key: cord-0265018-0610k0hj authors: Roberts, Steven A.; Lee, Chaebin; Singh, Shrishti; Agrawal, Nitin title: Versatile Encapsulation and Synthesis of Potent Therapeutic Liposomes by Thermal Equilibration date: 2021-10-24 journal: bioRxiv DOI: 10.1101/2021.10.22.465473 sha: 6950812ccc1122b1b0257214d439fb616ca840a7 doc_id: 265018 cord_uid: 0610k0hj The wide-scale use of liposomal delivery systems is hampered by difficulties in obtaining potent liposomal suspensions. Passive and active loading strategies have been proposed to formulate drug encapsulated liposomes, but are limited by low efficiencies (passive) or high drug specificities (active). Here, we present an efficient and universal loading strategy for synthesizing therapeutic liposomes. Integrating a thermal equilibration technique with our unique liposome synthesis approach, co-loaded targeting liposomes can be engineered in an efficient and scalable manner with potencies 200-fold higher than typical passive encapsulation techniques. We demonstrate this capability through simultaneous co-loading of hydrophilic and hydrophobic small molecules and through targeted delivery of liposomal Doxorubicin to a metastatic breast cancer cell line MDA-MB-231. Molecular dynamic simulations are used to explain interactions between Doxorubicin and liposome membrane during thermal equilibration. By addressing the existing challenges, we have developed an unparalleled approach that will facilitate the formulation of novel theranostic and pharmaceutical strategies. Decades of research have resulted in a variety of liposomal formulations for use in both therapeutic [1] [2] [3] and theranostic [4] [5] [6] applications. Advances in synthesis procedure and membrane functionalization have led to targeting and stealth capabilities, which can increase bioavailability at disease sites several fold with minimal side effects, compared to systemically administered drugs 7, 8 . These improvements have in turn led to the FDA approval and commercial use of several liposomal drugs, including Doxil® for cancer 9 , Abelcet® for fungal infections 10 , and DepoDur® for pain management 11 , and more recent Moderna/Pfizer vaccines for Covid-19 12 . While substantial progress has been made in the development of novel liposomal formulations, considerably fewer strategies to increase the potency and concentration of therapeutics within the vehicles have been developed. Sequestration of therapeutic compounds has traditionally been accomplished by either passive or active encapsulation approaches. The passive approach can sequester molecules irrespective of their polarity 13 , where lipophilic molecules reside within the lipid bilayer membrane and hydrophilic molecules are contained within the core. During vesiculation, liposomes form within the media containing drug, encapsulating the drug in the process. This approach is typically done with dilute solutions, and therefore the encapsulation efficiencies are relatively low 14 . This leaves much of the drug to be lost during secondary purification steps. Furthermore, organic solvents (e.g. chloroform and ethanol) are typically used, that can inactivate certain molecules 15 . Active loading techniques have been developed that drastically improve encapsulation efficiencies and minimize the amount of non-encapsulated drug. The sequestration is driven by transmembrane pH or ion gradients that cause the influx of hydrophilic molecules to the core of liposomes [16] [17] [18] . This approach has been successfully utilized to sequester weakly basic drugs with encapsulation efficiencies of nearly 100%. Once inside the liposome, drug molecules precipitate out of the solution, increasing their retention time. While this technique overcomes many of the barriers to passive encapsulation, it is limited to the molecules that can pass through the membrane. Furthermore, decorating the membrane with targeting moieties may require a second step to remove any unbound targeting molecules that may result in the loss of encapsulated drugs. We have previously demonstrated a simple approach for producing purified drug loaded liposomes 19 . Using this method, we rapidly synthesize concentrated populations of small unilamellar liposomes (SULs) with a very narrow polydispersity index (typically near 0.13). The approach relies on passive encapsulation to sequester small molecules, and as such, the encapsulation efficiency is relatively low. Therefore, a technique that can enhance the encapsulation efficiency utilizing small amount of compounds is desired. Here, we demonstrate a new approach for creating drug-loaded targeting liposomes using thermal equilibration, where drug molecules are sequestered at high concentrations via passive diffusion at the membrane transition temperature (T m ). We demonstrate the universal and translational appeal of this approach by simultaneously co-loading both hydrophilic and lipophilic small molecules. We further demonstrate the therapeutic potential of this technique by encapsulating Doxorubicin (DXR), without utilizing ion gradients typically required for active methods. Bioactivity of these liposomes was observed by actively targeting cancer cells in vitro. Finally, using molecular dynamic simulations, we explain the thermodynamics that underlie the encapsulation and model the effects of temperature on structural properties of the liposome membrane and DXR diffusion. The thermal equilibration approach presented here is a powerful universal strategy for rapidly encapsulating small molecule drugs. It offers significantly greater encapsulation capabilities than conventional passive encapsulation techniques and represents a feasible solution to producing therapeutic grade liposomes in a scalable manner. Cholesterol, 1,2-Distearoyl-sn-glycero-3-phosphocholine (DSPC), 5,5′-Dithiobis(2-nitrobenzoic acid) (Ellan's reagent), L-Cysteine, and all fluorescent dyes were purchased from Sigma Aldrich (St. Louis, MO). 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[maleimide(polyethylene glycol)-2000] (ammonium salt) (DSPE-PEG-Mal) was purchased from Avanti Polar lipids. Isopropyl Alcohol (IPA, 99% Pure), Ethanol (EtOH, 99% pure), Dimethyl Sulfoxide (DMSO), desalting columns, dye removal columns, and Trypsin-EDTA were purchased from Fisher Scientific (Hampton, NH). Glass syringes were purchased from Hamilton (Reno, NV) and luer lock dispensing needles were purchased from Jensen Global (Santa Barbara, CA). Amicon Ultra 100 kDa centrifugal filter units were purchased from EMD-Millipore (Billerica, MA). A NanoJet syringe pump from Chemyx Inc. (Stafford, TX) was used for all experiments. Dulbecco's Modified Eagle Medium (DMEM), Fetal Bovine Serum (FBS), and Penicillin-Streptomycin antibiotic were purchased from VWR (Radnor, PA). Doxorubicin (DXR) and WST-1 proliferation assay were purchased from Cayman Chemical (Ann Arbor, Michigan). MDA-MB-231 metastatic breast cancer cell line was acquired from ATCC (Manassas, VA). Liposomes were synthesized as previously described 19 . Briefly, DSPC and Cholesterol were dissolved in isopropyl alcohol at a constant 2:1 molar ratio to a final concentration of 10 mM and 5 mM, respectively. The alcoholic solution was injected using a syringe pump at 100 µL . min -1 into a glass vial containing distilled water preheated at 55 C to obtain a homogeneous population of liposomes. Typical synthesis volumes would be a total of 10 mL (1 mL alcoholic lipid suspension to 9 mL warm distilled water). The aqueous solution was constantly stirred at 600 rpm during infusion, and stirring was continued for 3 minutes following infusion. After the liposomal suspension was cooled down to room temperature with continuous stirring, liposomes were concentrated 20 fold to a volume of 500 µL and a final lipid concentration of 20 mM using filter centrifugation (100 kDa, 6000 x g) at 4 C for 30 minutes. The particles were stored at 4 C until use. Antibody coupling was accomplished through the addition of functionalized PEGylated lipids (DSPE-PEG-Mal). DSPE-PEG-Mal was dissolved to a concentration of 25mg/mL in DMSO and added to the liposomal suspension containing small molecules (e.g. DXR, Nile Red, or Fluorescein) to a final molar ratio of 1% (DSPE/DSPC). The suspension was incubated for one hour at 55 C to increase fluidity of the membrane. Simultaneously, antibodies targeting CD44 were thiolated by incubation with Traut's reagent at a ratio of 3:1 (protein: Traut) for 1 hour. Unreacted Traut's reagent was removed by passing antibody suspensions through a desalting column and the resulting thiolated antibody was added to the liposomal suspension at a molar ratio of 1:1 (DSPE:Protein). The coupling reaction occurred overnight at 4 C . Finally, nonreacted maleimide groups were neutralized by a 15 minute incubation with β-mercaptoethanol at a molar concentration of 3:1 (β-mercaptoethanol:DSPE). The resulting suspension was then filtered as previously described to remove all non-encapsulated and unreacted molecules. The amount of DSPE-PEG-maleimide on liposomes was evaluated following Ellman's test 20 . Briefly, 120 µL of the functionalized liposomes were mixed with 30 µL of L-Cysteine (0.36 mM) using a thermomixer at 700 rpm for 2 hours at room temperature. The particles were removed by filter centrifugation for 30 minutes at 6000 x g. In a transparent 96-well plate, 100 µL of the supernatant and 50 µL of 5,5′-Dithiobis(2nitrobenzoic acid) (0.144 mM) were added and mixed for 30 minutes at room temperature. The absorbance at 409 nm was measured using a microplate reader, and the unreacted L-Cysteine was determined using a standard curve. The amount of conjugated DSPE-PEG-maleimide was calculated by subtracting unreacted L-Cysteine from the total amount of L-Cysteine added to the functionalized liposomes, and the DSPE-PEG-MAL to DSPC ratio was evaluated. In this work, a hydrophilic (Fluorescein sodium salt) and two hydrophobic compounds (Nile red and Doxorubicin) were selected for encapsulation using thermal equilibration. Fluorescein sodium salt was prepared in water at a concentration of 1.5 M while Nile Red and Doxorubicin HCl were prepared in DMSO at concentrations of 30 mM and 172 mM, respectively. The concentrations of small molecule are detailed in the results section. The small molecule solution was added to liposome suspension and incubated at desired temperatures and times. After liposomes were cooled to 4 C, non-encapsulated DXR was removed from suspension using a Zeba Spin desalting column with a 7 kDa cutoff, while excess fluorescent dyes were removed using dye removal columns. For targeted liposomes, DSPE-PEG-Mal was dissolved to a concentration of 25 mg/mL in DMSO and added to a final mol ratio of 1% (DSPE/DSPC). Encapsulated compound concentrations were analyzed using a Thermo Scientific Varioskan Flash multimode reader. DXR concentration was quantified using an excitation of 525 nm and an emission of 580 nm. Liposomes were solubilized in 0.1% Triton x-100 before measuring the fluorescence. The fluorescence intensity was compared to the standard curve which was prepared by measuring the fluorescence of DXR in 0.1% Triton x-100. To quantify the concentration of co-loaded samples, after dissolving liposomes in 0.1% Triton x-100, the fluorescence was measured at an excitation of 488 nm and emission of 512 nm in water for fluorescein. The samples were then desiccated for several hours under vacuum and redissolved in IPA. The fluorescence of Nile Red was measured using an excitation of 550 nm and an emission of 625 nm. The fluorescence intensities of fluorescein and Nile Red from the sample were compared to standard curves prepared with either triton x-100 in water (fluorescein) or IPA (Nile Red) to evaluate encapsulated compound concentrations. DXR release was investigated by diluting samples 20-fold in PBS with 10% FBS and incubating for up to 48 hours at 37 C. At indicated time points, liposome suspension was purified by multiple filtrations to remove released DXR. First, the liposome suspension was filtered through a 7 kDa desalting column to remove free doxorubicin followed by another filtration using a 100 kDa Amicon ultra centrifugal filter unit to remove any serum bound DXR. Liposomes were dissolved in 0.1% Triton x-100, and the fluorescence of each sample was measured using an excitation of 525 nm and an emission of 580 nm. The remaining DXR in liposomes was calculated by comparing the fluorescence intensities to a standard curve prepared with DXR in 0.1% triton x-100. The effect of serum on liposome stability was tested by monitoring the particle size change in either H2O, PBS, or PBS with 10% FBS for up to 24 hours at 37 C. Briefly, 25 µL of liposome solution was dispersed into either H2O, PBS, or PBS with 10% FBS and mixed by thermomixer at 37 C. Samples were collected at different time points (0, 1, 4, 16, and 24 hours), and 20 µL of each sample was introduced into 500 µL of fresh H2O for size measurement using DLS. Phospholipid concentration was empirically determined via Stewart Assay 21 . Briefly, liposomal suspensions were dehydrated at 70 C in polypropylene microcentrifuge tubes. The lipids were then dissolved in 500uL chloroform, and an equal volume of Stewart's reagent (100 mM FeCl 3 and 400 mM NH 4 SCN in DI water) was added and vortexed for 20 seconds. The tubes were centrifuged for 10 minutes at 1000 x g to create a phase separation. The chloroform layer was then analyzed in a quartz cuvette for its absorbance at 472 nm and compared to a standard curve to determine the phospholipid concentration. The standard curve was made using a serial dilution of lipids in cholesterol at a 2:1 ratio. MDA-MB-231 cells were grown in DMEM containing 10% FBS and 1% Penicillin-Streptomycin solution. The cells were cultured at 70% confluence and plated in a 96 well plate at a density of 5000 cells/well. The cells were incubated undisturbed for two days prior to any experimentation. Liposomal suspensions were UV sterilized prior to introduction to cell cultures. On the second day, the liposomal carriers containing DXR were introduced to the cultures at the IC 50 of DXR (0.01 µM 22 ). Proliferating cells were quantified and normalized by the control using WST-1 proliferation assay following the manufactural protocol. Simulations are performed on a section of the lipid membrane consisting of DSPC and cholesterol held in a 2:1 molar ratio. The membrane section was constructed using Charmm-27 GUI Membrane Builder 23 , and designed to consist of 256 total lipid molecules and 128 cholesterol molecules. The membrane was hydrated using 30 water molecules per lipid. Nanoscale Molecular Dynamics (NAMD) was used to thermally equilibrate the bilayer system for 1 ns at designated temperatures. Production runs were also conducted at designated temperatures with a time step of 1 femtosecond and data collection every 20 femtoseconds. Every simulation was run for at least 100 ns, yielding at least 5000 data points per temperature. Trajectories of the simulations were visualized using Visual Molecular Dynamics (VMD). Further information on the simulation setup is available in the supplementary information. Statistical analysis was conducted using GraphPad Prism 5 (La Jolla, CA). For each condition, at least three samples were independently prepared, and each sample was analyzed three times. Statistical analysis was performed using either a student's unpaired t-test or a One-Way Analysis of Variance (ANOVA) with a Tukey post-test. Data was deemed statistically significant if p values were less than 0.05. Graphs show the mean and the standard error of the mean (SEM) of sample groups. Typical liposome synthesis techniques 15, 24, 25 require extensive secondary processing (e.g. sonication 26 , extrusion 27 freeze thawing 28 ) to generate monodisperse SULs. An alternative method that has been utilized to generate SUL's without extensive post processing is through the injection method (Figure 1a) . The injection method is advantageous in its simplicity but typically yields low encapsulation efficiencies 19 . To address this, we have developed a novel approach to efficiently and rapidly sequester small molecules. Small unilamellar liposomes are difficult to concentrate without losing part of the population in the process due to their low densities 29, 30 . Therefore, we proposed to use filter centrifugation instead of ultracentrifugation. To compare the efficiency of ultracentrifugation and filter centrifugation, raw liposome without loading any compound were tested. Empty liposomes were homogeneous with a polydispersity of less than 0.2. By filter centrifuging for thirty minutes at 6000 x g, we effectively concentrated the liposomes 20-fold while retaining the characteristic size and lipid concentration ( Figure 1b and 1c) . The mean population diameter of liposomes increased from 86.6 nm (± 4.5 nm) following synthesis to 135.3 nm (± 7.7 nm) following ultracentrifugation at 60000 x g for two hours. Furthermore, by analyzing the lipid concentration, we found that only 42% (± 2.1%) of the lipids were retained, suggesting that ultracentrifugation can only pellet larger liposomes within the population. As further proof of this, we analyzed supernatant of the ultracentrifuged sample using DLS and routinely found liposomes with diameters of 80 nm (data not shown). Notably, there was a non-significant change in liposome diameter following filter centrifugation (85.3 ± 4.3 nm), and 79.2% (±1.3%) of the lipids were retained. While increased centrifugation speeds (>60,000 x g) and longer centrifugation durations using a specialized ultracentrifugation system may improve liposome retention, the proposed filter centrifugation approach is significantly more efficient and accessible. We evaluated the maximum equilibration concentration (C max ), where the number of drug molecules within the liposomes (C E ) in solution will be proportional to the number of drug molecules in the surrounding media (C T ). The constant of proportionality (α) here is the ratio of the respective volumes, where V Excluded is the cumulative core volume of all liposomes and V Total is the total volume of the solution. V Excluded can be estimated by determining the inner volume of one liposome and extrapolating for the entire liposome population within the sample. Furthermore, if the drug is lipophilic, V Excluded can be substituted for the volume of the lipid membrane. Using our maximum equilibration model, we can predict the D:L ratio as a linear function of concentration and liposome size (Figure 1d ). This function is linear due to the constant lipid molarity, meaning as the liposome size grows, it utilizes lipids that could otherwise be used to form additional liposomes. Additionally, if the density of the lipid membrane does not change as liposome size changes, the maximum equilibration of lipophilic drugs is solely dependent on C T . Therefore, when equilibrating liposomes with a narrow size distribution, we can use C max to hypothesize the liposome potency based on extra-liposomal drug concentration (dependent on drug solubility) as well as the number of liposomes present. By concentrating the samples using filter centrifugation, we can increase the C max and therefore the potency of liposomes. Several factors can affect the diffusion of small molecules across the membrane, including temperature of the system and solvents used. Alcohols such as isopropanol and ethanol offer convenient methods to increase the solubility of many drugs, however, these solvents can also deactivate or precipitate some drugs. DMSO is an alternative that is widely used to enhance the solubility of nonpolar chemicals for drug delivery. To investigate the effect of organic solvent on liposomal stability particularly at temperatures near T m , we individually incubated aqueous solutions of isopropanol, ethanol, and DMSO with concentrated liposomal slurries at 55 C and analyzed their size change. We found that 10% DMSO had the least effect on liposome size (Figure 2a) , with an average shift from 104 nm (± 0.35 nm) to 107 nm (± 0.55 nm). Membrane components immediately precipitated out of the solution when introduced to alcohol concentrations of greater than 40%. Samples that include 10% DMSO were stable for at least 24 hours at 55 C (Figure 2b) . In all cases, we found that the stability of liposomes diminishes rapidly when exposed to temperatures above T m . After characterizing the stability of liposomes during thermal equilibration, we evaluated the liposomal sequestration of therapeutic compounds. Thermal equilibration (Figure 3a) relies on the incubation of concentrated liposomal solutions, where the ratio of drug:lipid (D:L) can be optimized to enhance the potency of the vehicle without using large amounts of therapeutic compounds. Liposomes were incubated with DXR being 10% of the final concentration (D:L ratios of liposomal suspension were at 1.72:1). To validate our hypothesis of temperature affecting thermal equilibration, liposomal slurries were incubated with DXR at varying temperatures ranging from 4 to 55 C (Figure 3b) . We also compared this approach with another commonly used approach of passive encapsulation 14 . When held at identical initial D:L levels, equilibration at T m produced nearly 200 times higher D:L ratios compared to passive encapsulation (0.000239 and 0.041, respectively). While this is lower than current active forms of doxorubicin sequestration, which can achieve D:L of nearly 0.3, it is also less prohibitive and can be applicable to a larger library of molecules. Interestingly, we found that a significant amount of DXR had associated with liposomes equilibrated below T m (D:L ratio of 0.011), which we believe may be due to advantageous binding. We observed a clear exponential trend with time during the initial one-hour incubation (Figure 3c) . We found no difference in liposome diameter following thermal equilibration with DXR solubilized in DMSO (Supplementary Figure 1) . A hallmark of drug delivery systems such as liposomes is their stability and slow release of therapeutics under physiological conditions. This is especially important for the delivery of chemotherapeutic compounds such as DXR which can have severe toxic side effects 31 . Furthermore, extravasation into thick tumors may require 24-48 hours [32] [33] [34] , and therefore formulations that release compounds too quickly are not capable of reaching the required therapeutic index upon reaching the target site. Furthermore, interaction with serum is known to induce liposome degradation. We found that the liposomes remain stable in water and PBS without change in size at least for 24 hours at 37 C, whereas an increase in size of approximately 10% from 105.9 ± 10.2 nm to 117.8 ± 15.3 nm was observed when incubated with 10% FBS. The stability of thermally equilibrated liposomes under 10% FBS condition was confirmed by slow release of DXR with 75% of the encapsulated compound being retained after 48 hours (Figure 3d, Supplementary Figure 2) . A distinguishing characteristic of drug delivery systems is the capability to provide therapeutic benefits through cellular uptake. To demonstrate this, we created drug loaded liposomes using the thermal equilibration technique. Liposomes were incubated with DXR (17.2mM) and DSPE-PEG-MAL (1 mol%) for 1 hour 35,36 at 55 C followed by purification using spin columns. However, it is likely that part of DSPE-PEG-MAL will remain free, forming micelles that would be difficult to filter and subsequently interfere with antibody coupling. Therefore, retention of DSPE-PEG-MAL was calculated by determining the ratio of DSPE-PEG-MAL to DSPC lipid by Ellman's test (Supplementary Figure 3) . The DSPE-PEG-MAL to DSPC ratio of 0.009 ± 0.001 was observed, as compared to the starting ratio of 0.01 based on 1 mol% concentration suggesting that >90% of DSPE-PEG-MAL was retained within the solution following the insertion protocol. The subsequent CD44 conjugation and cancer cell targeting with fluorescein loaded liposomes displayed a strong signal of the liposomal uptake further validating that neither the DSPE-PEG-MAL molecules nor the CD44 antibodies were lost as micelles or during filtration (Figure 5c ). CD44 is highly expressed on MDA-MB-231 cells 37 and was thus used to target DSPE-PEG-MAL incorporated liposomes after functionalization with the thiolated anti-CD44 to the maleimide group. Targeting liposomal DXR (LDXR) was incubated with MDA-MB-231 (at IC 50 ) for 8 or 24 hours prior to being washed away with PBS. To ensure DMSO did not play a role in cell death, vehicles were generated by incubating pre-formed liposomes with DMSO in the absence of DXR. These liposomes were also functionalized with anti-CD44 IgG (HCAM-L). Free DXR was used as a positive control. Cell viability was analyzed over 72 hours using WST-1 as an indicator of proliferation and normalized to the media only control (Figure 3e) . Cells demonstrated a time dependent decline in proliferation, as expected, when incubated with any form of DXR. Both 8 and 24 hours LDXR exposures eventually ended with nearly equivalent (p=0.173 at 48 hours and p=0.603 at 72 hours, n>3) viabilities. Based on the predicted maximum equilibration ratios, liposomes with an average size of 100 nm, should have D:L ratios of 0.4 (Figure 1d) . The prediction assumes that the inner core of the liposome contains the same concentration of compound as the external environment at equilibrium. Since the encapsulation curtails at 1 hour (Figure 3c) , the liposomes are assumed near equilibrated by then, with D:L ratios at 68% of the maximum equilibrium concentration after which it plateaued. We were interested in what might be causing the difference between predicted and observed D:L ratios, and what forces cause the association of DXR with the liposome membrane at temperatures below T m . We hypothesized that the observed association between DXR and liposomes at temperatures T m . Over the 18 C difference between gel phase (37 C) and the T m (55 C), the conformational energy is increased 2.51%. Whereas, a smaller 10 C change in the temperature from 55 C to 65 C netted a similar 2.52% increase in conformational energy. Taken together with our experimental results, we believe that this non-proportional increase in conformational energy results in the observed decrease in liposome stability. While liposomes can mitigate the initial increase in conformational energy, for at least our observation window (24 hours) (Figure 2c and 3b) , they are unable to remain stable with the additional increase in energy. Therefore, while increasing the temperatures beyond T m may increase DXR movement (Figure 4c) , it also greatly decreases the membrane stability and thus does not offer a feasible approach to increase equilibration. The Universal Nature of the Thermal Equilibration Approach Many popular encapsulation approaches are drug specific and are thus not versatile. However, we foresee using passive equilibration as a universal encapsulation technique that could be applicable to a variety of molecules. To demonstrate this, we chose two small molecules that vastly vary in their polarity, Fluorescein, and Nile Red. These molecules are similar in size and have compatible excitation and emission spectra, which allows us to view the co-loading capabilities of our approach (Figure 5a) . Fluorescein and Nile Red (solubilized in DMSO) were introduced to liposomal suspensions at concentrations of 750 mM and 20 mM, respectively. Samples showed D:L levels of 0.03 (±0.013) and 0.003 (±0.000099), respectively (Figure 5b) . Furthermore, we demonstrate efficient targeting of liposomes to cancer cells and in vitro cellular uptake using the proposed technique (Figure 5c) . Here, PEGylated liposomes were created through "Preinsertion" (DSPE-PEG-Mal was added into the IPA solution pre-synthesis), or "Post-insertion" (DSPE-PEG-Mal was integrated in following the synthesis via thermal equilibration along with fluorescein). Both approaches yielded functionalized targeting liposomes, as demonstrated by the presence of fluorescein when examining cells exposed to PEGylated liposomes, and comparing their fluorescence to cells introduced to non-PEGylated liposomes. We show here a universal approach to sequester small molecules within liposomes for drug delivery, irrespective of the drug polarity. We hypothesized that by incubating liposomes with small volumes of high concentration drug solution, the internal environment of liposomes would eventually equilibrate with the external environment, with localization largely depending on the partitioning coefficient of the compound (Figure 3a) . We further hypothesized that this effect could be enhanced by increasing the fluidity of the membrane (via thermal incubation), which would also allow the introduction of targeting moieties within the membrane. A lipid vesicle's T m is a physical characteristic that dictates the order of the lipid membrane at varying temperatures, and it is a crucial variable to consider in engineering liposomal delivery systems, where synthetic lipids have a higher T m than natural lipids 40 . DSPC is a synthetic, saturated, 18 carbon lipid with a T m of 55 C. This elevated T m imparts several favorable characteristics on the liposome, including enhanced in vivo stability and reduced clearance rates compared to natural lipids 41 . We synthesized liposomes via a modified form of our previously published SPIN method (Figure 1a) 42 , and concentrated using filter centrifugation, which we found was significantly superior to ultracentrifugation (Figure 1b and 1c) . Liposomes display high stability when incubated at the T m in up to 10% DMSO (as a vehicle of lipophilic molecules). We found that above this temperature, and when incubated with isopropyl or ethyl alcohol, liposomes rapidly fell apart resulting in the precipitation of lipids. However, previous reports by Balley et. al. have demonstrated that DSPC liposomes were stable for up to one hour at 70 C in up to 30% ethanol 43 . This discrepancy may be due to the increased levels of cholesterol (45%) in their liposomes as compared to 33% cholesterol in our experiments, which is reported to be the most stable 41, 44, 45 . Cholesterol has a stabilizing effect, creating order in the hydrophobic region at increased temperatures that could stabilize liposomes at these increased temperatures 46, 47 . In our simulations, we found no interaction between the DXR and cholesterol (Supplementary Figure 5) , but simulations to determine the effect of increased cholesterol on conformational energy may explain this discrepancy. In our simulated leaflet containing 33% cholesterol, we found little difference in the interactions between DXR and DSPC once the temperature was >T m . Electrostatic, VdW, and hydrogen bonding interactions (Figures 4a and 4b) were similar at both 55 C and 65 C. However, there was a larger than expected increase in conformational energy when comparing 55 C and 65 C membrane simulations. This increase in conformational energy also affects other membrane dynamics which can lead to loss of stability, such as the solvent accessible surface area (Supplementary Figure 6) . We showed equilibration as a robust approach to create potent targeting liposomes. The liposomes quickly equilibrated within an hour (Figure 3b ) and can provide a targeted therapeutic effect (Figure 3e) . Interestingly, we found that even at lower temperatures (i.e. in vivo activity Liposome: methods of preparation and applications Pharmacokinetics and antitumor effect of doxorubicin carried by stealth and remote loading proliposome Remote loading of doxorubicin into liposomes driven by a transmembrane phosphate gradient Remote loading of preencapsulated drugs into stealth liposomes SPIN: Rapid Synthesis, Purification, and Concentration of Small Drug Loaded Liposomes Quantification of PEG-Maleimide Ligands and Coupling Efficiencies on Nanoparticles with Ellman's Reagent Colorimetric determination of phospholipids with ammonium ferrothiocyanate Doxorubicin induces drug resistance and expression of the novel CD44st via NF-κB in human breast cancer MCF-7 cells CHARMM: The Biomolecular Simulation Program Liposome: classification, preparation, and applications Comparison of Liposomes Formed by Sonication and Extrusion: Rotational and Translational Diffusion of an Embedded Chromophore Evaluation of Extrusion Technique for Nanosizing Liposomes Formation of unilamellar vesicles by repetitive freeze-thaw cycles: characterization by electron microscopy and 31P-nuclear magnetic resonance Effect of cholesterol and ethanol on dermal delivery from DPPC liposomes Remote loading of doxorubicin into liposomes by transmembrane pH gradient to reduce toxicity toward H9c2 cells Size-dependent extravasation and interstitial localization of polyethyleneglycol liposomes in solid tumor-bearing mice Pharmacokinetics and in vivo drug release rates in liposomal nanocarrier development Mechanistic model and analysis of doxorubicin release from liposomal formulations Influence of poly(ethylene glycol) grafting density and polymer length on liposomes: Relating plasma circulation lifetimes to protein binding An immobilized nanoparticle-based platform for efficient gene knockdown of targeted cells in the circulation CD44 isoforms are heterogeneously expressed in breast cancer and correlate with tumor subtypes and cancer stem cell markers Quantification of the hydrophobic interaction by simulations of the aggregation of small hydrophobic solutes in water Fluidity enhancement: a critical factor for performance of liposomal transdermal drug delivery system A review on phospholipids and their main applications in drug delivery systems Liposomes with prolonged circulation times: factors affecting uptake by reticuloendothelial and other tissues SPIN: rapid synthesis, purification, and concentration of small drug-loaded liposomes A simple passive equilibration method for loading carboplatin into pre-formed liposomes incubated with ethanol as a temperature dependent permeability enhancer Ligand-targeted liposomes Influence of cholesterol on liposome stability and on in vitro drug release Influence of Cholesterol on the Phase Transition of Lipid Bilayers: A Temperature-Controlled Force Spectroscopy Study Influence of cholesterol inclusion on the doxorubicin release characteristics of lysolipid-based thermosensitive liposomes Liquid-ordered phases induced by cholesterol: a compendium of binary phase diagrams Liposome co-encapsulation of synergistic combination of irinotecan and doxorubicin for the treatment of intraperitoneally grown ovarian tumor xenograft Co-loaded paclitaxel/rapamycin liposomes: Development, characterization and in vitro and in vivo evaluation for breast cancer therapy Lung cancer combination therapy: co-delivery of paclitaxel and doxorubicin by nanostructured lipid carriers for synergistic effect The authors are grateful to Dr. Rohan Fernandes at The George Washington University for allowing the use of their analytical equipment and other resources for this project.