mRNA Lipid nanoparticle (LNP) encapsulation and purification

Scaled-down process development using NanoAssemblr™ Blaze nanoparticle formulation system

To determine the formulation conditions used in the production scale batch, scaled-down formulation work was carried out using the NanoAssemblr™ Blaze nanoparticle formulation system. Two batches (20 and 40 mg) of mRNA-LNP encapsulation were performed successfully.

mRNA-LNP formulation was performed using a NxGen™ 500D cartridge by mixing a proprietary lipid cocktail (25 mM, organic phase) and mRNA (diluted to 0.17 g/L in 0.1 M sodium acetate, pH 4.0, aqueous phase) with a 2:1 in-line dilution in 1× PBS, pH 7.4. The mRNA payload used for encapsulation in LNPs was purified and sterile-filtered bulk drug substance in sodium citrate, pH 6.5 at a concentration of 1.9 g/L. The total flow rate (TFR) of the aqueous and organic streams in these development batches was 115 mL/min and the flow rate ratio of the aqueous to organic phases was 3:1.

Production scale encapsulation using the NanoAssemblr™ commercial formulation system

To produce sufficient mRNA-LNPs for further downstream filtration studies, uncapped eGFP2 coding mRNA in aqueous buffer (100 mM sodium acetate, pH 4) was encapsulated with a proprietary lipid mix. Prior to use, the lipid mix was filtered using a Supor™ Prime sterilizing grade filter (effective filtration area: 240 cm²) and stored at 2°C to 8°C until the day of formulation. The lipid mix was analyzed before and after filtration by high-performance liquid chromatography (HPLC) to confirm the stock concentration and assess any impacts of filtration on individual lipid concentrations.

The day before formulation, the eGFP2 mRNA payload was gradually thawed (from -80°C) at 2°C to 8°C, then placed in a water bath at room temperature on the day of formulation to accelerate the thawing process. The payload was subsequently diluted to 0.17 mg/mL in 100 mM sodium acetate pH 4 buffer and the concentration was assessed by microvolume spectrophotometer measurement. A 250 mM sodium acetate pH 4 buffer was used to prepare the aqueous phase.

Fig 2. NCFS fluid path layout showing aqueous, organic, and dilution inlets, and system connections made during operation.

The NCFS equipped with an engineering prototype of the NxGen™ commercial formulation system flow kit 48 L/h fluid path was utilized to perform the encapsulation, with a total output formulation volume of 49 L. Suspended bioprocess containers (BPCs) were used to house the aqueous priming solution, aqueous phase, and waste. The product was collected into a 50 L ReadyCircuit™ bag. 5 L glass media bottles with a hose barb outlet were utilized to house the organic priming solution and lipid mix. A custom holder was constructed to tilt the lipid mix bottle to allow full drainage and secure the unit from disturbance during filtration. The 1× PBS, pH 7.4 dilution buffer, was housed in a 20 L surge BPC; a peristaltic pump was utilized to maintain the bag level during formulation by drawing additional PBS from a stock drum. Custom tubing assemblies were designed, constructed, and sanitized with 0.25 M NaOH to make all required process connections. As the NCFS flow path was an internally produced engineering grade prototype, it was sanitized using 0.25 M NaOH and neutralized with WFI to pH < 8 prior to use.

A pre-dilution total flowrate (aqueous + organic phases) of 800 mL/min was utilized. The mRNA-LNP particles were diluted with 1× PBS, pH 7.4 buffer using an in-line dilution pump to bring the bulk ethanol concentration below 8% v/v due to facility considerations. The viscosity of the aqueous phase was set to 1 cSt and the organic phase viscosity was set to 1.5 cSt. Table 1 outlines additional formulation conditions.

Table 1. Formulation conditions for production scale encapsulation using the NanoAssemblr™ commercial formulation system

Post-formulation, a fraction of the LNP product was diverted for other work, while the remainder was transferred to the TFF system for downstream processing.

Production scale buffer preparation – cryobuffer

Post-formulation, LNPs were stored in an appropriate cryoprotectant buffer. A total of 60 L of proprietary cryobuffer was prepared and adjusted to pH 7.4 in a 200 L Xcellerex™ XDUO Single-Use Mixing System using Xcellerex™ XDM single-use mixer bag with Fortem™ film (200 L standard and ReadyMate™ connector). The entire buffer solution was filtered through a 0.2 µm Supor™ ECV capsule filter (effective filtration area 220 cm²), dispensed into ReadyCircuit bioprocess containers and stored at 2°C to 8°C until the day of formulation.

Production scale downstream purification via TFF (UF/DF)

The mRNA-LNP drug product was purified via TFF immediately after formulation to introduce the cryoprotectant storage buffer; hold-time was minimized in the formulation buffer to prevent degradation of the liquid nanoparticles. This typically represents the third application of UF/DF in the mRNA-LNP manufacturing workflow, hence UF/DF3. An ÄKTA readyflux™ TFF system was used for the product concentration and buffer exchange. This utilized a 0.5" Flow Kit Plus with ReadyMate connectors. The TFF filter was constructed of two 100 kDa, 0.5 m² T-series Centrasette™ cassette stacked in an LV holder for a total surface area of 1 m² and a target mRNA loading of 1.6 g/m². A 20 L ReadyCircuit™ bag was used as the feed reservoir. All TFF activities were executed using pre-defined methods within UNICORN™ software.

Prior to loading product onto the filter, it was flushed with WFI, sanitized using 0.25 M NaOH, neutralized with WFI, then equilibrated using PBS buffer. A normalized water permeability (NWP) measurement was also taken using WFI post-sanitization.

The TFF operating conditions were set at a feed pump flowrate of 6 LMM (6 LPM) and a transmembrane pressure (TMP) of 1 bar (14.5 psi, 0.1 MPa). These conditions were previously determined to be the optimal condition for this formulation through prior experimental work (1). An initial concentration factor (5×) was set and the particles were diafiltered with four diavolumes (DV) of cryobuffer. The product was collected into a 20 L ReadyCircuit™ bag, then a product flush was completed to improve product recovery.

Production scale aliquoting of final material and freezing to -80°C

Post-TFF, the product was dispensed into 50 and 200 mL fractions for downstream filtration tests. The bottles were individually sealed into plastic bags and then initially frozen using dry ice. Once frozen, the bottles were removed from dry ice and transferred to a -80°C freezer.

Analytics

Dynamic light scattering (DLS) is an LNP characterization technique for measuring size and polydispersity index (PDI). The Zetasizer Ultra was the instrument used to determine particle size and distribution by analyzing the frequency and intensity of the scattered light of the LNPs. All samples were analyzed using an attenuator between 7 and 8 to ensure measurement accuracy.

RiboGreen Assay is a fluorescence-based technique to measure RNA concentration and Encapsulation Efficiency (EE%) of LNPs. This assay utilizes RiboGreen, a dye that binds to RNA to emit fluorescence signals. RiboGreen was first added to intact LNPs (with Tris-EDTA buffer) to measure the unencapsulated RNA concentration. Then, RiboGreen was added to disrupted LNPs to quantitate total RNA (encapsulated and unencapsulated). An RNA standard curve was also generated to compare it with the fluorescence samples to provide accurate measurements. Encapsulated RNA was calculated by subtracting total RNA and unencapsulated RNA.

HPLC is an analytical technique used to separate, identify, and quantify specific lipids within the LNPs. Lipids are first extracted from the LNPs using organic solvents. The lipid extract is injected into the HPLC system where lipids are separated based on their hydrophobicity. The detectors measure the concentration of each lipid by analyzing the area under the corresponding peaks in the chromatogram. This method provides precise lipid composition data for ensuring the stability of LNP formulations.

Capillary electrophoresis (CE) is an analytical technique for ensuring the integrity of mRNA within LNP formulations, using the SCIEX PA800 plus instrument. CE operates by separating mRNA molecules based on their size and charge within a capillary tube filled with a fluorescent dye that binds specifically to RNA. An electric field is applied which causes the mRNA to migrate through the capillary giving an electropherogram. The electropherogram is analyzed to give the mRNA integrity percentage of the LNPs. mRNA was extracted from LNPs by silica affinity purification and integrity was assessed by capillary electrophoresis employing the SCIEX RNA9000 kit according to the manufacturer's protocol.

Process development work with NanoAssemblr™ Blaze nanoparticle formulation system

To determine the scaled-up formulation conditions, 20 and 40 mg batches of mRNA were encapsulated into LNPs with 74% to 79% final product recovery. After the encapsulation, the hydrodynamic diameter and polydispersity index were in the range of 73 to 76 nm and 0.11 to 0.18 respectively, whereas the mRNA EE% and integrity in the LNPs > 99%, as shown in Table 2. Process development scale data on Nanoassemblr™ Blaze using the lipid mix and purified mRNA at a total flow rate of 115 mL/min, with final 8.3% ethanol in the formulation, provided reliable performance and the desired LNP product critical quality attributes (CQAs) post-encapsulation, to proceed to large-scale batch manufacture.

Table 2. Analytical results for scaled-down development batches

*based on main peak area from electropherograms of capillary gel electrophoresis

Production scale organic and aqueous phase preparation

For the organic phase, a total of 4.1 L of proprietary lipid mix was generated. HPLC results indicate that the overall lipid mix concentration was 15.62 mg/mL pre-filter or 88% of the target value. Post-filtration the overall lipid mix concentration was 15.24 mg/mL. This indicates that the filtration did not significantly impact the lipid concentration and that lipids were not retained on the Supor™ Prime filter used in this step.

The mRNA payload was removed from -80°C storage the day before the formulation. On the day of formulation, the material was fully thawed at room temperature using a water bath to accelerate the final thawing time. The integrity of the thawed mRNA stock was measured by CE at 94.4%.

A total mass of 11.7 kg aqueous phase was generated. Thawed eGFP2 payload (1.0 kg) was combined with 250 mM sodium acetate buffer (4.7 kg) and WFI (6.0 kg) and mixed using a stir plate. After mixing, the aqueous phase was transferred into a 20 L bioprocess container with a peristaltic pump and a sanitized tubing assembly. An initial microvolume spectrophotometer measurement taken of the stock mRNA concentration was 2.2 mg/mL (within 20% of the target value). The final aqueous phase concentration was assessed by microvolume spectrophotometer measurement at 0.152 mg/mL (within 10% of the target).

Production scale encapsulation and downstream purification via TFF (UF/DF) results

The NCFS instrument successfully executed the formulation conditions as outlined in Table 1 with no error messages for a total run time of 19.2 min. A maximum flowrate error of 0.8% was measured during the formulation. The combined flowrate errors led to a maximum mixing rate error of 2.7%. The maximum error values on all 3 system parameters were well within the system accuracy specifications of ±5%.

The PBS surge bag was successfully filled via a pump throughout the duration of the run to ensure constant supply of dilution buffer to the system. NCFS logs indicate that 49.0 L of product and 0.5 L of waste were generated during the run, as expected and verified by independent weight measurement of the product bag. The excess volumes of aqueous phase and lipid mix prepared for the run were sufficient to allow the run to completion without fully emptying the BPC bags, ensuring the full formulation could be completed.

Fig 3. Particle sizing and DLS results for in-process mRNA-LNP samples.

DLS sizing results indicate that the mRNA-LNP particle size immediately after formulation was 69.7 nm with a PDI of 0.05. This is generally consistent with previous results for this formulation, however it is noted that the location and DLS instrument was different for the scale-up results. Average sizing results for the in-process samples run in duplicate are displayed in Figure 3. A description of the sampling points is found in Table 3.

Table 3. List of sampling points for mRNA-LNP in-process samples

The concentration of the mRNA-LNPs was measured at 0.037 mg/mL post-formulation by Ribogreen assay, which was within 4% of the theoretical outlet concentration. Including the excess payload volume (250 mL, or ~40 mg of payload) that was produced but not formulated, the overall step yield is 94%. The encapsulation efficiency was > 99%. Ribogreen results for the in-process samples are displayed in Figure 4. The integrity of the formulation post-chip was measured by CE at 94.3%, indicating minimal degradation.

Fig 4. Concentration and encapsulation efficiency (EE%) results for in-process mRNA-LNP samples.

Post-formulation, 42.6 L of product was loaded onto the ÄKTA readyflux™ tangential flow filtration system. The remaining formulation volume was diverted for other work. Based on the post-formulation Ribogreen values, the TFF filter loading was 1.6 g payload/m². The UF/DF method included an initial recirculation vessel fill to 18 kg, fed-batch concentration until the product bag was empty, followed by concentration to a target mass of 8.2 kg (including hold-up volume). Once the product concentration was completed, the system performed a 4× diafiltration function with cryobuffer. Process data for the TFF step is found in Table 4.

Table 4. Process data for production scale UF/DF purification of mRNA-LNP formulation

Flux and TMP data for the UF/DF process is shown in Figure 5. The flux was observed to decline through the concentration stage from a starting flux of ~160 LMH, as the mRNA-LNP concentration increased and stabilized at ~71 LMH during diafiltration. After the completion of each process stage (fed-batch concentration, concentration, and diafiltration), the feed pump was briefly stopped and a momentary reduction in flux and TMP is observed at the 13 and 19 min timepoints. During the product recovery stage, a buffer flush with additional cryobuffer was used to increase product recovery.

Fig 5. Permeate flux and TMP over time for mRNA-LNP product during concentration to 0.2 g/L and diafiltration into 4× DV of cryobuffer.

Analytical results for the post-TFF samples indicate minor impacts of the TFF process on the LNP properties. This may be attributed to the impacts of shear forces experienced by the LNPs during processing. After TFF, the average particle size was not impacted (post-formulation: 70 nm, post-TFF: 68 nm), however the PDI of the formulation increased (post-formulation: 0.05, post-TFF: 0.14). An increase in PDI during TFF is a common occurrence when processing mRNA-LNP formulations and the post-TFF results are consistent with the process development batches DLS results are shown in Figure 3. The small difference in PDI between the TFF final and Conc Adj samples may be attributed to instrument variability and a time delay from when the samples were collected and analyzed. The post-formulation sample was analyzed the same day as the batch was formulated; all other samples were tested four to five days after formulation and were stored between 2°C to 8°C until analysis.

The TFF process had minimal impact on the encapsulation efficiency of the product, the pre-TFF and post-TFF EE% values were both 99%. The mRNA-LNP concentration was 0.206 mg/mL after diafiltration – within 10% of the theoretical TFF concentration target of 0.193 mg/mL. Based on the mass of product recovered, the TFF recovery was 102% - it is noted that some variability in the Ribogreen measurement may have attributed to this result. The integrity of the formulation post-TFF was measured by CE at 93.1%, a minor decrease.

The pre-use NWP of the filter stack was 22.9 LMH/psi. After use, the NWP value was 21.2 LMH/psi, indicating that the filter was not damaged or clogged by the application. The overall TFF process results indicate that the 100 kDa T-series Centrasette™ cassettes are a suitable membrane for processing similar mRNA-LNP materials.

The recovered material was further diluted with cryobuffer to a target concentration of 0.12 mg/mL. Cryobuffer was added to the collected product bag and mixed well prior to further sampling. The actual value of the concentration-adjusted, post-TFF material was determined by Ribogreen to be 0.11 mg/mL, for a final product volume of approximately 15 L. The material was subsequently aliquoted into 50 mL and 200 mL fractions for downstream testing using a peristaltic pump and a single-use dispensing set-up.

After concentration adjustment, the product aliquots were required to be stored at -80°C prior to further downstream processing activities. This presented a logistical challenge as the existing ultra-low cold storage units at the manufacturing site were not suitable to handle the significant heat load of the large volume (> 10 L) generated during the batch. Due to the properties of the specific LNP formulation, it was deemed to be undesirable to utilize an intermediate freezing step at -20°C. As a result of these constraints, an alternative freezing method using dry ice was investigated. As dry ice sublimates at -78.5°C, it is commonly used during the shipping of temperature sensitive biologics and other materials.

Prior to the manufacturing scale batch, freezing trials with dry ice were conducted to assess the freezing time of various container sizes of LNPs. A temperature probe was installed into a bottle containing the target volume of surrogate LNPs in cryobuffer, to examine the temperature profile during freezing. Surrogate LNPs were utilized solely to mimic fluid composition and thermodynamic behavior of the manufacturing batch; the analytical properties were not assessed. Two bottle sizes and aliquot volumes were tested: 125 mL PETG bottles containing 50 mL of LNP or 250 mL PETG bottles containing 200 mL of LNP. The bottles were packed into a container with dry ice and the temperature was tracked over time. Figure 6 indicates that the surrogate bottles of both sizes reached temperatures below -75°C within 3 hours.

Fig 6. Freezing process development study results using surrogate LNP aliquots packed in dry ice.

We employed this methodology to freeze the manufacturing batch. During the manufacturing batch, the 50 and 200 mL aliquots were dispensed into 60 and 250 mL PETG bottles respectively. All bottles were individually sealed in bags and placed into large tubs. The tubs were filled with dry ice, then placed in a large dry ice chest to freeze overnight. After approximately 11 hours of storage time, the bottles were removed and transferred to a -80 °C freezer. All product bottles were observed to be fully frozen at the time of transfer.

After one freeze/thaw cycle, the size and PDI displayed a modest increase across both aliquot sizes. This increase was expected as some particle aggregation will occur during a freeze/thaw cycle. Based on the minimal size differences between the Filt F/T 50 mL aliquot (size: 73 nm, PDI: 0.20) and the BRT F/T 200 mL aliquot (size: 75 nm, PDI: 0.21), the aliquot size and corresponding freezing/thawing rate did not appear to impact the final particle properties. The mRNA integrity of the final freeze/thawed aliquots was also assessed at 93.8% for the 50 mL Filt fraction and 93.3% for the 200 mL BRT fraction. Overall, this indicates that mRNA integrity (starting stock integrity: 94.4%) was not impacted by the formulation and downstream processing steps.

1. Application note: T-series TFF cassettes with Delta 100 kDa membranes for RNA and LNP applications. Cytiva. com/en/us/solutions/bioprocessing/knowledge-center/tff-membranes-for-rna-and-lnp-applications.
2. Application note: Process development considerations for RNA-LNP therapeutics. Cytiva. com/en/us/knowledge-center/advance-technology-rna-lnp-process-development

CY47795