INTRODUCTION
Lipid Nanoparticles (LNPs) became widely known in the COVID-19 pandemic by playing a major role in the development of SARS-CoV-2 vaccines (1).
Nowadays, LNPs are one of the most commonly used tools for xRNA (2, 3) and xDNA delivery (4). LNPs usually consist of five components, ionizable lipids, phospholipids, cholesterols, pegylated lipids, and the payload, each of them adding unique properties to the structure of the LNP. Therefore, synthetic lipids as well as lipid conjugates have been gaining significant importance in recent years as drug target, therapeutic molecules, and liposomes.
Traditionally such molecules are purified by normal phase chromatography on silica using large volume of toxic solvents, including chlorinated solvents. As CDMO and producer of numerous lipids CordenPharma accepted the challenge of reducing organic solvent usage and developing greener processes. Ways to achieve this goal include aqueous/non-aqueous reversed-phase chromatography (NARP) (5, 6) and super critical fluid chromatography (SFC) that can be considered to supplant organic solvents (7) with more benign alternatives. In SFC, the mobile phase consists of a supercritical fluid like carbon dioxide replacing commonly used organic solvents.
SFC shows several benefits compared to HPLC. The polarity of supercritical carbon dioxide is similar to n-heptane, and it is miscible with most organic solvents, allowing to tune the polarity of the solvent. Due to its low viscosity and high diffusivity chromatographic processes are much faster as in HPLC, reducing run times significantly. Also, in economic regard SFC plays an important role by decreasing solvent costs. CO2 is non-toxic, non-flammable, and readily available, making SFC an environmentally benign alternative to other solvent-based chromatography methods.
Three case studies based on such alternative technologies for purification of lipophilic compounds will be presented replacing traditional organic solvents in order to achieve greater sustainability and/or to reduce both environmental impact and operational costs.
MATERIALS AND METHODS
Introduction to relevant lipids for LNP formulation
The lipids used in formulation each add unique properties to the LNP. They contribute to the stabilization, encapsulation, and delivery of the xRNA/DNA.
Ionizable lipids play an important part in particle formation. Their positively charged moieties electrostatically interact with the negatively charged backbone of the RNA, leading to encapsulation in a self-assembly process. It also allows for the endosomal release of the drug substance once introduced into the target cells (2).
Phospholipids are built of one or more fatty acids attached to a glycerol “head”. Their function is the formation of a stable bilayer structure underneath the PEG surface. Besides, they also balance out the non-bilayer propensity of the ionizable lipids.
Cholesterols also contribute to the structure of LNPs. They stabilize the liposomes and prevent bilayer-leakage by interaction between its hydroxyl groups and the polar heads of the phospholipids. A main source for cholesterols in industrial application are extracts from animal sources (wool grease of sheep or animal tissues) which is a potential risk factor for animal-sourced contaminations like Transmissible Spongiform Encephalopathies (TSEs). Therefore, chemists at CordenPharma and the Otto-von-Guericke-University of Magdeburg developed non-animal origin cholesterol (8), which can be made through a short synthesis directly starting from plant-based material.
Last but not least, pegylated lipids have a vital part in the stability of LNPs. By formation of a protective hydrophilic coating, they increase storage stability, avoid recognition of the LNP by macrophages of the immune system, and decrease nonspecific binding to proteins. Therefore, they not simply increase storage stability, but also prolong in vivo circulation. Due to this “camouflage” ability, they are often referred to as “stealth” lipids (9).
The different classes of molecules present a variety of challenges regarding their purification. Phospholipids for example can often be purified via precipitation or crystallization. For others, like the molecules discussed in this article, a chromatographic approach is the method of choice.
Experimental section
Initial scouting experiments for the HPLC methods were done on an Agilent 1260 Infinity II LC System using a variety of 4.6mm ID columns from different suppliers. Ethanol (absolute, gradient HPLC grade) and 2-Propanol (LC-MS grade) were purchased from Scharlab. ACN (HPLC-S gradient grade) was purchased from Biosolve. Scale-up was done on a VWR LaPrep HPLC-system using a Novasep LC50.500VE100 column. Ethanol (absolute) for the scale-up was purchased from Reuss Chemie AG.
SFC methods were developed on a SFC-system from PIC-Solution (PIC-Lab Hybrid SFC 10-20). Carbon dioxide (4.5, 99.995 vol%) was purchased from Messer Schweiz AG.
RESULTS AND DISCUSSION
Applications for ionizable lipids
Lipid A§ (see figure 2) is a typical example for an ionizable lipid and the challenges associated with this class of molecules. Normal-phase flash chromatography on silica was investigated but no satisfactory purity could be achieved.
Therefore, a reversed phase method was developed; however, due to its highly non-polar side chains, the compound is insoluble in the common solvents for Reversed Phase chromatography (RP) like acetonitrile, methanol, ethanol, and aqueous mixtures commonly used in RP. Although the solubility issues in aqueous mixtures can be solved by addition of acid, likely through protonation of the tertiary amines, degradation occurs under these conditions. To overcome these challenges, a novel approach was chosen: non-aqueous reversed phase chromatography (NARP). In this technique, highly polar solvents like acetonitrile or methanol are used as solvent A. As solvent B, controlling the retention of the compound, organic solvents with weak polarity, like tetrahydrofuran, isopropyl alcohol, dichloromethane, or methyl tert-butyl ether are chosen.
A systematic screening of different stationary phases and eluents was performed. When methanol was tested as solvent A, a fast degradation of the target molecule was observed. Therefore, acetonitrile was chosen as the highly polar component of the eluent system. Due to the high solubility of lipid A in isopropanol and the chemical stability observed, isopropanol was chosen as solvent B. As a result, an acetonitrile/isopropanol gradient in combination with a C18 stationary phase gave satisfactory results for the purification of lipid A regarding yield and purity. In a subsequent study on loading and gradient, further optimization was done (see figure 2). With these conditions in hand, the feasibility of the method could be demonstrated on a Phenomenex Luna 10 µm PREP C18(3) (4.6x250mm) with 30mg loading.
Green chemistry is an on-going challenge in the industry (7), therefore, the ecological impact of processes needs to be addressed. The use of chlorinated solvents can be avoided
through non aqueous reversed phase chromatography as demonstrated for lipid A. NARP has become a standard purification method for several similar ionizable lipids.
Applications for “stealth” lipids
Lipid B (see Figure 3) is a typical example for a pegylated “stealth” lipid. After an initial systematic screening of standard stationary phases similar to lipid A, a phenyl-hexyl derivatized silica was found to be suitable for further purification of lipid B. In fact, the phenyl-hexyl phase used in this experiment turned out to be useful for various other lipids with pegylated moieties too. Further scouting experiments regarding the eluent system led to a water/ethanol gradient as mobile phase. Attempts with “classical” co-solvents for RP like acetonitrile and methanol failed due to stability, solubility and selectivity issues. Besides being a green solvent, ethanol prevents stability issues previously observed with methanol.
As in the previous example, subsequent studies regarding loading and gradient were performed further optimizing the conditions allowing baseline separation of the product peak with a loading of 30mg on a 4.6×250 mm Phenomenex Luna 10 µm PREP Phenyl-Hexyl column. In a final step, the conditions could be successfully scaled up to a 50mm ID column, yielding the target compound in >99% purity.
The resulting chromatogram can be seen in figure 3.
Replacement is one of the main principles of green chemistry (12) which should always be considered. In this regard, the case study of lipid B is an example how processes can be made “greener” by method design. Ethanol not only serves as a “greener” replacement for commonly used solvents like acetonitrile (10) but also reduces the potential toxicity in occupational exposure to human health compared to solvents like methanol.
Application for a cholesterol like lipid
Lipid C (see figure 4) is a cholesterol like lipid. The target purity of >99.5% could not be achieved via crystallization nor normal-phase chromatography, so an approach via Supercritical Fluid Chromatography (SFC) was investigated. SFC offers enhanced selectivity due to the tunable nature of supercritical fluids. By adjusting temperature, pressure, and the addition of co-solvents, the mobile phase can be tailored to achieve optimal separation of target compounds.
The development was again started with screening for a suitable stationary phase, resulting in the application of standard C18 packing material (Kromasil C18, 5 µm) used for many years in preparative HPLC. Different modifiers were evaluated for optimal separation. On small scale experiments (4.6x250mm), baseline separation could be achieved by addition of 9% (v/v) ethanol to the CO2. Scale-up experiments were done on a Kromasil C18 (7 µm) column with 76.5 mm ID. The results are illustrated in figure 4 a). With these promising results in hand, a stacked injection method was developed (see figure 4 b)). Given the separation time of 4 minutes shown in figure 4 a), a cycle time of 3.3 minutes was chosen. The concentration of the injection solution was 20%, allowing to purify 1.2g crude per injection (7mL injection volume). With the given parameters, roughly 500g per day could be purified (see Figure 5). Regarding mobile phase consumption, 35 L ethanol as well as 15 kg CO2 (with recycling) were used per 100g of final product. The chromatographic yield was approximately 91% with an excellent purity of ≥99.9% for the target molecule.
The separation of lipid C benefits from the general advantages of SFC. Carbon dioxide is an inexpensive, non-flammable gas (11). CO2 is a recovered industrial waste product and can even be recycled during SFC (11). Supercritical carbon dioxide has properties similar to hydrocarbons and can therefore be used as a replacement of normal-phase solvents like hexane or heptane. Unlike these organic solvents, the supercritical carbon dioxide can be easily removed due to its gaseous aggregate state once passing the outlet pressure regulator, directly impacting time and cost. Furthermore, the low viscosity allows for rather high flow rates, decreasing the overall time needed for the purification (12). Optimization of the method by stacked injections greatly increased the productivity of the process.
CONCLUSIONS
In this article, we presented three examples of novel purification strategies for lipid nanoparticle building blocks. Non-aqueous reversed phase chromatography (NARP) can be successfully employed to overcome commonly observed solubility and polarity issues common for ionizable lipids.
In the second example, a reversed phase chromatography with a less common organic modifier could be developed. A water/ethanol gradient was used for the
successful purification of the pegylated lipid B. Ethanol was used as a “greener” replacement for commonly used solvents like acetonitrile and methanol.
The last case study demonstrates a very efficient use of SFC. A stacked injection method could be developed which led to an overall great productivity concerning the purification of cholesterol like lipid C. Supercritical Fluid Chromatography proofs to be a versatile and innovative separation technique that due to its unique combination of efficiency, environmental sustainability, and compatibility with a wide range of compounds makes it a valuable tool for the purification of lipids.
All these methods are easily scalable and allow kg-scale purification of LNP building blocks to very high purity under GMP for early and late phases of LNP development.
ACKNOWLEDGEMENTS
The authors thank Nicolas Spinner and Robert Kastl, both CordenPharma Switzerland, and Sébastien Thomas, CordenPharma Chenôve, for their support and many ideas for the novel approaches described in this publication.
Figure 1. Overview of different lipid types used as LNP building blocks.
Figure 2. Chromatogram of Lipid A on a 4.6 x 250 mm Phenomenex Luna 10 µm PREP C18(3) column after optimization.
Figure 3. Scale-up of the optimized conditions for the purification of lipid B on a 50 x 253 mm Phenomenex Luna 10 µm PREP Phenyl-Hexyl column.
Figure 4. a) SFC purification of lipid C on a 76.5×250 mm Kromasil 7 µm C18 column.
b) Purification of lipid C in stacked injection mode.
Figure 5. SFC production of a commercial lipid to support mRNA LNP formulation at CordenPharma Chenôve.
REFERENCES AND NOTES
§ Full structure cannot be disclosed due to confidentiality.
- E. Dolgin. The tangled history of mRNA vaccines. Nature. 2021; 597: 318-324.
- N. Pardi, M. J. Hogan, F.W. Porter, D. Weissman. mRNA-a new era in vaccinology. Nat. Rev Drug Discov. 2018; 17: 261-279.
- A. Tam, J. Kulkarni, K. An, L. Li, DR Forscheid, GK. Singhera, et al. Lipid nanoparticle formulation for optimal RNA-based topical delivery to murine airways. European Journal of Pharmaceutical Sciences. 2022; 176: 106234.
- S. S. Scully, D. Laplaca, R. Pelly, R. G. Parmar, M. Maetani. Ionizable amine lipids and lipid nanoparticles. Patent WO/2020/219876. Published 29.10.2020.
- N. A. Parris. Non-aqueous reversed-phase liquid chromatography: A neglected approach to the analysis of low polarity samples. J. Chrom. A. 1978; 157: 161-170.
- A. Causevic, K. Olofsson, P. Adlercreutz, C. Grey. Non-aqueous reversed phase liquid chromatography with charged aerosol detection for quantitative lipid analysis with improved accuracy. J. Chrom. A. 2021; 1652: 462374.
- J.P. Taygerly, L. M. Miller, A. Yeec, E. A. Peterson. A convenient guide to help select replacement solvents for dichloromethane in chromatography. Green Chem. 2012; 14: 3020-3025.
- Scalable Synthesis of GMP Grade Plant-Based Cholesterol (BotaniChol®). https://cordenpharma.com/articles/facile-scalable-synthesis-of-plant-based-cholesterol-botanichol-in-gmp-grade/ (last accessed 15.03.2024)
- M. Rastogi, R. N. Saha, A. Alexander, G. Singhvi, A. Puri, S. K. Dubey. Role of stealth lipids in nanomedicine-based drug carriers. Chemistry and Physics of Lipids. 2021; 235: 105035.
- Y. Shen, B. Chen, T. A. van Beek. Alternative solvents can make preparative liquid chromatography greener. Green Chem. 2015; 17: 4073-4080.
- C. J. Welch. Challenges and opportunities for the greening of separation science in the pharmaceutical industry. Drug Discovery World. 2007; 8(4): 71-77.
- M. Dhoru, K. Shah, K. Detholia, M. Patel. Green chromatography: the eco-friendly and safer concept of green analytical chemistry. IJPSR. 2020; 11(3): 1022-1032.
