Lipid nanoparticles are developed using a microfluidic mixing platform approach for mRNA and DNA encapsulation.
Lipid-based drug carriers have been used for clinically and commercially available delivery systems due to their small size, biocompatibility, and high encapsulation efficiency. Use of lipid nanoparticles (LNPs) to encapsulate nucleic acids is advantageous to protect the RNA or DNA from degradation, while also promoting cellular uptake. LNPs often contain multiple lipid components including an ionizable lipid, helper lipid, cholesterol, and polyethylene glycol (PEG) conjugated lipid. LNPs can readily encapsulate nucleic acids due to the ionizable lipid presence, which at low pH is cationic and allows for complexation with negatively charged RNA or DNA. Here LNPs are formed by encapsulating messenger RNA (mRNA) or plasmid DNA (pDNA) using rapid mixing of the lipid components in an organic phase and the nucleic acid component in an aqueous phase. This mixing is performed using a precise microfluidic mixing platform, allowing for nanoparticle self-assembly while maintaining laminar flow. The hydrodynamic size and polydispersity are measured using dynamic light scattering (DLS). The effective surface charge on the LNP is determined by measuring the zeta potential. The encapsulation efficiency is characterized using a fluorescent dye to quantify entrapped nucleic acid. Representative results demonstrate the reproducibility of this method and the influence that different formulation and process parameters have on the developed LNPs.
Drug carriers are used to protect and deliver a therapeutic with typical favorable properties including low cytotoxicity, increased bioavailability, and improved stability1,2,3. Polymeric nanoparticles, micelles, and lipid-based particles have previously been explored for nucleic acid encapsulation and delivery4,5,6,7. Lipids have been used in different types of nanocarrier systems, including liposomes, and lipid nanoparticles, as they are biocompatible with high stability8. LNPs can readily encapsulate nucleic acids for gene delivery9,10. They protect the nucleic acid from degradation by serum proteases during systemic circulation11 and can improve delivery to specific sites, as the surface topography and physical properties of LNPs influence their biodistribution12. LNPs also improve tissue penetration and cellular uptake9. Previous studies have demonstrated the success of siRNA encapsulation within an LNP13, including the first commercially available LNP containing siRNA therapeutic for the treatment polyneuropathy of hereditary transthyretin-mediated amyloidosis14 treatment that was approved by United States Food and Drug Administration (FDA) and European Medicines Agency in 2018. More recently, LNPs are being studied for the delivery of larger nucleic acid moieties, namely mRNA and DNA9. As of 2018, there were ~ 22 lipid-based nucleic acid delivery systems undergoing clinical trials14. Additionally, mRNA containing LNPs are currently leading candidates and have been employed for a COVID-19 vaccine15,16. The potential success for these non-viral gene therapies requires forming small (~100 nm), stable, and uniform particles with high encapsulation of the nucleic acid.
Use of an ionizable lipid as a main component in the LNP formulation has shown advantages for complexation, encapsulation, and delivery effciciency14. Ionizable lipids typically have an acid dissociation constant (pKa) < 7; for example, dilinoleylmethyl-4-dimethylaminobutyrate (D-Lin-MC3-DMA), the ionizable lipid used in the FDA approved LNP formulation, has a pKa of 6.4417. At low pH, the amine groups on the ionizable lipid become protonated and positively charged, allowing for the assembly with negatively charged phosphate groups on mRNA and DNA. The ratio of amine, "N", groups to phosphate, "P", groups is used to optimize the assembly. The N/P ratio is dependent on the lipids and nucleic acids used, which varies depending on the formulation18. After formation, the pH can be adjusted to a neutral or physiological pH to allow for therapeutic administration. At these pH values, the ionizable lipid is also deprotonated which imparts neutral surface charge to the LNP.
The ionizable lipid also aids in endosomal escape19,20. LNPs undergo endocytosis during cellular uptake and must be released from the endosome in order to deliver the mRNA cargo into the cell cytoplasm or DNA cargo to the nucleus21. Inside the endosome is typically a more acidic environment than the extracellular medium, which renders the ionizable lipid positively charged22,23. The positively charged ionizable lipid can interact with negative charges on the endosomal lipid membrane, which can cause destabilization of the endosome allowing for the release of the LNP and nucleic acid. Different ionizable lipids are currently being studied for improving efficacy of both LNP distribution, as well as endosomal escape14.
Other typical components of an LNP include helper lipids, such as a phosphatidylcholine (PC) or phosphoethanolamine (PE) lipid. 1,2-Dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE), 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC), and 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) are commonly used helper lipids24,25. DOPE has been shown to form an inverted hexagonal II (HII) phase and enhance transfection by membrane fusion26, while DSPC has been thought to stabilize LNPs with its cylindrical geometry27. Cholesterol is also incorporated in the formulation in order to increase membrane rigidity, subsequently aiding in the stability of the LNP. Finally, lipid-conjugated polyethylene glycol (PEG) is included in the formulation to provide the necessary steric barrier to aid in particle self-assembly27. PEG also improves the storage stability of LNPs by preventing aggregation. Furthermore, PEG is often used as a stealth component and can increase the circulation time for the LNPs. However, this attribute can also pose challenges for recruitment of LNPs to hepatocytes through an endogenous targeting mechanism driven by apolipoprotein E (ApoE)28. Thus, studies have investigated the acyl chain length for diffusion of PEG from the LNP, finding that short lengths (C8-14) dissociate from the LNP and are more amenable to ApoE recruitment compared to longer acyl lengths28. Further, the degree of saturation of the lipid tail that PEG is conjugated to has been shown to influence the tissue distribution of LNPs29. Recently, Tween 20, which is a commonly used surfactant in biological drug product formulations and has a long unsaturated lipid tail, was shown to have high transfection in draining lymph nodes compared to PEG-DSPE, which largely transfected the muscle at the injection site29. This parameter can be optimized to achieve the desired LNP biodistribution.
Conventional methods of forming LNPs include the thin-film hydration method and ethanol-injection method27. While these are readily available techniques, they are also labor intensive, can result in low encapsulation efficiency, and are challenging to scale up27. Advancements in mixing techniques have resulted in methods more amenable to scale up, while developing more uniform particles27. These methods include T-junction mixing, staggered herringbone mixing, and microfluidic hydrodynamic focusing27. Each method has a unique structure, but all allow for rapid mixing of an aqueous phase containing the nucleic acid with an organic phase containing the lipid components, resulting in high encapsulation of the nucleic acid27. In this protocol, rapid and controlled mixing through a microfluidic cartridge is utilized, which employs the staggered herringbone mixing design. This protocol outlines the preparation, assembly, and characterization of nucleic acid containing LNPs.
A schematic of the overall process is provided in Figure 1.
1. Preparation of buffers
NOTE: Sterile filtering of the buffers is highly suggested here to remove any particulates which may impact the nucleic acid and LNP quality.
2. Preparation of lipid mix
3. Preparation of nucleic acid solution
NOTE: Preparation and handling of nucleic acid solutions is to be performed in a sterile and RNase-free environment wherever possible. Work in a biosafety cabinet whenever possible with the nucleic acid.
4. Priming the microfluidic channels
NOTE: This protocol is adapted from the instrument manufacturer's guidelines.
5. LNP formation
NOTE: This protocol is adapted from the instrument manufacturer's guidelines.
6. Buffer exchange
NOTE: Protocol for using ultra-centrifuge filters is provided. While this method results in a more time efficient exchange of buffers, dialysis may be substituted here.
7. Measure encapsulation efficiency
8. Concentration adjustments
9. Measure LNP hydrodynamic size and polydispersity
10. Measure LNP zeta potential
Multiple batches of LNPs with the same lipid formulation and N/P ratio of 6 were developed on separate days to demonstrate reproducibility of the technique. Batch 1 and 2 resulted in overlapping size distributions with similar polydispersity (Figure 2A) No significant difference was observed in the size or encapsulation efficiency between the two different batches (Figure 2B). The encapsulation efficiency was high for each batch (>98.5%) and the sizes were similar with a 77 nm LNP diameter. The particles were uniform with an average polydispersity index (PDI) of 0.15 for batch 1 and 0.18 for batch 2.
Changes in formulation parameters showed some small, yet statistically significant differences with respect to the N/P ratio, ionizable lipid used, and nucleic acid encapsulated. While differences are discussed, it is important to note that all LNPs formed resulted in encapsulation greater than 80%, with most formulations greater than 95%, and particle sizes less than 110 nm, making all formulations developed here desirable for gene delivery. First, ionizable lipid A was used to develop LNPs at an N/P of 10 and 36. Decreasing the N/P ratio resulted in a 4% decrease in encapsulation efficiency and an increase in the hydrodynamic diameter of the LNPs from 98 nm at N/P = 36 to 109 nm at N/P = 10 (Figure 3A). Comparing LNPs with ionizable lipid A to a different ionizable lipid B and maintaining N/P of 36 resulted in a significant change in encapsulation efficiency, where 100% of pDNA was encapsulated with LNPs formed using ionizable lipid A and 81% of pDNA was encapsulated with LNPs formed using ionizable lipid B (Figure 3B). Ionizable lipid B LNPs also resulted in slightly smaller particles with a hydrodynamic diameter of 95 nm. Finally, LNPs were formed using ionizable lipid A with both mRNA and pDNA. LNPs encapsulating pDNA resulted in larger particles with a 119 nm diameter compared with mRNA LNPs with a 91 nm diameter (Figure 3C). Both pDNA and mRNA LNPs resulted in similar encapsulation efficiency at ~91-94%.
Lastly, changes in the flow rate process parameter did not impact the LNPs developed at the flow rates tested here. At both 4 mL/min and 12 mL/min, LNPs were developed and characterized to have encapsulated 96% of pDNA and have a 110 nm diameter (Figure 4). All LNPs regardless of process parameter or formulation parameter resulted in charge neutral zeta potential measurements.
Figure 1: LNP development and characterization workflow. First, lipid mix and nucleic acid solutions are made (1 and 2). The lipid mix contains the ionizable lipid, helper lipid, cholesterol, and PEG in ethanol, while the nucleic acid solution contains either mRNA or DNA in buffer. Solutions are mixed using a microfluidic cartridge (3), which forms LNPs (4). Next, a buffer exchange is required to remove the ethanol and increase the solution pH to neutral (5). Characterization of LNPs is performed to determine encapsulation efficiency and particle size, polydispersity, and zeta potential using a fluorescence microplate assay and zetasizer, respectively (6 and 7). Please click here to view a larger version of this figure.
Figure 2: Batch to batch reproducibility of LNPs formed on separate days. (A) Size distributions for batch 1 vs. batch 2 (B) Encapsulation efficiency (%) and hydrodynamic diameter (nm) for each batch with mRNA and N/P = 6. Error bars note standard deviation. Statistical analysis using two-way ANOVA with α = 0.05 shows no significance. Please click here to view a larger version of this figure.
Figure 3: Variations of formulation parameters. (A) LNPs formed at N/P = 10 and 36 both using ionizable lipid A with pDNA. (B) LNPs formed with ionizable lipid A and ionizable lipid B both at N/P = 36 with pDNA. (C) LNPs formed with either mRNA or pDNA both using ionizable lipid C at N/P = 6. Error bars note standard deviation. Statistical analysis was performed using two-way ANOVA with α = 0.05; *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. Please click here to view a larger version of this figure.
Figure 4: Variations of process parameters. LNPs formed at a flow rate of 4 and 12 mL/min using ionizable lipid A with pDNA at N/P = 10. Error bars note standard deviation. Statistical analysis using two-way ANOVA with α = 0.05 shows no significance. Please click here to view a larger version of this figure.
Lipid | Molar Ratio | Stock lipid concentration (mM) | Concentration in ethanol for lipid mix (mM) | Volume (µL) |
Ionizable lipid | 50 | 38.9 | 5 | 68.5 |
Helper lipid | 10 | 10 | 1 | 53.3 |
Cholesterol | 39 | 20 | 3.9 | 103.9 |
C-14 PEG | 1 | 1 | 0.1 | 53.3 |
EtOH | 254 | |||
Total | 533 |
Table 1: Example lipid mix to prepare 1 mL of LNPs. The lipid stock concentrations in ethanol provided have been shown to allow for the lipids to solubilize in ethanol, but other stock concentrations may be utilized and will not affect the outcome as long as the lipid is solubilized. Example concentrations of lipids in ethanol for microfluidic mixing are also provided. These concentrations are based on the molar ratio, which can be varied based on the desired LNP preparation.
Priming | Formulation | |
Volume (mL) | 2 | 1.37 |
Flow Rate Ratio (Aqueous:EtOH) | 3:1 | 3:1 |
Total Flow Rate (mL/min) | 12 | 4 |
Left Syringe Size (mL) | 3 | 3 |
Right Syringe Size (mL) | 1 | 1 |
Start Waste Volume (mL) | 0.35 | 0.25 |
End Waste Volume (mL) | 0.05 | 0.05 |
Table 2: Microfluidic Mixing Benchtop Instrument Software Priming and LNP Formulation Example Parameters
Reproducibility, speed, and low volume screening are significant advantages of using microfluidic mixing to form LNPs compared to other existing methods (e.g., lipid film hydration and ethanol injection). We have demonstrated the reproducibility of this method with no impact on encapsulation efficiency or particle size observed with different LNP batches. This is an essential criterion for any therapeutic, including LNPs, to become clinically available.
The technique described here employs staggered herringbone microfluidic mixing, which results in LNP formation on the time scale of only a few minutes. This mixing uses chaotic advection which is advantageous for mixing control and shortened time27. This mixer enables the aqueous and organic phases to effectively wrap around each other27. Using the staggered herringbone mixing, previous studies have shown that the particles form at the smallest thermodynamically stable size32, which means that the composition tends to influence the size and polydispersity of the LNPs27,32,33. This was observed in the representative results, where the N/P ratio, ionizable lipid used, and nucleic acid encapsulated were the impacting factors on changes in encapsulation efficiency and particle size. Operating parameters, such as flow rate and mixing ratio can also influence the size above a certain threshold, where afterwards the particle size is at its smallest stable size27,33. No change in encapsulation efficiency or particle size was observed when a flow rate of 4 mL/min vs. 12 mL/min was used. Thus, likely both flow rates are above the threshold that would impact the LNP outcome. The example experiment, and results described above, used lipid A and pDNA. It is possible that different ionizable lipids and nucleic acids could have more influence on LNP characteristics with respect to flow rate. Other types of microfluidic mixing include the T-junction, which uses turbulent flow and the microfluidic hydrodynamic focusing method that is based on convective-diffusive mixing27. Compared to these other types of microfluidic mixing techniques for LNP development, the staggered herringbone mixing enables the combination of three important criteria: rapid mixing, minimizing batch to batch variability, and is commercially available27. All three of the microfluidic mixing methods do allow for higher encapsulation efficiency and controlled size compared to conventional lipid film hydration or ethanol injection methods27.
Finally, the ability to produce low volumes of various LNP formulations at the research & development stage is a significant advantage. One challenge of developing LNPs is the number of variables that can be tested and optimized per formulation to achieve the desired outcome and efficacy. Lipids and nucleic acids can be cost prohibitive to screen, troubleshoot, and modify many formulation parameters (e.g., molar ratios, N/P ratios, process parameters, etc.) to find the most suitable LNP for a given application. While low volumes could be a limitation for producing a final formulation at a large scale, the ability to scale up the technique with larger microfluidic mixing instruments is commercially available.
Critical steps of the protocol start with proper storage of lipid stock solutions at the manufacturer's recommendation. LNPs should then be stored at 2-8 °C until further use. For the nucleic acid preparation, the results presented demonstrate that citrate buffer and malic acid buffer are effective at successfully forming LNPs with high nucleic acid encapsulation34, 35. Other buffers may be used instead if desired. If another buffer is chosen, it is important to maintain the pH below the pKa of the ionizable lipid to ensure that the lipid is cationic and can complex with the nucleic acid. When using the microfluidic mixing instrument, it is important to prime the cartridge prior to LNP formation, not to exceed the the cartridge use as recommended by the manufacturer, and to change the cartridge in between different formulation compositions. The most common flow ratio for formation of the aqueous: organic solution is 3:1; however, this can be changed if needed. The flow rate can also be adjusted as desired. Finally, it is important when working with mRNA to ensure an RNase-free environment throughout the entire process. If the desired size or encapsulation efficiency is not achieved, some places to begin troubleshooting include changing the N/P ratio used or the lipid molar percentages. The instrument process described here uses a benchtop model that has a maximum volume limit of 12 mL, although this process is scalable to larger volumes using different microfluidic mixing models.This process can be adapted to changes in lipid mixtures and nucleic acids for use in developing LNPs for various clinical indications. With this flexibility, numerous future applications can be achieved with LNPs to produce different desired formulations. This technique has also been used for developing other types of nanoparticles, including liposomes and polymeric nanoparticles. With some parameter changes, this method can be used for a variety of nanoparticle formulations.
The protocol detailed here describes a reproducible method for achieving mRNA or DNA encapsulated LNPs. In addition to process parameters, additional considerations can influence the LNP outcome. Previous work has also used similar methods to produce LNPs with various nucleic acids, ionizable lipids, N/P ratios, PEG linker length, etc. These parameters can influence the encapsulation efficiency, size, and charge of the particles. The instrument manufacturer has also noted similar changes depending on these parameters that can be optimized18,36. These parameters can further influence the biodistribution and efficacy of the nucleic acid. For example, studies have investigated hydrocarbon chain lengths (C14, C16, and C18) conjugated to PEG and found that the shorter acyl chain of C14 resulted in higher levels of liver uptake compared to the longer acyl chain, which remained in circulation for a longer period of time28. This protocol allows for the formation, optimization, and testing of LNPs with varied compositions, which makes this a versatile process.
The authors have nothing to disclose.
Thank you to Atul Saluja, Yatin Gokarn, Maria-Teresa Peracchia, Walter Schwenger, and Philip Zakas for their guidance and contributions towards LNP development.
1,2-dimyristoyl-rac-glycero-3-methoxypolyethylene glycol-2000 (C-14 PEG) | Avanti Polar Lipids | 880151P | |
10 µl Graduated Filter Tips (RNase-,DNase-, DNA-free) | USA Scientific | 1121-3810 | |
1000 µl Graduated Filter Tips (RNase-,DNase-, DNA-free) | USA Scientific | 1111-2831 | |
20 µl Beveled Filter Tips (RNase-,DNase-, DNA-free) | USA Scientific | 1120-1810 | |
200 µl Graudated Filter Tips (RNase-,DNase-, DNA-free) | USA Scientific | 1120-8810 | |
3β-Hydroxy-5-cholestene, 5-Cholesten-3β-ol (Cholesterol) | Sigma-Aldrich | C8667 | |
BD Slip Tip Sterile Syringes (1 ml syringe) | Thermo Fisher Scientific | 14-823-434 | |
BD Slip Tip Sterile Syringes (3 ml syringe) | Thermo Fisher Scientific | 14-823-436 | |
BD Vacutainer General Use Syringe Needles (BD Blunt Fill Needle 18G) | Thermo Fisher Scientific | 23-021-020 | |
Benchtop Centrifuge | Beckman coulter | ||
Black 96 well plates | Thermo Fisher Scientific | 14-245-177 | |
BrandTech BRAND BIO-CERT RNase-, DNase-, DNA-free microcentrifuge tubes (1.5mL) | Thermo Fisher Scientific | 14-380-813 | |
Citric Acid | Fisher Scientific | 02-002-611 | |
Corning 500ml Vacuum Filter/Storage Bottle System, 0.22 um pore | Corning | 430769 | |
Disposable folded capillary cells | Malvern | DTS1070 | |
Ethyl Alcohol, Pure 200 proof | Sigma-Aldrich | 459844 | |
Fisher Brand Semi-Micro Cuvette | Thermo Fisher Scientific | 14955127 | |
Invitrogen Conical Tubes (15 mL) (DNase-RNase-free) | Thermo Fisher Scientific | AM12500 | |
MilliporeSigma Amicon Ultra Centrifugal Filter Units | Thermo Fisher Scientific | UFC901024 | |
NanoAssemblr Benchtop | Precision Nanyosystems | ||
Nuclease-free water | Thermo Fisher Scientific | AM9930 | |
Phosphate Buffered Saline (PBS) | Thermo Fisher Scientific | AM9624 | |
Quant-iT PicoGreen dsDNA Assay Kit | Thermo Fisher Scientific | P7589 | |
Quant-iT RiboGreen RNA Assay Kit | Thermo Fisher Scientific | R11490 | |
Sodium Chloride | Fisher Scientific | 02-004-036 | |
Sodium Citrate, Dihydrate, granular | Fisher Scientific | 02-004-056 | |
SpectraMax i3x | Molecular Devices | ||
Zetasizer Nano | Malvern |