Reverse engineering of Onivyde^® – Irinotecan liposome injection

Abstract

Onivyde^® is an intravenous irinotecan liposomal injection approved by the FDA for the treatment of gemcitabine-refractory metastatic adenocarcinoma of the pancreas in combination with fluorouracil and leucovorin. In the Onivyde^® formulation, irinotecan is encapsulated in the inner compartment of the liposome using sucrose octasulfate as a trapping agent, and stabilized by a pegylated lipid membrane, resulting in prolonged circulation in the body. Due to its complex formulation design, there is limited information available regarding the critical quality attributes (CQAs) of Onivyde^® and suitable methods for evaluating these attributes. In this study, we have developed a series of analytical methods to comprehensively characterize Onivyde^®. These methods encompass particle size analysis, morphology and structure assessment, examination of physical and chemical properties, determination of drug and lipid contents, and evaluation of its release behavior in vitro.

1. Introduction

Onivyde^® (irinotecan liposome injection) is a liposomal formulation of the topoisomerase I inhibitor irinotecan indicated for the treatment of metastatic adenocarcinoma. The drug product was first approved by the FDA in October 2015 and garnered increasing sales (133 million dollars in 2022) (Ipsen, 2022). Prepared by intra-liposomal drug stabilization technique using sucrose octasulfate (SOS) as a trapping agent, irinotecan forms crystal bundles in the core with the phospholipid bilayer composed of 1,2-distearoyl-SN-phosphatidylcholine (DSPC), cholesterol and DSPE-PEG2k. The structure of pegylated liposomal irinotecan protects the drug substance from being metabolized to SN-38 by carboxylesterase (Milano et al., 2022). In addition, the PEG layer helps reduce the rapid clearance of the drug by the endoplasmic reticulum system and increases circulation time in the blood to enhance the passive targeting to tumor tissue. After the drug product crosses the blood vessel boundary and penetrates the tumor, the drug product is internalized by cells in the tumor and releases the active pharmaceutical ingredient (API) irinotecan. Then, irinotecan actives SN-38 that binds topoisomerase I and stabilizes the cleavable complex between topoisomerase I and DNA, resulting in DNA breaks, inhibition of DNA replication, and ultimately leads to cell apoptosis (FDA, 2015).

The product’s activity is dependent on characteristics such as composition, physical and chemical characteristics that drive release behavior and in vivo distribution. These characteristics are the end result of how the drug product was manufactured which is controlled by manufacturing parameters. Importantly, comprehensive data on liposome characterization is desired to provide understanding of what features are needed to ensure product quality, safety, and efficacy and, ultimately, positive patient outcomes (US Food and Drug Administration, 2022; US Food and Drug Administration, 2018).

In this study, we examined four batches of Onivyde^®, by performing a battery of characterization methods to better understand the product design and evaluate the critical quality attributes (Liu et al., 2020). Specifically, the morphology of Onivyde^® was observed using cryogenic transmission electron microscopy (cryo-TEM), and its size and size distribution were measured by ZetaSizer Nano ZS. The pH value of the Onivyde^® formulation was determined by using a pH meter. The phase transition temperature of the phospholipid bilayer was evaluated by nano-DSC. Ultra-performance liquid chromatography-evaporative light scattering detector (UPLC-ELSD), ultra-performance liquid chromatography-fluorescence detector (UPLC-FLR) and gas chromatography (GC) methods were established to quantify the concentrations of irinotecan, lipids (DSPC, cholesterol, DSPE-PEG2k) and other excipients in Onivyde^®. Small angle X-ray scattering (SAXS) was used to further evaluate the packing and uniformity of the nanostructure. Additionally, the in vitro drug release of Onivyde^® was studied to better understand crucial factors potentially impacting its in vivo behaviors.

2. Material and methods

2.1. Chemicals and reagents

Four batches of Onivyde^® (120518S, 200048A, 200147A, U16025) were purchased from the pharmacy department of the University of Michigan Health System. Irinotecan hydrochloride trihydrate (>98 %) was procured from AdooQ Bioscience. DSPC was purchased from Avanti. Cholesterol was purchased from Sigma-Aldrich. DSPE-PEG2k was purchased from NOF Corporation. Egg lysophosphatidylcholine (Lyso-PC, 830071P) was bought from Avanti. Dimethyl sulfoxide (DMSO) of GC grade, methanol, acetonitrile, isopropanol, methanol, acetonitrile of HPLC grade or grade were procured from Fisher Scientific. Poly-Prep chromatography columns were purchased from Bio-Rad Laboratories (Hercules, CA, USA). TOYOPEARL HW-55F resin in ethanol was purchased from Tosoh Bioscience. Spectra/Por^® dialysis membrane (regenerated cellulose) of molecular weight cut-off (MWCO) 8–10 kDa and Spectra/Por^® Float-A-Lyzer dialysis device of MWCO 300 kDa were bought from Spectrum Laboratories, Inc., Rancho Dominguez.

2.2. Particle size

Particle size and size distribution of liposomes were evaluated using dynamic light scattering (DLS) with a Malvern ZetaSizer Nano ZS. Approximately 40 μL of liposome formulation was placed in a disposable plastic micro cuvette, then placed inside the ZetaSizer instrument. Three measurements were performed for each sample at room temperature (25 °C) with a dispersant refractive index of 1.332 and a dispersant viscosity of 0.9024.

2.3. Particle morphology

The morphology of Onivyde^® liposomal formulation was determined by cryo-TEM (Jeol JEM 1400). Onivyde^® were loaded and tested without dilution, using glow-discharged Quantifoil copper grids with Holey carbon films (R 2/1, 200 mesh) prepared by a glow-discharged (EMS 150 T S glow discharger). Grids were then mounted in Vitrobot (Thermo Scientific) in which the grid was loaded with 3.0 μL of sample and then blotted for 10 s at 4 °C in 100 % humidity to remove excess liquid. After blotting, the grid was immediately plunged into a liquid ethane bath cooled by liquid nitrogen, maintained at − 175 °C to vitrify the material. The sample was then transferred to a liquid nitrogen bath and stored until imaging. When imaging, TEM grids were placed into the sample cartridge and transferred to the cryo-TEM for analysis. All images were acquired on a Talos Artica (Thermo Fisher Scientific, USA) operating at an acceleration voltage of 200 keV. Cryo-TEM images were processed in ImageJ software.

2.4. SAXS

SAXS data were collected in the high throughput mode (HT-SAXS) using the Advanced Light Source SIBYLS beamline 12.3.1 at the Lawrence Berkeley National Laboratory (CA, USA) (Dyer et al., 2014; Putnam et al., 2007). X-ray wavelength was set at λ = 1.216 Å, and the sample-to-detector distance was 2070 mm, resulting in a scattering vector, q, ranging from 0.01 Å-1 to 0.45 Å-1. The scattering vector is defined as q = 4πsinθ/λ, where 2θ is the scattering angle. Experiments were performed at 20 °C as described. Briefly, the sample was exposed for 10 s with the detector framing at 0.3 s to maximize the signal. All scattering curves presented in this manuscript were background subtracted.

2.5. pH measurement

The pH value of Onivyde^® was measured using a potentiometric standard glass electrode method at room temperature after the pH meter was calibrated with standard buffers at pH 4.01, pH 7.00, and 10.01. The glass electrode was immersed in the liposome formulation to obtain pH readings. Values were recorded at a minimum of three measurements for each batch.

2.6. Phase transition temperature

Onivyde^® was diluted fivefold with a reference solution, a pH 7.4 buffer containing 4.05 mg/mL 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) and 8.42 mg/mL sodium chloride. The resulting solutions were degassed and subjected to colorimetry analysis by a NanoDSC (TA Instruments, New Castle, DE, USA). The analysis involved heating within the temperature range of 10 °C to 90 °C at 1°C/min, and the results were subsequently analyzed using TATRIOS software.

2.7. Irinotecan analysis in Onivyde^®

2.7.1. Quantification of total irinotecan

Irinotecan stock solution was prepared using DMSO. The carboxylate form of irinotecan was obtained by diluting the stock solution with a mixture of acetonitrile and 0.02 M borate buffer (pH 9.0) in a 50:50 (v/v) ratio. The lactone form was prepared by diluting the stock solution with a mixture of acetonitrile and 0.01 M citric acid (pH 3.0) in a 50:50 (v/v) ratio (Yang et al., 2005). A mixed solution was prepared by combining equal volumes of the carboxylate and lactone form solutions. A series of standard solutions with various concentrations (0.1, 1.0, 10, 50, 100, and 200 μg/mL) were diluted using the mobile phase.

Onivyde^® was diluted with methanol and filtered through a 0.22 μm filter. Irinotecan quantification was performed using an Ultra-Performance Liquid Chromatography (UPLC) system, consisting of an Acquity Quaternary Solvent Manager, a Sample Manager-FTN, a column manager, and a fluorescence detector (Waters, Milford, MA, USA). Separation of irinotecan was achieved using an Acquity UPLC BEH C18 column (1.7 μm, 2.1 × 100 mm, Waters, Milford, MA, USA) with a gradient elution using 10 mM KH₂PO₄ at pH 6.4 as solvent A and acetonitrile as solvent B, at a flow rate of 0.5 mL/min. Mobile phase B changed from 20 % to 50 % in 2 min and returned to 20 % at 3.5 min (Table 1) The total run time was 5 min. Irinotecan concentration was detected by the fluorescence detector with excitation at 375 nm and emission at 500 nm. Quantification of irinotecan was performed for four batches of Onivyde^®, and the experiments were conducted in triplicate.

Table 1.

Gradient elution for irinotecan quantification.

Time/min	A%	B%
0	80	20
2	50	50
3.5	80	20
5.0	80	20

2.7.2. Determination of encapsulated irinotecan

To determine the encapsulated irinotecan concentration in Onivyde^®, poly-Prep columns were washed with methanol and water. TOYOPEARL HW-55F resin (200 μL) was added to the column and installed on a vacuum manifold processing station. Excess liquids were drained before loading the samples and the resin surface was kept flat. After pre-treatment, 5 μL of Onivyde^® containing 21.5 μg of irinotecan was added to the columns. The preparation was first washed with 300 μL of deionized water three times, followed by washing with 300 μL of a mixed solution of acetonitrile and methanol (50:50, v/v) containing 0.1 % formic acid, also three times. The collected fractions of water and organic solution were individually dried using nitrogen and then redissolved in methanol for detection by UPLC-FLR.

2.8. Quantitative analysis of DSPC, cholesterol and DSPE-PEG2k

The powder of DSPC, cholesterol, or DSPE-PEG2k was weighed and dissolved in methanol. Subsequently, a series of standard solutions were prepared by diluting with methanol accordingly. For the analysis, four batches of Onivyde^® with different batch numbers were diluted tenfold to determine DSPE-PEG2k and one hundredfold to determine DSPC and cholesterol. Each experiment was conducted in triplicate.

The quantification of DSPC, cholesterol and DSPE-PEG2k were determined by UPLC-ELSD. The UPLC system included an Acquity Quaternary Solvent Manager, a Sample Manager-FTN and a column manager, and an ELSD detector, (Waters, Milford, MA, USA). Separation of the lipids was achieved using a Waters Xbridge C18 column (3.5 μm, 2.1 × 50 mm, Waters, Milford, MA, USA) with a gradient elution shown in Table 2. Solvent A consisted of MeOH:10 mM ammonium acetate (pH 6.4) in a 40:60 (V: V) ratio, while solvent B was isopropanol: solvent A in a 90:10 (V:V) ratio. The flow rate was set at 0.3 mL/min. Lipid concentrations were detected by ELSD with gain set to 150, gas pressure at 40 psi, nebulizer at 45 %, and a drift tube temperature of 55 °C.

Table 2.

Gradient elution for lipids quantification.

Time/min	A%	B%
0	100	0
1.0	100	0
3.0	20	80
3.5	20	80
5.0	0	100
10.0	0	100
11.0	100	0

2.9. Analysis of other ingredients

2.9.1. Quantitative analysis of lyso-PC

Egg Lyso-PC was weighed and dissolved in methanol to prepare a 1 mg/mL stock solution. Various standard solutions with concentrations of Lyso-PC at 100, 50, 10, 5, 1, 0.1, and 0.01 ng/mL were prepared by diluting the stock solution with methanol. Quantification of Lyso-PC was performed using the same UPLC/ELSD method with other lipids as described in 2.8.

2.9.2. Quantitative analysis of triethylamine

Triethylamine (TEA) in Onivyde was identified and quantified using gas chromatography (GC) on a Trace 1310 gas chromatograph (Thermo Fisher Scientific Inc., Waltham, MA, USA) equipped with a TG-624 column (30 m × 0.53 mm × 3.00 μm, Thermo Fisher Scientific Inc.). Nitrogen was utilized as the carrier gas at a flow rate of 2 mL/min. A series of TEA standard solutions (500 ppm, 100 ppm, 10 ppm, 1 ppm, 0.1 ppm, and 0.01 ppm) were prepared by dissolving and diluting TEA in DMSO. For analysis, 50 μL of Onivyde was diluted in DMSO, sealed in GC headspace vials, and subjected to GC analysis. TEA quantification was performed using external standards. Each sample was agitated at 80 °C for 20 min, after which 1 mL of the headspace sample was injected into the GC inlet in split mode. The inlet temperature was maintained at 140 °C, with a split flow of 40 mL/min and a split ratio of 20. The GC conditions were as follows: an initial temperature of 40 °C held for 15 min, followed by a 10 °C/min ramp to 240 °C, which was maintained for 2 min. The detector temperature was set at 240 °C, with an airflow of 350 mL/min, a makeup gas flow of 25 mL/min, and a hydrogen flow of 35 mL/min.

2.10. Release kinetics in vitro

Drug release in PBS buffer

The drug release study of Onivyde^® (Batch # 200048A) was conducted using a Float-A-Lyzer Dialysis Device (MWCO 300 kD) in a release medium consisting of phosphate-buffered saline (PBS) at pH 8.0, with varying concentrations of NH₄HCO₃, at 55°C. 50 μL of Onivyde^® was diluted with each release medium to a final volume of 1 mL. The solution was then transferred to the inner chamber of the Float-A-Lyzer device, which was placed in 39 mL of the external medium. At specified time intervals (0.5, 1, 2, 4, 8, 12, and 24 h), 1 mL of the release medium was sampled from external medium and analyzed using a fluorescence detector. To maintain a consistent volume, an equal amount of fresh release medium was replenished after each sampling.

2.11.1. Drug release in 50 % human plasma

The stability of Onivyde^® (Batch # 200048A) in plasma was investigated by measuring the release of irinotecan from liposomes at 37 °C using a dialysis method with shaking at 80 rpm. 5 μL of Onivyde^® was diluted 40 times in diluted human plasma (50 % in normal saline) and then added to a dialysis bag (Spectra/Por, 8–10 kDa MWCO). The bag was subsequently immersed in 10 mL of diluted human plasma (50 % in normal saline). At specified time intervals (0.5, 1, 2, 4, 8, 12, 24, 48, 72, 96, and 120 h), 10 mL samples were drawn from the dialysis setup, and an equal volume of fresh media was immediately replenished. Standard irinotecan solutions were prepared at concentrations of 0, 0.1, 0.5, 1.0, 5.0, 10.0, 25.0, and 50.0 μg/mL. Samples were diluted and precipitated with methanol before mixing and centrifuging at 15000 rpm for 5 min. The supernatant was measured using UPLC-FLR method. The injection volume of samples was 10 μL.The cumulative release (%) was calculated by dividing the amount of irinotecan in the release samples at each time point by the total amount of irinotecan initially added.

$$Cumulativerelease\left(\mathrm{\%}\right)=\frac{\mathit{Released}\mathit{irinotecan}}{\mathit{Total}\mathit{irinotecan}}\times 100$$

2.12. Statistical analysis

Statistical analysis was conducted by ordinary one-way ANOVA multiple comparisons using GraphPad Prism 8.0.2 software. The level of significance was set at a 95 % confidence interval (α = 0.05).

3. Results and discussion

In this study, we focused on the characterization, quantitative methods, and drug release behavior of Onivyde^® to establish a set of potential evaluation parameters that could be used to assure the quality of the drug product. In addition, four batches of Onivyde^® were analyzed to assess sample-to-sample variability in the marketed products we obtained for this study.

3.1. Particle size and external pH

In Fig. 1 and Table 3, Onivyde^® from different batches exhibited an intensity-weighted mean particle size of 110 nm, which matched the label value on the package of Onivyde^®. Furthermore, the particle size distribution of the product had a notably narrow range, indicating a high degree of homogeneity. This observation was further supported by the low value of the polydispersity index (PDI), e.g., 0.03–0.16. The mean external pH value of the Onivyde^® buffer in Table 3 fell within the range of 7.1 to 7.2, which matched the label value.

Fig. 1.

Particle size distribution of Onivyde^®. The particle size distribution of Onivyde^® was measured in triplicate using a Zetasizer Nano instrument. Representative curve of size distribution of Onivyde^® from each batch was generated using GraphPad Prism 8.0.2 software.

Table 3.

Physiochemical and structure parameters of four batches Onivyde^®. Intensity size and PDI were obtained by DLS. Diameter polar and diameter equatorial were obtained by model fitting of SAXS data. The data are presented as mean ± SD, n = 3.

Batch No.	Published Values	120518S	200048A	200147A	U16025
Expire date		Feb. 2022	May 2022	May 2023	Apr. 2024
Date of analysis		Nov. 2021	Nov. 2021	Jun. 2022	Sep.2022
Intensity size (nm)	110	110.0 ± 1.8	112.9 ± 2.5	113.2 ± 1.0	112.5 ± 0.8
PDI		0.05 ± 0.01	0.05 ± 0.01	0.16 ± 0.02	0.03 ± 0.02
External pH	7.2	7.12 ± 0.01	7.10 ± 0.01	7.17 ± 0.01	7.12 ± 0.01
Diameter polar (nm)		63.0 ± 3.2	68.6 ± 5.5	58.1 ± 4.7	68.0 ± 3.3
Diameter equatorial (nm)		105.3 ± 0.9	109.4 ± 0.7	103.3 ± 12.2	110.1 ± 1.3
Aspect ratio		1.67 ± 0.08	1.60 ± 0.14	1.78 ± 0.17	1.62 ± 0.08

3.2. Particle morphology

In the cryo-TEM images (Fig. 2), Onivyde^® appeared as either a prolate spheroid or a spheroid, with irinotecan forming an electron- dense precipitate in the core of the liposome. Unlike the traditional spherical liposome structure, this prolate spheroid (indicated in green arrow) was primarily attributed to the formation of long bundles of fibers in the core of Onivyde^® during the encapsulation of irinotecan using the gradient remote loading method. Additionally, the appearance of a round shape (indicated in yellow arrows) in the cryo-TEM view may be a result of the isotropic orientation of particles on the TEM grid. In 2D projection images, particles oriented in the direction of the electron beam may appear circular, but they were, in fact, prolate spheroids in three-dimensional space (Noble et al., 2019).

Fig. 2.

Cryo-TEM images of four batches of Onivyde^®. Green arrows indicate the prolate spheroid with crystal bundle inside. Yellow arrows point out round shape particles. White arrows refer to complex structures with both liposome and drug crystal inside. Red arrows indicate symmetric multilamellar vehicles. Blue arrows point out empty liposomes. A1-A2, B1-B2, C1-C2 and D1-D2 are representative cryo-TEM images of batch #120518S, batch #200048A, batch #200147A, batch #U16025 of Onivyde^® respectively. Scale bar, 100 nm.

Meanwhile, certain particles manifested as empty liposomes (indicated by blue arrows), and a few exhibited a bilamellar structure (highlighted by red arrows). A considerable number of liposomes underwent substantial deformation due to co-existing drug crystal bundle (depicted by white arrows) and another liposomal structure within. Intriguingly, in certain distorted liposomes, over two clusters of drug crystal bundles from distinct directions were observed.

The structure complexity of Onivyde^® is attributed to the formulation method and low content of DSPE-PEG2k. Nele, et al., (Nele et al., 2019) have reported that the addition of 0.1 mol% PEG (where the PEG was expected to be in the mushroom configuration) could change the proportion of liposomal lamellarity from 40 to 0 vol% because the mushroom configuration of PEG hindered lipid overpacking during hydration. In addition, Nele et al., noted that the reverse-phase evaporation vesicle method produced better unilamellar vesicles than the freeze–thaw cycles method and shaker agitation method.

Analysis of the cryo-TEM image data of over 500 particles showed that the membrane thickness of liposome was about 4.9 ± 0.5 nm (Fig. 3). The Feret-Max of four batches of Onivyde^® had a range from 71 nm to 307 nm with an average value of 109.3 ± 17.7 nm, 113.4 ± 17.5 nm, 112.4 ± 5.7 nm, and 110.0 ± 15.6 nm, respectively. The Feret-Min had a range from 31 nm to 130 nm, with an average value of 60.4 ± 13.2 nm, 56.1 ± 13.5 nm, 63.5 ± 4.9 nm, and 71.5 ± 7.8 nm correspondingly. Feret-Max and Feret-Min measurements of Onivyde^® did not differ across the samples measured in this study. The aspect ratio (Feret-Max/Feret-Min) fell in the range of 1.62–1.78 nm and showed no difference among different samples.

Fig. 3.

Cryo-TEM analysis of 4 batches of liposomal Onivyde^® formulations. A1-A4, Feret-Max distribution of 1st to 4th batch of Onivyde^® respectively. B1-B4, Feret-Min distribution of 1st to 4th batch of Onivyde^® respectively. C, cryoTEM analysis. The number of observed and analyzed particles is over 500. All values are presented as mean ± SD(n ≥ 3).

While comparing the Feret-Max of Onivyde^® from cryo-TEM data with the intensity size of Onivyde^®, there was no meaningful difference, indicating that the intensity size may primarily reflect the Feret-Max of Onivyde^®. In addition, the percentage of multi-lamellar (including bi- lamellar) of four batches of Onivyde^®, was 28.5 %, 29.2 %, 25.7 % and 36.2 %, respectively.

Overall, the formulation appeared less homogeneous compared to coffee-bean shaped Doxil (Wibroe et al., 2016) and displayed a subtle complexity in the nanostructure. The potential influence of the structure complexity on the efficacy of Onivyde^® is still unknown.

3.3. SAXS

SAXS is an analytical technique used in the characterization of liposomes and other nanoparticles and can determine the size and shape of liposomes in solution. By analyzing the scattering patterns of X-rays, SAXS data provides information about the dimensions of liposomes, including the diameter, lamellarity, core–shell structure and beyond. These measurements are potentially important for understanding the physical characteristics of liposomes and their quality.

As depicted in Fig. 4, the four batches of Onivyde^® exhibited similar patterns in the SAXS results. Upon modeling and fitting using Sasview with an ellipsoid/ellipsoid model, the polar and equatorial diameters were determined as shown in Table 3. The diameter polar of four batches of Onivyde^® fell within the range of 58.1–68.6 nm, with individual value 63.0 ± 3.2 nm, 68.6 ± 5.5 nm, 58.1 ± 4.7 nm and 68.0 ± 3.3 nm, respectively. And the diameter equatorial fell within the range of 103.3–110.1 nm, with individual value 105.3 ± 0.9 nm, 109.4 ± 0.7 nm, 103.3 ± 12.2 nm and 110.1 ± 1.3 nm, respectively. The values of diameter polar and diameter equatorial showed no difference among different batches.

Fig. 4.

Small angle X-ray scattering of Onivyde^®. X-ray scattering data were collected from four batches of Onivyde^®: (A), #120518S, (B), #200048A, (C), #200147A, and (D), #U16025, and the background-subtracted radially integrated scattering intensity is depicted. Data were obtained from triplicate measurements (n = 3).

Comparison between the cryo-TEM data with the SAXS data indicated that the Feret-Max obtained from cryo-TEM images and the diameter equatorial from SAXS data in all batches of Onivyde^® had no difference. The Feret-Min from cryo-TEM images and diameter polar from SAXS data also had no meaningful difference. Because cryo-TEM images can characterize the structural details of Onivyde^® and SAXS data used a different physical principle to assess the same characteristics of the liposomes, cryo-TEM and SAXS were orthogonal approaches to evaluate the quality of Onivyde^®. Thus, the practical size values from cryo-TEM images contributed to the precise model fitting of SAXS data.

3.4. Phase transition temperature

DSPC, DSPE-PEG2k and cholesterol form the lipid bilayer of Onivyde^® and the lipid phase plays a role in stability, drug release behavior and ultimately the efficacy of the product. Typically, phospholipids like DSPC and DSPE-PEG2k have gel-phase (solid) and liquid-crystalline phase transition temperatures, and cholesterol can modulate the phase transition temperature of phospholipids by making membranes more fluid at lower temperatures and less fluid at higher temperatures.

The phase transition profiles of four batches of Onivyde^® were similar, as indicated in Fig. 5, with a platform temperature range between 52 to 57 °C. Upon fitting with Gaussian models, three phase transition peaks emerged in each batch of Onivyde^®, in which, the amplitudes of batch #120518S, batch #200147A, and batch #U16025 shared a similar tendency, namely 55 °C > 56 °C > 53 °C and batch #200048A featured slightly higher individual T_m values of 54 °C, 56°C, and 57 °C, with tendency 56 °C > 57 °C > 54 °C.

Fig. 5.

The phase transition temperature of Onivyde^® (#120518S (A), #200048A (B), #200147A (C), # U16025 (D)).

In sum, the T_m of all four batches of Onivyde^® ranged from 52 °C to 57 °C, where the phase transition temperature of DSPC is 55 °C and that of PEG (350–5000) is 55–64 °C (Kenworthy et al., 1995). Additionally, the broad transition peak might suggest a non-isotropic transfer of heat across the liposome as all the chains are not arranged in a uniform array (Malekar et al., 2016), in which batch #200048A may show a slightly different transition behavior of phospholipids reflecting the complexity of Onivyde^®’s structure and calling for deep investigation of interactions among drug and lipids.

3.5. Analysis of irinotecan concentration

As shown in S1, the representative chromatogram of carboxylate form of irinotecan appeared at 1.3 min and the lactone form of irinotecan eluted at 2.7 min. In Table 4, the total concentration of irinotecan in Onivyde^® was 4.3 ± 0.1 mg/mL, which matched the labeled value. The percentage of encapsulated irinotecan fell within the range of 96.2 % to 98.0 % and the percentage of free drug in Onivyde^® was between 1.9 % to 3.5 %. Although the free drug level in the 3rd batch of Onivyde is slightly higher than other batches, the free irinotecan content 3.5 ± 0.2 % (<5.0 %), is still within the acceptable threshold in the liposome formulation.

Table 4.

Total irinotecan and encapsulated irinotecan contents in four batches of Onivyde^®. All values are presented as mean ± SD, n = 3.

Batch No.	Total irinotecan (mg/mL)	Free irinotecan (%)	Encapsulated irinotecan (%)
Published Values	4.3	-	-
120518S	4.3 ± 0.1	1.95 ± 0.01	97.8 ± 0.4
200048A	4.2 ± 0.1	2.06 ± 0.39	98.0 ± 0.3
200147A	4.3 ± 0.1	3.48 ± 0.24	96.2 ± 0.2
U16025	4.3 ± 0.1	1.91 ± 0.07	98.1 ± 2.6

The gradient remote loading method contributed to the high encapsulation efficiency of irinotecan. During manufacturing, TEA-SOS in the core of liposome dissociated into [8 TEA + 8H^± + SOS⁸⁻ ], of which lipid soluble TEA freely diffused across the phospholipid bilayer. The remaining protons left in the core resulted in acidification of the intra-liposomal environment (Patel and Patel, 2020). Meanwhile, irinotecan HCl in the external medium neutralized with TEA and converted to free base irinotecan. Irinotecan free base then diffused through the lipid bilayer into the inner liposome core and was subsequently protonated in the core. Protonated irinotecan further bound to SOS⁸⁻ via electrostatic interaction and formed a crystalline precipitate which improved the encapsulation of the drug (Lewis and Hall, 2019). Polyanion SOS with 8 sulfates could bind to multiple irinotecan and facilitate inter-fiber crosslinking which was indispensable to the final sterically stable liposomal structure (Liu et al., 2016).

3.6. Quantitative analysis of DSPC, cholesterol and DSPE-PEG2k

DSPC, cholesterol, and DSPE-PEG2k were effectively separated in the gradient elution program, with retention times of 6.33 min, 5.56 min, and 4.95 min, respectively (Fig. 6). The concentration of DSPC ranged from 6.37 to 6.77 mg/mL, cholesterol ranged from 2.21 to 2.28 mg/mL, and DSPE-PEG2k ranged from 0.12 to 0.13 mg/mL (Table 5). The concentrations of lipids are almost the same as label values (shown in Tables 5).

Fig. 6.

Representative chromatograms of DSPC (A), cholesterol (B), DSPE-PEG2k (C) standard solution, their mixture (D) and Onivyde^® (E) by UPLC/ELSD.

Table 5.

Lipid contents in four batches of Onivyde^® formulation. The quantification was performed using UPLC with an ELSD detector. All values were presented as mean ± SD (n = 3).

Batch No.	DSPC (mg/mL)	Cholesterol (mg/mL)	DSPE-PEG2k (mg/mL)
Published Values	6.81	2.22	0.12
120518S	6.37 ± 0.44	2.25 ± 0.06	0.12 ± 0.02
200048A	6.38 ± 0.12	2.21 ± 0.04	0.13 ± 0.004
200147A	6.77 ± 0.18	2.27 ± 0.07	0.12 ± 0.003
U16025	6.62 ± 0.05	2.28 ± 0.03	0.13 ± 0.01

3.7. Analysis of other ingredients

The content of TEA was determined using GC, and the TEA amount was found to be less than 11 ppm in all batches of Onivyde^® (Table 6), which mirrored the high encapsulation efficiency of irinotecan ensuring the sufficient drug compound in each injection unit and ultimate efficacy. Additionally, UPLC/ELSD was used to determine that Lyso-PC (Figure S2) was undetectable, reflecting the stability of DSPC in Onivyde^® during the storage at 4 °C. Lyso-PC is a glycerylphosphocholine fatty acid monoester, produced through the hydrolysis of phospholipids. Excessive hydrolytic breakdown of phospholipids in irinotecan liposomes can affect the stability of the formulation by altering the release profile of irinotecan from the liposomes during storage, ultimately shortening the product’s shelf life.

Table 6.

Levels of other components in four batches of Onivyde^®. Triethylamine was detected by GC, and Lyso-PC was detected by UPLC/ELSD. ND refers to not detectable. All values are presented as mean ± SD, n = 3.

Batch No.	Triethylamine (ppm)	Lyso-PC (μg/mL)
120518S	<11	ND
200048A	<11	ND
200147A	<11	ND
U16025	<11	ND

3.8. Release kinetics in vitro

3.8.1. Drug release of Onivyde^® in PBS with different concentrations of ammonium bicarbonate

In Fig. 7, the cumulative release of irinotecan from Onivyde^® in pure PBS was the slowest, showing a gradual increase. To be more specific, the total drug release was approximately 10 % at 12 h, and it remained less than 20 % at 24 h. In the presence of 0.1 mM NH₄HCO₃, the cumulative release of Onivyde^® slightly improved, reaching 35 % at 12 h and 45 % at 24 h. With 0.5 mM NH₄HCO₃, the cumulative release reached 90 % at 12 h, almost achieving complete release at 24 h. For the free drug, more than 80 % was released in 1 h, and the cumulative release reached 90 % at 12 h, ultimately up to 92.4 % at 24 h.

Fig. 7.

Drug release of Onivyde^® in PBS with different concentrations of ammonium bicarbonate (0 mM, 0.1 mM, 0.5 mM NH₄HCO₃).

The solid crystalline drug in the core of Onivyde^® and the closely packed phospholipid bilayer affect the drug release profiles (Redondo-Morata et al., 2012) because the lipids in Onivyde^® are in an ordered gel phase at 37 °C, which hinders the diffusion of the drug across the lipid bilayer membrane. Further, drug in a crystalline form primarily contributes to the slow drug release of Onivyde^®. The dissolution of crystal aggregates was a rate-limiting step for the release evaluation in this scenario. However, when ammonium bicarbonate was added, a notable increase in the cumulative release of Onivyde^® was observed. The higher the concentration of ammonium bicarbonate was, the faster the release rate. This enhancement in release rate could be caused by the generation of ammonia in the buffer. Ammonia penetrated across the liposome’s phospholipid bilayer and de-protonated the irinotecan salt, promoting the conversion of irinotecan to its free base form. This free base form of irinotecan could then readily cross the liposome membrane, leading to an accelerated drug release (Barenolz et al.; Yechezkel Barenolz).

3.8.2. Drug leakage of Onivyde^® in 50 % human plasma

As depicted in Fig. 8, Onivyde^® exhibited the fastest release of irinotecan in 50 % human plasma within the first hour, with the cumulative release percentage reaching 15 %. Subsequently, the release rate decreased, and the cumulative release approached 30 % at 24 h, ultimately reaching 100 % at 120 h.

Fig. 8.

Drug release of Onivyde^® in 50 % human plasma. All values are presented as mean ± SD, n = 3.

When comparing the drug release rate in pure PBS (Fig. 7) and that in 50 % human plasma (Fig. 8), the cumulative release of Onivyde^® at 24 h was greater in human plasma. This enhanced release in human plasma may result from the presence of various components in human plasma that facilitate drug release, including albumin, coagulation factors, fibrinolytic proteins, immunoglobulins, and other proteins (Opanasopit et al., 2002). Human serum albumin, the most abundant protein in circulation, has been shown in many studies to partially penetrate liposomes due to hydrophobic interactions. This interaction leads to the formation of larger complexes that can disrupt the lipid bilayer’s packing order, resulting in the leakage of molecules from the liposome (Thakur et al., 2014). Apolipoproteins found in plasma can form amphipathic helices with both polar and nonpolar faces, which can perturb lipid bilayers (Pownall et al., 2016; Liu et al., 1990). They also serve as opsonin for macrophages and ligands between receptors on cells and remnant particles. Additionally, purified fibronectin has been observed to bind to liposomes of various compositions (Rossi and Wallace, 1983), potentially increasing liposome uptake (Halter et al., 2005). However, when evaluating drug release behavior in human plasma using a dialysis bag certain limitations need to be considered. Firstly, the precipitation of proteins in human plasma can block the pores in the dialysis bag, hindering the diffusion of free drug. Secondly, some components in human plasma may absorb small-molecule drugs and retain them inside the dialysis bag. Therefore, new research methods to evaluate drug release need to be studied to further optimize in vitro release testing of this formulation (Juang et al., 2024).

4. Conclusions

Onivyde^®, with an average hydrated diameter around 110 nm, is a prolate spheroid liposome dispersed in buffer with a pH value around 7.2. The drug’s polar diameter was measured to be within the range of 58–68 nm with the equatorial diameter between 103–110 nm as demonstrated by cryo-TEM images and SAXS data analysis. In nanostructure, the API in this study, irinotecan, forms crystal bundles inside the liposome with a crystal length of around 60 nm. DSPC, cholesterol and DSPE-PEG2k arrange into phospholipid bilayer around irinotecan crystal and had Tm range around 52–57 °C. The total concentration of irinotecan, DSPC, cholesterol and DSPE-PEG2k was consistent to label claims. In vitro, ammonium bicarbonate was shown to promote the release of irinotecan, and the drug release was much faster in human plasma than in pure PBS.

We established a series of analytical methods to characterize Onivyde^® including particle size analysis, morphology, examination of physical and chemical properties, determination of drug and lipid contents, and evaluation of its release behavior in vitro. Furthermore, the results provide some insight on the potential batch to batch variability of Onivyde^®. Overall, the test parameters evaluated were consistent with label claims.

Data availability

Data will be made available on request.