentrations (≥ × − ; equivalent to q-ratio of

equal to or greater than .). The morphological structures

in the dierent cases were conﬁrmed by TEM imaging and it

was further veriﬁed that the disk-like bicelle structures exhib-

ited interfacial membrane properties that are characteristic of

ordered bicelles (as opposed to those of mixed micelles[])

(FigureC,D). Systematic experiments identiﬁed that hetero-

geneous vesicles, homogenous vesicles, or bicelles could be

formed depending on the total lipid concentration and q-ratio,

and that a suciently high total lipid concentration was needed

to form bicelles (FigureE). The diusional speciﬁcs of the sol-

vent mixing process also inﬂuenced the self-assembly outcome

while controlling the ﬂow conditions along with the q-ratio ena-

bled ﬁne-tuning of the bicelle size (FigureF).

A follow-up study explored the general utility of the micro-

ﬂuidic chip based on hydrodynamic focusing to produce

bicelles composed of DHPC short-chain phospholipid with

other long-chain phospholipids that have dierent gel-to-ﬂuid

phase transition temperatures (Tm).[] Interestingly, the eect

of long-chain phospholipid concentration on lipid nanoparticle

morphology varied depending on the speciﬁc type of long-

chain phospholipid that was used in the microﬂuidic produc-

tion process. While heterogeneous vesicles and bicelles were

formed at low and high DMPC concentrations, respectively,

replacing DMPC (Tm≈  °C) with -palmitoyl--oleoyl-sn-

glycero--phosphocholine (POPC; Tm≈ − °C) lipid led to a

nearly opposite trend. In that case, bicelles were formed at low

POPC concentrations, which was attributed to strong hydro-

phobic interactions between POPC molecules, whereas large

vesicles were formed at high POPC concentrations. Additional

experiments with ,-dipalmitoyl-sn-glycero--phosphocholine

(DPPC; Tm≈°C) lipid further revealed how lipid phase prop-

erties can aect self-assembly outcomes. When the microﬂu-

idic production process was carried out with DPPC at  °C,

large vesicles were formed at all tested DPPC concentrations

due to strong interactions between DPPC molecules, which is

consistent with the gel-phase properties of DPPC at that tem-

perature. In marked contrast, when the microﬂuidic production

process was carried out with DPPC at °C, DPPC was in the

ﬂuid-phase state during production and bicelles were formed

at all tested DPPC concentrations. Of note, it was mentioned

that the resulting DPPC/DHPC bicelles had high membrane

Figure 11. Microﬂuidic production of lipid bicelles. A) Schematic illustration of microﬂuidic setup whereby long-chain and short-chain phospholipids

are dispersed in organic solvent and aqueous buer, respectively, and mix to form lipid bicelles according to the hydrodynamic focusing concept. In

this example, DMPC and DHPC are model long-chain and short-chain phospholipids, respectively. B) Eect of long-chain phospholipid (DMPC) con-

centration on the size of lipid nanoparticles. At low DMPC concentration, the size distribution indicates formation of heterogeneous vesicles (indicated

by label ii), whereas bicelle formation (indicated by label i) occurred at higher DMPC concentrations. C) Transmission electron microscopy (TEM)

image of microﬂuidic-fabricated vesicles and inset corresponds to a cryo-TEM image of the vesicles. D) TEM image of microﬂuidic-fabricated bicelles.

E) Morphology of microﬂuidic-fabricated lipid nanoparticles depending on long-chain phospholipid (DMPC) and short-chain phospholipid (DHPC)

concentrations. i–iii) Bicelles, heterogeneous vesicles, and homogenous vesicles, respectively. F) Size of microﬂuidic-fabricated bicelles depending

on the eective q-ratio of long-chain phospholipid to short-chain phospholipid concentrations. All panels are adapted with permission.[223] Copyright

2021, American Chemical Society.

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

ordering after cooling down to ambient conditions, which is

indicative of the gel-phase properties of DPPC in that tempera-

ture range. Together, these ﬁndings establish that microﬂuidic

systems are capable of producing bicelles with diverse lipid

compositions, and it is possible to rationally control bicelle size

and lipid composition.

From a broader perspective, continued progress on the man-

ufacturing side highlights how it is possible to produce bicelles

using low-cost lipid and lipid-like materials and to shift from

batch to continuous production methods. However, there is an

outstanding need to expand eorts in this direction. Until now,

such eorts have focused solely on producing lipid bicelles

by considering short-chain and long-chain phospholipids and

related surfactants, and it would be beneﬁcial to also focus on

how to fabricate bicelles containing nucleic acids based on the

strategies described above along with adding other important

components such as PEGylated lipids.

4.4. Current Applications and Future Possibilities

Until now, there have been only a few cases that utilize nucleic-

acid-loaded bicelles for applications. However, bicelles have

emerged as a promising type of lipid nanoparticle for various

biomedical applications due to their much smaller size than

LNPs, liposomes, and lipoplexes so that they can eectively

penetrate intercellular spaces (-nm) and due to their unique

morphological diversity depending on the lipid composition

and molar ratio of components. Among numerous potential

applications, bicelles are already known to reinforce the skin’s

biophysical properties and complement drug penetration. We

critically review the latest progress in those ﬁelds, whereby par-

ticular focus is placed on advantageous features that could be

useful for nucleic acid delivery applications in future work.

4.4.1. Skin Repair and Delivery

Disk-like bicelles initially received attention for topical skin

applications because they are smaller than liposomes (which

cannot penetrate intracellular spaces of the stratum corneum

(SC),[] which is the outermost layer of skin) and contain less

irritating molecular components than micelles. Early studies

veriﬁed that disk-like bicelles are useful for their permeability

enhancing eects on skin while not causing irritation,[] and

can also penetrate the SC[] and inﬂuence the lipid composi-

tion/phase there.[] Follow-up studies revealed that adding

speciﬁc molecular components such as cholesterol sulfate could

improve the distribution and penetration of bicelles throughout

the SC,[] and such ﬁndings heightened interest in utilizing

bicelles as lipid nanoparticles to improve skin properties and to

deliver other medically useful molecules.[] Indeed, continued

research delineated how tuning bicelle morphology could

expand the range of skin-related biological functionalities,[]

leading to the use of bicelles as dermal carriers and enhancers

for treating impaired skin.

In terms of applying bicelles as carriers, it was reported

that bicelle-encapsulated diclofenac diethylamine (DDEA) had

lower skin penetration than aqueous DDEA suspensions.[]

However, when skin was pretreated with empty bicelles prior

to adding aqueous DDEA, the degree of skin penetration

was enhanced. Similar retarding eects were observed when

ﬂufenamic acid was encapsulated in bicelles, which was sug-

gested to be a potentially advantageous feature for safely deliv-

ering drugs that have otherwise rapid percutaneous absorp-

tion[] (see also related structural studies[]). Further investi-

gation supported that bicelle-encapsulated DDEA can be deliv-

ered to damaged skin and help to repair SC barrier function

while minimizing systemic side eects.[]

While relevant application studies have been limited to small

molecule drugs so far, there continues to be high interest in uti-

lizing bicelles for dermal applications.[] Several eorts have

been made to evaluate the biophysical properties of healthy

as well as irritated human skin in vivo before and after bicelle

treatment.[,] While conventional bicelles composed of

DMPC/DHPC increased transepidermal water loss, skin elas-

ticity, and skin dehydration without aecting the phase behavior

of SC lipidic microstructures in human subjects,[] incorpora-

tion of long-chain lipids with longer alkyl chain lengths might

tune the extent of skin reinforcement rather than penetrating

into the SC lipid structures.[b] Another study utilizing encap-

sulated bicelles in liposomes (bicosomes) to mimic epidermal

lamellar bodies showed that treatment with bicosomes rein-

forced the skin barrier function in both healthy and irritated

human skin, probably by reestablishing the lamellar lipid struc-

ture of theSC.[]

Looking forward, there is an outstanding need to investi-

gate how bicelle carriers might be useful carriers for delivering

nucleic acids to the skin, especially in intradermal applications.

Recent discussions have suggested that intradermal adminis-

tration of RNA-based vaccines may reduce dose requirements

and be possible with needle-free approaches.[] In these

respects, bicelles stand out among possible lipid nanoparticle

carrier options because they have well-validated properties for

skin delivery applications.

4.4.2. Cancer Therapy

There has also been extensive interest in utilizing disk-like

bicelles to deliver small-molecule and peptide drugs for anti-

cancer therapy. For example, PEGylated lipid bicelles have been

used to encapsulate the anticancer drug doxorubicin (DOX),

with up to % encapsulation eciency and >-times longer

circulation time in mice.[] Compared to free DOX, the bicelle-

encapsulated DOX was more eective at accumulating in

tumors and the bicelles could be internalized by breast cancer

cells. Anionic lipid bicelles have also been developed to coen-

capsulate two anticancer drugs, paclitaxel (PTX) and parthe-

nolide (PTL), for in vivo therapeutic treatment.[] To improve

tumor targeting, bicelles have been further functionalized with

RGD peptides and this strategy was able to increase tumor pen-

etration of bicelle-encapsulated melittin peptide compared to

non-functionalized bicelles, which in turn improved antitumor

activity in vivo.[] Motivated by these early results, the ﬁeld has

continued to progress toward clinical applications, especially

in terms of utilizing PEGylated lipids to enhance functional

performance. In one case, the addition of a PEGylated lipid to

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

adisk-like bicelle composition triggered the formation of small,

spherical micelles that, when loaded with the anticancer drug

docetaxel (DTX), not only had superior stability and cell and

tumor uptake properties compared to PEGylated liposomes,

but also demonstrated improved antitumor activity in vivo.[]

There have also been other reports that show PEGylated lipids

can help to sterically stabilize bicelles in order to enhance

uptake and anticancer drug delivery to human cancer cells.[]

Expanding on these concepts, there have also been eorts to

utilize more durable, organic–inorganic hybrid bicelles for anti-

cancer applications.[] In one interesting example, DOX was

loaded together with indocyanine green (ICG)—a ﬂuorescent

dye used for photothermal therapy—for temperature-regulated

DOX release and photothermal-triggered cancer cell cytotox-

icity, which could be achieved in the presence of near-infrared

ﬂuorescence light.[] The bicelles exhibited high levels of

tumor targeting and exhibited anticancer activity in vivo, which

was superior to either the chemotherapy or photothermal

therapy alone. Together with advances in developing partially

silica-coated hybrid bicelles as described above,[] such pos-

sibilities open the door to advanced cancer theragnostics and

it has been suggested the hybrid bicelles could also be useful

for gene delivery applications as well.[] Indeed, one study uti-

lizing hybrid bicelles for siRNA delivery has been reported for

anticancer treatment in vivo[] while highlighting the future

opportunities that lie ahead to use various types of bicelles,

including inorganic-organic hybrid bicelles, for nucleic acid

delivery applications.

5. Conclusion and Outlook

The development of lipid-based nucleic acid delivery vehi-

cles has increasingly gained traction over the past decade

and shifted from innovative laboratory concepts to practically

eective and scalable clinical technologies, most recently cul-

minating in the large-scale manufacturing and worldwide

distribution of mRNA–LNP vaccines to prevent COVID-

infection. These advances have been largely enabled by ration-

ally designing the LNP structure to achieve maximum perfor-

mance based on the nanoarchitectonics concept along with

utilizing microﬂuidic production technologies, which not only

paved the way for production scale-up, but more importantly

allowed precise control over production parameters to tune the

nanoarchitecture features of manufactured LNPs in a precise

and reproducible manner.

Along this line, there exist numerous opportunities to explore

numerous combinations of lipid excipients and nucleic acid

payloads, and the role of nanoarchitectonics-based research will

become increasingly crucial to spur future innovation. We iden-

tify several key areas in which nanoarchitectonics is expected to

play a central role. First, there is an opportunity to explore dif-

ferent routes of administration, especially by taking advantage

of dierent lipid nanoarchitectures such as bicelles that can be

potentially useful for intradermal applications. Secondly, there

is an urgent demand to overcome challenges related to the in

vivo behavior of LNPs, for example, due to issues related to the

immune response. Speciﬁcally, recent eorts have been made

to identify appropriate alternatives to PEG lipids. In doing so,

it is important to apply the concept of nanoarchitectonics to

modulate interactions of the PEG lipid (or alternatives) with

other lipid excipients and select appropriate candidates that will

retain favorable interactions with the rest of the components.

Another area of interest is the expansion of microﬂuidic

capabilities, especially to emerging classes of lipid-based nano-

particles such as bicelles for not only proof-of-concept dem-

onstration as done so far but also for practical application (as

in the LNP case). To date, improvements to microﬂuidic plat-

forms have been achieved incrementally and are mostly aimed

at obtaining better control over the physical properties of the

resultant lipid-based vehicle (i.e., mainly size and size distri-

bution). Enhancing the capability of microﬂuidic platforms

to precisely control the morphology and architecture of the

lipid-based delivery vehicle needs guidance from nanoarchitec-

tonics and could help to improve nucleic acid encapsulation in

tandem as well.

Finally, while clinically approved LNP-based drug and vac-

cines are composed of a single formulation (i.e., delivering a

single type of nucleic acid payload) there is growing interest in

multiple- or coformulations. For example, a recent study dem-

onstrated successful in vivo CRISPR/Cas gene editing for

the transthyretin gene in mice and rats using LNPs that were

coformulated to include single guide RNA (sgRNA) and Cas

mRNA.[] Formulation optimization was performed for in vitro

and in vivo gene editing with necessary modiﬁcations to incor-

porate sgRNA. The LNPs were formulated using microﬂuidics

and they achieved >% transthyretin protein knockdown for

months. This study highlights the need for nucleic acid LNPs

to be carefully engineered to match their therapeutic applica-

tions; speciﬁcally, how optimized, co-formulated LNPs can be

applied for cell engineering and genome editing. More notably,

it emphasizes the vast potential of LNPs as nucleic acid delivery

systems, which allows coformulations to be optimized with high

eciencies, paving the way for novel nucleic acid-based thera-

peutics for a wide range of diseases in the near future.

As a closing remark, we may also add that the three types

of lipid-based nanoparticles covered herein were selected due

to their distinct nanoarchitectures and breadth of development

stages, spanning conceptual innovations to full-ﬂedged com-

mercial products. Across this spectrum, the nanoarchitectonics

concept is a useful approach to analyze the progress and trends

and to build a rational understanding of how to eectively

design lipid-based nanoparticles for various biomedical appli-

cations. Until now, the impact of nanoarchitectonics on lipid-

based nanoparticles for nucleic acid delivery applications had

been largely implicit, but we foresee that nanoarchitectonics

will play an increasingly prominent and visible role to drive for-

ward future innovations both in terms of scientiﬁc discussion

and research progress.

Acknowledgements

A.R.F and S.P. contributed equally to this work. This work was supported

by the Ministry of Education (MOE) in Singapore under grant AcRF

TIER1-2020-T1-002-032 (RG111/20) and by a sponsored research

agreement from LUCA AICell Inc. (RCA- LUCA AICell REQ0239282 –

LUCA), and by the National Research Foundation of Korea (NRF) grants

funded by the Korean government (MSIT) (Nos. 2020R1C1C1004385

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

and 2021R1A4A1032782). In addition, this work was supported by

the International Research & Development Program of the National

Research Foundation of Korea (NRF) funded by the Ministry of Science

and ICT (2020K1A3A1A39112724). This work was also supported by a

grant of the Korea Health Technology R&D Project through the Korea

Health Industry Development Institute (KHIDI), funded by the Ministry

of Health & Welfare, Republic of Korea (Grant Number: HI19C1328).

H.T. was supported by an SINGA graduate scholarship from the A*STAR

Graduate Academy, Singapore. Schematic illustrations were created with

BioRender.com under an academic lab subscription.

Conﬂict of Interest

N.-J.C. and J.A.J. are listed as coinventors on patents and patent

applications that are related to lipid bilayer technologies. In addition,

N.-J.C. is a founder of and J.A.J. is a scientiﬁc advisor to LUCA AICell Inc,

which is developing lipid-related technologies. The other authors declare

no conﬂict of interest.

Keywords

drug delivery, lipid nanoparticle, microﬂuidics, nanoarchitectonics,

vaccine

Received: March 31, 2022

Revised: May 9, 2022

Published online: June 16, 2022

[1] a) J. A. Kulkarni, D. Witzigmann, S. B. Thomson, S. Chen,

B. R. Leavitt, P. R. Cullis, R. van der Meel, Nat. Nanotechnol.

2021, 16, 630; b) H. Yin, R. L.Kanasty, A. A.Eltoukhy, A. J.Vegas,

J. R.Dorkin, D. G.Anderson, Nat. Rev. Genet. 2014, 15, 541.

[2] S.Chesnoy, L.Huang, Annu. Rev. Biophys. Biomol. Struct. 2000, 29, 27.

[3] P.Guo, J. Nanosci. Nanotechnol. 2005, 5, 1964.

[4] Y.Eygeris, S.Patel, A.Jozic, G.Sahay, Nano Lett. 2020, 20, 4543.

[5] Y. Zhang, H. Li, J. Sun, J. Gao, W. Liu, B. Li, Y. Guo, J. Chen,

Int. J. Pharm. 2010, 390, 198.

[6] E.Dolgin, Nat. Biotechnol. 2022, 40, 283.

[7] P. F. McKay, K. Hu, A. K. Blakney, K. Samnuan, J. C. Brown,

R.Penn, J.Zhou, C. R.Bouton, P.Rogers, K.Polra, Nat. Commun.

2020, 11, 3523.

[8] C. Webb, S. Ip, N. V. Bathula, P. Popova, S. K. Soriano,

H. H.Ly, B.Eryilmaz, V. A.Nguyen Huu, R.Broadhead, M.Rabel,

Mol. Pharmaceutics 2022, 19, 1047.

[9] a) K. S.Corbett, D. K.Edwards, S. R.Leist, O. M.Abiona, S.Boyoglu-

Barnum, R. A. Gillespie, S. Himansu, A. Schäfer, C. T. Ziwawo,

A. T. DiPiazza, Nature 2020, 586, 567; b) L. Schoenmaker,

D. Witzigmann, J. A. Kulkarni, R. Verbeke, G.Kersten, W. Jiskoot,

D. J.Crommelin, Int. J. Pharm. 2021, 601, 120586.

[10] a) X. Hou, T. Zaks, R. Langer, Y. Dong, Nat. Rev. Mater.

2021, 6, 1078; b) R. Tenchov, R. Bird, A. E. Curtze, Q. Zhou,

ACS Nano 2021, 15, 16982; c) B. Z.Igyártó, S.Jacobsen, S.Ndeupen,

Curr. Opin. Virol. 2021, 48, 65; d) E. Kon, U. Elia, D. Peer,

Curr. Opin. Biotechnol. 2022, 73, 329.

[11] a) M. Aono, K. Ariga, Adv. Mater. 2016, 28, 989; b) K. Ariga,

J. A. Jackman, N. J. Cho, S. h. Hsu, L. K. Shrestha, T. Mori,

J. Takeya, Chem. Rec. 2019, 19, 1891; c) K. Ariga, Y. Yamauchi,

Chem.- Asian J. 2020, 15, 718.

[12] a) K.Ariga, S.Watanabe, T.Mori, J.Takeya, NPG Asia Mater. 2018,

10, 90; b) K.Ariga, M.Nishikawa, T.Mori, J.Takeya, L. K.Shrestha,

J. P. Hill, Sci. Technol. Adv. Mater. 2019, 20, 51; c) K.Ariga, X. Jia,

J.Song, J. P.Hill, D. T. Leong, Y. Jia, J.Li, Angew. Chem., Int. Ed.

2020, 59, 15424; d) K.Ariga, Nanoscale Horiz. 2021, 6, 364.

[13] a) J. Geall Andrew, A. Verma, R.Otten Gillis, A.Shaw Christine,

A. Hekele, K. Banerjee, Y. Cu, W. Beard Clayton, A. Brito Luis,

T. Krucker, T. O’Hagan Derek, M. Singh, W. Mason Peter,

M. Valiante Nicholas, R. Dormitzer Philip, W. Barnett Susan,

R.Rappuoli, B.Ulmer Jerey, W.Mandl Christian, Proc. Natl. Acad.

Sci. USA 2012, 109, 14604; b) E.Samaridou, J.Heyes, P.Lutwyche,

Adv. Drug Delivery Rev. 2020, 154-155, 37; c) Y. Zhang, C. Sun,

C.Wang, K. E.Jankovic, Y.Dong, Chem. Rev. 2021, 121, 12181.

[14] a) A. C. Anselmo, S. Mitragotri, Bioeng. Transl. Med. 2016, 1, 10;

b) A. C. Anselmo, S. Mitragotri, Bioeng. Transl. Med. 2019, 4,

e10143; c) S.Yonezawa, H.Koide, T. Asai, Adv. Drug Delivery Rev.

2020, 154–155, 64.

[15] a) T. T.Thi, E. J. A. Suys, J. S. Lee, D. H. Nguyen, K. D. Park,

N. P.Truong, Vaccines 2021, 9, 359; b) B. N.Aldosari, I. M.Alfagih,

A. S.Almurshedi, Pharmaceutics 2021, 13, 206; c) A. C. Anselmo,

S.Mitragotri, Bioeng. Transl. Med. 2021, 6, e10246.

[16] a) M.Cornebise, E.Narayanan, Y.Xia, E. Acosta, L.Ci, H.Koch,

J.Milton, S. Sabnis, T.Salerno, K. E.Benenato, Adv. Funct. Mater.

2022, 32, 2106727; b) G. Basha, T. I. Novobrantseva, N. Rosin,

Y. Y. C. Tam, I. M. Hafez, M. K. Wong, T. Sugo, V. M. Ruda,

J.Qin, B. Klebanov, M.Ciufolini, A.Akinc, Y. K. Tam, M. J.Hope,

P. R. Cullis, Mol. Ther. 2011, 19, 2186; c) N.Maurer, K. F.Wong,

H.Stark, L.Louie, D.McIntosh, T.Wong, P.Scherrer, S. C.Semple,

P. R.Cullis, Biophys. J. 2001, 80, 2310; d) S. C.Semple, S. K.Klimuk,

T. O. Harasym, N. Dos Santos, S. M. Ansell, K. F. Wong,

N.Maurer, H.Stark, P. R.Cullis, M. J.Hope, P.Scherrer, Biochim.

Biophys. Acta, Biomembr. 2001, 1510, 152.

[17] a) S. J. Shepherd, D. Issadore, M. J. Mitchell, Biomaterials 2021,

274, 120826; b) M.Maeki, S.Uno, A.Niwa, Y.Okada, M.Tokeshi,

J. Controlled Release 2022, 344, 80.

[18] a) K. Ward, Z. H.Fan, J. Micromech. Microeng. 2015, 25, 094001;

b) C.-Y. Lee, C.-L. Chang, Y.-N.Wang, L.-M. Fu, Int. J. Mol. Sci.

2011, 12, 3263; c) M. J. W.Evers, J. A.Kulkarni, R. van der Meel,

P. R. Cullis, P. Vader, R. M. Schielers, Small Methods 2018, 2,

1700375.

[19] a) K. M. McMahon, M. P. Plebanek, C. S. Thaxton, Adv. Funct.

Mater. 2016, 26, 7824; b) K. M. McMahon, R. K. Mutharasan,

S.Tripathy, D.Veliceasa, M.Bobeica, D. K.Shumaker, A. J.Luthi,

B. T.Helfand, H. Ardehali, C. A.Mirkin, O.Volpert, C. S.Thaxton,

Nano Lett. 2011, 11, 1208; c) K. M. McMahon, C. S. Thaxton,

Expert Opin. Drug Delivery 2014, 11, 231; d) J.Wang, A. E.Calvert,

N. Kaplan, K. M. McMahon, W. Yang, K. Q. Lu, H. Peng,

C. S.Thaxton, R. M.Lavker, Adv. Ther. 2020, 3, 2000138.

[20] a) K. C.Vickers, B. T.Palmisano, B. M.Shoucri, R. D.Shamburek,

A. T.Remaley, Nat. Cell Biol. 2011, 13, 423; b) F.Tabet, K. C.Vickers,

L. F. Cuesta Torres, C. B. Wiese, B. M. Shoucri, G. Lambert,

C. Catherinet, L. Prado-Lourenco, M. G. Levin, S. Thacker,

P.Sethupathy, P. J.Barter, A. T.Remaley, K.-A.Rye, Nat. Commun.

2014, 5, 3292.

[21] M.Ouimet, T. J.Barrett, E. A.Fisher, Circ. Res. 2019, 124, 1505.

[22] a) K. C. Vickers, A. T. Remaley, Curr. Opin. Lipidol. 2012, 23, 91;

b) K. C.Vickers, A. T.Remaley, J. Lipid Res. 2014, 55, 4; c) S.Nazir,

V. Jankowski, G. Bender, S. Zewinger, K.-A. Rye, E. P. C. van

derVorst, Adv. Drug Delivery Rev. 2020, 159, 94.

[23] X. Chen, L. S. Mangala, L. Mooberry, E. Bayraktar, S. K. Dasari,

S. Ma, C. Ivan, K. A. Court, C. Rodriguez-Aguayo, R. Bayraktar,

S. Raut, N. Sabnis, X. Kong, X. Yang, G. Lopez-Berestein,

A. G.Lacko, A. K.Sood, Oncogene 2019, 38, 6095.

[24] A. G.Lacko, N. A.Sabnis, B. Nagarajan, W. J.McConathy, Front.

Pharmacol. 2015, 6, 247.

[25] M. M. K. Shahzad, L. S. Mangala, H. D. Han, C. Lu,

J. Bottsford-Miller, M. Nishimura, E. M. Mora, J.-W. Lee,

R. L. Stone, C. V. Pecot, D. Thanapprapasr, J.-W. Roh, P. Gaur,

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

M. P.Nair, Y.-Y.Park, N.Sabnis, M. T.Deavers, J.-S.Lee, L. M.Ellis,

G. Lopez-Berestein, W. J. McConathy, L. Prokai, A. G. Lacko,

A. K.Sood, Neoplasia 2011, 13, 309.

[26] T. M.Allen, P. R.Cullis, Adv. Drug Delivery Rev. 2013, 65, 36.

[27] a) J. S. Suk, Q. Xu, N. Kim, J. Hanes, L. M. Ensign, Adv. Drug

Delivery Rev. 2016, 99, 28; b) Z.Wang, E. D.Hood, J.Nong, J.Ding,

O. A.Marcos-Contreras, P. M.Glassman, K. M.Rubey, M.Zaleski,

C. L. Espy, D.Gullipali, T. Miwa, V. R. Muzykantov, W.-C. Song,

J. W.Myerson, J. S.Brenner, Adv. Mater. 2022, 34, 2107070.

[28] M.Longmire, P. L.Choyke, H.Kobayashi, Nanomedicine 2008, 3,

703.

[29] a) J. Lu, Y. Zhao, X. Zhou, J. H. He, Y. Yang, C. Jiang, Z. Qi,

W.Zhang, J.Liu, Biomacromolecules 2017, 18, 2286; b) A. Orellana

Esteban, S. Tenneti, L. Rangasamy, L. T. Lyle, S. Low Philip,

L.Kasinski Andrea, Sci. Transl. Med. 2017, 9, eaam9327.

[30] a) C.Wolfrum, S.Shi, K. N.Jayaprakash, M.Jayaraman, G.Wang,

R. K.Pandey, K. G.Rajeev, T.Nakayama, K.Charrise, E. M.Ndungo,

T. Zimmermann, V. Koteliansky, M. Manoharan, M. Stoel,

Nat. Biotechnol. 2007, 25, 1149; b) M. Egli, M. Manoharan,

Acc. Chem. Res. 2019, 52, 1036.

[31] a) Y. Yang, J. Wang, H. Shigematsu, W. Xu, W. M. Shih,

J. E. Rothman, C. Lin, Nat. Chem. 2016, 8, 476; b) Z. Zhang,

Y.Yang, F.Pincet, M. C.Llaguno, C.Lin, Nat. Chem. 2017, 9, 653.

[32] a) J.Guan, Q. Shen, Z.Zhang, Z.Jiang, Y.Yang, M.Lou, J.Qian,

W. Lu, C. Zhan, Nat. Commun. 2018, 9, 2982; b) N.AlSawaftah,

W. G. Pitt, G. A. Husseini, ACS Pharmacol. Transl. Sci. 2021, 4,

1028.

[33] C.Kelly, C.Jeeries, S.-A.Cryan, J. Drug Delivery 2011, 2011, 727241.

[34] a) P.-A. Monnard, T.Oberholzer, P. Luisi, Biochim. Biophys. Acta,

Biomembr. 1997, 1329, 39; b) S.Dow, Expert Opin. Drug Delivery

2008, 5, 11; c) C. R.Saﬁnya, K. K. Ewert, R. N.Majzoub, C.Leal,

New J. Chem. 2014, 38, 5164.

[35] A. A. Barba, G. Lamberti, C. Sardo, B. Dapas, M. Abrami,

M. Grassi, R.Farra, F.Tonon, G. Forte, F. Musiani, M. Licciardi,

G. Pozzato, F. Zanconati, B.Scaggiante, G.Grassi, G.Cavallaro,

Curr. Drug Metab. 2015, 16, 427.

[36] a) E. H.Pilkington, E. J. A.Suys, N. L. Trevaskis, A. K.Wheatley,

D.Zukancic, A.Algarni, H.Al-Wassiti, T. P.Davis, C. W.Pouton,

S. J.Kent, N. P.Truong, Acta Biomater. 2021, 131, 16; b) Y.Eygeris,

M.Gupta, J.Kim, G.Sahay, Acc. Chem. Res. 2022, 55, 2.

[37] N. R. Larson, G. Hu, Y. Wei, A. D. Tuesca, M. L. Forrest,

C. R.Middaugh, J. Pharm. Sci. 2022, 111, 690.

[38] A. Dilliard Sean, Q. Cheng, J. Siegwart Daniel, Proc. Natl. Acad.

Sci. USA 2021, 118, e2109256118.

[39] M. Schlich, R. Palomba, G. Costabile, S.Mizrahy, M. Pannuzzo,

D.Peer, P.Decuzzi, Bioeng. Transl. Med. 2021, 6, e10213.

[40] a) S. M.Lee, Q.Cheng, X.Yu, S.Liu, L. T.Johnson, D. J.Siegwart,

Angew. Chem., Int. Ed. 2021, 60, 5848; b) M. Jayaraman,

S. M. Ansell, B. L. Mui, Y. K. Tam, J. Chen, X. Du, D. Butler,

L. Eltepu, S. Matsuda, J. K. Narayanannair, K. G. Rajeev,

I. M. Hafez, A. Akinc, M. A. Maier, M. A. Tracy, P. R. Cullis,

T. D.Madden, M.Manoharan, M. J.Hope, Angew. Chem., Int. Ed.

2012, 51, 8529; c) O. S.Fenton, K. J.Kauman, R. L. McClellan,

E. A.Appel, J. R.Dorkin, M. W.Tibbitt, M. W.Heartlein, F.DeRosa,

R.Langer, D. G.Anderson, Adv. Mater. 2016, 28, 2939.

[41] a) S. Liu, Q. Cheng, T. Wei, X. Yu, L. T. Johnson, L. Farbiak,

D. J.Siegwart, Nat. Mater. 2021, 20, 701; b) N.Meng, D. Grimm,

Signal Transduction Targeted Ther. 2021, 6, 206.

[42] K. A. Hajj, R. L. Ball, S. B. Deluty, S. R. Singh, D. Strelkova,

C. M.Knapp, K. A.Whitehead, Small 2019, 15, 1805097.

[43] a) M. A. Maier, M. Jayaraman, S. Matsuda, J. Liu, S. Barros,

W. Querbes, Y. K.Tam, S. M.Ansell, V.Kumar, J.Qin, X.Zhang,

Q. Wang, S. Panesar, R. Hutabarat, M. Carioto, J. Hettinger,

P. Kandasamy, D. Butler, K. G. Rajeev, B. Pang, K. Charisse,

K.Fitzgerald, B. L.Mui, X.Du, P.Cullis, T. D.Madden, M. J.Hope,

M.Manoharan, A.Akinc, Mol. Ther. 2013, 21, 1570; b) Y. Suzuki,

K. Hyodo, T. Suzuki, Y. Tanaka, H. Kikuchi, H. Ishihara, Int. J.

Pharm. 2017, 519, 34; c) H. Tanaka, T. Takahashi, M. Konishi,

N. Takata, M. Gomi, D. Shirane, R. Miyama, S. Hagiwara,

Y.Yamasaki, Y.Sakurai, K.Ueda, K.Higashi, K.Moribe, E.Shinsho,

R.Nishida, K.Fukuzawa, E.Yonemochi, K.Okuwaki, Y.Mochizuki,

Y.Nakai, K.Tange, H.Yoshioka, S.Tamagawa, H.Akita, Adv. Funct.

Mater. 2020, 30, 1910575; d) H. Tanaka, T.Takahashi, M.Konishi,

N. Takata, M. Gomi, D. Shirane, R. Miyama, S. Hagiwara,

Y.Yamasaki, Y.Sakurai, Adv. Funct. Mater. 2020, 30, 1910575.

[44] a) O. S.Fenton, K. J.Kauman, R. L.McClellan, J. C. Kaczmarek,

M. D. Zeng, J. L. Andresen, L. H. Rhym, M. W. Heartlein,

F.DeRosa, D. G.Anderson, Angew. Chem., Int. Ed. 2018, 57, 13582;

b) M. J.Carrasco, S.Alishetty, M.-G.Alameh, H. Said, L. Wright,

M.Paige, O. Soliman, D.Weissman, T. E.Cleveland, A.Grishaev,

M. D.Buschmann, Commun. Biol. 2021, 4, 956.

[45] F. Ding, H. Zhang, J. Cui, Q. Li, C. Yang, Biomater. Sci. 2021, 9,

7534.

[46] M. Kim, M. Jeong, S. Hur, Y. Cho, J. Park, H. Jung, Y. Seo,

H. A.Woo, K. T.Nam, K.Lee, H.Lee, Sci. Adv. 2021, 7, eabf4398.

[47] a) S.Mochizuki, N. Kanegae, K.Nishina, Y.Kamikawa, K.Koiwai,

H.Masunaga, K.Sakurai, Biochim. Biophys. Acta, Biomembr. 2013,

1828, 412; b) J. A.Kulkarni, D.Witzigmann, J.Leung, Y. Y. C.Tam,

P. R. Cullis, Nanoscale 2019, 11, 21733; c) R.Zhang, R. El-Mayta,

T. J. Murdoch, C. C.Warzecha, M. M. Billingsley, S. J.Shepherd,

N.Gong, L.Wang, J. M.Wilson, D. Lee, M. J.Mitchell, Biomater.

Sci. 2021, 9, 1449; d) J.Kim, A. Jozic, G. Sahay, Cell. Mol. Bioeng.

2020, 13, 463.

[48] S. Patel, N. Ashwanikumar, E. Robinson, Y. Xia, C. Mihai,

J. P. Grith, S. Hou, A. A. Esposito, T. Ketova, K. Welsher,

J. L.Joyal, Ö.Almarsson, G.Sahay, Nat. Commun. 2020, 11, 983.

[49] H.Tanaka, R.Miyama, Y.Sakurai, S.Tamagawa, Y.Nakai, K.Tange,

H.Yoshioka, H.Akita, Pharmaceutics 2021, 13, 2097.

[50] F. Sebastiani, M. Yanez Arteta, M. Lerche, L. Porcar, C. Lang,

R. A. Bragg, C. S. Elmore, V. R. Krishnamurthy, R. A. Russell,

T. Darwish, H. Pichler, S. Waldie, M. Moulin, M. Haertlein,

V. T.Forsyth, L.Lindfors, M.Cárdenas, ACS Nano 2021, 15, 6709.

[51] a) P. Sellaturay, S. Nasser, S. Islam, P.Gurugama, P. W. Ewan,

Clin. Exp. Allergy 2021, 51, 861; b) A. Troelnikov, G. Perkins,

C. Yuson, A. Ahamdie, S. Balouch, P. R. Hurtado, P. Hissaria,

J. Allergy Clin. Immunol. 2021, 148, 91; c) M.Picard, J.-P. Drolet,

M.-S. Masse, C. A. Filion, F. Almuhizi, M. Fein, A. Copaescu,

G. A. C. Isabwe, M. Blaquière, M.-N. Primeau, J. Allergy Clin.

Immunol.: Pract. 2022, 10, 620; d) J. A. Cahill, M. Kan, Allergy

2022, 77, 337; e) P.Sellaturay, P.Gurugama, V.Harper, T.Dymond,

P.Ewan, S.Nasser, Clin. Exp. Allergy 2022, 52, 12.

[52] B.Cabanillas, N.Novak, C. A.Akdis, Allergy 2021, 76, 1938.

[53] a) A. Akinc, W. Querbes, S. De, J. Qin, M. Frank-Kamenetsky,

K. N. Jayaprakash, M. Jayaraman, K. G. Rajeev, W. L. Cantley,

J. R. Dorkin, J. S. Butler, L. Qin, T. Racie, A. Sprague, E. Fava,

A. Zeigerer, M. J. Hope, M. Zerial, D. W. Y. Sah, K. Fitzgerald,

M. A. Tracy, M. Manoharan, V. Koteliansky, A. d. Fougerolles,

M. A. Maier, Mol. Ther. 2010, 18, 1357; b) S. Sabnis,

E. S. Kumarasinghe, T. Salerno, C. Mihai, T.Ketova, J. J. Senn,

A. Lynn, A. Bulychev, I. McFadyen, J. Chan, Ö. Almarsson,

M. G.Stanton, K. E.Benenato, Mol. Ther. 2018, 26, 1509.

[54] a) D. W.Lee, X.Banquy, K.Kristiansen, Y.Kaufman, J. M. Boggs,

J. N. Israelachvili, Proc. Natl. Acad. Sci. USA 2014, 111, E768;

b) A. H.Pande, S.Qin, S. A.Tatulian, Biophys. J. 2005, 88, 4084.

[55] a) A. Wittrup, A. Ai, X. Liu, P. Hamar, R.Trifonova, K. Charisse,

M. Manoharan, T. Kirchhausen, J. Lieberman, Nat. Bio-

technol. 2015, 33, 870; b) J. Gilleron, W. Querbes, A. Zeigerer,

A.Borodovsky, G.Marsico, U. Schubert, K.Manygoats, S. Seifert,

C.Andree, M. Stöter, H.Epstein-Barash, L.Zhang, V. Koteliansky,

K.Fitzgerald, E.Fava, M.Bickle, Y.Kalaidzidis, A.Akinc, M.Maier,

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

M.Zerial, Nat. Biotechnol. 2013, 31, 638; c) G.Sahay, W.Querbes,

C. Alabi, A. Eltoukhy, S. Sarkar, C. Zurenko, E. Karagiannis,

K.Love, D. Chen, R.Zoncu, Y. Buganim, A.Schroeder, R.Langer,

D. G.Anderson, Nat. Biotechnol. 2013, 31, 653.

[56] a) N. Hartl, F. Adams, O. M. Merkel, Adv. Ther. 2021, 4,

2000092; b) P. Sabourian, G. Yazdani, S. S. Ashraf, M. Frounchi,

S.Mashayekhan, S.Kiani, A.Kakkar, Int. J. Mol. Sci. 2020, 21, 8019.

[57] R.Gaspar, F.Coelho, B. F. B.Silva, Molecules 2020, 25, 5006.

[58] R. L. Ball, K. A. Hajj, J. Vizelman, P. Bajaj, K. A. Whitehead,

Nano Lett. 2018, 18, 3814.

[59] a) J. K.Watts, G. F.Deleavey, M. J.Damha, Drug Discovery Today

2008, 13, 842; b) N. J. Caplen, E. W. Alton, P. G. Mddleton,

J. R. Dorin, B. J. Stevenson, X. Gao, S. R. Durham, P. K. Jeery,

M. E.Hodson, C.Coutelle, Nat. Med. 1995, 1, 39.

[60] a) B. Lewin, Genes 9, Jones & Bartlett Learning, Burlington, MA

2008; b) B.Alberts, Molecular Biology of the Cell, Garland Science,

New York 2004.

[61] A. Wu, Y. Peng, B. Huang, X. Ding, X. Wang, P. Niu, J. Meng,

Z. Zhu, Z. Zhang, J. Wang, J. Sheng, L. Quan, Z. Xia, W. Tan,

G.Cheng, T.Jiang, Cell Host Microbe 2020, 27, 325.

[62] a) M. Yanez Arteta, T. Kjellman, S. Bartesaghi, S. Wallin, X. Wu,

J. Kvist Alexander, A. Dabkowska, N. Székely, A. Radulescu,

J. Bergenholtz, L. Lindfors, Proc. Natl. Acad. Sci. 2018, 115, E3351;

b) J. A.Kulkarni, M. M.Darjuan, J. E.Mercer, S.Chen, R.van derMeel,

J. L.Thewalt, Y. Y. C.Tam, P. R.Cullis, ACS Nano 2018, 12, 4787.

[63] Z. Hejdankova, V. Vanek, F. Sedlak, J. Prochazka, A. Diederichs,

S.Kereïche, B.Novotna, M.Budesinsky, G.Birkus, K.Grantz Saskova,

P.Cigler, Adv. Funct. Mater. 2021, 31, 2101391.

[64] a) K. J. Kauman, J. R. Dorkin, J. H. Yang, M. W. Heartlein,

F. DeRosa, F. F. Mir, O. S. Fenton, D. G. Anderson, Nano Lett.

2015, 15, 7300; b) M. P. Lokugamage, D. Vanover, J.Beyersdorf,

M. Z. C. Hatit, L. Rotolo, E. S. Echeverri, H. E. Peck, H. Ni,

J.-K. Yoon, Y.Kim, P. J.Santangelo, J. E.Dahlman, Nat. Biomed.

Eng. 2021, 5, 1059; c) S. C.Semple, A.Akinc, J.Chen, A. P.Sandhu,

B. L. Mui, C. K. Cho, D. W. Y. Sah, D. Stebbing, E. J. Crosley,

E.Yaworski, I. M.Hafez, J. R.Dorkin, J.Qin, K.Lam, K. G.Rajeev,

K. F.Wong, L. B.Jes, L.Nechev, M. L.Eisenhardt, M.Jayaraman,

M.Kazem, M. A.Maier, M.Srinivasulu, M. J. Weinstein, Q.Chen,

R. Alvarez, S. A. Barros, S. De, S. K. Klimuk, T.Borland, etal.,

Nat. Biotechnol. 2010, 28, 172; d) M. M. Billingsley, N. Singh,

P. Ravikumar, R. Zhang, C. H. June, M. J. Mitchell, Nano Lett.

2020, 20, 1578; e) T. Terada, J. A.Kulkarni, A. Huynh, S. Chen,

R.van derMeel, Y. Y. C.Tam, P. R.Cullis, Langmuir 2021, 37, 1120.

[65] K. T. Love, K. P.Mahon, C. G. Levins, K. A. Whitehead, W.Querbes,

J. R.Dorkin, J.Qin, W.Cantley, L. L.Qin, T.Racie, M.Frank-Kamenetsky,

K. N. Yip, R. Alvarez, D. W. Y. Sah, A. de Fougerolles, K. Fitzgerald,

V.Koteliansky, A.Akinc, R.Langer, D. G.Anderson, Proc. Natl. Acad. Sci.

USA 2010, 107, 1864.

[66] a) R. A. Petros, J. M. DeSimone, Nat. Rev. Drug Discovery 2010,

9, 615; b) T.Nakamura, M.Kawai, Y.Sato, M.Maeki, M. Tokeshi,

H.Harashima, Mol. Pharmaceutics 2020, 17, 944.

[67] S. Chen, Y. Y. C. Tam, P. J. C. Lin, M. M. H. Sung, Y. K.Tam,

P. R.Cullis, J. Controlled Release 2016, 235, 236.

[68] A. U.Andar, R. R.Hood, W. N.Vreeland, D. L.DeVoe, P. W.Swaan,

Pharm. Res. 2014, 31, 401.

[69] K. J. Hassett, J. Higgins, A. Woods, B. Levy, Y. Xia, C. J. Hsiao,

E. Acosta, Ö. Almarsson, M. J. Moore, L. A. Brito, J. Controlled

Release 2021, 335, 237.

[70] D.Chen, K. T.Love, Y.Chen, A. A.Eltoukhy, C.Kastrup, G.Sahay,

A.Jeon, Y.Dong, K. A. Whitehead, D. G.Anderson, J. Am. Chem.

Soc. 2012, 134, 6948.

[71] K.Ariga, Molecules 2021, 26, 1621.

[72] a) A.Tan, L. Hong, J. D. Du, B. J.Boyd, Adv. Sci. 2019, 6, 1801223;

b) M. R. Molla, P. A. Levkin, Adv. Mater. 2016, 28, 1159;

c) S.Banerjee, J.Pillai, Expert Opin. Drug Metab. Toxicol. 2019, 15, 499.

[73] A.Wagner, K.Vorauer-Uhl, G.Kreismayr, H.Katinger, J. Liposome

Res. 2002, 12, 259.

[74] a) G. M.Whitesides, Nature 2006, 442, 368; b) J. Zhang, S. Yan,

D.Yuan, G.Alici, N.-T. Nguyen, M. EbrahimiWarkiani, W.Li, Lab

Chip 2016, 16, 10.

[75] a) T. A. Duncombe, A. M.Tentori, A. E.Herr, Nat. Rev. Mol. Cell

Biol. 2015, 16, 554; b) S.Damiati, U. B. Kompella, S. A. Damiati,

R.Kodzius, Genes 2018, 9, 103; c) Y.Bai, M.Gao, L. Wen, C.He,

Y.Chen, C.Liu, X.Fu, S. Huang, Biotechnol. J. 2018, 13, 1700170;

d) L.Zhang, Q.Chen, Y.Ma, J.Sun, ACS Appl. Bio Mater. 2020, 3,

107; e) M.Maeki, N.Kimura, Y.Sato, H.Harashima, M.Tokeshi,

Adv. Drug Delivery Rev. 2018, 128, 84.

[76] a) E. C.Jensen, B. P.Bhat, R. A. Mathies, Lab Chip 2010, 10, 685;

b) B.Schuster, M.Junkin, S. S.Kashaf, I.Romero-Calvo, K.Kirby,

J. Matthews, C. R. Weber, A.Rzhetsky, K. P. White, S. Tay, Nat.

Commun. 2020, 11, 5271.

[77] H. Wang, K. Liu, K.-J. Chen, Y. Lu, S. Wang, W.-Y. Lin, F. Guo,

K.-i. Kamei, Y.-C. Chen, M. Ohashi, M. Wang, M. A. Garcia,

X.-Z.Zhao, C. K. F.Shen, H.-R.Tseng, ACS Nano 2010, 4, 6235.

[78] Y. Kim, B. Lee Chung, M. Ma, W. J. M. Mulder, Z. A. Fayad,

O. C.Farokhzad, R.Langer, Nano Lett. 2012, 12, 3587.

[79] P. M.Valencia, E. M.Pridgen, M.Rhee, R.Langer, O. C.Farokhzad,

R.Karnik, ACS Nano 2013, 7, 10671.

[80] P. M.Valencia, O. C.Farokhzad, R.Karnik, R.Langer, Nat. Nanotechnol.

2012, 7, 623.

[81] M. D. Buschmann, M. J. Carrasco, S. Alishetty, M. Paige,

M. G.Alameh, D.Weissman, Vaccines 2021, 9, 65.

[82] C. M. Bailey-Hytholt, P. Ghosh, J. Dugas, I. E. Zarraga,

A.Bandekar, J. Visualized Exp. 2021, 168, e62226.

[83] a) M. T.Abrams, M. L.Koser, J.Seitzer, S. C.Williams, M. A.DiPietro,

W. Wang, A. W. Shaw, X. Mao, V. Jadhav, J. P. Davide, P. A. Burke,

A. B.Sachs, S. M.Stirdivant, L.Sepp-Lorenzino, Mol. Ther. 2010, 18, 171;

b) T. S.Zimmermann, A. C. H.Lee, A.Akinc, B.Bramlage, D.Bumcrot,

M. N.Fedoruk, J.Harborth, J. A.Heyes, L. B.Jes, M.John, A. D.Judge,

K. Lam, K. McClintock, L. V.Nechev, L. R. Palmer, T. Racie, I. Röhl,

S. Seiert, S. Shanmugam, V. Sood, J. Soutschek, I. Toudjarska,

A. J. Wheat, E. Yaworski, W. Zedalis, V. Koteliansky, M. Manoharan,

H.-P.Vornlocher, I.MacLachlan, Nature 2006, 441, 111.

[84] D.Stroock Abraham, K. W.Dertinger Stephan, A.Ajdari, I.Mezić,

A.Stone Howard, M.Whitesides George, Science 2002, 295, 647.

[85] S. J. Shepherd, C. C. Warzecha, S. Yadavali, R. El-Mayta,

M.-G.Alameh, L.Wang, D.Weissman, J. M.Wilson, D. Issadore,

M. J.Mitchell, Nano Lett. 2021, 21, 5671.

[86] N. M. Belliveau, J. Huft, P. J. C. Lin, S. Chen, A. K. K. Leung,

T. J.Leaver, A. W.Wild, J. B.Lee, R. J.Taylor, Y. K.Tam, C. L.Hansen,

P. R.Cullis, Mol. Ther.–Nucleic Acids 2012, 1, e37.

[87] A.Jahn, W. N.Vreeland, M.Gaitan, L. E. Locascio, J. Am. Chem.

Soc. 2004, 126, 2674.

[88] C. Webb, N. Forbes, C. B. Roces, G. Anderluzzi, G. Lou,

S. Abraham, L. Ingalls, K. Marshall, T. J. Leaver, J. A. Watts,

J. W.Aylott, Y.Perrie, Int. J. Pharm. 2020, 582, 119266.

[89] N. Kimura, M. Maeki, Y. Sato, Y. Note, A. Ishida, H. Tani,

H.Harashima, M.Tokeshi, ACS Omega 2018, 3, 5044.

[90] K.Amreen, S.Goel, ECS J. Solid State Sci. Technol. 2021, 10, 017002.

[91] a) J. B. Lee, K. Zhang, Y. Y. C. Tam, Y. K. Tam, N. M. Belliveau,

V. Y. C. Sung, P. J. C. Lin, E. LeBlanc, M. A. Ciufolini, P. S.Rennie,

P. R. Cullis, Int. J. Cancer 2012, 131, E781; b) S. Rietwyk, D. Peer,

ACS Nano 2017, 11, 7572; c) Z.Liu, S. Wang, C.Tapeinos, G. Torrieri,

V. Känkänen, N. El-Sayed, A. Python, J. T. Hirvonen, H. A. Santos,

Adv. Drug Delivery Rev. 2021, 174, 576; d) A. del Pozo-Rodríguez,

M. Á.Solinís, A.Rodríguez-Gascón, Eur. J. Pharm. Biopharm. 2016, 109,

184.

[92] a) D.Bumcrot, M.Manoharan, V.Koteliansky, D. W. Y.Sah, Nat.

Chem. Biol. 2006, 2, 711; b) R. L.Setten, J. J.Rossi, S.-p.Han, Nat.

Rev. Drug Discovery 2019, 18, 421.

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

[93] a) A. Fire, S. Xu, M. K. Montgomery, S. A.Kostas, S. E.Driver,

C. C.Mello, Nature 1998, 391, 806; b) D. M.Dykxhoorn, D.Palliser,

J.Lieberman, Gene Ther. 2006, 13, 541.

[94] D. M.Dykxhoorn, J.Lieberman, Annu. Rev. Biomed. Eng. 2006, 8, 377.

[95] J. A. Kulkarni, D. Witzigmann, S. Chen, P. R. Cullis, R. van

derMeel, Acc. Chem. Res. 2019, 52, 2435.

[96] V. Kumar, J. Qin, Y. Jiang, R. G. Duncan, B. Brigham, S. Fishman,

J. K.Nair, A.Akinc, S. A.Barros, P. V.Kasperkovitz, Mol. Ther.–Nucleic

Acids 2014, 3, e210.

[97] C. B. Roces, G. Lou, N. Jain, S. Abraham, A. Thomas,

G. W.Halbert, Y.Perrie, Pharmaceutics 2020, 12, 1095.

[98] a) L. Schoenmaker, D. Witzigmann, J. A. Kulkarni, R. Verbeke,

G.Kersten, W.Jiskoot, D. J. A.Crommelin, Int. J. Pharm. 2021, 601,

120586; b) M.Gaviria, B.Kilic, Nat. Biotechnol. 2021, 39, 546.

[99] a) C. Wang, Y. Zhang, Y. Dong, Acc. Chem. Res. 2021, 54, 4283;

b) M.Qiu, Y. Li, H.Bloomer, Q.Xu, Acc. Chem. Res. 2021, 54, 4001;

c) L. R.Baden, H. M.El Sahly, B.Essink, K.Kotlo, S.Frey, R. Novak,

D.Diemert, S. A. Spector, N.Rouphael, C. B.Creech, J.McGettigan,

S. Khetan, N. Segall, J. Solis, A. Brosz, C. Fierro, H. Schwartz,

K.Neuzil, L.Corey, P.Gilbert, H.Janes, D.Follmann, M. Marovich,

J.Mascola, L. Polakowski, J.Ledgerwood, B. S.Graham, H.Bennett,

R. Pajon, C. Knightly, et al., New Engl. J. Med. 2020, 384, 403;

d) F. P. Polack, S. J. Thomas, N. Kitchin, J. Absalon, A. Gurtman,

S. Lockhart, J. L. Perez, G. Pérez Marc, E. D. Moreira, C. Zerbini,

R.Bailey, K. A.Swanson, S.Roychoudhury, K.Koury, P.Li, W. V.Kalina,

D.Cooper, R. W.Frenck, L. L.Hammitt, Ö.Türeci, H.Nell, A.Schaefer,

S. Ünal, D. B. Tresnan, S. Mather, P. R. Dormitzer, U. Şahin,

K. U.Jansen, W. C.Gruber, New Engl. J. Med. 2020, 383, 2603.

[100] a) D. Adams, A. Gonzalez-Duarte, W. D.O’Riordan, C.-C.Yang,

M. Ueda, A. V. Kristen, I. Tournev, H. H. Schmidt, T. Coelho,

J. L. Berk, K.-P. Lin, G. Vita, S. Attarian, V. Planté-Bordeneuve,

M. M. Mezei, J. M. Campistol, J. Buades, T. H. Brannagan,

B. J. Kim, J. Oh, Y. Parman, Y. Sekijima, P. N. Hawkins,

S. D. Solomon, M. Polydefkis, P. J. Dyck, P. J.Gandhi, S. Goyal,

J. Chen, A. L. Strahs, et al., New Engl. J. Med. 2018, 379, 11;

b) A.Akinc, M. A.Maier, M.Manoharan, K.Fitzgerald, M.Jayaraman,

S. Barros, S. Ansell, X. Du, M. J. Hope, T. D. Madden, B. L. Mui,

S. C.Semple, Y. K. Tam, M.Ciufolini, D.Witzigmann, J. A.Kulkarni,

R.van derMeel, P. R.Cullis, Nat. Nanotechnol. 2019, 14, 1084.

[101] M. L.Guevara, F.Persano, S.Persano, Front. Chem. 2020, 8, 963.

[102] a) R.Homan, L.Margolis, L.Bergelson, FEBS Lett. 1978, 93, 365;

b) R. Fraley, S. Subramani, P.Berg, D. Papahadjopoulos, J. Biol.

Chem. 1980, 255, 10431.

[103] F. Martinon, S. Krishnan, G. Lenzen, R. Magné, E. Gomard,

J. G.Guillet, J. P.Lévy, P.Meulien, Eur. J. Immunol. 1993, 23, 1719.

[104] R.Fraley, D.Papahadjopoulos, Trends Biochem. Sci. 1981, 6, 77.

[105] P.Felgner, Y.Barenholz, J.Behr, S.Cheng, P.Cullis, L.Huang, J.Jessee,

L.Seymour, F.Szoka, A.Thierry, Hum. Gene Ther. 1997, 8, 511.

[106] 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 1987, 84, 7413.

[107] R. W.Malone, P. L.Felgner, I. M.Verma, Proc. Natl. Acad. Sci. USA

1989, 86, 6077.

[108] U. Sahin, P. Oehm, E. Derhovanessian, R. A. Jabulowsky,

M. Vormehr, M. Gold, D. Maurus, D. Schwarck-Kokarakis,

A. N.Kuhn, T.Omokoko, Nature 2020, 585, 107.

[109] a) T.Wang, L. M.Larcher, L.Ma, R. N.Veedu, Molecules 2018, 23,

2564; b) X.Yu, X.Liang, H.Xie, S. Kumar, N.Ravinder, J.Potter,

X.deMollerat du Jeu, J. D.Chesnut, Biotechnol. Lett. 2016, 38, 919;

c) A.Masotti, G.Mossa, C.Cametti, G.Ortaggi, A.Bianco, N.Del

Grosso, D.Malizia, C.Esposito, Colloids Surf., B 2009, 68, 136.

[110] a) F. Barthel, J.-S. Remy, J.-P. Loeer, J.-P. Behr, DNA Cell Biol.

1993, 12, 553; b) J. H.Felgner, R.Kumar, C.Sridhar, C. J.Wheeler,

Y. J. Tsai, R.Border, P. Ramsey, M. Martin, P. L. Felgner, J. Biol.

Chem. 1994, 269, 2550.

[111] a) F. Ponti, M.Campolungo, C. Melchiori, N.Bono, G.Candiani,

Chem. Phys. Lipids 2021, 235, 105032; b) D. Zhi, S.Zhang, S.Cui,

Y. Zhao, Y. Wang, D. Zhao, Bioconjugate Chem. 2013, 24, 487;

c) D. Zhi, Y. Bai, J. Yang, S. Cui, Y. Zhao, H. Chen, S. Zhang,

Adv. Colloid Interface Sci. 2018, 253, 117.

[112] M. L.Tilkins, P.Hawley-Nelson, P.Battista, Focus 1994, 16, 117.

[113] S.Guan, J.Rosenecker, Gene Ther. 2017, 24, 133.

[114] a) E.Galanis, E.Hersh, A.Stopeck, R.Gonzalez, P.Burch, C.Spier,

E. Akporiaye, J. Rinehart, J. Edmonson, R. Sobol, J. Clin. Oncol.

1999, 17, 3313; b) J. Buck, P. Grossen, P. R. Cullis, J. Huwyler,

D.Witzigmann, ACS Nano 2019, 13, 3754.

[115] a) J.-S. Remy, C. Sirlin, P.Vierling, J.-P.Behr, Bioconjugate Chem.

1994, 5, 647; b) D. A.Balazs, W.Godbey, J. Drug Delivery 2011, 2011,

326497.

[116] a) C.Srinivasan, D. J. Burgess, J. Controlled Release 2009, 136, 62;

b) E.Dodds, M.Dunckley, K.Naujoks, U. Michaelis, G.Dickson,

Gene Ther. 1998, 5, 542; c) B. Maiti, M. Kamra, A. A. Karande,

S. Bhattacharya, Org. Biomol. Chem. 2018, 16, 1983; d) J. Yang,

L.Huang, Gene Ther. 1998, 5, 380.

[117] N. F. Bouxsein, C. S. McAllister, K. K. Ewert, C. E. Samuel,

C. R.Saﬁnya, Biochemistry 2007, 46, 4785.

[118] X.Gao, L.Huang, Biochem. Biophys. Res. Commun. 1991, 179, 280.

[119] a) S. Li, X.Gao, K.Son, F. Sorgi, H.Hoﬂand, L.Huang, J. Con-

trolled Release 1996, 39, 373; b) C.Esposito, J.Generosi, G.Mossa,

A.Masotti, A. C.Castellano, Colloids Surf., B. 2006, 53, 187.

[120] N. J. Zuidam, Y. Barenholz, Biochim. Biophys. Acta, Biomembr.

1997, 1329, 211.

[121] D. Pozzi, C. Marchini, F. Cardarelli, F. Salomone, S. Coppola,

M. Montani, M. E. Zabaleta, M. Digman, E. Gratton,

V.Colapicchioni, Biochim. Biophys. Acta, Biomembr. 2014, 1838, 957.

[122] a) Y.Tan, L.Huang, J. Drug Targeting 2002, 10, 153; b) G. J.Nabel,

E. G.Nabel, Z.-Y.Yang, B. A.Fox, G. E.Plautz, X.Gao, L.Huang,

S.Shu, D.Gordon, A. E. Chang, Proc. Natl. Acad. Sci. USA 1993,

90, 11307.

[123] G. Settanni, W. Brill, H. Haas, F. Schmid, Macromol. Rapid

Commun. 2021, 2100683.

[124] a) D. Habrant, P. Peuziat, T. Colombani, L. Dallet, J. Gehin,

E. Goudeau, B. Evrard, O. Lambert, T. Haudebourg, B. Pitard,

J. Med. Chem. 2016, 59, 3046; b) L. Desigaux, M. Sainlos,

O.Lambert, R.Chevre, E.Letrou-Bonneval, J.-P.Vigneron, P.Lehn,

J.-M.Lehn, B.Pitard, Proc. Natl. Acad. Sci. USA 2007, 104, 16534.

[125] N. Tateshita, N. Miura, H. Tanaka, T. Masuda, S. Ohtsuki,

K. Tange, Y. Nakai, H. Yoshioka, H. Akita, J. Controlled Release

2019, 310, 36.

[126] S.Akhter, M.Berchel, P.-A.Jarès, P.Midoux, C.Pichon, Pharma-

ceutics 2022, 14, 581.

[127] D. P. Vangasseri, Z. Cui, W. Chen, D. A. Hokey, L. D. FaloJr,

L.Huang, Mol. Membr. Biol. 2006, 23, 385.

[128] Z.Du, M. M. Munye, A. D.Tagalakis, M. D.Manunta, S. L. Hart,

Sci. Rep. 2014, 4, 7107.

[129] H. Farhood, N. Serbina, L. Huang, Biochim. Biophys. Acta,

Biomembr. 1995, 1235, 289.

[130] D. Simberg, D. Danino, Y. Talmon, A. Minsky, M. E. Ferrari,

C. J.Wheeler, Y.Barenholz, J. Biol. Chem. 2001, 276, 47453.

[131] a) Y. Liu, L. C. Mounkes, H. D. Liggitt, C. S. Brown, I. Solodin,

T. D. Heath, R. J. Debs, Nat. Biotechnol. 1997, 15, 167;

b) B. Sternberg, K. Hong, W. Zheng, D. Papahadjopoulos,

Biochim. Biophys. Acta, Biomembr. 1998, 1375, 23.

[132] D. Pozzi, C. Marchini, F. Cardarelli, H. Amenitsch, C. Garulli,

A.Bifone, G.Caracciolo, Biochim. Biophys. Acta, Biomembr. 2012,

1818, 2335.

[133] L.Xu, T. J.Anchordoquy, Biochim. Biophys. Acta, Biomembr. 2008,

1778, 2177.

[134] G. Caracciolo, D. Pozzi, A. L. Capriotti, C. Cavaliere, A. Laganà,

J. Nanopart. Res. 2013, 15, 1498.

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

[135] R. Koynova, B. Tenchov, in Nucleic Acid Transfection, Vol. 296

(Eds: W. Bielke, C. Erbacher), Springer, Berlin 2010, p. 51.

[136] J.Wang, X.Guo, Y.Xu, L.Barron, F. C.Szoka, J. Med. Chem. 1998,

41, 2207.

[137] M.Kang, H.Kim, C.Leal, Curr. Opin. Colloid Interface Sci. 2016, 26,

58.

[138] S. Eastman, C. Siegel, J. Tousignant, A. Smith, S. Cheng,

R.Scheule, Biochim. Biophys. Acta, Biomembr. 1997, 1325, 41.

[139] C. R.Saﬁnya, Curr. Opin. Struct. Biol. 2001, 11, 440.

[140] J. O.Rädler, I.Koltover, T.Salditt, C. R. Saﬁnya, Science 1997, 275,

810.

[141] I.Koltover, T.Salditt, J. O.Rädler, C. R.Saﬁnya, Science 1998, 281, 78.

[142] A. J. Lin, N. L. Slack, A. Ahmad, C. X. George, C. E. Samuel,

C. R.Saﬁnya, Biophys. J. 2003, 84, 3307.

[143] a) K. Ewert, A.Ahmad, H. M.Evans, C. R.Saﬁnya, Expert Opin.

Biol. Ther. 2005, 5, 33; b) G. Caracciolo, D. Pozzi, R. Caminiti,

A.Congiu Castellano, Eur. Phys. J. 2003, 10, 331.

[144] B. Ma, S. Zhang, H. Jiang, B. Zhao, H. Lv, J. Controlled Release

2007, 123, 184.

[145] a) R. Koynova, L. Wang, R. C. MacDonald, Proc. Natl. Acad.

Sci. USA 2006, 103, 14373; b) R.Hulst, I. Muizebelt, P.Oosting,

C.van derPol, A.Wagenaar, J. Šmisterová, E.Bulten, C.Driessen,

D.Hoekstra, J. B.Engberts, Eur. J. Org. Chem. 2004, 2004, 835.

[146] B. M. Bruininks, P. C.Souza, H.Ingolfsson, S. J.Marrink, eLife

2020, 9, e52012.

[147] K. K.Ewert, H. M.Evans, A.Zidovska, N. F.Bouxsein, A.Ahmad,

C. R.Saﬁnya, J. Am. Chem. Soc. 2006, 128, 3998.

[148] a) H.Kim, C.Leal, ACS Nano 2015, 9, 10214; b) H.Kim, J. Sung,

Y.Chang, A.Alfeche, C.Leal, ACS Nano 2018, 12, 9196.

[149] M.Riaz, Pak. J. Pharm. Sci. 1996, 9, 65.

[150] J. W.Meisel, G. W.Gokel, Sci. Rep. 2016, 6, 27662.

[151] S.Sriwongsitanont, M.Ueno, Open Colloid Sci. J. 2010, 4, 1.

[152] S.Batzri, E. D.Korn, Biochim. Biophys. Acta, Biomembr. 1973, 298,

1015.

[153] P.Ross, S.Hui, Gene Ther. 1999, 6, 651.

[154] M. R. Almofti, H. Harashima, Y. Shinohara, A. Almofti, W. Li,

H.Kiwada, Mol. Membr. Biol. 2003, 20, 35.

[155] a) Y.Maitani, S.Igarashi, M.Sato, Y. Hattori, Int. J. Pharm. 2007,

342, 33; b) S. Yang, J. Chen, D. Zhao, D. Han, X. Chen, Int. J.

Pharm. 2012, 434, 155; c) I.Tranchant, B.Thompson, C.Nicolazzi,

N.Mignet, D.Scherman, J. Gene Med. 2004, 6, S24.

[156] C.Krienke, L.Kolb, E. Diken, M. Streuber, S.Kirchho, T.Bukur,

Ö. Akilli-Öztürk, L. M. Kranz, H. Berger, J. Petschenka, Science

2021, 371, 145.

[157] a) V. Pector, J. Backmann, D. Maes, M. Vandenbranden,

J.-M.Ruysschaert, J. Biol. Chem. 2000, 275, 29533; b) E.Pozharski,

R. C.MacDonald, Biophys. J. 2003, 85, 3969.

[158] S. Even-Chen, R. Cohen, Y. Barenholz, Chem. Phys. Lipids 2012,

165, 414.

[159] J. Kim, J. Y. Kim, H. Kim, E. Kim, S. Park, K.-H. Ryu, E. G. Lee,

Int. J. Mol. Sci. 2021, 22, 12344.

[160] O. Zelphati, C.Nguyen, M.Ferrari, J. Felgner, Y.Tsai, P.Felgner,

Gene Ther. 1998, 5, 1272.

[161] M. T. Kennedy, E. V. Pozharski, V. A. Rakhmanova,

R. C.MacDonald, Biophys. J. 2000, 78, 1620.

[162] L.Wasungu, D.Hoekstra, J. Controlled Release 2006, 116, 255.

[163] J. M.Barichello, T.Ishida, H.Kiwada, in Liposomes: Methods and

Protocols (Ed: V. Weissig), Springer, Berlin 2010, p. 461.

[164] a) T. A. Balbino, N. T. Aoki, A. A. Gasperini, C. L. Oliveira,

A. R.Azzoni, L. P.Cavalcanti, G.Lucimara, Chem. Eng. J. 2013, 226,

423; b) T. A.Balbino, A. R.Azzoni, L. G.deLa Torre, Colloids Surf.,

B 2013, 111, 203; c) X.Liu, B.Bahloul, R. L. Kuen, K. Andrieux,

C.Roques, D.Scherman, Int. J. Pharm. 2021, 605, 120772.

[165] a) D. Llères, J. M. Weibel, D. Heissler, G. Zuber, G. Duportail,

Y.Mély, J. Gene Med. 2004, 6, 415; b) T. A.Balbino, J. M. Seraﬁn,

A. A. Malfatti-Gasperini, C. L. de Oliveira, L. P. Cavalcanti,

M. B. de Jesus, L. G. de La Torre, Langmuir 2016, 32, 1799;

c) A. T.-H. Hsieh, N. Hori, R.Massoudi, P. J.-H.Pan, H.Sasaki,

Y. A.Lin, A. P.Lee, Lab Chip 2009, 9, 2638.

[166] V. Escriou, C. Ciolina, F. Lacroix, G. Byk, D. Scherman, P. Wils,

Biochim. Biophys. Acta, Biomembr. 1998, 1368, 276.

[167] P.Foroozandeh, A. A.Aziz, Nanoscale Res. Lett. 2018, 13, 339.

[168] a) S. Grabbe, H. Haas, M. Diken, L. M. Kranz, P. Langguth,

U. Sahin, Nanomedicine 2016, 11, 2723; b) J. C. Birchall,

I. W.Kellaway, S. N.Mills, Int. J. Pharm. 1999, 183, 195.

[169] L. M.Kranz, M.Diken, H.Haas, S.Kreiter, C.Loquai, K. C.Reuter,

M.Meng, D.Fritz, F.Vascotto, H.Hefesha, Nature 2016, 534, 396.

[170] a) A. Peletta, E. Prompetchara, K. Tharakhet, P. Kaewpang,

S. Buranapraditkun, T. Techawiwattanaboon, T. Jbilou,

P. Krangvichian, S. Sirivichayakul, S. Manopwisedjaroen,

Vaccines 2021, 9, 874; b) Z.Heidari, J. S.Arora, D.Datta, V. T.John,

N.Kumar, G. P.Bansal, Pharm. Res. 2017, 34, 1796.

[171] Q. Cheng, T. Wei, L. Farbiak, L. T. Johnson, S. A. Dilliard,

D. J.Siegwart, Nat. Nanotechnol. 2020, 15, 313.

[172] A. K. Blakney, P. F. McKay, B. I. Yus, Y. Aldon, R. J. Shattock,

Gene Ther. 2019, 26, 363.

[173] T. A.Balbino, J. M.Seraﬁn, A.Radaic, M. B.deJesus, G.Lucimara,

Colloids Surf., B 2017, 152, 406.

[174] D. Simberg, S. Weisman, Y. Talmon, A. Faerman, T. Shoshani,

Y.Barenholz, J. Biol. Chem. 2003, 278, 39858.

[175] W.-Z. Zhou, D. Hoon, S. Huang, S. Fujii, K. Hashimoto,

R.Morishita, Y.Kaneda, Hum. Gene Ther. 1999, 10, 2719.

[176] A. De Beuckelaer, C. Pollard, S. Van Lint, K. Roose, L. Van

Hoecke, T. Naessens, V. K.Udhayakumar, M.Smet, N.Sanders,

S.Lienenklaus, Mol. Ther. 2016, 24, 2012.

[177] K. Okumura, M. Nakase, M. Inui, S. Nakamura, Y. Watanabe,

T.Tagawa, J. Gene Med. 2008, 10, 910.

[178] L.Van Hoecke, K.Roose, M.Ballegeer, Z.Zhong, N. N. Sanders,

S.De Koker, X.Saelens, S.Van Lint, Mol. Ther.– Nucleic Acids 2020,

22, 373.

[179] W. De Haes, J. Rejman, C. Pollard, C. Merlin, M. Vekemans,

E.Florence, S. C. De Smedt, J.Grooten, G. Vanham, S.De Koker,

Nanomedicine 2013, 8, 77.

[180] C. Pollard, J. Rejman, W. De Haes, B. Verrier, E. Van Gulck,

T. Naessens, S. De Smedt, P. Bogaert, J. Grooten, G. Vanham,

Mol. Ther. 2013, 21, 251.

[181] D. Christensen, K. S.Korsholm, I.Rosenkrands, T.Lindenstrøm,

P.Andersen, E. M.Agger, Expert Rev. Vaccines 2007, 6, 785.

[182] T. R. Smith, A. Patel, S. Ramos, D. Elwood, X. Zhu, J. Yan,

E. N.Gary, S. N.Walker, K.Schultheis, M.Purwar, Z.Xu, J.Walters,

P. Bhojnagarwala, M. Yang, N. Chokkalingam, P. Pezzoli,

E.Parzych, E. L.Reuschel, A.Doan, N.Tursi, M.Vasquez, J.Choi,

E.Tello-Ruiz, I.Maricic, M. A.Bah, Y.Wu, D.Amante, D. H.Park,

Y.Dia, A. R.Ali, etal., Nat. Commun. 2020, 11, 2601.

[183] D. J. Crommelin, T. J. Anchordoquy, D. B. Volkin, W. Jiskoot,

E.Mastrobattista, J. Pharm. Sci. 2021, 110, 997.

[184] K. Kubota, K. Onishi, K. Sawaki, T. Li, K. Mitsuoka, T. Sato,

S.Takeoka, Int. J. Nanomed. 2017, 12, 5121.

[185] A. K. Blakney, P. Deletic, P. F. McKay, C. R. Bouton, M.Ashford,

R. J.Shattock, A.Sabirsh, J. Controlled Release 2021, 330, 1250.

[186] A. A.De Angelis, S. J.Opella, Nat. Protoc. 2007, 2, 2332.

[187] L.Lin, X.Wang, Y. Guo, K.Ren, X.Li, L. Jing, X. Yue, Q.Zhang,

Z.Dai, RSC Adv. 2016, 6, 79811.

[188] T. N.Sut, E. R.Valle-González, B. K.Yoon, S.Park, J. A.Jackman,

N.-J.Cho, Appl. Mater. Today 2021, 22, 100947.

[189] a) K. Kolahdouzan, J. A. Jackman, B. K. Yoon, M. C. Kim,

M. S.Johal, N.-J. Cho, Langmuir 2017, 33, 5052; b) J. A.Jackman,

N.-J.Cho, Langmuir 2020, 36, 1387.

[190] E. J. Dufourc, Biochim. Biophys. Acta, Biomembr. 2021, 1863,

183478.

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

[191] a) N. E. Gabriel, M. F. Roberts, Biochemistry 1984, 23, 4011;

b) C. R. Sanders, G. C. Landis, Biochemistry 1995, 34, 4030;

c) R. R.Vold, R. S.Prosser, J. Magn. Reson. 1996, 113, 267.

[192] C. R.Sanders, J. P.Schwonek, Biochemistry 1992, 31, 8898.

[193] T. A. Harroun, M. Koslowsky, M.-P. Nieh, C.-F. de Lannoy,

V.Raghunathan, J.Katsaras, Langmuir 2005, 21, 5356.

[194] U. H.Dürr, R.Soong, A.Ramamoorthy, Prog. Nucl. Magn. Reson.

Spectrosc. 2013, 69, 1.

[195] a) N. L. Yamada, M. Hishida, N. Torikai, Phys. Rev. E 2009,

79, 032902; b) U. H. Dürr, M. Gildenberg, A. Ramamoorthy,

Chem. Rev. 2012, 112, 6054.

[196] S.Ghosh, A.Ray, N.Pramanik, Biophys. Chem. 2020, 265, 106429.

[197] C.-H.Yang, T.-L.Lin, U.-S.Jeng, Langmuir 2019, 35, 9483.

[198] L. van Dam, G. Karlsson, K. Edwards, Biochim. Biophys. Acta,

Biomembr. 2004, 1664, 241.

[199] H. S.Cho, J. L.Dominick, M. M. Spence, J. Phys. Chem. B 2010,

114, 9238.

[200] M. N.Triba, P. F.Devaux, D. E.Warschawski, Biophys. J. 2006, 91, 1357.

[201] M.Li, W. T.Heller, C.-H.Liu, C. Y.Gao, Y.Cai, Y.Hou, M.-P.Nieh,

Biochim. Biophys. Acta, Biomembr. 2020, 1862, 183315.

[202] J.Struppe, J. A.Whiles, R. R.Vold, Biophys. J. 2000, 78, 281.

[203] Y. Liu, M. Li, Y. Yang, Y. Xia, M.-P.Nieh, Biochim. Biophys. Acta,

Biomembr. 2014, 1838, 1871.

[204] Z.Shen, H.Ye, M.Kröger, S.Tang, Y.Li, Nanoscale 2019, 11, 15971.

[205] a) W.Aresh, Y. Liu, J.Sine, D. Thayer, A.Puri, Y.Huang, Y.Wang,

M.-P. Nieh, J. Biomed. Nanotechnol. 2016, 12, 1852; b) X.Wang,

L.Lin, R.Liu, M. Chen, B.Chen, B.He, B. He, X. Liang, W. Dai,

H.Zhang, Adv. Funct. Mater. 2017, 27, 1700406.

[206] K. Yasuhara, S. Miki, H. Nakazono, A. Ohta, J.-i. Kikuchi, Chem.

Commun. 2011, 47, 4691.

[207] K.Yasuhara, H.Hayashi, J.-i.Kikuchi, Chem. Lett. 2012, 41, 1223.

[208] L. Lin, X. Wang, X. Li, Y. Yang, X. Yue, Q. Zhang, Z. Dai,

Bioconjugate Chem. 2017, 28, 53.

[209] P.-W. Yang, T.-L. Lin, T.-Y. Lin, C.-H. Yang, Y. Hu, U.-S. Jeng,

Soft Matter 2013, 9, 11542.

[210] P.-W.Yang, T.-L.Lin, Y.Hu, U.-S.Jeng, Soft Matter 2014, 10, 2313.

[211] A. T. Rad, S. Malik, L. Yang, T. K. Oberoi-Khanuja, M.-P. Nieh,

R.Bahal, Nanoscale 2019, 11, 12517.

[212] C. Janich, C. Wölk, S. Taßler, S. Drescher, A. Meister,

G. Brezesinski, B. Dobner, A. Langner, Eur. J. Lipid Sci. Technol.

2014, 116, 1184.

[213] C.Janich, S.Taßler, A.Meister, G.Hause, J.Schäfer, U.Bakowsky,

G.Brezesinski, C.Wölk, Soft Matter 2016, 12, 5854.

[214] C. Wölk, C. Janich, A. Meister, S. Drescher, A. Langner,

G.Brezesinski, U.Bakowsky, Bioconjugate Chem. 2015, 26, 2461.

[215] C. Wölk, C. Janich, D. Pawlowska, S. Drescher, A. Meister,

G.Hause, B. Dobner, A.Langner, G.Brezesinski, ChemPhysChem

2015, 16, 2115.

[216] Q. Chen, G. Guan, F. Deng, D. Yang, P. Wu, S. Kang, R. Sun,

X.Wang, D.Zhou, W.Dai, Biomaterials 2020, 251, 120008.

[217] H. Zhang, in Liposomes: Methods and Protocols (Ed: V. Weissig),

Springer, Berlin 2017, p. 17.

[218] Y. Watanabe, K. Aramaki, Y. Kadomatsu, K. Tanaka, Y. Konno,

Chem. Lett. 2016, 45, 558.

[219] I. K. Mkam Tsengam, M. Omarova, E. G. Kelley, A. McCormick,

G. D.Bothun, S. R.Raghavan, V. T.John, J. Phys. Chem. B 2022,

126, 2208.

[220] K. Aramaki, C. Iwata, J. Mata, T. Maehara, D. Aburano,

Y.Sakanishi, K.Kitao, Phys. Chem. Chem. Phys. 2017, 19, 23802.

[221] K. Aramaki, K. Adachi, M. Maeda, J. Mata, J. Kamimoto-Kuroki,

D.Tsukamoto, Y.Konno, Materials 2020, 13, 3066.

[222] J. Kamimoto-Kuroki, M. Yamashita, K. Tanaka, Y. Kadomatsu,

D. Tsukamoto, K. Aramaki, K. Adachi, Y. Konno, Colloids Surf., A

2020, 606, 125418.

[223] S.Choi, B.Kang, S.Taguchi, H.Umakoshi, K.Kim, M. K.Kwak,

H.-S.Jung, Langmuir 2021, 37, 12255.

[224] S. Taguchi, K. Suga, K. Hayashi, Y. Okamoto, H.-S. Jung,

H.Nakamura, H.Umakoshi, Colloids Interfaces 2018, 2, 73.

[225] S. Choi, B. Kang, T. Shimanouchi, K. Kim, H. Jung, Micro Nano

Syst. Lett. 2021, 9, 7.

[226] H. A.Benson, Curr. Drug Delivery 2005, 2, 23.

[227] L.Barbosa-Barros, C.Barba, M.Cócera, L.Coderch, C.López-Iglesias,

A.deLa Maza, O.López, Int. J. Pharm. 2008, 352, 263.

[228] L.Barbosa-Barros, A.De La Maza, J.Estelrich, A.Linares, M.Feliz,

P.Walther, R.Pons, O.López, Langmuir 2008, 24, 5700.

[229] a) G.Rodriguez, L. Barbosa-Barros, L.Rubio, M. Cocera, A.Diez,

J.Estelrich, R. Pons, J.Caelles, A. D. l.Maza, O.Lopez, Langmuir

2009, 25, 10595; b) G.Rodriguez, L.Rubio, M.Cócera, J.Estelrich,

R.Pons, A.de laMaza, O.López, Langmuir 2010, 26, 10578.

[230] G.Rodríguez, M.Cócera, L.Rubio, C.Alonso, R. Pons, C.Sandt,

P.Dumas, C.López-Iglesias, A.de laMaza, O.López, Phys. Chem.

Chem. Phys. 2012, 14, 14523.

[231] L. Barbosa-Barros, G. Rodríguez, C. Barba, M.Cócera, L.Rubio,

J.Estelrich, C.López-Iglesias, A.de laMaza, O.López, Small 2012,

8, 807.

[232] a) G.Rodriguez, M.Cocera, L.Rubio, C.Lopez-Iglesias, R. Pons,

A. de la Maza, O. López, Mol. Pharmaceutics 2012, 9, 482;

b) G. Rodríguez, L. Rubio, C. Barba, C. López-Iglesias,

A.de laMaza, O.López, M.Cócera, Eur. Biophys. J. 2013, 42, 333.

[233] L.Rubio, C. Alonso, G.Rodríguez, L. Barbosa-Barros, L.Coderch,

A.De la Maza, J.Parra, O.López, Int. J. Pharm. 2010, 386, 108.

[234] L. Rubio, C.Alonso, G.Rodríguez, M.Cócera, C. López-Iglesias,

L.Coderch, A.De la Maza, J.Parra, O.López, Int. J. Pharm. 2013,

444, 60.

[235] L. Rubio, G.Rodriguez, C.Alonso, C.Lopez-Iglesias, M. Cocera,

L.Coderch, A.De la Maza, J.Parra, O.Lopez, Soft Matter 2011, 7,

8488.

[236] L.Rubio, C.Alonso, G. Rodríguez, M.Cócera, L. Barbosa-Barros,

L.Coderch, A.de laMaza, J.Parra, O.López, Eur. J. Pharm. Biop-

harm. 2014, 86, 212.

[237] a) G. Rodríguez, L. Barbosa-Barros, L. Rubio, M. Cócera,

F. Fernández-Campos, A.Calpena, E. Fernández, A.De La Maza,

O.López, J. Biomed. Nanotechnol. 2015, 11, 282; b) M. S.Kapoor,

S.GuhaSarkar, R.Banerjee, Ther. Delivery 2017, 8, 701.

[238] V. Moner, E. Fernández, A. Del Pozo, G. Rodríguez, M. Cócera,

A.de laMaza, O.López, Contact Dermatitis 2017, 77, 25.

[239] G. Anderluzzi, G. Lou, S. Woods, S. T. Schmidt, S. Gallorini,

M. Brazzoli, R. Johnson, C. W. Roberts, D. T. O’Hagan,

B. C.Baudner, J. Controlled Release 2022, 342, 388.

[240] W. Zhang, J. Sun, Y. Liu, M. Tao, X. Ai, X. Su, C. Cai, Y.Tang,

Z.Feng, X.Yan, Mol. Pharmaceutics 2014, 11, 3279.

[241] A. T. Rad, D. Hargrove, L. Daneshmandi, A. Ramsdell, X. Lu,

M.-P.Nieh, Adv. NanoBiomed Res. 2022, 2, 2100080.

[242] J.Gao, C.Xie, M. Zhang, X.Wei, Z.Yan, Y. Ren, M.Ying, W. Lu,

Nanoscale 2016, 8, 7209.

[243] P. E.Saw, M.Yu, M.Choi, E.Lee, S.Jon, O. C.Farokhzad, Bioma-

terials 2017, 123, 118.

[244] V. M. Stees, Z. Zhang, S. MacDonald, J. Crowe, K. K. Ewert,

B.Carragher, C. S.Potter, C. R.Saﬁnya, ACS Appl. Mater. Interfaces

2019, 12, 151.

[245] L.Lin, X.Liang, Y. Xu, Y. Yang, X.Li, Z. Dai, Bioconjugate Chem.

2017, 28, 2410.

[246] S.Hameed, P.Bhattarai, Z.Dai, Front. Chem. 2018, 6, 127.

[247] J. D. Finn, A. R. Smith, M. C. Patel, L. Shaw, M. R. Youniss,

J. van Heteren, T. Dirstine, C. Ciullo, R. Lescarbeau, J. Seitzer,

R. R. Shah, A. Shah, D. Ling, J. Growe, M. Pink, E. Rohde,

K. M. Wood, W. E. Salomon, W. F.Harrington, C.Dombrowski,

W. R.Strapps, Y.Chang, D. V.Morrissey, Cell Rep. 2018, 22, 2227.

Adv. Funct. Mater. 2022, 32, 2203669

www.afm-journal.dewww.advancedsciencenews.com

Abdul Rahim Ferhan, Ph.D. is a research fellow in the Translational Materials Innovation Group in

the School of Materials Science and Engineering at Nanyang Technological University. He received

his B.Eng. degree in Chemical and Biomolecular Engineering and Ph.D. degree in Biomedical

Engineering from the School of Chemical and Biomedical Engineering at Nanyang Technological

University. His research focuses on lipid membrane interactions and the incorporation of lipid

membrane models within novel nanoplasmonic sensing platforms for fundamental investigations

into biomacromolecular interaction processes.

Soohyun Park, Ph.D. is a research fellow in the School of Materials Science and Engineering,

Nanyang Technological University. She earned her B.S. in Biomaterials Engineering from Seoul

National University in 2016 and Ph.D. in Materials Science and Engineering from Nanyang

Technological University in 2021. Her research interests include lipid membrane-targeting antiviral

peptides, lipid nanoparticles, model lipid membranes, and biomaterials.

Joshua A. Jackman, Ph.D. is an assistant professor in the School of Chemical Engineering and

Director of the Translational Nanobioscience Research Center at Sungkyunkwan University. He

earned his B.S. degree in Chemistry from the University of Florida in 2010 and his Ph.D. degree in

Materials Science and Engineering from Nanyang Technological University in 2015, and completed

postdoctoral studies at the Stanford University School of Medicine. His research focuses on lipid

membrane biotechnology, membrane biophysics, and the development of membrane-targeting

strategies for infectious disease and cancer applications.

Nam-Joon Cho, Ph.D. is the MRS-Singapore Chair Professor in the School of Materials Science

and Engineering at Nanyang Technological University. He earned his B.S. degree in Civil

Engineering from the University of California, Berkeley in 1996 and his M.S. degree in Materials

Science and Engineering and Ph.D. degree in Chemical Engineering from Stanford University

in 2003 and 2007, respectively, and completed postdoctoral studies at the Stanford University

School of Medicine. His research focuses on biomaterial strategies to develop new classes of

anti-infective drugs, to engineer advanced lipid nanoparticle technology, and to create natural

materials to replace plastics.

Adv. Funct. Mater. 2022, 32, 2203669

A preview of this full-text is provided by Wiley.

Learn more

Content available from Advanced Functional Materials

This content is subject to copyright. Terms and conditions apply.

Composite Nanoarchitectonics Towards Method for Everything in Materials Science

Article

Full-text available

Apr 2024
J INORG ORGANOMET P

Katsuhiko Ariga

The characteristic feature of a biofunctional system is that components with various functions work together. These multi-components are not simply mixed together, but are rationally arranged. The fundamental technologies to do this in an artificial system include the synthetic chemistry of the substances that make the component unit, the science and techniques for assembling them, and the technology for analyzing their nanoostructures. A new concept, nanoarchitectonics, can play this role. Nanoarchitectonics is a post-nanotechnology concept that involves building functional materials that reflect the nanostructures. In particular, the approach of combining and building multiple types of components to create composite materials is an area where nanoarchitectonics can be a powerful tool. This review summarizes such examples and related composite studies. In particular, examples are presented in the areas of catalyst & photocatalyst, energy, sensing & environment, bio & medical, and various other functions and applications to illustrate the potential for a wide range of applications. In order to show the various stages of development, the examples are not only state-of-the-art, but also include those that are successful developments of existing research. Finally, a summary of the examples and a brief discussion of future challenges in nanoarchitectonics will be given. Nanoarchitectonics is applicable to all materials and aims to establish the ultimate methodology of materials science.

A careful look at lipid nanoparticle characterization: analysis of benchmark formulations for encapsulation of RNA cargo size gradient

Article

Full-text available

Jan 2024

With the recent success of lipid nanoparticle (LNP) based SARS-CoV-2 mRNA vaccines, the potential for RNA therapeutics has gained widespread attention. LNPs are promising non-viral delivery vectors to protect and deliver delicate RNA therapeutics, which are ineffective and susceptible to degradation alone. While food and drug administration (FDA) approved formulations have shown significant promise, benchmark lipid formulations still require optimization and improvement. In addition, the translatability of these formulations for several different RNA cargo sizes has not been compared under the same conditions. Herein we analyze “gold standard” lipid formulations for encapsulation efficiency of various non-specific RNA cargo lengths representing antisense oligonucleotides (ASO), small interfering RNA (siRNA), RNA aptamers, and messenger RNA (mRNA), with lengths of 10 bases, 21 base pairs, 96 bases, 996 bases, and 1929 bases, respectively. We evaluate encapsulation efficiency as the percentage of input RNA encapsulated in the final LNP product (EEinput%), which shows discrepancy with the traditional calculation of encapsulation efficiency (EE%). EEinput% is shown to be < 50% for all formulations tested, when EE% is consistently > 85%. We also compared formulations for LNP size (Z-average) and polydispersity index (PDI). LNP size does not appear to be strongly influenced by cargo size, which is a counterintuitive finding. Thoughtful characterization of LNPs, in parallel with consideration of in vitro or in vivo behavior, will guide design and optimization for better understanding and improvement of future RNA therapeutics.

Elucidating Structural Configuration of Lipid Assemblies for mRNA Delivery Systems

Article

Apr 2024
ACS NANO

Surface engineering of lipid nanoparticles: targeted nucleic acid delivery and beyond

Article

Full-text available

Oct 2023

Harnessing surface engineering strategies to functionalize nucleic acid-lipid nanoparticles (LNPs) for improved performance has been a hot research topic since the approval of the first siRNA drug, patisiran, and two mRNA-based COVID-19 vaccines, BNT162b2 and mRNA-1273. Currently, efforts have been mainly made to construct targeted LNPs for organ- or cell-type-specific delivery of nucleic acid drugs by conjugation with various types of ligands. In this review, we describe the surface engineering strategies for nucleic acid-LNPs, considering ligand types, conjugation chemistries, and incorporation methods. We then outline the general purification and characterization techniques that are frequently used following the engineering step and emphasize the specific techniques for certain types of ligands. Next, we comprehensively summarize the currently accessible organs and cell types, as well as the other applications of the engineered LNPs. Finally, we provide considerations for formulating targeted LNPs and discuss the challenges of successfully translating the “proof of concept” from the laboratory into the clinic. We believe that addressing these challenges could accelerate the development of surface-engineered LNPs for targeted nucleic acid delivery and beyond.

2D Materials Nanoarchitectonics for 3D Structures/Functions

Article

Full-text available

Feb 2024

Katsuhiko Ariga

It has become clear that superior material functions are derived from precisely controlled nanostructures. This has been greatly accelerated by the development of nanotechnology. The next step is to assemble materials with knowledge of their nano-level structures. This task is assigned to the post-nanotechnology concept of nanoarchitectonics. However, nanoarchitectonics, which creates intricate three-dimensional functional structures, is not always easy. Two-dimensional nanoarchitectonics based on reactions and arrangements at the surface may be an easier target to tackle. A better methodology would be to define a two-dimensional structure and then develop it into a three-dimensional structure and function. According to these backgrounds, this review paper is organized as follows. The introduction is followed by a summary of the three issues; (i) 2D to 3D dynamic structure control: liquid crystal commanded by the surface, (ii) 2D to 3D rational construction: a metal–organic framework (MOF) and a covalent organic framework (COF); (iii) 2D to 3D functional amplification: cells regulated by the surface. In addition, this review summarizes the important aspects of the ultimate three-dimensional nanoarchitectonics as a perspective. The goal of this paper is to establish an integrated concept of functional material creation by reconsidering various reported cases from the viewpoint of nanoarchitectonics, where nanoarchitectonics can be regarded as a method for everything in materials science.

Confined Space Nanoarchitectonics for Dynamic Functions and Molecular Machines

Article

Full-text available

Feb 2024

Katsuhiko Ariga

Nanotechnology has advanced the techniques for elucidating phenomena at the atomic, molecular, and nano-level. As a post nanotechnology concept, nanoarchitectonics has emerged to create functional materials from unit structures. Consider the material function when nanoarchitectonics enables the design of materials whose internal structure is controlled at the nanometer level. Material function is determined by two elements. These are the functional unit that forms the core of the function and the environment (matrix) that surrounds it. This review paper discusses the nanoarchitectonics of confined space, which is a field for controlling functional materials and molecular machines. The first few sections introduce some of the various dynamic functions in confined spaces, considering molecular space, materials space, and biospace. In the latter two sections, examples of research on the behavior of molecular machines, such as molecular motors, in confined spaces are discussed. In particular, surface space and internal nanospace are taken up as typical examples of confined space. What these examples show is that not only the central functional unit, but also the surrounding spatial configuration is necessary for higher functional expression. Nanoarchitectonics will play important roles in the architecture of such a total system.

Lipid‐Based Nanoparticles

Chapter

Feb 2024

Modulating Lipid Nanoparticles with Histidinamide‐Conjugated Cholesterol for Improved Intracellular Delivery of mRNA

Article

Full-text available

Feb 2024

Recently, mRNA‐based therapeutics, including vaccines, have gained significant attention in the field of gene therapy for treating various diseases. Among the various mRNA delivery vehicles, lipid nanoparticles (LNPs) have emerged as promising vehicles for packaging and delivering mRNA with low immunogenicity. However, while mRNA delivery has several advantages, the delivery efficiency and stability of LNPs remain challenging for mRNA therapy. In this study, an ionizable helper cholesterol analog, 3β[L‐histidinamide‐carbamoyl] cholesterol (Hchol) lipid is developed and incorporated into LNPs instead of cholesterol to enhance the LNP potency. The pKa values of the Hchol‐LNPs are ≈6.03 and 6.61 in MC3‐ and SM102‐based lipid formulations. Notably, the Hchol‐LNPs significantly improve the delivery efficiency by enhancing the endosomal escape of mRNA. Additionally, the Hchol‐LNPs are more effective in a red blood cell hemolysis at pH 5.5, indicating a synergistic effect of the protonated imidazole groups of Hchol and cholesterol on endosomal membrane destabilization. Furthermore, mRNA delivery is substantially enhanced in mice treated with Hchol‐LNPs. Importantly, LNP‐encapsulated SARS‐CoV‐2 spike mRNA vaccinations induce potent antigen‐specific antibodies against SARS‐CoV‐2. Overall, incorporating Hchol into LNP formulations enables efficient endosomal escape and stability, leading to an mRNA delivery vehicle with a higher delivery efficiency.

Lipid Nanoparticle-Based Delivery System─A Competing Place for mRNA Vaccines

Article

Full-text available

Jan 2024

mRNA, as one of the foci of biomedical research in the past decade, has become a candidate vaccine solution for various infectious diseases and tumors and for regenerative medicine and immunotherapy due to its high efficiency, safety, and effectiveness. A stable and effective delivery system is needed to protect mRNAs from nuclease degradation while also enhancing immunogenicity. The success of mRNA lipid nanoparticles in treating COVID-19, to a certain extent, marks a milestone for mRNA vaccines and also promotes further research on mRNA delivery systems. Here, we explore mRNA vaccine delivery systems, especially lipid nanoparticles (LNPs), considering the current research status, prospects, and challenges of lipid nanoparticles, and explore other mRNA delivery systems.

Construction Strategy of Functionalized Liposomes and Multidimensional Application

Article

Full-text available

Jan 2024
SMALL

Liposomes are widely used in the biological field due to their good biocompatibility and surface modification properties. With the development of biochemistry and material science, many liposome structures and their surface functional components have been modified and optimized one by one, pushing the liposome platform from traditional to functionalized and intelligent, which will better satisfy and expand the needs of scientific research. However, a main limiting factor effecting the efficiency of liposomes is the complicated environmental conditions in the living body. Currently, in order to overcome the above problem, functionalized liposomes have become a very promising strategy. In this paper, binding strategies of liposomes with four main functional elements, namely nucleic acids, antibodies, peptides, and stimuli‐responsive motif have been summarized for the first time. In addition, based on the construction characteristics of functionalized liposomes, such as drug‐carrying, targeting, long‐circulating, and stimulus‐responsive properties, a comprehensive overview of their features and respective research progress are presented. Finally, the paper critically presents the limitations of these functionalized liposomes in the current applications and also prospectively suggests the future development directions, aiming to accelerate realization of their industrialization.

Transformation of Lipid Vesicles into Micelles by Adding Nonionic Surfactants: Elucidating the Structural Pathway and the Intermediate Structures

Article

Full-text available

Mar 2022

The phospholipid lecithin (L) and the nonionic surfactant Tween 80 (T) are used together in various contexts, including in drug delivery and oil spill remediation. There is hence a need to elucidate the nanostructures in LT mixtures, which is the focus of this paper. We study these mixtures using cryogenic transmission electron microscopy (cryo-TEM), coupled with dynamic light scattering and small-angle neutron scattering. As the concentration of Tween 80 is increased, the vesicles formed by lecithin are transformed into spherical micelles. We identify bicelles (i.e., disc-like micelles) as well as cylindrical micelles as the key stable nanostructures formed at intermediate L/T ratios. The bicelles have diameters ∼13-26 nm, and the bicelle size decreases as the Tween 80 content increases. We propose that the lecithin lipids form the body of the discs, while the Tween 80 surfactants occupy the rims. This hypothesis is consistent with geometric arguments because lecithin is double-tailed and favors minimal curvature, whereas the single-tailed Tween 80 molecules prefer curved interfaces. In the case of cylindrical micelles, cryo-TEM reveals that the micelles are short (length < 22 nm) and flexible. We are able to directly visualize the microstructure of the aggregates formed by lecithin-Tween 80 mixtures, thereby enhancing the understanding of morphological changes in the lecithin-Tween 80 system.

mRNA Lipoplexes with Cationic and Ionizable α-Amino-lipophosphonates: Membrane Fusion, Transfection, mRNA Translation and Conformation

Article

Full-text available

Mar 2022

Cationic liposomes are attractive carriers for mRNA delivery. Here, mRNA lipoplexes (LX) were prepared with the cationic lipids α-aminolipophosphonate (3b) or imidazolium lipophosphoramidate (2) associated with various α-aminolipophosphonates co-lipids comprising protonable groups (imidazole or pyridine) and DOPE. Physicochemical parameters of liposomes and their membrane fusion activity were measured. LXs comprising either 3b- or 2- allowed transfection of ~25% and 40% of dendritic cells with low cytotoxicity, respectively; the efficiency increased up to 80% when 2 was combined with the imidazole-based co-lipid 1. The transfections were high with 3b/1, 3b/DOPE, 2/1 and 2/DOPE LXs. We observed that the transfection level was not well correlated with the acid-mediated membrane fusion activity of liposomes supposed to destabilize endosomes. The mRNA release from LXs and its translation capacity after release were studied for the most efficient LXs. The results showed that the more mRNA was condensed, the poorer the translation efficiency after release was. In contrast to DNA, circular dichroism performed on mRNA complexed with 2/DOPE revealed the presence of denatured mRNA in LXs explaining this lack of translation efficiency. This is an important parameter that should be stressed for the preparation of mRNA LXs with a conserved mRNA translation activity.

Current Status and Future Perspectives on MRNA Drug Manufacturing

Article

Full-text available

Mar 2022

The coronavirus disease of 2019 (COVID-19) pandemic launched an unprecedented global effort to rapidly develop vaccines to stem the spread of the novel severe acute respiratory syndrome coronavirus (SARS-CoV-2). Messenger ribonucleic acid (mRNA) vaccines were developed quickly by companies that were actively developing mRNA therapeutics and vaccines for other indications, leading to two mRNA vaccines being not only the first SARS-CoV-2 vaccines to be approved for emergency use but also the first mRNA drugs to gain emergency use authorization and to eventually gain full approval. This was possible partly because mRNA sequences can be altered to encode nearly any protein without significantly altering its chemical properties, allowing the drug substance to be a modular component of the drug product. Lipid nanoparticle (LNP) technology required to protect the ribonucleic acid (RNA) and mediate delivery into the cytoplasm of cells is likewise modular, as are technologies and infrastructure required to encapsulate the RNA into the LNP. This enabled the rapid adaptation of the technology to a new target. Upon the coattails of the clinical success of mRNA vaccines, this modularity will pave the way for future RNA medicines for cancer, gene therapy, and RNA engineered cell therapies. In this review, trends in the publication records and clinical trial registrations are tallied to show the sharp intensification in preclinical and clinical research for RNA medicines. Demand for the manufacturing of both the RNA drug substance (DS) and the LNP drug product (DP) has already been strained, causing shortages of the vaccine, and the rise in development and translation of other mRNA drugs in the coming years will exacerbate this strain. To estimate demand for DP manufacturing, the dosing requirements for the preclinical and clinical studies of the two approved mRNA vaccines were examined. To understand the current state of mRNA-LNP production, current methods and technologies are reviewed, as are current and announced global capacities for commercial manufacturing. Finally, a vision is rationalized for how emerging technologies such as self-amplifying mRNA, microfluidic production, and trends toward integrated and distributed manufacturing will shape the future of RNA manufacturing and unlock the potential for an RNA medicine revolution.

Successful administration of second dose of BNT162b2 COVID‐19 vaccine in two patients with potential anaphylaxis to first dose

Article

Full-text available

Dec 2021
ALLERGY

Combating Complement's Deleterious Effects on Nanomedicine by Conjugating Complement Regulatory Proteins to Nanoparticles

Article

Full-text available

Jan 2022
ADV MATER

Complement opsonization is among the biggest challenges facing nanomedicine. Nearly instantly after injection into blood, nanoparticles are opsonized by the complement protein C3, leading to clearance by phagocytes, fouling of targeting moieties, and release of anaphylatoxins. While surface polymers such as polyethylene glycol (PEG) partially decrease complement opsonization, most nanoparticles still suffer from extensive complement opsonization, especially when linked to targeting moieties. To ameliorate the deleterious effects of complement, we have conjugated two of mammals’ natural regulators of complement activation (RCAs), Factors H & I, to the surface of nanoparticles. In vitro, Factor H or I conjugation to PEG-coated nanoparticles decrease their C3 opsonization, and markedly reduce nanoparticle uptake by phagocytes. In an in vivo mouse model of sepsis-induced lung injury, Factor I-conjugation abrogates nanoparticle uptake by intravascular phagocytes in the lungs, allowing the blood concentration of the nanoparticle to remain elevated much longer. For nanoparticles targeted to the lung's endothelium by conjugation to anti-ICAM antibodies, Factor I-conjugation shifts the cell type distribution away from phagocytes and towards endothelial cells. Finally, Factor I-conjugation abrogates the severe anaphylactoid responses common to many nanoparticles, preventing systemic capillary leak and preserving blood flow to visceral organs and the brain. Thus, conjugation of RCAs, like Factor I, to nanoparticles is likely to help in nanomedicine's long battle against complement, improving several key parameters critical for clinical success. This article is protected by copyright. All rights reserved

tRNA therapeutics burst onto startup scene

Article

Feb 2022

Elie Dolgin

Companies advance tRNA therapeutics to overcome mutant stoppages in protein synthesis shared by thousands of genetic diseases and cancers.

Microfluidic technologies and devices for lipid nanoparticle-based RNA delivery

Article

Feb 2022
J CONTROL RELEASE

In 2021, mRNA vaccines against COVID-19 were approved by the Food and Drug Administration. mRNA vaccines are important for preventing severe COVID-19 and returning to normal life. The development of RNA-delivery technology, including mRNA vaccines, has been investigated worldwide for ~30 years. Lipid nanoparticles (LNPs) are a breakthrough technology that stably delivers RNA to target organs, and RNA-loaded LNP-based nanomedicines have been studied for the development of vaccines and nanomedicines for RNA-, gene-, and cell-based therapies. Recently, microfluidic devices and technologies have attracted attention for the production of LNPs, particularly RNA-loaded LNPs. Microfluidics provides many advantages for RNA-loaded LNP production, including precise LNP size controllability, high reproducibility, high-throughput optimization of LNP formulation, and continuous LNP-production processes. In this review, we summarize microfluidic-based RNA-loaded LNP production and its applications in RNA-based therapy and genome editing.

On the mechanism of tissue-specific mRNA delivery by selective organ targeting nanoparticles

Article

Dec 2021

Significance Liver accumulation represents a significant barrier in the development of therapeutically efficacious nanoparticle drug delivery systems. Using a series of lipid nanoparticles with distinct organ-targeting properties, we provide evidence for a plausible mechanism of action for nanoparticle delivery to non-liver tissues. Following intravenous injection, specific proteins in the blood are recruited to the nanoparticle’s surface based on its molecular composition and they endow it with a unique biological identity that governs its ultimate fate in the body. An innovative paradigm emerges from this mechanistic understanding of nanoparticle delivery—endogenous targeting—wherein the molecular composition of a nanoparticle is rationally engineered to interact with specific proteins in the blood to overcome liver accumulation and to target specific organs.

Safety of COVID-19 vaccination in patients with polyethylene glycol allergy: A case series

Article

Dec 2021

The role of nanoparticle format and route of administration on self-amplifying mRNA vaccine potency

Article

Dec 2021
J CONTROL RELEASE

The efficacy of RNA-based vaccines has been recently demonstrated, leading to the use of mRNA-based COVID-19 vaccines. The application of self-amplifying mRNA within these formulations may offer further enhancement to these vaccines, as self-amplifying mRNA replicons enable longer expression kinetics and more potent immune responses compared to non-amplifying mRNAs. To investigate the impact of administration route on RNA-vaccine potency, we investigated the immunogenicity of a self-amplifying mRNA encoding the rabies virus glycoprotein encapsulated in different nanoparticle platforms (solid lipid nanoparticles (SLNs), polymeric nanoparticles (PNPs) and lipid nanoparticles (LNPs)). These were administered via three different routes: intramuscular, intradermal and intranasal. Our studies in a mouse model show that the immunogenicity of our 4 different saRNA vaccine formulations after intramuscular or intradermal administration was initially comparable; however, ionizable LNPs gave higher long-term IgG responses. The clearance of all 4 of the nanoparticle formulations from the intramuscular or intradermal administration site was similar. In contrast, immune responses generated after intranasal was low and coupled with rapid clearance for the administration site, irrespective of the formulation. These results demonstrate that both the administration route and delivery system format dictate self-amplifying RNA vaccine efficacy.

Lipid Nanoparticle Technologies for Nucleic Acid Delivery: A Nanoarchitectonics Perspective

Abstract and Figures

Recommended publications

Lab-scale siRNA and mRNA LNP manufacturing by various microfluidic mixing techniques -an evaluation...

Evaluating how cationic lipid affects mRNA-LNP physical properties and biodistribution

Single-Particle Spectroscopic Chromatography Reveals Heterogeneous RNA Loading and Size Correlations...

Negatively charged phospholipids doped liposome delivery system for mRNA with high transfection effi...