The encapsulation of Ubiquinone-10 in diﬀerent lipid nanocarriers

(1)

nanocarriers

Degree Project C in Chemistry Molly ˚ Angstr¨ om

Supervisor: Katarina Edwards Department of Chemistry - BMC

Uppsala university

March 2020

(2)

ability has been shown by encapsulation of ubiquinone-10 in lipid nanocarriers. In this work, the effect of different phospholipid compositions on the loading capacity and structures of two different types of carriers, liposomes and lipodisks, were investigated for two ubiquinone- 10:phospholipid molar ratios (1:10 and 1:4). The aim was to study whether lipodisk formulations could load sufficient Q10 to be of therapeutic interest. The techniques used to determine the LC and the carriers structures were: phosphorus analysis, ubiquinone-10 extraction and cryogenic transmission electron microscopy (cryo-TEM). Interestingly, both liposome formulations tested indicated significantly lower loading capacity even at the same ubiquinone-10 ratios compared to corresponding lipodisk formulations. Furthermore, it was found that the loading capacity was significantly different between the same lipodisk formulations with different ubiquinone-10 content - indicating that there was no proof that the maximum loading capacity had been reached. It was found that the loading capacities for lipodisk formulations with different phospholipid compositions (but identical phospholipid:ubiquinone-10 ratios) showed no significant difference.

(3)

2 Introduction 2

2.1 Methods . . . 3

3 Experimental details 4 3.1 Materials . . . 4

3.2 Preparation of stock solutions . . . 5

3.3 Lipodisk and liposome preparation . . . 5

3.3.1 Lipid film preparation . . . 5

3.3.2 Lipid film hydration and sonication . . . 5

3.4 Ubiquinone-10 extraction . . . 6

3.5 Phosphorus analysis . . . 6

3.6 Characterisation of nanoparticles . . . 6

4 Result and discussion 6 4.1 Reference samples . . . 6

4.2 Lipodisk formulations . . . 8

4.2.1 Samples devoid of Q10 . . . 8

4.2.2 Samples with a low Q10 content . . . 10

4.2.3 Samples with a high Q10 content . . . 13

4.3 Liposome formulations . . . 16

5 Summary and concluding remarks 20 6 Acknowledgements 21 7 Appendix 1 7.1 Phosphorus analysis . . . 1

7.2 Q10 extraction . . . 1

7.2.1 The theoretical amount of Q10 . . . 2

7.3 Statistical tests . . . 3

7.3.1 t -tests between two experimental means . . . 3

7.3.2 Uncertainty interval . . . 5

(4)

1 Abbreviations

CL - cardiolipin (Heart, Bovine) (sodium salt) CPP- Critical packing parameter

Cryo-TEM- cryogenic transmission electron microscopy DLS - Dynamic Light Scattering

DOPG - 1,2-Dioleoylphosphatidylglycerol

DSPE-PEG₂₀₀₀- 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy(polyethylene glycol)- 2000](ammonium salt)

EE- encapsulation efficiency, explained further in section ”2.1 Methods”

HBS- HEPES Buffered Saline

HEPES- (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) HSPC - 1,2-Distearoyl-sn-glycero-3-phosphocholine

LC- loading capacity, explained further in section ”2.1 Methods”

PEG - polyethylene glycol Q10 - ubiquinone-10 std- standard deviation

(5)

2 Introduction

Ubiquinone-10, or Q10 for short, is a physiologically interesting compound. It is used as an electron transporter in the respiratory chain[1], a membrane stabiliser[2] and as an antioxidant (in its reduced form ubiqiuinol-10)[3]. Q10 is most concentrated in tissues and structures strongly con- nected to energy production and usage (such as mitochondria in the heart tissue)[4]. The structure of Q10 can be seen in Figure 1 with its characteristic ten unit isoprenyl chain. Other lengths are also possible, e.g. ubiquinone-8 and ubiquinone-6, but mainly ubiquinone-10 is present in human cells[4]. Many different diseases and conditions have been linked to a decreased concentration of Q10 in serum and/or tissue. Examples include; ageing, diabetes, cancer, treatment with drugs belonging to the statin class and genetic mutations impeding the biosynthesis. Unfortunately, it has been determined that the natural occurrence of Q10 in different foods cannot significantly raise the level of Q10 available in serum, not without supplementation[5]. This makes investigation of potential therapeutic formulations an important topic.

O

MeO Me

MeO

Figure 1: The structure of ubiquinone-10[6]

In order to effectively supplement Q10, a couple of issues must be overcome to improve its poor bioavailability. The main one is the extreme hydrophobicity of the compound, a reported solubility in water of 0.7 ng/mL at physiological temperature[7]. This can be improved by loading the Q10 into lipid nanoparticles[8]. Lipid nanoparticles form when lipid surfactants in an aqueous environment self assemble to different structures. The formation is driven by the hydrophobic effect and is thus dependent on some part of the lipid being sufficiently hydrophobic. The shape of the particle formed is dependent on the geometrical structure, or critical packing parameter (CPP), of the lipids. The CPP is defined as the volume of the hydrophobic chain divided by the product of the optimal hydrophilic headgroup area and the length of the hydrophobic chain. The formation of the lipid particles is not initiated until a specific concentration is reached in the bulk solution[9], complicating therapeutic formulations that often must withstand intense dilution upon administration to the patient.

Two of the most relevant structures for this report are lipodisks and liposomes (to the right and left in Figure 2 respectively). Liposomes are nanoparticles consisting of a spherical lipid bilayer encapsulating an aqueous core, schematically illustrated to the left in Figure 2. They form upon agitation of lipid bilayer forming lipids in an aqueous environment, and sonication or extrusion can be used to produce unilamellar liposomes of different sizes[9].

Lipodisks, first described in 1997 [10], are formed when enough of the micelle forming PEG- lipids are introduced into some bilayer forming formulations in aqueous solution[11]. When the formulation is agitated, the resulting structures are disks consisting of a lipid bilayer center with its highly curved edges stabilised by PEG-lipids. Increasing the molar fraction of PEG-lipid in a formulation decreases the size of the lipodisks. If the amount of PEG-lipid is high enough, spherical micelles are the most dominant structure present[12]. A schematic illustration of a lipodisk can be seen to the right in Figure 2.

(6)

Figure 2: Left: A schematic illustration of a liposome. Right: An illustration of a lipodisk with the PEG-corona and lipid bilayer center.

Images adapted from figures by G¨oran Karlsson and Malin Morin, Edwards Group.

For a nanocarrier to be effective, the amount of matrix materials should be low - meaning by extension that a high loading capacity (LC) is better than a low one[13]. Although many different liposome formulations have already proven effective as carriers for Q10[14]-[17], it has also been shown that liposomes have a limited LC[18]. This work seeks to investigate whether lipodisks may prove beneficial as carriers for Q10 from a loading capacity point of view, by comparison with the LC of a liposome formulation.

2.1 Methods

Figure 3: The 3D to 2D rendition of lipodisks and liposomes in cryo-TEM. Adapted from a figure by G¨oran Karlsson, Edwards Group.

To determine the type of structures present in the samples, cryogenic transmission electron microscopy (cryo-TEM) was used. In this technique, an electron beam is used to get a 2D micrograph from a 3D sample (a graphical de- ciption is shown in Figure 3). In cryo-TEM images, liposomes appear as circular objects with an enhanced rim. This is because they are hol- low structures, which means that the electron beam have to pass through more mass on the edges compared to the centres. Lipodisks will look like a dark line (when fully standing on its edge, called ”edge-on” orientation) or as a much less contrasted liposome without the dark rim (when fully laying down perpendicular to the beam, called ”face-on” orientation). It can also orient in a way that is in between these two extremes, producing a rather oval shape.

For the purpose of this report, loading capacity (LC) will be calculated as the amount of trapped drug divided by the amount of phospholipid present and be reported as a ratio. As for encapsulation efficiency (EE), this will be calculated by dividing the amount of trapped drug by the total amount of added drug and be represented in percentage. The total amount of phospholipid in the sample will be determined

through phosphorus quantification with a UV-vis spectrophotometer. This will result in the to-

(7)

tal amount of phosphorus, and (as can be seen at the bottom in Figure 4) each cardiolipin (CL) molecule contains two phosphorus atoms while HSPC (at the top of Figure 4) and DOPG (in the middle of Figure 4) only contain one. An appropriate correction is necessary to find the total amount of lipids in batches containing CL.

Figure 4: Top: The structure of the most abundant lipid in the HSPC lipid mixture used in this work. Middle: The structure of DOPG. Bottom: The structure of Cardiolipin (sodium salt)[19]

To compare the LC of different samples, a uncertainty interval of the difference between the two LC-values in question (at the 95% confidence level) will be calculated. The standard deviation of mentioned difference can be calculated through the standard deviations of the experimental means and error propagation of random errors[20]. Should the produced uncertainty interval contain zero, it cannot be concluded that the two values, e.g. loading capacities, are distinctly different.

However, neither can it be concluded that they are not.

3 Experimental details

3.1 Materials

Sulfuric acid ( 95-97%, H₂SO₄), chloroform (pro analysis), methanol (pro analysis), potassium antimony tartrate hemihydrate (K(SbO)C₄H₄O₆·¹₂H₂O) and L(+)-Ascorbic acid were products from MERCK KGaA in Darmstadt, Germany. The spectroscopic grade ethanol used (99.7%) was from SOLVECO in Roserborg, Sweden. n-Hexane (>85%) was from Fisher Scientific and the acetone from VWR BDH Chemicals (Fontenay-Sous-Bois, France). The ubiquinone-10 (Coenzyme Q₁₀, >98%) and (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) (HEPES) used came from SIGMA-ALDRICH (St. Louis, Missouri, US). Ammonium molybdate (99.99%, (NH₄)₆Mo₇O₂₄· 4 H₂O) was obtained from Acros Organics, New Jersey, USA. Cardiolipin (CL) sodium salt from bovine heart and 1,2-Dioleoylphosphatidylglycerol (DOPG) were from Avanti Polar Lipids (Al- abaster, USA). For all aqueous solutions, Milli-Q water (18.2 MΩcm) obtained from a Millipore (Bedford, USA) Milli-Q plus system was used. 1,2-distearoyl-sn-glycero-3-phosphoethanolamine- N-[methoxy(polyethylene glycol)-2000](ammonium salt) (DSPE-PEG₂₀₀₀) were kindly gifted by Lipoid GmbH (Ludwigshafen, Germany). A quartz cuvette (Quartz SUPRASIL^®, Hellma Ana- lytics, M¨uhlheim, Germany) and disposable plastic cuvettes (PMMA cuvettes from VWR, Leuven, Belgium) was used for the spectroscopic measurements. Experiments were carried out at room temperature unless otherwise indicated.

(8)

3.2 Preparation of stock solutions

Stock solutions were prepared by weighing directly into an appropriately sized vial using an analytical grade scale. Aqueous solutions were diluted with Milli-Q water and non-aqueous solutions with a 2:1 mixture of chloroform:methanol or, for the Q10 case, chloroform:spectroscopic grade ethanol.

3.3 Lipodisk and liposome preparation

To prepare a sample containing lipodisks, two main steps were performed. First, a lipid film was prepared from either solid substances or from stock solutions. This was to ensure proper mixing of the components. Then, the lipid film was hydrated and sonicated in order to form lipodisks or liposomes. The compositions of the lipodisks studied in this report can be found in table 1 below.

Table 1: phospholipid compositions of the lipodisk samples studied in this report.

Composition Denotation

HSPC:DSPE-PEG₂₀₀₀:CL (60:20:20 molar ratios) CL lipodisk HSPC:DSPE-PEG₂₀₀₀(80:20 molar ratios) HSPC lipodisk HSPC:DSPE-PEG₂₀₀₀:DOPG (48:20:32 molar ratios) DOPG lipodisk

All lipodisk samples had a phospholipid concentration of either 5 mM or 10 mM. Regarding Q10, this was added in 1:10 or 1:4 Q10:phospholipid molar ratios, if at all.

Preparation of liposomes was conducted the same way as for lipodisks, but with a significant decrease or even omission of DSPE-PEG₂₀₀₀. The liposome phospholipid compositions can be found in table 2. If the liposome sample was to contain Q10, the addition led to a Q10:phospholipid molar ratio of 1:10. All liposome samples had a 5 mM phospholipid concentration.

Table 2: phospholipid compositions of the liposome samples studied in this report.

Composition Denotation

HSPC (100 mol%) HSPC liposome

HSPC:DSPE-PEG₂₀₀₀:DOPG (58.8:3:38.2 molar ratios) DOPG liposome

3.3.1 Lipid film preparation

The lipid films were prepared in two ways, either from stock solutions or from pure powders. The components were weighed or pipetted into a test tube using an analytical grade scale or automatic pipette respectively. Q10 was added to the lipid samples in a Q10:phospholipid ratio of either 1:4 or 1:10 (if at all) and all components co-dissolved in about 2 mL chloroform. The solvent was then removed under a gentle stream of nitrogen gas before the sample was put in a vacuum oven (Lab instruments IL, USA) to remove residual solvent. If hydration was not to be carried out right away, the test tube was filled with nitrogen gas and stored in a freezer (-20 C°).

3.3.2 Lipid film hydration and sonication

For hydration of the lipid film, 1000 µL of 20 mM HBS (pH 7.4) was added and the sample mixed for 1 minute before put in a 70 C° water bath for 15 minutes. This cycle of agitation and water bath immersion was repeated three times before the sample was sonicated at 70 °C for 5 to 7 minutes using a probe sonicator. If the sample had a concentration of 10 mM, the sample was sonicated at a lower intensity but for a longer time interval in order to avoid foaming. The sample was then centrifuged for 30 min at 6700g to remove metal debris from the titanium probe used for sonication, as well as excess Q10 (if present). The samples were stored in the fridge.

(9)

3.4 Ubiquinone-10 extraction

The sample volumes used for the Q10 extraction was 50 and 70 µL, and each volume was prepared in triplicates. To all Eppendorf tubes, 0.32 mL hexane and 480 µL of MeOH was added and the sample mixed for 1 minute. Then, 0.32 mL of acetone was added and the samples were mixed on the shaker for 30 minutes before centrifuged at 2100g for 2 min. The upper phase was transferred to another vial. After that, an additional 0.32 mL of hexane was added to the Eppendorf tubes and the sample mixed for 1 min before placed on the shaker for another 30 minutes. The samples were then centrifuged at 2100g for 2 minutes and the upper phase transferred to the same glass vials as earlier. Residual solvent was removed in a vacuum oven (lab instruments IL, USA). Then the samples were diluted with 2.5 mL of spectroscopic grade ethanol and mixed. Measurements were carried out at 275 nm (using an HP 84 UV-Vis spectrometer from Agilent Technologies, Santa Clara, USA) with a quartz cuvette (Quartz SUPRASIL^®, Hellma Analytics, M¨uhlheim, Germany) and spectroscopic grade ethanol as a blank. The molar extinction coefficient for Q10 in ethanol, used to convert the measured absorbance to concentration, was 12.6 mM⁻¹cm⁻¹. Q10 dissolved in spectroscopic ethanol was used as a control sample by pipetting 10 µL of the Q10-stock solution used for the samples, and then measure that the same way the samples were.

3.5 Phosphorus analysis

To determine the amount of phosphorus in the samples, sample aliquotes corresponding to between 10-60 nmol were placed in test tubes and calcinated at 500 C° for 4 hours. The samples were prepared in two sets of triplicates.

The test tubes were diluted with 2 mL Milli-Q water. A 1:3:10 mixture of 1 part K(SbO)C₄H₄O₆·

1

2H₂O (with a concentration of 1 mg Sb/mL), 3 parts (NH₄)₆Mo₇O₂₄·4H₂O (with a concentration of 4 wt% (NH₄)₆Mo₇O₂₄· 4 H2O) and 10 parts H₂SO₄(2.5 M) was prepared. This mixture was then combined 7:3 with 0.1 M ascorbic acid. The resulting mixture was immediately pipetted into the samples as well as the blank, 0.5 mL for every sample. Then the samples were mixed for 20 minutes before measured twice at 882 nm (using an HP 84 UV-Vis spectrometer from Agilent Technologies, Santa Clara, USA) in disposable plastic cuvettes (PMMA cuvettes from VWR, Leuven, Belgium).

The quantification of phosphorus was determined from a standard curve made through repeated dilution of a phosphorus standard solution (0.65 mM) from Sigma Aldrich (St. Louis, USA).

3.6 Characterisation of nanoparticles

The characterisation of structures present in the samples were done by cryo-TEM analysis. Shortly described, a small amount of sample was placed on a AGAR SCIENTIFIC 300 MESH copper grid (reinforced with a holey polymer coating) before vitrified in liquid ethane that is held just above its freezing point (-182.8°C) and then analysed in the microscope. The chamber in which the samples are prepared is carefully controlled both in temperature and humidity (25°C, >90%) and the sample is, during transfer to the microscope, protected against the atmosphere and kept below -160°C. The sample is also held at -160°C for the remainder of the analysis.

4 Result and discussion

4.1 Reference samples

A Q10 reference containing unencapsulated Q10 (corresponding to 1:4 Q10:phospholipid molar ratio for a 5 mM sample) prepared similar to the phospholipid formulations was found to contain a significant amount of Q10 compared to two measured blanks when tested with a t -test. This means that although the presence of Q10 in the sample is low (1.27·10⁻¹⁷mol Q10 with a standard deviation of 3.38 · 10⁻¹⁸ mol), it is not lower than the noise level of the instrument. Furthermore, the method evidently does not remove all Q10 not encapsulated by lipids. The measured Q10 in the Q10 reference sample corresponds to a concentration of 1.1 · 10⁻¹⁴ g/mL. This value can be compared to the 7 ng/mL value, which was the reported solubility of Q10 in water at 37°. Most likely, this measured lower value is the solubility at room temperature; the cetrifugation probably

(10)

does not have an impact on the solved Q10. Regardless, if one compare the measured Q10 for the Q10 reference sample with the theoretical addition made to even the lowest Q10 ratios of 5 mM phospholipid formulations (5 · 10⁻⁷mol) it is evident that the former is not high enough to interfere significantly with the end result of either the loading capacity or the encapsulation efficiency.

A cardiolipin reference, containing HSPC:DSPE-PEG₂₀₀₀:CL in molar ratios of 60:20:20 and no Q10, was subjected to the Q10 extraction procedure and measured. The calculated mean in the cuvette for that CL lipodisk batch was 8.61 nmol (for 50 nmol aliquots, standard deviation 0.23 nmol) and 10.69 nmol (for 70 nmol aliquots, standard deviation 0.27 nmol). This is rather high values compared to the 50 and 70 nmol which would correspond to the theoretical value present in CL batches that should contain Q10. This means that cardiolipin, or something else in the formulation, gives rise to a significant background that could alter the results so they may be perceived as to contain a higher concentration of Q10. The background could come from CL itself, since it contains double bonds (the bottom structure of Figure 4) which may have an absorption at 275 nm. The background may also come from small amounts of Q10 in the CL-powder. The powder is extracted from bovine heart - a tissue that is rich in Q10. Regardless of the reason, this would alter the calculated encapsulation efficiency and the loading capacity by making both appear higher than they are. Not surprisingly, two t -tests found that there was a significant difference between the investigated sample amounts (50 and 70 nmol, two of each amount) compared to the calculated mean of two blanks. Too further investigate, the 70 nmol aliquot extracted Q10 samples from the CL lipodisk batch, Q10 and HSPC powder solubilized in spectroscopic grade ethanol was measured in the uv-vis spectrophotometer to produce a spectra spanning over the relevant interval 240-290 nm. This spectra can be seen in Figure 5 and shows that Q10 (dashed line) has the highest absorbance at 275 nm, but that the extracted CL lipodisk sample still has a distinct absorbance while HSPC does not. However, note that because this was a preliminary experiment with time limitations, the exact concentrations of the analytes in the quartz cuvettes are unknown. Further experiments with relevent concentrations are needed to find a suitable correction for the apparent CL background.

Figure 5: A spectrum over the relevant wavelengths for the Q10 measurements. Absorbance on the y-axis and wavelength in nm on the x-axis. The dashed line is Q10, the continuous line CL and the dotted line HSPC.

(11)

The raw data for the t -tests mentioned in this section can be found in the appendix.

4.2 Lipodisk formulations

4.2.1 Samples devoid of Q10

Three samples, one each of CL, HSPC and DOPG lipodisk formulations, were prepared without any Q10 and analysed by both phosphorus analysis and cryo-TEM. The amount of phosphorus was in all cases within a reasonable range with respect to the theoretical value for a 1 mL 5 mM sample (results can be found in the appendix). Figure 6 shows a cryo-TEM image for the CL lipodisk formulation. There are many disks oriented face- and edge-on, but the sample also appears contain a population of small spherical particles that could be micelles made of PEG-lipids.

Figure 6: Cryo-TEM image of CL lipodisk sample without Q10, 5 mM phospholipid concentration.

The white and black arrows points at a lipodisk as viewed face- and edge-on respectively, and the white arrowhead at the edge of the polymer film. The size bar is 200 nm.

Figure 7 shows a cryo-TEM image for the HSPC lipodisk batch. This sample contains a lot of disks of seemingly two types, one population with larger size and one with smaller size. The smaller could be micelles made of PEG-lipids. The large chainlike cluster near the top of the image is thought to be ice present on the surface of the vitrified sample.

(12)

Figure 7: Cryo-TEM image of HSPC lipodisk sample without Q10, 5 mM phospholipid concentration. The white and black arrows points at a lipodisk as viewed face- and edge-on respectively.

The size bar is 200 nm.

Figure 8 shows a cryo-TEM image for the DOPG lipodisk batch. This sample contains a lot of disks, but much like the CL and HSPC batches, these seem to be rather poly-disperse in size.

(13)

Figure 8: Cryo-TEM image of DOPG lipodisk sample without Q10, 5 mM phospholipid concentration. The white and black arrows points at a lipodisk as viewed face- and edge-on respectively and the white arrow head points at the edge of the polymer film. The size bar is 200 nm.

From the cryo-TEM analysis it is clear that lipodisks were formed in all the samples under the conditions used. Further, all samples also showed formation of small spherical particles thought to be PEG-lipid micelles. No distinct differences could be concluded between phospholipid compositions.

4.2.2 Samples with a low Q10 content

Three samples, one each of CL, HSPC and DOPG lipodisk formulations, were prepared with a Q10 ratio of 1:10 and analysed by phosphorus analysis, Q10 extraction and cryo-TEM. The amount of phosphorus and Q10 was for all samples within a reasonable range with respect to the theoretical values for the samples (results can be found in the appendix). The calculated LC and EE can be found in table 3.

Table 3: LC and EE results for lipodisk batches with 1:10 Q10:phospholipid ratio, calculated from experimental values that can be found in the appendix.

Sample EE (%) Std EE (%) LC Std LC

CL lipodisk 102.62 2.76 0.114 0.005

HSPC lipodisk 95.32 3.43 0.117 0.006

DOPG lipodisk 93.27 2.72 0.117 0.005

The EE can be used to draw conclusions about individual samples, and since all of the percent- ages are very high it can be concluded that the maximum for the LC probably has not yet been reached. All the formulations are effective at this low Q10 amount. The CL lipodisks have an EE above 100%, which can be explained by the significant background found when the Q10 extraction was performed on CL lipodisks that contained no Q10. Further discussion about this can be found under subsection 4.1 Reference samples.

None of these loading capacities can be concluded to be significantly different from any of the other, as determined by calculation of an uncertainty interval (for the difference between the

(14)

loading capacities) that covers zero (see appendix for data).The LC was expected to vary more between compositions, and more data points for the LC would be beneficial to fully verify the obtained results. The high LC of the samples might be due to the formation of a type of Q10 rich particle discussed later in this section.

Figure 9 shows a cryo-TEM image for the CL lipodisk sample with low Q10 content. This sample does not contain any apparent discoidal structures, as concluded from the the lack of edge- on orientations of the very same. This does not mean that no disks are present, just that there is no clear evidence of ones. There thus seems to be less disks present than for the CL lipodisks that did not contain Q10. Some liposomes are visible in the picture, but the main bulk of structures appears to have more contrast than typically observed for disks or liposomes, meaning they are denser than those structures typically are. The appearance of these particles are much like those found in a study about the mechanical properties of Q10 on phospholipid membranes[18], in which they were judged to be ”Q10-rich oil-like structures”. Judging from the observation in the reference samples (Figures 6 to 8) that the formation of lipodisks are more common without Q10 addition, and that Q10 is likely in the hydrophobic part due to its extreme hydrophobicity, it is probable that Q10 adds to the hydrophobic part and thus promotes the formation of a spherical particle.

Figure 9: Cryo-TEM image of CL lipodisk sample with 1:10 Q10:phospholipid molar ratio (5 mM phospholipid concentration). The black arrow points at an unknown, most likely roughly spherical, particle thought to be rich in Q10. The white arrow heads points at ice on the surface of the vitrified sample and the black arrow head at the edge of the polymer film. The size bar is 200 nm.

Figure 10 shows a cryo-TEM image for the HSPC lipodisk sample. This sample has disk edges visible in the image as well as circular structures displaying low contrast that probably are disks with face-on orientation. Some liposomes are visible in the picture and the general size distribution is rather broad and contains particles of different structures. There seem to be less disks present

(15)

than for the HSPC lipodisks that did not contain Q10.

Figure 10: Cryo-TEM image of HSPC lipodisk sample with 1:10 Q10:phospholipid molar ratio (10 mM phospholipid concentration). The black arrow points at an disk edge and the white one at a disk face-on. The black arrow head points at a liposome. The size bar is 200 nm.

Figure 11 shows a cryo-TEM image for the DOPG lipodisk sample. This sample contains both disks edge-on as well as circular structures (displaying low contrast) that probably are disks with face-on orientation. Some liposomes are visible in the picture and the general size distribution is rather broad. There seem to be less disks present than for the DOPG lipodisks that did not contain Q10.

(16)

Figure 11: Cryo-TEM image of DOPG lipodisk sample with 1:10 Q10:phospholipid molar ratio (10 mM phospholipid concentration). The black arrow points at a disk edge-on and the white one at what is likely a large disk viewed face-on. The black arrow head points at a liposome and the white arrow head at ice. The scalebar is 100 nm and the dark structure at the bottom left is the polymer film.

Q10 seems to affect all preparations, but in different ways. The CL composition has no apparent disks in the cryo-TEM images for the 1:10 Q10:phospholipid molar ratio, and while HSPC and DOPG still seems to show disk structures, they contain more liposomes at this low Q10 amount than they did without any Q10 addition.

4.2.3 Samples with a high Q10 content

Three samples, one each of CL, HSPC and DOPG lipodisk formulations, were prepared with a Q10 ratio of 1:4 and analysed by phosphorus analysis, Q10 extraction and cryo-TEM. The amount of phosphorus and Q10 were for all samples within the range expected from the theoretical calculations (results can be found in the appendix). The calculated LC and EE can be found in table 4.

Table 4: LC and EE results for lipodisk batches with 1:4 Q10 ratio, calculated from experimental values that can be found in the appendix.

CL lipodisk 84.85 4.87 0.265 0.018

HSPC lipodisk 87.59 5.18 0.255 0.016

DOPG lipodisk 87.60 3.47 0.246 0.013

The EE can be used to draw a conclusion about individual samples and, like for the 1:10 Q10:phospholipid ratio, the EE values are rather high. This leads to the conclusion that the top limit for the Q10 loading capacity in these phospholipid compositions might not yet have been reached even with 1:4 Q10:phospholipid ratio. Further experiments with even higher Q10 ratios would be needed to find the maximal Q10 loading capacity. Each of the loading capacities reported for these Q10 rich samples are significantly higher than those reported for the same formulations with lower Q10 ratios, determined by calculation of an uncertainty interval (for the difference between the loading capacities) that does not cover zero (see appendix for data). Still, there is no evidence the maximum LC has been reached, since this would require a distinct levelling off in a

(17)

graph where LC is plotted against the Q10 content. Only two data points are not enough to know this. The same type of uncertainty interval as before indicate no significant difference between the formulations containing the same amount of Q10, however, more data points for the LC would be beneficial to fully verify this. When studying the cryo-TEM pictures generated for these samples (Figures 12 to 14), it is evident that the high EE values most likely are influenced by the large number of Q10 rich particles present. This raises the question whether the EE and LC reflects the encapsulation in lipid nanocarriers with distinct properties, or if it more reflects the formation of Q10 rich phospholipid stabilised nanoparticles. Whether or not it is important to make that distinction from a therapeutic point of view (regarding for instance the stability and bioavailability of that formulation) is for future studies to determine.

Figure 12 shows a cryo-TEM image for the CL lipodisk batch. No visible edge-on structures are seen in the image, and although this is not evidence that there are none, it is an indication that there might be few disks and that the sample is dominated by spherical particles. The spherical particles appear similar to those in the low Q10 content CL lipodisk sample (Figure 9), and are thought to be rich in Q10.

Figure 12: Cryo-TEM image of CL lipodisk sample with 1:4 Q10:phospholipid molar ratio (5 mM phospholipid concentration). The black arrow points at an unknown, most likely spherical, particle thought to be rich in Q10 and the white arrow head at the edge of the polymer film. The size bar is 200 nm.

Figure 13 shows a cryo-TEM image for the HSPC lipodisk sample. This sample display disks edge-on, as well as circular structures displaying low contrast that probably are disks with face-on orientation. The picture also shows the same darker Q10 structures as noted in the CL lipodisk

(18)

formulation in Figure 12. The size distribution is rather poly-disperse and there seem to be less disks present than for the HSPC lipodisks with less Q10, strengthening the earlier conclusion that the particles observed are rich in Q10.

Figure 13: Cryo-TEM image of the HSPC lipodisk sample with 1:4 Q10:phospholipid molar ratio (5 mM phospholipid concentration). The black arrows point at two disks, one face-on and one edge-on. The white arrow points at the same unknown particle as present in the CL sample in Figure 12 and the white arrow head at the edge of the polymer film. The size bar is 200 nm.

Figure 14 shows a cryo-TEM image for the DOPG lipodisk sample. This sample has both edge-on oriented disks and disks with face-on orientation. The image also shows the same darker Q10 structures as noted in the CL lipodisk formulation in Figure 12. In the image, there are also some liposomes and ice visible. The liposomes are much larger compared to the disks. There also seems to be less disks present than for the DOPG lipodisks with less Q10.

(19)

Figure 14: Cryo-TEM image of the DOPG lipodisk sample with 1:4 Q10:phospholipid molar ratio (5 mM phospholipid concentration). The black arrows point at two disks, one face-on and one edge-on. The white arrow points at the same unknown particle as present in the CL sample in Figure 12. The small white arrow head points at ice while the large points at the edge of the polymer film. The black arrow head at a liposome and the size bar is 200 nm.

In all of the samples with the higher Q10:phospholipid ratios, the amount of lipodisks and liposomes is lower than in the samples with the low content while the opposite is true for the Q10 rich particles observed. This leads to the conclusion that high amounts of Q10 hinders the formation of well structured lipid lipodisks and promotes formation of a Q10 rich phospholipid stabilised nanoparticle.

4.3 Liposome formulations

Three samples, one HSPC liposome formulation with 1:10 Q10:phospholipid molar ratio and two DOPG lipodisk formulations (one with 1:10 and one without Q10) were prepared and analysed by phosphorus analysis, Q10 extraction (if it contained Q10) and cryo-TEM. The amount of phosphorus and Q10 were for all samples within a reasonable range with respect to the theoretical values for a 1 mL 5 mM sample (results can be found in the appendix). The calculated LC and EE can be found in table 3.

Table 5: LC and EE results for liposome batches with 1:10 Q10:phospholipid molar ratio, calculated from experimental values that can be found in the appendix.

HSPC liposome 6.31 0.41 0.024 0.002

DOPG liposome 95.36 2.04 0.104 0.003

Since EE is not comparable between samples (due to differing phospholipid amounts), one

(20)

can only use this value to draw conclusions about individual samples. However, because EE is much lower for the HSPC liposomes compared to both the DOPG liposomes and all of the lipodisk formulations, it can be concluded that the top limit for the encapsulation of Q10 in this phospholipid composition has likely been reached. Further experiments with even higher Q10 ratios would be needed to find the maximal Q10 loading capacity for the DOPG liposomes.

The LC for the DOPG liposomes is significantly lower than that for the DOPG lipodisks with the same Q10 content, determined by calculation of an uncertainty interval (for the difference between the loading capacities) that does not cover zero (see appendix for data). The low LC for the HSPC liposomes indicate that this formulation is not effective for therapeutic use when looking solely at the drug per phospholipid ratio. However, the low cost of HSPC might still warrant it fit for commercial applications.

Figure 15 shows a cryo-TEM image for the HSPC lipodisk sample. The picture also shows the same darker Q10 structures as noted in the CL lipodisk formulations containing Q10. The sample is very poly-dispersed with a broad range of structures present, something that adds to the conclusion that this composition is unfit for therapeutic use. The angular appearance of the structures is because the HSPC in the sample is below its transition temperature, meaning it is in gel phase, at room temperature.

Figure 15: Cryo-TEM image of the HSPC liposome sample with 1:10 Q10:phospholipid molar ratio (5 mM phospholipid concentration). The black arrows points at a liposome and the white arrow at a Q10 rich particle. The white arrow head point at the edge of the polymer film. The size bar is 200 nm.

An interesting observation for the HSPC liposome sample is that there was a significant amount of yellow precipitate after the centrifugation step, a picture of this is found in Figure 16. This

(21)

helps explain the large decline in both Q10 and phosphorus observed for the sample. Since the cryo-TEM image in Figure 15 is produced from the already centrifuged sample, it is not known what the structures that were spun down looked like.

Figure 16: The precipitate left in the tube after the centrifugation step of the HSPC liposomes.

The white arrow points at the yellow precipitate and the white line indicates the surface level of the sample volume.

Figure 17 shows a cryo-TEM image for the DOPG liposome sample without Q10. This sample displayed, apart from liposomes, also lipodisks. This was somewhat surprising because of the low DSPE-PEG₂₀₀₀amount in the formulation and might warrant further investigation on the effect of high DOPG ratios on lipid structures. The presence of a few disks could perhaps be explained by the preparation method and the usage of a low mol fraction of PEG-lipids - but the large number of disks visible in Figure 17 is surprising.

(22)

Figure 17: Cryo-TEM image of the DOPG liposome sample, 5 mM phospholipid concentration and no Q10. The black arrows points at two lipodisks and the white arrow points at a liposome.

The white arrow head points at the edge of the polymer film. The size bar is 200 nm.

Figure 18 shows a cryo-TEM image for the DOPG liposome sample with Q10. Also here there are a lot of liposomes present, but no apparent disks. The smaller dark particles without a distinct rim might be the same Q10 rich particles as observed for the Q10 rich lipodisk samples.

(23)

Figure 18: Cryo-TEM image of the DOPG liposome sample 1:10 Q10:phospholipid molar ratio (5 mM phospholipid concentration). The black arrow points at a liposome and the white arrow at something that might be Q10 rich particles. The white arrow head points at the edge of the polymer film. The size bar is 200 nm.

5 Summary and concluding remarks

The encapsulation efficiencies calculated in this work do not compare well between samples, since the samples contained different amounts of lipids even when they theoretically should be identical. The loading capacities were found, based on uncertainty intervals of the difference between them covering zero (95 % certainty level), to not differ significantly between the lipodisk formulations tested. On the other hand, the loading capacities did differ significantly between a high and a low ubiquinone-10 addition, with the higher loading capacity observed for batches with a high ubiqinone-10 addition - indicating that it might still be possible to further increase the amount of ubiquinone-10 loaded into the formulations. Additional testing is necessary to find the maximum loading capacity, and more repetitions should be made to strengthen the previous conclusion. However, there might not be a clear line between actually encapsulating ubiquinone-10 into lipid nanocarriers and just creating ubiquinone-10 rich particles that might be stabilised by lipids. Whether or not it is important to make this distinction, from a colloidal stability and bioavailability point of view, is for future studies to determine. The HSPC liposome had both a very low encapsulation efficiency and loading capacity, which might be due to the 1:10 ubiquinone- 10:phospholipid molar ratio being over its maximal loading capacity. Furthermore, in case of preparations with ubiquinone-10:phospholipid ratio 1:10, the DOPG liposome formulation did significantly differ from the DOPG lipodisk with the lipodisk’s loading capacity being higher. It is unknown whether this has a structural or compositional explanation, but shows promise for the

(24)

possibility of lipodisks being effective carriers. The amount of visible lipodisks in the cryo-TEM images of all the lipodisk formulations decreased as the ubiquinone-10 content was increased. This indicate that increasing ubiquinone-10 hinders the formation of well defined lipodisk structures.

The CL lipodisk formulation with ubiquinone-10 contained no apparent lipodisk structures, and although this is not conclusive evidence that none exist somewhere else in the sample; it is an indication they might be scarce. It would be interesting from a therapeutic point of view to find at what ubiquinone-10:phospholipid ratio the samples mostly consisted of the ubiquinone-10 rich particles observed instead of well defined structures, and whether this structural difference affect the colloidal stability. The loading capacities observed for the lipodisk formulations are promising, and although many questions remain to be investigated, a small step towards determining the suitability of lipodisk ubiquinone-10 carriers has nevertheless been achieved.

6 Acknowledgements

I would like to show my sincere appreciation to Philipp Grad for the many hours he spent giving guidance and answering my questions throughout this entire project. I am also thankful to Lu´ıs Silva and V´ıctor Agmo Hern´andez for their willingness in sharing their expertise and wisdom with me. A great thank you to Lars Gedda for performing all the cryo-TEM experiments and finding time to show me the process. Finally, a great thank you to Katarina Edwards for her valuable feedback and insights on the project along the way. I am very grateful I got the opportunity to work on this interesting project and for the things I have learned during it.

References

[1] M. Alc´azar-Fabra, P. Navas, G. Brea-Calvo, Coenzyme Q biosynthesis and its role in the respiratory chain structure, BBA, V. 1857 issue 8, August 2016 pg. 1073-1078

[2] E. K. Eriksson, V. Amgo Hern´andez, K. Edwards, Effect of Ubiquinone-10 on the stability of biomimetic membranes of relevance for the inner mitochondrial membrane, BBA, February 2018, pg. 1205-1215

[3] B. Frei, M.C Kim, BN Ames, Ubiquinol-10 is an Effective Lipid-soluble Antioxidant at Physi- ological Concentrations, Proc Natl Acad Sci U S A. 1990 Jun, 87(12):4879-83

[4] F. ˚Aberg, E-L. Appelkvist, G. Dallner, L. Ernster, Distribution and redox state of ubiquinones in rat and human tissues, Archives of Biochemistry and Biophysics, V. 295 issue 2, June 1992 pg. 230-234

[5] L. C. Frederick, Journal of the American College of Nutrition, 20, 591-598.

[6] Drawn by Philipp Grad on ChemDraw

[7] Persson, E.M., Gustafsson, A., Carlsson, A.S. et al. The Effects of Food on the Dissolution of Poorly Soluble Drugs in Human and in Model Small Intestinal Fluids. Pharm Res 22, 2141–2151 (2005)

[8] Noha M. Zaki (2016) Strategies for oral delivery and mitochondrial targeting of CoQ10, Drug Delivery, 23:6, 1868-1881

[9] R. M Pashley, M.E Karaman, Applied Colloid and Surface Chemistry, 2004, 61-73

[10] K. Edwards, M. Johnsson, G. Karlsson, and M. Silvander, Effect of Polyethyleneglycol- Phospholipids on Aggregate Structure in Preparations of Small Unilamellar Liposomes, Bio- physical J., 73, (1997) 258-266

[11] M. C. Sandstr¨om, E. Johansson, K. Edwards, Structure of Mixed Micelles Formed in PEG- Lipid/Lipid Dispersions, Langmuir 2007 23 (8), 4192-4198

(25)

[12] M. Johnsson, K. Edwards, Liposomes, Disks, and Spherical Micelles: Aggregate Structure in Mixtures of Gel Phase Phosphatidylcholines and Poly(Ethylene Glycol)-Phospholipids, Biophys- ical J., V. 85 (2003), P3839-3847

[13] Jaspart, S., Piel, G., Delattre, L., Evrard, B. (2005). Solid lipid microparticles: Formulation, preparation, characterisation, drug release and applications. Expert Opinion on Drug Delivery, 2(1), 75–87.

[14] Verma DD, Hartner WC, Thakkar V, Levchenko TS, Torchilin VP (2007) Protective effect of coenzyme Q10-loaded liposomes on the myocardium in rabbits with an acute experimental myocardial infarction. Pharm Res 24(11):2131–2137

[15] Lee WC, Tsai TH (2010) Preparation and characterization of liposomal coenzyme Q10 for in vivo topical application. Int J Pharm 395(1–2):78–83

[16] Niibori K, Yokoyama H, Crestanello JA, Whitman GJR (1998) Acute administration of liposomal coenzyme Q 10 increases myocardial tissue levels and improves tolerance to ischemia reperfusion injury. J Surg Res 79(2):141–145

[17] Xia F, Jin H, Zhao Y, Guo X (2012) Preparation of coenzyme Q10 liposomes using supercritical anti-solvent technique. J Microencapsul 29(1):21–29

[18] V. A. Hern´andez, E. K. Eriksson, K. Edwards, Ubiquinone-10 alters mechanical properties and increases stability of phospholipid membranes, BBA (2015) pg 2233-2243

[19] The two structures on the top are drawn on Chem-space.com by the author and the bottom one is drawn on ChemDraw by Philipp Grad

[20] J.N Miller, J. C. Miller, Statistics and Chemiometrics for Analytical Chemistry, 6th edition, 2010, pg 31-33

(26)

7 Appendix

7.1 Phosphorus analysis

Tables 6 to 8 contains the result of the phosphorus analysis for the different batches. The correction for the batches containing CL consisted of removing the percentage of phosphorus stemming from CL (based on the mol ratio) and dividing it by two before adding it back again. When the molar ratio CL added is 20, this corresponds to dividing the measured phosphorus by 1.2.The value listed in the the tables is the corrected value.

Table 6: Phosphorus analysis results for lipodisk batches with 1:10 Q10:phospholipid molar ratio.

Composition (1:10 Q10:phospholipid) Concentration Mean P¹ Standard deviation

(µmol) (µmol)

HSPC:DSPE-PEG₂₀₀₀:CL (60:20:20 molar ratio) 5 mM 4.9 0.15

HSPC:DSPE-PEG₂₀₀₀(80:20 molar ratio) 10 mM 8.1 0.3

HSPC:DSPE-PEG₂₀₀₀:DOPG (48:20:32 molar ratio) 10 mM 7.9 0.2

1Mean phosphorus

Table 7: Phosphorus analysis results for lipodisk batches with 1:4 Q10 ratio.

Composition (1:4 Q10:phospholipid) Concentration Mean P¹ Standard deviation

(µmol) (µmol)

1Mean phosphorus

Table 8: Phosphorus analysis results for liposome batches and Q10 reference.

Composition Concentration Mean P¹ Standard deviation

(µmol) (µmol)

HSPC:DSPE-PEG₂₀₀₀:DOPG (58.8:3:38.2 molar ratio) 5 mM 5.0 0.001 (1:10 Q10:phospholipid)

HSPC:DSPE-PEG₂₀₀₀:DOPG (58.8:3:38.2 molar ratio) 5 mM 4.8 0.007 (No Q10)

HSPC (100 mol%) (1:10 Q10:phospholipid) 5 mM 1.4 0.066

No lipids, Q10 corresponding to 1:4 5 mM 0.0003 0.0001

1Mean phosphorus

7.2 Q10 extraction

Tables 9 to 11 contains the result of the Q10 extraction for the different batches.

Table 9: Q10 extraction results for lipodisk batches with 1:10 Q10:phospholipid molar ratio.

Composition (1:10 Q10:phospholipid) Concentration Mean Q10 Standard deviation

(µmol) (µmol)

(27)

Table 10: Q10 extraction results for lipodisk batches with 1:4 Q10 ratio.

Composition (1:4 Q10:phospholipid) Concentration Mean Q10 Standard deviation

(µmol) (µmol)

Table 11: Q10 extraction results for liposome batches, Q10 reference and CL lipodisk reference.

Composition Concentration Mean Q10 Standard deviation

(µmol) (µmol)

HSPC:DSPE-PEG₂₀₀₀:DOPG (58.8:3:38.2 molar ratio) 5 mM 0.5 0.01

(1:10 Q10:phospholipid)

HSPC (100 mol%) (1:10 Q10:phospholipid) 5 mM 0.034 0.0002

No lipids, Q10 corresponding to 1:4 5 mM 1.3 · 10⁻¹¹ 3.4 · 10⁻¹²

(No Q10)

7.2.1 The theoretical amount of Q10

Tables 12 to 14 shows the measured concentration of the Q10 stock solution at the time for preparing the samples. From this and the knowledge about how large volume was pipetted into the samples, the theoretical mol Q10 in the samples was calculated. This value was then used to calculate the EE.

Table 12: The measured concentration of the Q10 stock solution at the time of sample preparation, as well as the volume and mol added to the sample for the 1:10 lipodisk batches.

Composition (1:10 Q10:phospholipid) Stock concentration Volume added Q10 added

(M) (µL) (mol)

HSPC:DSPE-PEG2000:CL (60:20:20 molar ratio) 0,004058 135 5, 48 · 10⁻⁷

HSPC:DSPE-PEG2000(80:20 molar ratio) 0,00372 269 1, 00 · 10⁻⁶

HSPC:DSPE-PEG2000:DOPG (48:20:32 molar ratio) 0,00372 269 1, 00 · 10⁻⁶

Table 13: The measured concentration of the Q10 stock solution at the time of sample preparation, as well as the volume and mol added to the sample for the 1:4 lipodisk batches.

(M) (µL) (mol)

HSPC:DSPE-PEG2000:CL (60:20:20 molar ratio) 0,00406 337 1, 37 · 10⁻⁶

HSPC:DSPE-PEG2000(80:20 molar ratio) 0,00406 337 1, 37 · 10⁻⁶

HSPC:DSPE-PEG2000:DOPG (48:20:32 molar ratio) 0,00406 337 1, 37 · 10⁻⁶

Table 14: The measured concentration of the Q10 stock solution at the time of sample preparation, as well as the volume and mol added to the sample for the 1:10 liposome batches.

(M) (µL) (mol)

HSPC (100 mol%) 0,00406 135 5, 48 · 10⁻⁷

HSPC:DSPE-PEG2000:DOPG (58.2:3:38.2 molar ratio) 0,00392 135 5, 30 · 10⁻⁷

(28)

7.3 Statistical tests

7.3.1 t -tests between two experimental means

All t-tests reported in this section were preformed with Origin 2008 software (from the OriginLab corporation).

Figure 19 demonstrate a picture of the t-test between the mean value of the measured Q10 amount in the 50 nmol aliquots of the CL lipodisk formulation that (in reality) contained no added Q10 and blanks containing only spectroscopic grade ethanol. As the Prob> |t| value is less than 0.05 (both when equal variance is and is not assumed), it can be concluded that the amount of Q10 measured in the CL lipodisk formulation is significantly different from a blank when a 50 nmol aliquot is used.

Figure 19: A t-test between the mean Q10 value of the 50 nmol aliquots from the CL lipodisk formulation (without Q10) and pure spectroscopic grade ethanol blanks. These correspond to 1st and 2nd (Summarized) Data respectively, visible in the middle of the image.

Figure 20 demonstratete a picture of the t-test between the mean value of the measured Q10 amount in the 70 nmol aliquots of the CL lipodisk formulation that (in reality) contained no added Q10 and blanks containing only spectroscopic grade ethanol. As the ”Prob> |t|” value is less than 0.05 (both when equal variance is and is not assumed), it can be concluded that the amount of Q10 measured in the CL lipodisk formulation is significantly different from a blank when a 70 nmol aliquot is used.

(29)

Figure 20: A t-test between the mean Q10 value of the 70 nmol aliquots from the CL lipodisk formulation (without Q10) and pure spectroscopic grade ethanol blanks. These two samples correspond to 1st and 2nd (Summarized) Data respectively, visible in the middle of the image.

Figure 21 demonstrate a picture of the t-test between the measured Q10 amount in the Q10 reference sample and blanks containing only spectroscopic grade ethanol. As the ”Prob” value is less than 0.05 (both when equal variance is and is not assumed), it can be concluded that the amount of Q10 measured in the Q10 reference formulation is significantly different from a blank.

(30)

Figure 21: A t-test between the mean Q10 value of the Q10 reference sample and pure spectroscopic grade ethanol blanks. These two samples correspond to 1st and 2nd (Summarized) Data respectively, visible in the middle of the image.

7.3.2 Uncertainty interval

Figure 22 demonstrates the calculations made for the uncertainty intervals of the difference between the loading capacity of the formulations.

Figure 22: The calculations for the confidence intervals of the differences in loading capacities between the batches, made in Excel.