For our project, we needed a carrier system which protects the siRNA from degradation and transports it safely and efficiently to the brain via the nose-to-brain route. For this task we have chosen liposomes, which are small spherical vesicles with a diameter of under 1000 nm consisting of at least one lipid bilayer (Akbarzadeh et al., 2013). They encompass an aqueous solution with the drug inside, and can release their content into a target cell by either directly fusioning with its membrane, or by endocytosis and subsequent degradation of the lipid bilayers in the endosome. Another commonly used type of lipid-based nanoparticles are lipid nanoparticles (LNPs). They have a diameter of 10-1000 nm and consist of a lipid core and the encapsulated drug (Saupe & Rades, 2006). Both, LNPs and liposomes, have already been proven to transport drugs retrogradely along the olfactory and trigeminal nerves to the brain (Lochhead et al., 2015), and both have been successfully used as delivery vehicles for nucleic acids. For instance, the COVID-19 vaccines by BioNTech/Pfizer and Moderna use solid LNPs to encapsulate mRNA (Schoenmaker et al., 2021). Lipid based carrier systems allow a wide range of uses and targets, depending on the individual lipids and surface modifications that are being used. Additionally, lipid-based nanoparticles provide advantages such as biocompatibility, extracellular transport and reduced side effects by limiting the drug distribution to non-targeted areas (Battaglia et al., 2018, Lee & Minko, 2021). Due to similar efficacies, we have decided on liposomes instead of (solid) lipid nanoparticles as our drug delivery system, because the latter are considerably more laborious to produce.
Our siRNA-encapsulating liposomes consist of three lipids and cholesterol. The lipids are 1,2-Dioleoyl-3-trimethylammonium propane (DOTAP), 1,2-Dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE), and N-(carbonyl-methoxy polyethyleneglycol-2000)-1,2-distearoyl-sn-glycero-3-phosphoethanolamine (mPEG2000-DSPE, hereafter: mPEG2000).
DOTAP is a cationic lipid with a quaternary ammonium group, and has been used previously in in vitro and in vivo cationic lipid carriers (Simberg et al., 2004). DOPE is a neutral helper lipid and has been used together with other cationic lipids to enhance cellular uptake by destabilising the endosome after uptake into the cell (Koltover et al., 1998; Mochizuki et al., 2013). The lipid mPEG2000 is used for enhanced circulation time, because it protects the liposome from clearance out of the blood (Karmakar et al., 2018). The function of cholesterol in liposomes is the reduction of membrane permeability to increase the halflife and improve particle stability (Nakhaei et al., 2021).
The lipids are used in an mPEG:cholesterol:DOTAP:DOPE ratio of 0.3:2:1:1. This ratio has been used in other siRNA-loaded liposomes before (Jbara-Agbaria et al., 2022).
Liposomes can be produced in special centrifuges with two turntables, one for regular rotation and another for rotation around the axis of the vials. This method is called dual centrifugation (DC). The shear forces resulting from the two contra-rotating movements vigorously mix the sample, consisting of a lipid film covered with an aqueous solution, and produce liposomes (Massing et al.Two step dual centrifugation vs VPG method, 2008). DC is ideal for entrapping sensitive drugs in liposomes, because it reduces the risk of lipid breakdown products (Koehler et al., 2021), and therefore poses the perfect method for encapsulating somewhat fragile siRNA too.
Liposomes have different characteristics that influence their stability and effectiveness and can be used to evaluate their quality. One of these characteristics is their size, which can be measured via different methods. We decided to use dynamic light scattering (DLS), which is a commonly used method to determine the size of nanoparticles including liposomes. DLS uses the scattering of a laser by the particles in the sample to calculate the hydrodynamic radius dH, the size of the particle including the hydration shell around it, with the Stokes-Einstein equation (Varga et al., 2020). In order to achieve the easiest and therefore most effective endocytosis into the target cells, the liposomes intended for our drug delivery system should be as small as possible. We set the threshold for the size at 200 nm, because larger liposomes have been shown to be disadvantageous for cellular uptake (Feng et al., 2013).
The polydispersity index is another characteristic and used to describe the degree of non-uniformity of a size distribution of particles (Danaei et al., 2018). Generally, PDIs below 0.3 are deemed acceptable for liposomes used for drug delivery (Danaei et al., 2018). Since it was intended as a therapeutic, we wanted to establish a high standard for our liposomes and decided on a cutoff for ideal liposomes at a PDI of 0.2.
The zeta potential of a liposome is a third characteristic which can be measured to approximate the general surface charge at the hydrodynamic shear boundary, the end of the hydration shell of liposomes. It can be positive, negative or neutral and influences, among other things, the uptake in cells and the stability of the liposomes (Smith et al., 2017). Regarding the uptake in cells, the liposomes have to be either neutral or the opposite charge compared to the targeted cell’s surface potential. Because most cells, and especially neurons, are negatively charged towards the outside, we needed liposomes with a positive zeta potential. Stability is also dependent on the zeta potential, because the larger its charge, the more the liposomes will repulse each other and will therefore not merge into bigger, less effective liposomes.
Before starting with the production of our desired liposomes, we decided to conduct preliminary tests to familiarise ourselves with the methods and machines used. Furthermore, we wanted to determine how the amount of added beads influences the size of liposomes, whether SiLi Beads or steel beads are more suitable for liposome production, and whether nucleotides have any affinity towards the beads.
Beads were needed to ensure efficient liposome formation for our chosen production method in the dual centrifuge, as these beads greatly amplify the shear forces exerted during dual centrifugation. Ceramic beads are commonly used, more specifically SiLi Beads. Before we could start conducting our experiments, we needed to ensure that the SiLi Beads (type ZY-E 1.0-1.2 mm) did not interact with our siRNA, as suspected due to the general high affinity of oligonucleotides to silicate. We reasoned that a high affinity of nucleotides to SiLi Beads could lead to reduced loading efficiency.
For the initial experiments, we used DNA primers as substitutes instead of siRNA to effectively determine the ideal bead amount and establish the final production method. They were of comparable size (18 to 24 nt) and therefore also of a comparable total charge. Since our final siRNA consists of many different sequences, the specific sequence of the DNA primers used was not relevant to the outcomes of our experiment. For these reasons, we decided to load leftover primer stocks from previous iGEM Heidelberg teams into our liposomes for preliminary tests.
To evaluate oligonucleotide binding affinities to the SiLi Beads, we prepared three 100 µl solutions. One contained only PBS as a negative control, one additionally contained DNA primers (0.78 mg/ml), and one contained DNA primers (0.78 mg/ml) as well as 142 g SiLi Beads. We measured the DNA concentration in the solution at the NanoDrop immediately after combining the reagents, as well as after 5 minutes, 10 minutes, 30 minutes, 1 hour, 2 hours and 1 day.
If a clear deviation between DNA primer concentration in PBS and concentration in PBS with SiLi Beads occurs, an interaction between nucleotides and beads would have been very likely. However, there was no considerable difference between the two concentrations, suggesting that nucleotides do not have a significant affinity to SiLi Beads, as can be seen in Figure 1.
Since we proved that hereby, there is no detectable affinity between SiLi Beads and nucleotides, we could proceed with our experiments as planned.
The next step was considering different bead types and compositions. Because SiLi Beads are quite expensive, we started looking into less costly alternatives. Here, we discovered steel beads, which are of the same size and only cost a fraction of the price of SiLi Beads. Because steel is not known to interact with nucleotides in any manner, we did not have to test for nucleotide affinity again.
To determine whether steel beads acted the same way as SiLi Beads, we produced empty liposomes using the two-step-centrifugation method with both bead types each. We used a lipid concentration of 100 mM and two different lipid:bead ratios, namely 1:1 and 1:10. We compared the sizes and PDIs of each version of liposomes we created, as can be seen in Figure 2.
Overall, the sizes are comparable and do not vary significantly between the production with SiLi Beads and steel beads. The sizes of liposomes with the lipid:bead ratio 1:1 only vary slightly between steel and SiLi Beads. The PDI is larger in liposomes produced with steel beads, but still well below 0.3. With the lipid:bead ratio of 1:10 sizes and PDI of the steel bead liposomes are smaller.
Judging by hydrodynamic diameter and PDI, steel beads should be suitable as a substitute for SiLi Beads. However, the liposomal solutions started to take on an unexpected orange tint after one day of storage in the fridge, which can be seen in Figure 3.
Massing et al. found that the same was true in their conducted experiments. There was no significant difference in size between liposomes produced with stainless steel beads and those produced with SiLi Beads. Additionally, they also noticed the yellow discoloration in their liposome solution in the production with steel beads (Massing et al., 2008).
The yellowish coloration led us to the conclusion that metals from the beads were somehow solubilised and dissolved into the solution. As contamination with an unknown metal is unacceptable for the given safety standards in therapeutics development, we decided that steel beads were not suitable. We conducted all the following experiments with SiLi Beads.
One of the first suspects of contaminations was iron, the main component of the steal beads, so we conducted ion detections for Fe3+ as well as Fe2+ to find the cause of the discoloration, as we suspected different metals in the beads getting oxidised and then falling off into the solution. Although the test for Fe3+ remained negative, the Fe2+ ion detection yielded positive results, as can be seen in Figure 4. This indicates abrasion of the bead material into the liposome solution. These results cemented us in our belief that carrying on with steel beads would add unwanted new elements to the final drug and would therefore be extremely irresponsible.
After determining the right type of beads, the next step was to examine the effect of the amount of added beads on liposome quality.
First, we produced several 50 mM, 100 mM and 200 mM lipidfilms. We then created empty liposomes out of the lipid films while keeping the amount of added beads constant, and compared the sizes. By doing so, we intended to select the most promising concentration for testing the relation between beads and size, and achieving the smallest liposomes possible. After initial experiments, we chose to focus on 100 mM lipid films, which we used for all further experiments.
After establishing 100 mM lipid films as superior, we began testing how the amount of beads used in the dual centrifuge affect the sizes of liposomes. To do that, we produced 100 mM liposomes and varied the amount of beads used in each experiment. We tested the lipid:beads ratios 1:1, 1:5, 1:10, 1:20 and 1:25. We measured one sample of each ratio three times in the Zetasizer to determine size and PDI. The results of the experiments are illustrated for size and PDI respectively in Figure 5.
There is a vast difference in liposome sizes that occur across varying amounts of beads. The sizes range from 172.1 nm to 148.9 nm, as can be seen in Figure 5A. While the ratios 1:1, 1:5, 1:10 and 1:25 fluctuate around 170 nm, the lipid:bead ratio of 1:20 has the smallest size with 148.93 nm.
By increasing the bead amount, liposome sizes decrease until they reach a plateau. Additional shear forces from the collision of beads with each other and with the walls of the vial are responsible for the reduction in size. After this plateau, mean sizes increase again, creating an inverted bell curve. This is consistent with the literature (Massing et al., 2008).
All liposomes from 1:1 to 1:20 have a PDI that is smaller than 0.2, which is illustrated in Figure 5B. Only the liposomes created with the 1:25 lipid:bead ratio have a PDI of 0.252. While this PDI is still acceptable, it is higher than our set cutoff point at 0.2.
The results of this experiment suggest that an optimal amount of beads for producing liposomes can be determined for individual lipid concentrations. When diverging from this ideal ratio, the liposomes change drastically in size and might even have an increased PDI.
We tested three different procedures to find the ideal production method for liposomes. The research group at our university produced their liposomes by a two step centrifugation protocol. While researching, we discovered a method involving vesicular phospholipid gels (VPG) first published by Dr. Hirsch in 2009. Finally, we developed our own three step centrifugation protocol, which is based on the production method that was already established in the lab.
First, we decided to compare the two step centrifugation method to the VPG method. For preparing liposomes according to the VPG procedure, we pursued two ways of production.
Since we had already dissolved our lipids in CHCl3:MeOH, but in the VPG method, the lipids are first dissolved in ethanol instead, we had to first remove the solvent by drying the lipids under N2. Then, the lipids were dissolved in ethanol, and then a lipid film was created again by evaporating the ethanol, and the standard VPG protocol continued. For our second way of production, we simply used the lipids dissolved in CHCl3:MeOH instead for creating a lipid film. By doing so, we intended to establish whether a difference in solvents would influence the size and PDI of liposomes produced.
We measured the sizes and PDIs of the liposomes created with the VPG method for ethanol and CHCl3:MeOH each, as well as our previously tested 100 mM liposomes with 20 beads. The results of these measurements can be seen in Figure 6.
Liposomes produced with both VPG methods yielded sizes of approximately 335 nm with PDIs larger than 0.4. In contrast, the liposomes created with the two step centrifugation were at the expected range of around 150 nm with a PDI smaller than 0.2.
This showed unequivocally that the two step centrifugation method was superior for our chosen lipid composition.
We then started loading our liposomes with DNA primers, and varied certain parameters of the two step centrifugation method (2-fold DC), hoping to find an even better procedure to yield the smallest loaded liposomes possible. For instance, we varied the volumes added in between certain centrifugation steps or added our primer DNA in the last centrifugation step instead of the first. While most of those reduced the size of our liposomes to some extent, only one variation decreased the liposome sizes significantly. In the end, we found out that introducing an additional centrifugation step in the beginning with just the lipid film and oligonucleotide solution, before adding the beads and then continuing with the standard 2-fold DC centrifugation steps yields liposomes with the smallest size and PDI among all variations. For this reason, our newly developed method will be called three step centrifugation (3-fold DC).
At this point, we adjusted the amount of lipids to the yield of siRNA per round of production and conducted all further experiments with 1 mM total lipid concentration.
The results of this novel technique in comparison to the two step centrifugation can be seen in Figure 7.
The liposomes that were produced using the three step centrifugation were in the range of 125 nm and were significantly smaller than those produced with the two step centrifugation at 160 nm. The PDI of three step centrifugation liposomes was well below 0.2 and therefore ideal.
Following the discovery of our new centrifugation method, we decided to focus on the three step centrifugation for our final encapsulation of siRNA.
After establishing the characterisation and production protocol for liposomes, we used it to produce loaded liposomes. We had to adjust the final concentration of the liposomes from 100 mM to a lower concentration, because we were limited by the siRNA production yield per round of production.
Another characteristic of loaded liposomes is their N/P ratio. It describes the ratio between the total number of positively charged nitrogen atoms (amines) in the cationic lipids of the liposome and the number of negatively charged phosphate groups in the encapsulated oligonucleotides. It can be used as a property to compare different oligonucleotide-loaded liposomes across different protocols. In our case, we tried out two different N/P ratios: 5/1 and 20/1.
We tested different amounts of beads for the N/P ratios 5/1 and 20/1 in two step centrifugation as well as three step centrifugation to find the ideal production method. For our liposomes that were based on N/P ratios we did not use the lipid:bead ratio as we did with 100 mM, but varied the number of beads. This was because the weight of one bead is too high to weigh the exact lipid:bead ratio. We therefore tested liposome production with 1, 5, 10, 20, 30 and 40 beads.
Our comparison of the bead number for a N/P ratio of 5/1 and 20/1 for the 2-fold DC protocol can be seen in Figure 8 A, B, C, D. The results for the N/P ratio of 20/1 for the 3-fold DC protocol is illustrated in Figure 8 E, F.
The two step dual centrifugation with a very small number of SiLi Beads (1 to 5 beads) yielded liposomes with a dH of larger than 200 nm. The liposomes with 10 or more beads had a dH between 160 nm and 185 nm. In contrast, the 3-fold DC with 20 beads created liposomes with a dH of only 130 nm. This makes it the smallest liposome formulation that we were able to produce.
But not only the size is important, but the PDI as well, which can be found in Figure 8 B, D, F. The PDI of the N/P ratios of 5/1 liposomes is over 0.3 and therefore over our set cutoff point. The two liposome formulations with a PDI below 0.2 are N/P = 20/1, 2-fold protocol, 30 beads (PDI = 0.198) and N/P = 20/1, 3-fold protocol, 20 beads (PDI = 0.179).
In both quality determining criteria, size and PDI, the same production method was shown to be the best. After careful consideration, we chose the N/P ratio of 20/1 with the 3-fold DC protocol and 20 beads, which will be used for all following experiments.
After determining the ideal N/P ratio and bead amount, it was time to produce and characterise empty liposomes as well as loaded ones with primer DNA, eGFP siRNA and VP5 siRNA, all produced with the 3-fold DC protocol with 20 SiLi Beads and a N/P ratio of 20/1. RNase inhibitors were added to siRNA solutions directly after production. The measured dH, PDI and zeta potential can be found in Table 1.
The liposomes loaded with eGFP and VP5 siRNA depicted in Table 1 were used for our cell culture experiments, on which you can read more here.
| dH | PDI | zeta potential | |
| empty | 170 ± 32 nm | 0.259 ± 0.059 | -7.3 ± 1.1 mV |
| loaded with primer DNA | 131.5 ± 6.2 nm | 0.187 ± 0.041 | -7.6 ± 2.3 mV |
| loaded with eGFP siRNA | 195.5 ± 36 nm | 0.320 ± 0.097 | -6.2 ± 2.5 mV |
| loaded with VP5 siRNA | 700.0 ± 85 nm | 0.649 ± 0.015 | -7.5 ± 2.7 mV |
Surprisingly, the empty liposomes were larger than the ones loaded with DNA primer. The additional cargo should lead to an increase in size. Instead, the opposite was the case and the loading with DNA primer led to smaller liposomes.
When siRNA was used instead of DNA primers, the properties of the liposomes changed. The sizes increased from 131.5 nm to 195.5 nm for eGFP and to 700.0 nm for VP5. This is an unexpected result, since we had expected DNA and siRNA to behave similarly, because they share a multitude of prominent characteristics. Out of our siRNA loaded liposomes, only sample 2 of the eGFP loaded liposomes had a PDI below 0.3 and was therefore within our set cutoff point. The properties of the three liposomes with eGFP siRNA can be seen individually in Table 2.
| dH | PDI | zeta potential | |
| eGFP siRNA loaded liposomes 1 | 238.4 nm | 0.421 | -8.68 mV |
| eGFP siRNA loaded liposomes 2 | 149.8 nm | 0.189 | n.a. |
| eGFP siRNA loaded liposomes 3 | 198.1 nm | 0.351 | -3.67 mV |
We were not able to produce siRNA loaded liposomes consistently. This is an issue that needs to be fixed in the future. Our plan is to tackle this hurdle with the microfluidic device for lipid nanoparticles that we have designed.
Drug loaded liposomes have another crucial property besides dH, PDI and zeta potential: the encapsulation efficiency (EE). It describes how much of the initially added cargo (in our case oligonucleotides) is encapsulated in the liposomes. To examine the EE of oligonucleotide loaded liposomes, agarose gels can be used (Li et al., 2021). Free oligonucleotides run faster through the gel than the larger liposomes. In order to quantify the intensity of the nucleic acid bands, we used different amounts of the respective oligonucleotide as a standard. The higher the EE is, the less DNA or siRNA is found free in the liposome solution and therefore on the gel as a separate band.
We tested the EE of our liposomes with DNA primer as a substitute for siRNA. The band indicating free DNA of the DNA primer-loaded liposomes is very faint, indicating a high EE of over 90 % for DNA loaded liposomes when compared to the standard concentration bands (Figure 9). The bright band above in the same lane arises from the DNA that is encapsulated or bound to the liposomes, which travel slower through the gel.
The siRNA loaded liposomes did not turn out as expected regarding size and PDI, but nevertheless, we also characterised their EE. The siRNA standard and samples were run on separate gels, but were observed with the same settings for comparability. The samples in Figure 10 and the standard can be seen in Figure 11.
The free siRNA ran directly below the dye, which is visible as a dark dot in Figure 11. In Figure 10, as a negative control, each siRNA loaded liposome had an equal amount of empty liposomes directly in the chamber to their right. The dye was only clearly visible for the VP5 siRNA Liposome 1 and its empty counterpart, which was enough to estimate where the siRNA should be. In the first and fifth column, a faint band is visible. In comparison to the standard this indicated an EE of over 90 % (Figure 11, first column). The gel was cut to improve visibility of the bands.
Overall we observed very high EEs for all our loaded liposomes. Not only the liposomes with sufficient quality criteria have an EE of over 90 %, but the siRNA loaded liposomes with large sizes and high PDIs (see Table 1 and 2) as well.
The previous characterisation has shown that all liposomes have a negative zeta potential. However, this is not ideal for cellular uptake, which we intend to achieve with our liposomes. Especially the cell membrane of neurons is negatively charged and equal charges generally repel each other (Nishino et al., 2020), leading to little uptake of liposomes into neurons.
Additionally, unspecific endocytosis is another problem when looking at intranasal drug delivery for the nose-to-brain route (Aderibigbe & Naki, 2019), as simple liposomes without any further surface modifications do not exclusively adhere to neurons to be retrogradely transported to their target areas in the brain and the trigeminal ganglion, but often fuse with adjacent epithelial cells, releasing the contained drug in the nose already. Therefore, we wanted to modify the surface of our liposomes in such a way that it has both a positive zeta potential as well as a higher affinity towards neurons and thereby a higher chance of being transported retrogradely to the brain. Fortunately, both of these desired changes could be achieved by one single surface modification: Coating the liposomes with Chitosan.
Chitosan is a linear polysaccharide made out of beta-1,4-linked D-glucosamine. At neutral pH, it is in its protonated form, leading to a net positive charge of the molecule. For this reason, it can be used to help make the zeta potential of liposomes positive by coating the surface with chitosan (Mady & Darwish, 2010). Furthermore, chitosan has been proven to increase the mucoadhesivity of liposomes and enhance nose-to-brain delivery (Aderibigbe & Naki, 2019).
To first analyse the effect of chitosan in solution on zeta potential measurement, we tested different concentrations of chitosan in PBS measuring their size, PDI and their zeta potential, which can be seen in Figure 12. The PBS on its own had a zeta potential of -5.1 mV. With increasing concentration of chitosan, the zeta potential shifts to a positive charge. Surprisingly, a 1/200 ratio yielded a less positive charge than a 1/400 ratio. However, here a measuring or pipetting error could be the cause.
Then we coated our DNA primer-loaded liposomes with Chitosan by adapting a protocol from Mady & Darwish in 2010, and wanted to find out which ratios of chitosan to liposomes lead to a chitosan coating of the liposomes. Chitosan and liposome solutions (both empty and DNA loaded) were combined in different chitosan:lipid volume ratios: 1:5, 1:4, 1:3, 1:2, 1:1, 2:1, 3:1, 4:1 and 5:1. The empty liposomes increased in size by 50 nm and the PDI increased from 0.2 or 0.3 to over 0.35 when a larger volume of chitosan in relation to liposomes was used, indicating a larger diameter of the liposomes and therefore a chitosan coating on top. Loaded liposomes changed only slightly in size and the PDI varied between 0.2 and 0.3. There was no sign that the size had significantly increased or decreased depending on the chitosan concentration.
However, the zeta potential changed drastically with an increasing amount of chitosan. At a ratio of 1:5, the loaded liposomes had a zeta potential of -4.9 mV, but with increasing chitosan, the zeta potential rose to -2.2 mV at a 1:1 ratio and even to 3.3 mV at 5:1. Empty liposomes have the highest zeta potential at a 3:1 ratio with 5.7 mV, which could also indicate a successful coating of the liposomes with chitosan.
Our experiments have demonstrated that it is possible to produce liposomes loaded with oligonucleotides using DC. It has been shown that there are some aspects which have worked well and others which need further improvement.
When comparing the methods for producing liposomes, we could show that size and size distribution (PDI) using the VPG method are significantly higher than those using the 2-fold DC method. When discussing the results with Prof. Dr. Massing, we realised that insufficient homogenization of the lipids could be the reason for this. In the two step centrifugation method, homogenization is achieved by allowing the lipids to dissolve in the non-polar methanol/chloroform mixture. However, this is not the case with the VPG method, since ethanol is used as the solvent here and is too polar. As a result, homogenization in this method occurs merely due to the shear forces exerted on the lipid mixture during dual centrifugation. Since the centrifuge available to us could not apply the forces indicated in the publication, we reasoned it was likely that the insufficient homogenization of the lipids was responsible for the significantly increased size and size distribution of liposomes.
The other two methods, the 2-fold DC and 3-fold DC, were continued to be used for loaded liposomes. They were tested for size and PDI as well with varying bead numbers. In Figure 8 it is particularly striking that for the same N/P ratio with the 2-fold DC protocol, the size of the liposomes decreases as the amount of bead increases, but this trend is exactly reversed with the 3-fold DC protocol. This opposite behaviour with different bead quantities can probably be attributed to the fact that with three centrifugation steps, the total amount of heat generated by friction is greater and thus the temperature within the solution is higher. This in turn leads to an increased motility of the lipid molecules, resulting in liposome instability. This assumption will be clarified in the future by further experiments in which the N/P ratio as well as the bead amount and the number of centrifugation steps are kept constant, but the temperature is varied. If the liposomes increase in size with increasing temperature, this would support our conjecture.
Comparing the sizes between the unloaded and DNA primer loaded liposomes, it is noticeable that the loaded liposomes are significantly smaller. This is unusual, since in the literature so far only a stable or increasing size when loading liposomes has been described (Uhl et al., 2017). This decrease in size could indicate a denser packing of the lipids and thus the formation of lipid nanoparticles. In this process, a lipid shell would first form around the DNA due to the interaction of the polar lipid with the nucleotides. This shell is then tightly enveloped by a second lipid layer due to interaction of the lipids with each other. However, verifying whether lipid nanoparticles actually form during the process is only possible using electron microscopy.
When comparing the size of the differently loaded liposomes, it can be seen from Table 1 that the liposomes loaded with DNA primer are smaller than the liposomes loaded with siRNA. This increase in size is most likely caused by the RNase inhibitor added to the siRNA solution. With its molecular mass of 50 kDa, this inhibitor demands significantly more space, resulting in the enlarged liposomes. In the future, this inhibitor should be replaced by a smaller molecule that does not affect the size of the liposomes as much, meets the requirements for ingredients of a therapeutic agent, and guarantees sufficient stability of the siRNA.
As for the results of our experiments, we have shown that we can achieve an encapsulation efficiency of over 90 % with our liposome manufacturing process. This can be attributed to the fact that we use dual centrifugation for the production of our liposomes. This process can achieve a much higher encapsulation efficiency compared to other liposome manufacturing methods such as extrusion. Thus, the method is much more efficient and accordingly requires smaller volumes for the manufacturing process. Furthermore, our liposomes have an N/P ratio of 20/1, which means that we have a 20-fold excess of positive charge, resulting in an efficient interaction between the nucleic acid and the lipid molecules.
All measured liposomes loaded with siRNA had an encapsulation efficiency of more than 90 %. This makes them comparable to liposomes loaded with DNA primers. However, due to the high PDI, which is as high as 0.6 for VP5 siRNA, these liposomes are not suitable for use as therapeutics, where the size must be precisely defined. From the six siRNA loaded liposome batches, only one met our quality standards.
Additionally, the zeta potentials of any given liposomes - empty, loaded with DNA primer or siRNA - were measured as negative, as can be seen in Tables 1 and 2. This was surprising, since we consciously decided on cationic lipids to avoid this problem. Neurons are always slightly negatively charged, which would repel our therapeutic instead of promoting its absorption. Prof. Massing proposed one hypothesis which could explain this is the possibility of the DNA/siRNA not only being encapsulated inside the liposome, but also adhering to the positively charged lipid bilyaer surface of it, making the finalzeta potential of the liposome negative again. To solve this, we conducted experiments involving chitosan. We were able to drastically change the zeta potential from -4.9 mV to 3.3 mV with a chitosan:lipid ratio of 5:1 in liposomes loaded with DNA primer. For empty liposomes, we managed to increase the zeta potential from -2.3 mV to 5.7 mV with the ratio 3:1. This shows that reduction of the zeta potential is possible and a viable solution for our initial problem of the negative zeta potential.
Finally, all samples of liposomes loaded with siRNA were assessed according to the quality objectives that we had previously established. These targets included an PDI of below 0.3, ideally below 0.2, among others. Although we achieved great encapsulation efficiencies in all batches, only sample 2 out of the eGFP siRNA loaded liposomes and no samples out of the VP5 siRNA loaded liposomes was able to meet the expectations for our therapeutic. This result leads us to conclude that we need to further optimise our manufacturing process. Our next steps would include the search for a suitable replacement of the RNase inhibitor as well as the establishment of a process for the standardised production of the desired liposomes, since any drug production method should be consistently reproducible. A possible solution in the future would be the production of liposomes by our self-developed microfluidic device.
The produced siRNA loaded liposomes were then used for following experiments in cell culture. From the eGFP siRNA loaded liposomes, Nr. 2 was chosen, and from the VP5, all three batches were used.