Liposome-based mucus-penetrating particles (MPP) for mucosal theranostics: Demonstration of diamagnetic chemical exchange saturation transfer (diaCEST) magnetic resonance imaging (MRI)

Abstract

Mucus barriers lining mucosal epithelia reduce the effectiveness of nanocarrier-based mucosal drug delivery and imaging (“theranostics”). Here, we describe liposome-based mucus-penetrating particles (MPP) capable of loading hydrophilic agents, e.g., the diaCEST MRI contrast agent barbituric acid (BA). We observed that polyethylene glycol (PEG)-coated liposomes containing ≥7 mol% PEG diffused only ~10-fold slower in human cervicovaginal mucus (CVM) compared to their theoretical speeds in water. 7 mol%-PEG liposomes contained sufficient BA loading for diaCEST contrast, and provided improved vaginal distribution compared to 0 and 3 mol%-PEG liposomes. However, increasing PEG content to ~12 mol% compromised BA loading and vaginal distribution, suggesting that PEG content must be optimized to maintain drug loading and in vivo stability. Non-invasive diaCEST MRI illustrated uniform vaginal coverage and longer retention of BA-loaded 7 mol %-PEG liposomes compared to unencapsulated BA. Liposomal MPP with optimized PEG content hold promise for drug delivery and imaging at mucosal surfaces.

Background

Mucosal drug delivery via nano-carriers holds potential to improve detection and treatment of numerous diseases.^{1, 2} For efficient mucosal delivery, nano-carriers must first bypass the highly protective mucus linings that rapidly remove most foreign particles from the mucosae.³ To overcome the mucus barrier, we have previously developed polymer- and pure drug-based nanoparticulates that possess dense coatings with polyethylene glycol (PEG) that effectively avoid mucoadhesion, thus allowing rapid penetration through mucus.⁴ As a result, these mucus-penetrating particles (MPP) provide more uniform distribution and sustained delivery of therapeutics at various mucosal sites.^5–7

Liposomes were the first nano-carrier system to be developed and translated for clinical use.⁸ Although liposomal systems have been explored for mucosal delivery,^{9, 10} there has not been a focus on directly observing the interactions of liposomal formulations with mucus, and how these interactions impact mucosal distribution. Here, we varied the composition of PEG-conjugated lipids to investigate the effect of PEG surface density on liposome mobility in human cervicovaginal mucus (CVM) and vaginal distribution in vivo. Chemical Exchange Saturation Transfer (CEST) Magnetic Resonance Imaging (MRI) contrast agents have shown great promise, particularly due to their capability of being distinguished from each other through artificial color contrast.¹¹ As a demonstration, we loaded the liposomal MPP with barbituric acid (BA), a water-soluble, highly biocompatiblediamagnetic CEST (diaCEST) contrast agent,¹² and monitored the vaginal distribution and retention of the liposomes via MRI.^{11, 13}

Methods

Liposomes composed of 1,2-disteoroyl-sn-glycero-3-phosphatidylcholine (DSPC), cholesterol, and 1,2-distearoyl-sn-glycerophosphoethanolamine poly(ethylene glycol)₂₀₀₀ (DSPE-PEG_2k) were prepared and characterized following procedures adapted from previous reports.^{5, 12, 14, 15} Human CVM was collected following a protocol approved by the Johns Hopkins School of Medicine Institutional Review Board; all patients provided written informed consent. All animal studies were performed in accordance to protocols approved by the Institutional Animal Care and Use Committee (IACUC) at the Johns Hopkins University; humane care of all animals used in the studies was provided. Full methods, including liposome tracking in mucus, animal studies, and MRI protocols, are provided in Supporting Information. Data represent mean ± standard error of the mean (S.E.M.).

Results

We first formulated DSPC liposomes containing 6 different ratios of DSPE-PEG_2k (Table 1). We used extrusion to reduce the mean diameters of all formulations to below the average mesh size of human CVM (~340 nm)⁴ to minimize steric hindrance. The PEGylated formulations were relatively uniform in size (low polydispersity index, or PDI), whereas non-PEGylated liposomes displayed high PDI, implying aggregation occurred. We also measured the actual molar fraction of DSPE-PEG_2k and estimated the PEG surface density. The Γ/SA ratios (theoretical area covered by unconstrained PEG chains vs. total surface area of a liposome,¹⁴ see Supporting Information) suggest that liposomes with ≥ 7 mol%-PEG were coated with brush-like PEG chains forming effective surface shielding, whereas those with ≤ 5 mol%-PEG were covered with mushroom-like PEG chains and, thus, less effectively shielded.¹⁴

Table 1.

Characterization of DSPC liposomes at different PEGylation levels^[a]

Sample	Number Mean Diameter (nm)	Polydispersity (PDI)	Actual Mol% of DSPE- PEG2k	PEG Density (Chains/ 100 nm2)	[Γ/SA][b]	MSDw/<MSD>m[c]
						0h	3h
0 mol%-PEG	129 ± 18	0.37 ± 0.06	NA	NA	NA	110	270
3 mol%-PEG	134 ± 9	0.09 ± 0.01	3.2 ± 0.1	7.2	0.6	13	25
5 mol%-PEG	121 ± 9	0.06 ± 0.02	4.9 ± 0.1	10.9	0.9	14	31
7 mol%-PEG	139 ± 4	0.06 ± 0.01	6.2 ± 0.1	13.9	1.2	8	15
10 mol%-PEG	147 ± 9	0.03 ± 0.01	8.5 ± 0.2	18.8	1.6	8	15
12 mol%-PEG	149 ± 5	0.04 ± 0.01	10.6 ± 0.3	23.7	2.0	6	7
PS-COOH	91 ± 1	0.04 ± 0.01	NA	NA	NA	1,400	NA

^[a]Containing 3-(trimethyl-silyl)-1-propanesulfonic acid sodium salt for NMR measurements.

^[b]Ratio of theoretical area covered by unconstrained PEG chains vs. total surface area of a liposome; detailed calculations are provided in Supporting Information.

^[b]Ratio of theoretical MSD in water vs. <MSD> measured in CVM; detailed calculations are provided in Supporting Information.

We then investigated the diffusion of liposomes immediately (0 h) and 3 h after addition to CVM via multiple particle tracking (MPT).¹⁵ PEGylated liposomes diffused overall faster than the non-PEGylated liposomes, exhibiting more diffusive trajectories and ~10-fold higher ensemble-averaged mean-squared displacement (<MSD>) (Figure 1, A-B). A significant population of immobilized non-PEGylated liposomes was revealed in the logarithmic distribution of individual liposome MSD (Figure 1, C). The <MSD> of PEGylated and non-PEGylated liposomes was ~10- and 110-fold slower than their theoretical MSD in water (t = 1 s), respectively (Table 1). After 3 h incubation in CVM, liposomes with lower PEG content (0 – 5 mol%) displayed more restricted trajectories and ~2-fold decrease in <MSD>, with an increased immobilized fraction (Figure 1, A-C). Overall, liposomes with ≥ 7 mol% PEG diffused similarly in CVM compared to polymeric MPP (MSD_w/<MSD>_m ~10)^{5, 14} and the mobility remained stable over time. Next, we prepared BA-loaded liposomes for diaCEST MRI (Table 2). BA encapsulation minimally affected the liposome size and PDI. The loading capacity (BA:lipid ratio) correlated inversely with PEG content, with a substantial decrease at 12 mol%-PEG, likely due to reduced free volume associated with high PEG content on the inner surface of the liposomal shell, and the increased permeability of the lipid bilayer.¹⁶ The in vitro CEST contrast was generally consistent with the BA loading level.

Figure 1.

Mobility of PEGylated and non-PEGylated DSPC liposomes 0 h or 3 h post addition to CVM. A. Representative liposome trajectories over 1 s. B. <MSD> as a function of time scale. C. Distributions of the logarithms of individual liposome MSD.

Table 2.

Characterization of BA-loaded liposomes

Sample	Number Mean Diameter (nm)	Polydispersity (PDI)	BA:Lipid Ratio (%)	In vitro CEST Contrast at 5 ppm (%)
0 mol%-PEG	113 ± 12	0.28 ± 0.06	23 ± 4	32 ± 2
3 mol%-PEG	130 ± 5	0.05 ± 0.01	23 ± 3	28 ± 2
7 mol%-PEG	126 ± 7	0.06 ± 0.01	21 ± 1	21 ± 5
12 mol%-PEG	130 ± 3	0.08 ± 0.01	13 ± 4	13 ± 3

We next investigated the vaginal distribution of BA-loaded liposomes in the vaginas of mice in the estrus phase of their estrous cycle.⁵ Particle mobility in mucus has been demonstrated to correlate with in vivo mucosal distribution.^5–7 Similarly, we observed non-uniform distribution of mucoadhesive, non-PEGylated liposomes, which appeared to outline mucin bundles (Figure 2, A). This non-uniform distribution was also reflected by a high variance-to-mean ratio (VMR, increased VMR reflects decreased uniformity) (Figure 2, E). While all PEGylated liposomes provided improved vaginal distribution, 7 mol%-PEG liposomes demonstrated the most uniform coverage with the lowest VMR. Additionally, individual cell outlines were observed, implying that the 7 mol%-PEG liposomes were able to reach the vaginal epithelium (Figure 2, C). Liposomes with less PEG content may be insufficiently shielded to avoid mucoadhesion in vivo. Despite rapid diffusion in CVM, the 12 mol%-PEG liposomes also distributed suboptimally in vivo (Figure 2, D), perhaps due to their disassembly via micellization in vivo.¹⁷ Therefore, PEG content must be optimized to eliminate mucoadhesion while maintaining stability in vivo.

Figure 2.

Distribution of rhodamine-labeled (red) BA-loaded liposomes on flattened mouse vaginal tissues. A-D. Representative fluorescence images for different PEGylation levels. E. Variance-to-mean ratio of fluorescence intensity (Lower values indicate increased uniformity).

Finally, we monitored the vaginal retention of BA-loaded liposomes via diaCEST MRI. 7 mol%-PEG liposomes were selected as liposomal MPP given their sufficient loading and retention of BA (Figure S1) and most uniform vaginal distribution. Liposomal MPP displayed good vaginal coverage with prolonged CEST contrast (at least 90min; highest relative MTR_asym^5ppm~4%); much shorter vaginal retention time was observed for unencapsulated BA (~30 min; highest relative MTR_asym^5ppm ~1%) (Figure 3, A-B). The increase in CEST contrast over time for liposomal MPP was likely due to initial spreading throughout the vaginal tract, followed by liposome concentration at the epithelial surface as fluid was absorbed by the epithelium (Figure 3, B). At 90 min, images of liposomal MPP exhibited a significant fraction of high contrast pixels (MTR_asym^5ppm~5%) (Figure 3, C).

Figure 3.

Monitoring of intravaginally administered BA-loaded liposomal MPP and unencapsulated BA via MRI in mice. A. Representative T2-weighted and CEST images. B. Relative MTR_asym over time. C. Histogram of pixelated MTR_asym at 90min.

Discussion

The <MSD> of non-PEGylated liposomes in CVM was slightly higher than that of the mucoadhesive polystyrene nanoparticles,⁵ implying that the zwitterionic DSPC headgroups may have reduced adhesive (hydrophobic, electrostatic, etc.) interactions with mucus.⁴ Nevertheless, adhesion of the liposomes to mucus was apparent, and aggregation of non-PEGylated liposomes may have resulted in further immobilization. PEG-shielding was required to allow effective mucus-penetration; this need for PEG-shielding would likely be higher for liposomes composed of more mucoadhesive lipids, e.g., comprising non-neutral headgroups, such as cationic lipids commonly used for nucleic acid delivery. The gradually decreased mobility of insufficiently PEGylated liposomes in CVM may also result from dissociation of PEGylated lipids over time.

CEST MRI has been previously used to monitor liposomes administered intratumorally and systemically,^{12, 18} and here we demonstrate the capability of diaCEST MRI for non-invasive monitoring of liposomes administered intravaginally. Different liposome formulations could also be loaded with CEST agents having distinct chemical shifts, thereby allowing simultaneous monitoring for direct comparison and evaluation of interactions between the formulations.¹¹ Additionally, diaCEST liposomal MPP may be co-loaded with therapeutics via carefully designed schemes for treatment.¹² These capabilities could enable clinical evaluation of nanocarrier based vaginal therapies, especially when combined with new imaging methods.^{19, 20}

In summary, we developed liposomal MPP with optimized surface PEG shielding and capable of loading hydrophilic agents like BA. PEGylation, particularly at levels ≥ 7 mol%, enhanced the mobility of liposomes in human CVM. However, increasing PEGylation to ~12 mol% compromised drug encapsulation and in vivo distribution. Moderately PEGylated liposomes (~7 mol%) maintained encapsulation efficiency while distributing most uniformly in the mouse vagina. Using non-invasive diaCEST MRI, we showed that liposomal MPP provided uniform vaginal coverage and retained BA for ≥ 90 min in vivo. These results demonstrate the potential of liposomal MPP for mucosal delivery and imaging, and suggest that liposomal MPP formulations may be suitable for theranostics in mucosal surfaces, like that of the vagina.

Abbreviations

BA

barbituric acid

CVM

cervicovaginal mucus

diaCEST

diamagnetic chemical exchange saturation transfer

MPP

mucus-penetrating particles

MPT

multiple particle tracking

MRI

magnetic resonance imaging

PEG

polyethylene glycol

DSPC

1,2-disteoroyl-sn-glycero-3-phosphatidylcholine

DSPE-PEG_2k

1,2-distearoyl-sn-glycerophosphoethanolamine poly(ethylene glycol)₂₀₀₀