The ALC-0315 lipid (Figure 1) gained worldwide recognition as a component of the SARS-CoV-2 vaccine [BNT162b2, developed by BioNTech and Pfizer]. Today, it is one of several lipid-based components used in formulating lipid nanoparticles (LNPs) for gene therapies and cancer treatments. In this study, we demon- strate the purification of ALC-0315 from synthetic protocols using flash chromatography. By optimizing solvent composition, purification methods, stationary phase, and loading, we show that high-performance liquid chromatography (HPLC) is not always necessary for lipid purification. Our methodology offers a scalable approach to flash purification, suitable for multigram to kilogram scales.
Cationic lipids are used for the targeted delivery of nucleic acids, making high purity crucial1. However, lipids like ALC-0315 are difficult to purify with classical methods such as distilla- tion, due to their non-crystalline, colorless nature, weak UV absorbance, and presence of structurally similar impurities. Consequently, chromatography is often the preferred method. Chemical synthesis of ALC-0315 has been published (2), but potential purity and yield issues during stepwise synthesis can be expected due to several factors: 1) complex and lengthy synthetic pathways leading to cumulative inefficiencies; 2) impure or poorly soluble raw materials causing unknown peaks and variability; 3) suboptimal synthetic chemistry and experi- mental techniques; and 4) compound degradation in complex systems.
With these variances in mind, it is important to ensure that robust purification protocols are in place to ensure high purity for the final product. Using high-quality flash silica and robust purification parameters, the required lipid purity can be achieved without expensive HPLC.
Biotage® Selekt flash chromatography system equipped with Biotage® Selekt ELSD evaporative light-scattering detector was used with Biotage® 10 g Sfär Silica D – Duo 60 µm, Biotage® 5 g Sfär Silica HC – Duo 20 µm, Biotage® 11g Sfär Amino D – Duo 50 µm, and Biotage® 11g Sfär KP-Amino D – Duo 50 µm columns. Samples were loaded onto the columns by liquid loading or by dry loading with Biotage ISOLUTE® HMN-R media in Biotage® Sfär Dry Load Vessel (DLV) for 10 g columns. Fraction identity was verified by direct infusion into a Biotage® Isolera™ Dalton 2000 mass detector with an electrospray ionization (ESI) source.
In this study we performed the previously published synthesis of ALC-0315 (Scheme 1)2.
While the chemistry is relatively straightforward, the process produces several byproducts, resulting in reduced yield and quality. This places additional expectations on the ultimate purification strategy.
In our product mixtures we expected to find ALC-0315 along with an impurity profile consisting of at least four byproduct peaks:
Mono/dialkylated mixture of final product
Carry forward - transesterification of intermediate
Over oxidation product of first stages
Step 2 trimer from starting materials
Two identical syntheses were carried out, resulting in lipid batches A and B. Despite adhering to the same protocol, the batches displayed different impurity profiles and batch B yielded less product. However, both batches contained the desired product (Figure 2). Although we did not further inves- tigate the nature of the by-products, we determined that these batches are well-suited for developing our purification methods.
Figure 2. Purification of ALC-0315 from two batches. [Silica (Sfär HC, 5g), linear gradient (in black), 0-50%, DCM/MeOH, 4% wt/wt silica, 7 mL/min (4.33cm/min), UV (very weak signals just above baseline) λ-all 198-810 nm, 200 nm, 205 nm. ELSD parameters: solvent: Acetone, temperature:36 °C, N2 pressure:1.5 bar (tan line)]
The crude ALC-0315 mixtures were first purified by normal phase methods using a Biotage® Selekt system equipped with Biotage® Selekt ELSD4. Lipids are often poorly detected by UV3, but purification with complementary evaporative light scattering detection (ELSD) improves sensitivity significantly4. In our examples, the UV response for ALC-0315 was low (approx- imately 25 mAU) even at its absorbance max (205 nm). Lambda All is a useful feature that sums up wavelengths, amplifying the overall signal. However, when the initial signal is minimal, the collective UV signal strength can remain weak. Biotage® Selekt ELSD ALC-0315 response was about 800mV. This enhanced sensitivity supported a complete visualization of the product elution profile, facilitating robust method development and validation, product identification and collection, and improved recovery and purity particularly at larger scale.
The mixtures were compared to a control sample injection of commercially sourced pure ALC-0315, which confirmed the presence of the desired product (Figure 3).
To avoid adding a modifier in the mobile phase, which is typically employed when purifying basic compounds with a silica stationary phase, an amino-functionalized silica was used. This stationary phase highlighted the differences between the two synthetic batches A and B clearly (Figure 4).
Methyl tertbutyl ether (MTBE) was initially chosen as the solvent due to its similar purification efficiency and elution profiles to ethyl acetate (EtOAc). Unlike EtOAc, it does not absorb UV at low wavelengths, thus increasing UV sensitivity for our compound of interest. Although MTBE is interesting for research-scale applications, it is neither practical due to storage concerns nor environmentally sustainable on a larger scale. Consequently, we replaced MTBE with EtOAc for further purification method development.
In addition to Sfär KP-Amino (40–65-micron, spherical silica with average diameter of 60 microns), a second amine- functionalized stationary phase, Sfär Amino (40–63-micron, irregular shaped silica with average diameter of 50 microns) was also tested. These stationary phases are known to have slightly different selectivity due to their different physicochemical structures (Figure 5).
While both amine-functionalized stationary phases supported better resolution of these lipid samples, given the excellent separation afforded by the step gradient5, we also looked at whether normal phase silica (without a modifier) could provide an acceptable separation for this type of sample.
In our small-scale studies, we used Biotage® Selekt platforms which interface well with external detectors such as ELSD.
As larger sample loads are used when scaling methods, the process development platform Biotage® Isolera LS has powerful software features (GO – gradient optimization) which help to convert linear gradients to step gradients automatically, helping to reduce solvent consumption by up to 50%5. Combining these two powerful strategies can be very advantageous when developing large scale purification methods.
After screening several mobile phase conditions (data not shown), the best separation was obtained using a heptane/2- propanol step gradient with Sfär 60 Silica. The chromatographic method provided a large separation (ΔCV =4) between cationic ALC-0315 and various lipophilic byproducts that eluted in 10% 2-propanol at a 4% sample/wt silica load (200 mg) (Figure 6).
The mass identity of the ALC-0315 fraction was verified by direct infusion mass analysis, using an Isolera™ Dalton 2000 mass detector with an electrospray ionization (ESI) source (Figure 7).
Process-scale purification typically employs higher sample loading percentages to maximize throughput and yield compared to analytical applications. To verify the robustness of our purification method, we increased the sample load by 2.5 times (to 10% wt/wt silica, 500 mg) and repeated the purification under identical conditions (Figure 8).
Figure 8. ALC-0315 10% wt/wt silica (500 mg) purification using a 5g Sfär 60 pre-packed column. [Silica (Sfär 60, 5g), step gradient, 10% (4CV) then 20% (7CV) heptane / propanol, 10% wt/wt silica, 7 mL/min, (4.33cm/ min), UV λ-all 198-810 nm, 200 nm, 205 nm, ELSD parameters: solvent: Acetone, temperature:36 °C, N2 pressure:1.5 bar]
The results show that ALC-0315 remained well-resolved from its less polar byproducts, even at a high 10% sample load.
These results highlight the importance of developing robust purification methods during the research and development phase of a project. While there may be several potential purifi- cation strategies to choose from, setting the correct goals and managing experimental bias in process method development ensure the highest possible yields and recoveries.
Our data shows that even in sequential synthesis, chemistry can sometimes go wrong, leading to differences in supposedly identical samples for downstream for batch pooling and purifi- cation. A well-designed and robust purification method can accommodate such batch-to-batch variation. Even in extreme cases of column overload, pure material can still be recovered through peak cutting and reprocessing of mixed fractions.
In our studies, the relatively large particle size ensured backpressures remained within the design limits of the flash purification platforms used, typically staying below seven bars.
The large-scale Biotage® Flash 400 purification platform supports up to 4 kg of sample purification per batch injection. However, there are several intermediate platforms that can handle up to 500 g of lipid sample loads. Biotage flash purifi- cation systems are designed for seamless transfer to larger Biotage platforms, as they use the same linear flow velocity.
At a 10 percent sample load, even the small-scale method development runs in the above series would correspond to a purification capability of 150 g of lipid per purification using Biotage® SNAP XL columns with Sfär 60 μm media. For the next step up in the scale-up workflow, the Biotage® Flash 150 purification platform supports efficient batch purification of up to 500 g of lipid per run.
It is not always necessary to use expensive HPLC techniques to purify complex or valuable compounds. Desired purification outputs can be achieved using relatively simple flash chroma- tography systems and standard stationary phases.
If the target molecule lacks a significant UV chromophore, method development and optimization can be performed using alternative detectors such as Biotage® Selekt ELSD at small scale, with purification method characteristics transferable seamlessly to larger scale. In our study, a step gradient was designed in to optimized purification of a synthetic cationic lipid (ALC-0315) at very high (10%) sample loads, using only ELS detection.
Our simple methodology provides a directly scalable route to multigram and kilogram scale purification, which has been used globally in batch style purifications6. Such platforms with plug- and-play workflows and transferrable purification conditions, make scaling up to isolate kilograms of the target compound in a single run a practical reality.
Literature number: AN1008