Fast and Straightforward Lipid Quantification in Pharmaceutical Compositions Using NMR

Abstract

Liposomes are versatile nanoscale delivery systems used extensively in pharmaceuticals. Accurate quantification of lipids in liposome formulations is critical for ensuring the product quality and efficacy. This study evaluates the use of NMR spectroscopy applied to lipid quantification within liposomes. The investigation focused on lipids commonly found in liposomal products and other lipid-based delivery systems, such as DSPC, DOPC, cholesterol, and DMPE-PEG2K. Lipid quantification was performed using both an internal reference and an external standard through PULCON (pulse length-based concentration determination) methods. Both methods showed high accuracy, precision, and reproducibility. Additionally, the PULCON NMR technique offered enhanced consistency and faster analyses, making it particularly suitable for industrial needs involving rapid analysis across numerous samples. The quantification of liposome solutions was also conducted using a UPLC-ELSD method for a comparison of the two techniques. These findings support the refinement of rapid, accurate lipid quantification methods, essential for quality control in large-scale manufacturing processes and optimizing the drug product formulation process.

Introduction

Liposomes are spherical vesicles composed of one or more phospholipid bilayers that allow the encapsulation of both hydrophilic and hydrophobic substances. Over the years, liposomes have found numerous applications in the fields of drug delivery, cosmetics, and medical diagnostics. − Their versatility, tunable physicochemical properties, and ability to mimic biological membranes make them ideal vehicles for targeted therapeutic delivery, enhancing the efficacy, and reducing the side effects of drugs. Notably, they are employed in anticancer treatments, with marketed examples such as Doxil (liposomal doxorubicin) and Myocet (nonpegylated liposomal doxorubicin), both used for treating various types of cancer, including metastatic breast cancer. − In the realm of antifungal treatments, liposomal formulations are used for amphotericin B delivery, with notable examples being Nomex Liposome and AmBisome. − Additionally, liposome application extends beyond pharmaceuticals to the field of vaccines. For instance, AS01 is used as an adjuvant to boost immune responses in vaccines for malaria and shingles. −

Lipid composition in liposomes is a key determinant in the effectiveness of cargo encapsulation and release, as well as in the stability and potency of the formulation, and is often considered a critical quality attribute (CQA) under the Quality by Design Framework. − The initial concentration of lipids may differ from that of the final product due to potential losses during the multiple steps required for the formulation (such as mixing, filtration, centrifugation, etc.). Accurately quantifying these losses with a precise and rapid method can aid in formulation optimization and support large-scale process development. − Additionally, degradative processes such as hydrolysis and oxidation can significantly alter the phospholipid membrane properties (e.g., permeability) and lead to the loss of encapsulated material. − Thus, implementing analytical methods that allow for accurate and rapid quantification of lipids in the final product is essential for both quality control and optimization of preparation processes to minimize losses.

Lipid-based systems have historically posed analytical challenges due to the limited presence of functional groups that differentiate individual lipids, which primarily share a hydrocarbon backbone. The traditional colorimetric molybdate method to quantify phospholipids is still frequently used. Despite its high sensitivity, the method is error-prone and naturally cannot be applied to nonphosphorus lipids. , In recent years, alongside the development of mass spectrometry methods, chromatography techniques have gained importance, particularly those using charged aerosol detection (CAD) and evaporative light scattering detection (ELSD). − These techniques do not require a chromophore, which is typically absent in lipids, as occurs for UV–vis detection.

Nuclear magnetic resonance (NMR) is widely used for quantifying pharmaceutical components due to its atomic resolution, which allows us to solve the complexity of mixtures, analyzing single components without overlaps. − Solution NMR has been utilized to study liposomes, focusing on various key features, supported by well-established knowledge of the assignment of lipids commonly used in formulations. − Franz et al. already showed the feasibility of using NMR for lipid quantification in liposomes through ¹H and ³¹P experiments using triphenyl phosphate (Ph₃PO₄), tritolyl phosphate (Tol₃PO₄), and triphenylphosphine oxide (Ph₃PO) as references. This study aims to refine NMR methods for lipid quantification tailored for industrial applications, seeking faster and reproducible analyses through the comparison with a more standard ultraperformance liquid chromatography coupled with the evaporative light scattering detection (UPLC-ELSD) method. The method presented herein aims to be versatile across various lipid delivery systems, testing formulations that are prepared by mimicking commercial products. The investigated lipids are typical lipids found in liposome formulations. Specifically, it involves phospholipids (e.g., DSPC, DOPC), which facilitate the interface between lipid and aqueous phases; cholesterol, which modulates interaction with cellular membranes; and PEGylated lipids (e.g., DMPE-PEG2K), which control particle size and circulation via the “stealth effect”, thereby preventing premature clearance by the immune system. ,,, Liposomes were prepared using microfluidics, a method that has gained popularity for its practicality, high encapsulation efficiency, and applications in the formulation of liposomes and lipid nanoparticles (LNPs). − Developing a fast and precise method to measure lipid concentrations throughout the manufacturing process enables the evaluation of potential losses during formulation. Additionally, it supports the optimization of production processes to maintain the expected concentration levels of components in the formulations. NMR offers a distinct advantage over chromatography and mass spectrometry by providing a direct correlation between the signal intensity and the molar concentration of nuclei for each component. This eliminates the need for analyte calibration curves, which are essential for chromatography and mass spectrometry. It is sufficient to utilize a reference standard compound of known concentration that does not interfere with the analyte and is miscible in the solution under analysis, enhancing both the speed and reproducibility of analyses. The reference standard can be internal (mixed with the analyte in the NMR tube) or external (in a separate NMR tube). The use of an external reference compound for NMR quantification was explained first by Wider et al., referring to the method as pulse length-based concentration determination (PULCON). This technique does not require the addition of an internal standard directly into the sample mixture, which is advantageous when a standard may interfere with the sample or is impractical to use. In addition, once a reference tube is prepared at an exact concentration and analyzed to measure its proton signal integral, it is possible to quickly analyze multiple batches, eliminating the variability associated with the addition of a reference to each tube and saving considerable time. To the best of our knowledge, this method has not been tested for lipid quantification in complex mixtures so far. − In our study, we demonstrated its applicability and highlighted its benefits for this purpose in comparison to the NMR analytical method, which uses an internal standard.

Finally, the proposed method has been compared to a protocol optimized for lipid quantification using UPLC-ELSD. Comparing NMR to UPLC methods, which are commonly used and well-established in the industry for similar applications, provides a comprehensive evaluation of the advantages and limitations of each technique, enabling more informed decision-making about their suitability for specific applications.

Results and Discussion

NMR Quantification of Single Lipid Solutions

A well-resolved specific signal for each lipid that could be used for integral analysis also in the spectrum of liposomes was selected. The chosen signals and the corresponding assigned proton are indicated in Figure . Three tubes from the same stock were analyzed to assess the variability of the NMR tube preparation and the analysis itself. Additionally, investigating three different concentrations allowed us to identify the range in which the method responds linearly and to determine the detection limit for accurate analysis. The intermediate concentration analyzed for each lipid is on the same order of magnitude as the one the lipid has in the prepared liposomes. Standard single lipid solutions were analyzed by performing triplicate measurements of each of the investigated concentrations, always adding the same quantity of DMF reference in each tube. The same analysis was also performed using the PULCON method to compare the two techniques for this application. The PULCON method is based on the reproducibility of the NMR signal as a function of analyte concentration to quantify an analyte in an NMR tube using a separate NMR tube for reference. All the parameters such as temperature, pulse calibration, and relaxation delay that directly influence the signal intensity and integration accuracy are taken into account for concentration determination by comparing integrals of analyte and reference in two separate tubes, allowing for the use of an external standard (refer to the “Methods” section for the equation). The reference integral of DMF was determined by averaging the measurements from three tubes, each prepared at a concentration of 36.796 mM.

1.

NMR signals of lipids used for the integration. (A) 1D ¹H NMR spectrum of a standard lipid mix solution, containing DSPC, cholesterol, and DMPE-PEG2K; the integrals of the signals used for quantification across all tested samples are displayed in red. Each signal was identified based on individual lipid spectrum recordings and chemical shift predictions using ChemDraw Professional software and literature refs and . (B) The signals used for the integration of each component are highlighted in the molecular structures: (1) the 9 N-methyl protons for DSPC, which share the same chemical shift as DOPC when present, resonating at 3.35 ppm; (2) the 180 protons from the repeating units of PEG (–[–O–CH₂–CH₂–]–) in DMPE-PEG2K at 3.66 ppm; (3) the 3 protons of the methyl group at 0.69 ppm for cholesterol; and (4) the aldehydic proton of the standard reference DMF, integrated at 8.03 ppm.

Table summarizes the results, showing the concentration calculated from the mean of three replicates and the corresponding percentage of relative standard deviation (% RSD) to express the variability of the method among replicates. Additionally, the recovery value, expressed as the percentage of the experimental concentration relative to the theoretical concentration, serves as an indicator of the method accuracy in each case.

1. Summary of NMR Results for DSPC, Cholesterol, and DMPE-PEG2K Standard Solutions across Different Concentrations.

lipid	theoretical (mg/mL)	experimental (mg/mL)	recovery %	% RSD replicates (%)
DSPC	10.28	10.15	98.8	2.6
DSPC	1.03	1.00	97.5	5.2
DSPC (PULCON)	1.03	1.02	99.2	0.6
DSPC	0.100	0.095	95.0	8.0
cholesterol	5.56	5.40	97.1	3.0
cholesterol	0.56	0.53	94.9	2.8
cholesterol (PULCON)	0.56	0.53	95.0	1.6
cholesterol	0.054	0.031	52.4	6.2
DMPE-PEG2K	2.34	2.15	92.0	2.1
DMPE-PEG2K	0.23	0.22	95.6	1.9
DMPE-PEG2K (PULCON)	0.23	0.22	95.6	3.0
DMPE-PEG2K	0.023	0.021	91.3	0.2

^aThe ratio of experimental to theoretical concentrations, expressed as a percentage.

^bThe ratio between the standard deviation and the mean, expressed as a percentage, calculated from the three replicates of each sample.

^cAn additional decimal is retained because omitting it would have yielded excessive approximations for the lowest concentrations, inconsistent with the uncertainty of the method.

For each lipid, a different level of accuracy and precision was obtained, and it has been summarized as follows:

• DSPC: the experimental concentrations were in good agreement with the theoretical values throughout the tested range. As expected, variability among replicates increases with decreasing concentration. Determination with PULCON at 1.03 mg/mL concentration demonstrated excellent accuracy as well, accompanied by even lower variability among replicates, highlighting the advantage of this method in terms of consistency.

• Cholesterol: experimental values close to the theoretical ones, with low variability between replicates, were obtained at concentrations of 5.56 and 0.56 mg/mL. These results were consistent across both internal and external standard methods. However, for the solution with the largest dilution of cholesterol (0.054 mg/mL), a value nearly half of what was expected was obtained. The discrepancy between the experimentally calculated concentration and the theoretical one is likely due to the weak signal intensity. However, it is possible to proceed to liposome analysis with sample concentrations higher than 0.56 mg/mL, which is typically the case for pharmaceutically relevant liposomes. Sufficiently accurate and precise results were, indeed, obtained above this concentration without needing to increase the number of scans or the acquisition times, which would have improved the signal-to-noise ratio and the accuracy for low concentrations.

• DMPE-PEG2K: the method exhibited good linearity even at the concentration of 0.023 mg/mL, thanks to the high number of protons (180) composing the polymeric repeating unit, which led to an intense NMR signal. Recovery values ranged from 91% to 95%, slightly lower than those observed for other lipids. However, considering the minimal absolute differences in concentration between theoretical and experimental values, these discrepancies do not impact the purpose of the analysis. It is important to note that the molecular weight of commercial pegylated lipids provided by the supplier is an average value due to the polydispersity of PEG. Therefore, calculating molarity concentrations requires careful consideration.

Overall, the experimental concentrations closely matched the theoretical values, though they tended to be slightly lower, possibly due to the purity of the materials.

NMR Quantification of Lipid Mixtures

Quantification of lipid concentrations was also performed on lipid mixtures with the same lipid composition used for liposome preparation to further confirm that the lipids can be analyzed simultaneously. The results are reported in Table . The recovery values for all lipids in the investigated mixtures range from 94.2% to 100.7% with an average % RSD equal to 6.8%, indicating accurate and precise quantification. This demonstrates that the simultaneous presence in solution of different lipids, mimicking liposome composition, does not impair the efficiency of the analysis.

2. Summary of NMR Results for Lipid Mixtures Tested with the Internal Standard Method.

lipid mix	lipid	theoretical (mg/mL)	experimental (mg/mL)	recovery %	% RSD replicates(%)
(A) DSPC/Chol (2:1)	DSPC	1.11	1.08	97.3	8.0
	cholesterol	0.56	0.54	96.4	6.0
(B) DSPC/Chol/DMPE-PEG2K (10.5:14.2:1)	DSPC	1.05	1.03	98.1	6.7
	cholesterol	1.42	1.39	97.9	3.9
	DMPE-PEG2K	0.100	0.098	94.2	9.1
(C) DOPC/Chol (4:1)	DOPC	2.96	2.98	100.7	8.0
	cholesterol	0.76	0.72	94.7	6.0

^aThe ratio of experimental to theoretical concentrations, expressed as a percentage.

^bThe ratio between the standard deviation and the mean, expressed as a percentage, calculated from the three replicates of each sample.

^cAn additional decimal is retained because omitting it would have yielded excessive approximations for the lowest concentrations, inconsistent with the uncertainty of the method.

With the aim of comparing the quantification using an internal or external standard compound, the lipid mix (D) was prepared to test the PULCON method. The same signals used for lipid integration with the method of an internal standard were employed here. As shown in Table , the experimental concentrations are in agreement with the theoretical values. Moreover, the variability among the replicates is very low, demonstrating the advantageous reproducibility of this method. Overall, the outcome matched that of the standard method, demonstrating that both approaches are fit-for-purpose for accurate quantification.

3. Summary of NMR Results for Lipid Mix (D) Obtained with the PULCON Method.

lipid mix	lipid	theoretical (mg/mL)	experimental (mg/mL)	recovery %	% RSD replicates (%)
(D) DSPC/Chol/DMPE-PEG2K (54.6:83.5:1)	DSPC	3.71	3.73	99.7	2.0
	cholesterol	5.67	5.78	98.2	0.3
	DMPE-PEG2K	0.0679	0.0677	99.4	1.7

^aThe ratio of experimental to theoretical concentrations, expressed as a percentage.

^bThe ratio between the standard deviation and the mean, expressed as a percentage, calculated from the three replicates of each sample.

^cAn additional decimal is retained because omitting it would have yielded excessive approximations for the lowest concentrations, inconsistent with the uncertainty of the method.

NMR Quantification of Liposomes

Liposomes with the same lipid composition analyzed in the previous section were formulated. Successful liposome formulation was confirmed by using dynamic light scattering (DLS) analysis (Table S1). To avoid quantification inaccuracies due to the rapid relaxation of high molecular weight liposomes or lipid micelles in water, liposome solutions were dried and resuspended in CDCl₃ for liposome disruption prior to analysis. The experimental concentrations obtained with NMR are compared to the theoretical concentrations of the formulation and are presented in Table . Theoretical liposome concentrations are calculated under the assumption that no losses occurred during preparation. The experimental analysis aimed to evaluate whether such losses happened. For all formulations, the experimental concentrations were below the theoretical values, indicating losses during the formulation steps. However, the recovery ranges between 89% and 99%, suggesting that the extent of the losses was low, demonstrating good efficiency of the manufacturing process.

4. Analysis of Liposome (A–C) Solutions with NMR.

liposomes	lipid	theoretical concentration (mg/mL)	experimental concentration (mg/mL)	recovery %	NMR mass ratio
(A) DSPC/Chol (2:1)	DSPC	1.20	1.13	94.2	1.9
	cholesterol	0.60	0.62	103.3	1.0
(B) DSPC/Chol/DMPE-PEG2K (10.9:14.6:1)	DSPC	1.09	1.00	91.7	10.5
	cholesterol	1.46	1.31	89.7	13.6
	DMPE-PEG2K	0.100	0.0096	96.0	1.0
(C) DOPC/Chol (4:1)	DOPC	3.18	2.98	93.7	3.9
	cholesterol	0.79	0.76	95.9	1.0

^aThe ratio of experimental to theoretical concentrations, expressed as a percentage.

^bAn additional decimal is retained because omitting it would have yielded excessive approximations for the lowest concentrations, inconsistent with the uncertainty of the method.

^cThe experimental NMR mass ratio of the lipids in each formulation.

Comparison of NMR and UPLC Performance in Lipid Quantification

UPLC analysis of a lipid mixture containing DSPC, cholesterol, and DMPE-PEG2K, at concentrations on the same order of magnitude as those tested with NMR, was first performed to assess the accuracy of the method using known concentrations. Experimental concentrations were in good agreement with theoretical values with deviations within 5% of the total (Table S2).

Lipid quantification of liposomes was conducted both on the solutions immediately after formulation and following the drying and resuspension steps required for NMR analysis. No significant variations were observed between the two conditions, indicating that no loss of material occurred during the drying and resuspension processes (Table S3).

The results obtained by UPLC-ELSD were compared with those of NMR on the same samples, and the recovery values are shown in Figure . The theoretical concentrations were calculated under the assumption that no lipid losses occurred throughout all formulation steps. The experimental concentrations derived from both NMR and UPLC-ELSD showed notable but slight differences, within a 7% variance, except for DSPC in liposomes (A), which shows a 10% difference. For liposomes (A) and DSPC in liposomes (B), the experimental values range between values near to 100% of the recovery. For the other components in liposome formulations (B) and (C), there is a good agreement between the two techniques with experimental concentrations being lower than theoretical values, confirming potential modest losses during the formulation process.

2.

NMR vs UPLC-ELSD lipid recovery. The percentage recovery of each component of liposome formulations is shown, indicating the experimental concentration as a percentage of the theoretical concentration (i.e., the expected lipid concentration in the absence of loss). The results obtained using NMR (displayed with black bars) and UPLC-ELSD (displayed with red bars) are compared. Error bars represent the relative standard deviation among the replicates. For liposomes (A), DPSC showed the highest difference between NMR and UPLC-ELSD results, with values close to theoretical concentrations, while for cholesterol, the experimental concentration was slightly higher than the theoretical value with both techniques. Overall, the findings suggest no significant losses over the formulation. DSPC in liposomes (B) is close to theoretical values with both techniques. For the other components in liposome formulations (B) and (C), the two techniques show a good agreement, with experimental concentrations being lower than the theoretical values. This suggests that there may be modest losses occurring during the formulation process.

Overall, both NMR and UPLC-ELSD methods have demonstrated high accuracy and precision, yielding experimental concentration values close to theoretical values in standard solutions with low variability among replicates. However, there are differences between the methods that should guide the choice of technique on a case-by-case basis. The main advantages of NMR lie in the use of a single reference compound as opposed to the time-consuming UPLC requirement of the calibration curve construction with each lipid of interest. NMR signal integrals are intrinsically proportional to the concentration of the nucleus under examination and rely on a standard (DMF in this work), which can be changed as long as it ensures a high degree of purity, chemical compatibility, and an isolated chemical shift and is paired with a sufficiently long interscan delay to ensure complete relaxation of the protons in both the reference and the samples. The advantage of using a single standard potentially applicable to the analysis of all lipids eliminates the need for highly pure certified standards for each lipid analyzed, which can be demanding in terms of both cost and effort, especially in early development phases. Additionally, as we have demonstrated, the use of an external reference with the PULCON method presents significant benefits; indeed, using an external standard abrogates the manipulation and variability associated with the introduction of an internal standard. Moreover, a single NMR tube containing the reference can be used for the quantification of lipids across different formulations in a single session. On the other hand, UPLC-ELSD offers advantages primarily related to sensitivity. Indeed, we observed a decrease in NMR accuracy for cholesterol concentrations close to 0.05 mg/mL, although this issue could be partially mitigated by increasing the number of scans at the cost of longer acquisition times. However, the sensitivity issue does not affect the pegylated lipid using NMR due to the enhanced signal from the repeating protons along the polymer. UPLC faces challenges when approaching the detection limit at the tested concentration of 0.10 mg/mL. Moreover, although UPLC-ELSD does not necessarily require liposomes to be disintegrated prior to injection into the column, drying and dissolving them in an organic solvent such as CDCl₃ was essential for NMR analysis. This step ensured that no supramolecular lipid structures formed, which could otherwise lead to quantification inaccuracies due to fast relaxation of protons and broadening of signals.

Conclusions

We demonstrated the successful application of NMR spectroscopy for lipid quantification in liposomal formulations. By using DMF as an internal standard, the method exhibited high levels of accuracy, precision, and reproducibility. Concurrently, the PULCON NMR method delivered results consistent with the standard approach, demonstrating enhanced reproducibility due to the use of a single common external reference. This streamlined approach is particularly advantageous for industrial applications that require rapid analysis of numerous batches. Notably, quantitative NMR benefits from the ease of access to commercially available standards. Only a single compound can be employed as a universal standard to quantify all lipid species, simplifying the analytical workflow in comparison to chromatographic methods.

This study focused on DSPC, DOPC, cholesterol, and DMPE-PEG2K as model lipids, which represent typical components of liposomal formulations. However, the developed methodologies are adaptable to other lipids by identifying isolated representative peaks for integration during quantification. The compatibility of the method with other classes of biomolecules present in the formulations, such as cargo, should be verified by ensuring that there is no significant overlap between the signals of these molecules and those used for the integration of lipids.

Comparative analyses with UPLC-ELSD revealed that both techniques are highly effective and accurate for lipid quantification in liposomes, with each method offering distinct advantages. The choice between NMR and UPLC-ELSD should be guided by the specific composition and concentration of the lipids being analyzed. When feasible, the combination of these techniques provides a more comprehensive and robust analytical framework.

Additionally, this method, optimized here for liposomes, can be readily applied to other lipid-based delivery systems, such as lipid nanoparticles. These nanoparticles, which often contain similar lipid compositions, are gaining significant attention as vehicles for mRNA-based vaccine delivery, further underscoring the versatility and relevance of the methods described for the broader pharmaceutical industry.

Methods

Materials

All reagents were purchased from common commercial sources and were used without additional purification. Distearoylphosphatidylcholine (DSPC), cholesterol, 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), and chloroform-d 99.8% atom enrichment were purchased from Sigma-Aldrich. DMPE-PEG2K was purchased from Avanti Polar. Ethanol was purchased from Merck. HyPure WFI quality water was purchased from Cytiva. N,N-Dimethylformamide anhydrous 99.8% (DMF) and trifluoroacetic acid (TFA) were purchased from Sigma-Aldrich. Isopropanol was purchased from Carlo Erba.

All the volumes with CDCl₃ as a solvent were handled with glass Hamilton precision syringes of 1 mL, 500 μL, and 100 μL capacity. Plastic pipet tips were not used because of potential plasticizer leaks.

Caution! CDCl₃ is a volatile halogenated solvent that is recognized as potentially carcinogenic and poses significant health risks upon inhalation or skin contact in addition to potential environmental impacts. Similarly, DMF is a polar aprotic solvent known for its reproductive toxicity and other health hazards. Both chemicals must be handled in a well-ventilated fume hood with appropriate personal protective equipment.

Lipid Standard Preparation

Solutions of each lipid at a known concentration were prepared in CDCl₃. DOPC was not considered in this step since the functional group used for the integration of its signal is the same as that of DSPC, without differences in chemical shift and intensity. Solutions were vortexed to ensure the complete dissolution of lipids. Each lipid was analyzed at three concentrations: (i) the concentration of the stock solution and (ii) after 1:10 and (iii) 1:100 dilutions. The intermediate concentration for each lipid was on the same order of magnitude as the lipid concentration in the final liposome formulation. Lipid mix solutions were prepared, aiming to prepare mixtures representative of the composition and concentration of each liposome formulation analyzed. Refer to Tables S5–S8 for detailed preparation.

Liposome Solution Preparation

Liposomes were prepared by using microfluidic mixing with the NanoAssemblr Ignite nanoparticle formulation system. For each liposome formulation, the organic and aqueous phases were prepared as detailed in Table . Lipids were dissolved in ethanol for the organic phase, while PBS pH 7.4 was used for the aqueous phase.

5. Lipid Composition of Liposome Formulations and Mixing Parameters.

liposome	lipid composition (w/w)	DSPC or DOPC concentration (mg/mL)	cholesterol concentration (mg/mL)	DMPE-PEG2Kconcentration (mg/mL)	flow rate ratio (AP/OP)	total flow rate (mL/min)
(A)	DSPC/cholesterol (2:1)	1.20	0.60	N/A	3:1	10
(B)	DSPC/cholesterol/DMPE-PEG2k (10.9:14.6:1)	1.09	1.46	0.1	2:1	10
(C)	DOPC/cholesterol (4:1)	3.18	0.79	N/A	2:1	2

^aAP = aqueous phase and OP = organic phase.

The resulting mixtures appeared as clear solutions, with the liposome (C) exhibiting a more opaque, whitish appearance likely due to its higher concentration. Each formulation was exchanged with Hypure WFI quality water using PD10 columns to remove ethanol.

Liposomes must be disrupted to enable accurate NMR quantification, free from proton relaxation issues caused by their large size. For each liposome solution, 800 μL was placed into glass vials, prepared in four replicates, two for NMR and the other pair for UPLC analysis. The resulting 12 vials, along with two additional vials each containing 800 μL of H₂O as controls, were placed in a SpeedVac set to 45 °C at a pressure of 1 mTorr for 3 h. At the end of this process, all water had evaporated, leaving only the dried lipid residue in the liposome vials. Finally, the content of each vial was resuspended in 800 μL of CDCl₃ for NMR analysis.

NMR Sample Preparation

DMF was chosen as an internal reference due to its inertness, miscibility in organic solvents, and the fact that its chemical shift occurs in a spectral region where no other proton resonates. A pure (>99.5%) DMF solution was diluted to achieve a concentration in the tube comparable to that of lipids, aiming for a similar signal-to-noise ratio. The dilution process was designed to ensure that the addition of DMF constituted roughly 10% of the total volume in the NMR tube and to avoid many dilution steps to reduce the risk of imprecision. Specifically, 40 μL of DMF solution (MW = 73.09, density = 0.944 g/mL) was diluted to a total volume of 1.040 mL in CDCl₃ to achieve a final concentration of 497 mM. Each NMR tube (Wilmad, 5 mm, 500 MHz precision) was filled with 500 μL of the analyzed solution and 40 μL of a 497 mM DMF solution, resulting in a final DMF concentration of 36.796 mM. In contrast, for the PULCON method, the analyte tubes were prepared by adding 540 μL of the sample solution in CDCl₃, while the reference tubes were prepared by adding 40 μL of the DMF solution and 500 μL of CDCl₃ to achieve the same DMF concentration of 36.796 mM.

NMR Measurements

All NMR experiments were recorded at 298 K on a Bruker AVANCE NEO spectrometer, operating at 500 MHz, ¹H Larmor frequency, and 11.7 T, equipped with a quadrupole resonance QCI H/P/C/N Cryoprobe. For each sample, the following procedures were applied: locking to CDCl₃, manual matching and tuning, shimming, optimization of the ¹H pulse, and automatic receiver gain optimization. 1D ¹H zg experiments were acquired using 16 scans, 64 K data points, an offset of 4.674 ppm, and an acquisition time of 4.2 s. To correctly evaluate the interscan delay, 1D ¹H experiments were performed, gradually increasing such a delay. An interscan delay of 35 s was found sufficient to allow complete spin relaxation, since no increase in signal intensity was observed for longer delays.

NMR Spectra Analysis

All of the spectra were processed with a Bruker Topspin 3.5pl6 software package. Fourier transformation with an exponential function with line broadening equal to 5.00 Hz was used to process all of the spectra, followed by manual phase optimization and baseline correction. Integration of the signals of interest was performed in manual mode. For each lipid, one well-solved specific signal was identified recording the spectra of the single lipids (Figure S1) and integrated for the quantification, while the well-solved aldehyde proton peak of DMF was chosen for the integration of the reference (Figure ). Equation was applied for lipid quantification using the method with an internal standard:

$$[\mathrm{L}\mathrm{i}\mathrm{p}\mathrm{i}\mathrm{d}]=\left(\frac{\mathrm{I}\mathrm{l}\mathrm{i}\mathrm{p}\mathrm{i}\mathrm{d}}{\mathrm{I}\mathrm{D}\mathrm{M}\mathrm{F}}\right)/n\mathrm{H}\mathrm{l}\mathrm{i}\mathrm{p}\mathrm{i}\mathrm{d}\times 36.796\times (540/500)$$ 1

The unknown concentration of the lipid is calculated from the ratio of the specific integral of the lipid (Ilipid) and the integral of the aldehyde proton of DMF (IDMF) divided by the number of protons contributing to the lipid signal (nHlipid). This ratio is then multiplied by the molar concentration of DMF (36.796 mM) and further adjusted by the dilution factor of the sample (from 500 to 540 μL).

The PULCON method consisted of recording the spectra of both the reference and the sample on the same spectrometer using the same pulse sequence, the same delay (d ₁) to ensure complete proton relaxation, and the same receiver gain. Sample concentration was determined using eq below, which correlates the concentration (C) of the reference (r) and the sample of interest (x) with the integration of the specified resonance (I), the number of protons associated with that resonance (A), the sample temperature in Kelvin (T), the 90° pulse length (θ), and the number of scans (N).

$$Cx=Cr\frac{IxArTx\theta xNr}{IrAxTr\theta rNx}$$ 2

The molar concentrations calculated from NMR integrals were converted into milligrams per milliliter concentrations using the molecular weight (g/mol) of each component.

UPLC-ELSD Analysis

UPLC analysis was performed using a Waters Acquity CSH C18 column (2.1 mm × 50 mm, 1.7 μm) maintained at 55 °C with samples held at 20 °C. A 5 μL injection volume and consistent flow rate of 0.6 mL/min were employed over a run time of 13 min. The mobile phase consisted of 0.1% v/v trifluoroacetic acid (TFA) in water (phase A) and 0.1% v/v TFA in isopropyl alcohol (IPA) (phase B), with a gradient of solvent compositions spanning throughout the run. Evaporative Light Scattering (ELS) detection was configured with an evaporation temperature set at 50 °C, a data collection rate of 2 Hz, and a filter setting of 3.6.

Standard solutions containing DSPC, DOPC, cholesterol, and DMPE-PEG2K were prepared, showing good resolution of the components (Figure S2). Calibration curves for each lipid were generated in triplicate, originating from the standard solution across the concentration ranges relevant for lipid mixtures and liposome analysis, taking into account a 1:5 dilution factor applied to samples (Figure S3). A lipid mix solution containing DSPC, cholesterol, and DMPE-PEG2K dissolved in CDCl₃ at known concentrations (Table S9) was analyzed to test the accuracy and precision of the method. The solution was analyzed in triplicate after dilution 1:5 in absolute ethanol.

The liposome solutions were first analyzed in their formulation buffer and subsequently after drying, as described for NMR analysis, followed by resuspension in 800 μL of ethanol. In both cases, the solutions were diluted to a 1:5 ratio in absolute ethanol and thoroughly vortexed to ensure homogeneity. Each sample was prepared and analyzed in triplicate.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c09329.

• Standard lipid solution preparation, dynamic light scattering data for liposome formulations, NMR spectra of individual lipid solutions, UPLC-ELSD elution profile, calibration curves, and results for both lipid mix and liposome samples (PDF)

⊥.

Asparia Glycomics, Paseo Miramón 170, Donostia-San Sebastián, Gipuzkoa, 20014, Spain

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

The authors declare the following competing financial interest(s): This work was sponsored by GlaxoSmithKline Biologicals SA. Francesco Curr is a student at the University of Florence and participates in a post graduate studentship program at GSK. Silvia Martini, Massimiliano Biagini, Daniela Stranges and Maxime Denis are employed by the GSK group of companies. Ander Eguskiza Bilbao was employed by the GSK group of companies at the time the experimental work was performed, while he was employed by Asparia Glycomics during writing and review processes.