The problem the researchers set out to solve is a stubborn one in pharmacology. Hydrophilic, or water-loving, drugs diffuse rapidly away from an injection site, making it difficult to maintain therapeutic concentrations locally without exposing the rest of the body. This is especially dangerous for drugs with narrow therapeutic indices, where the difference between an effective dose and a toxic one is small. Existing platforms—implantable devices, hydrogels, polymeric microparticles, and conventional liposomes—each carry drawbacks, from the need for surgical implantation to complex chemical modifications that complicate manufacturing.

Liposomes, spherical vesicles made of lipid bilayers enclosing aqueous compartments, have long been favored carriers because they can encapsulate water-soluble drugs and present a diffusion barrier. The standard strategy has been to stiffen their membranes with cholesterol and saturated phospholipids possessing high phase-transition temperatures, on the theory that a less fluid membrane slows drug escape. Unsaturated lipids, whose double bonds introduce kinks into the acyl chains and lower the phase-transition temperature, were assumed to produce more fluid, more permeable membranes and therefore faster release and poorer encapsulation.

To test this assumption rigorously, the team compared liposomes made from phospholipids of identical 18-carbon chain length but varying numbers of double bonds, prepared under matched conditions with 28.6 mol% cholesterol. The model drug was tetrodotoxin (TTX), an ultrapotent local anesthetic that is extremely hydrophilic—its octanol/water partition coefficient logarithm is −4.44—and notoriously difficult to encapsulate, typically achieving only about 25% efficiency with passive loading. The results were striking: liposomes made from 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), which carries two double bonds, achieved roughly 72% encapsulation efficiency, about three times higher than the saturated DSPC formulation, with drug loading also threefold greater.

Release kinetics told an even more surprising story. In the first hour after preparation, only a small fraction of TTX escaped from DOPC liposomes—11-fold less than from the saturated DSPC version—while the four-double-bond DLPC formulation suppressed initial release 14-fold. Over seven days at body temperature, DOPC liposomes released just 1.3% of their TTX payload, roughly four times less than the saturated analogue. Because smaller particles release drug faster due to their higher surface area-to-volume ratios, and the unsaturated liposomes were in fact smaller, the effect could not be explained by size. The team also confirmed that direct chemical interactions between TTX and DOPC were absent by proton nuclear magnetic resonance spectroscopy.

The mechanism, revealed through cryogenic electron microscopy and confocal laser scanning microscopy, lay in ultrastructure. As the number of double bonds increased, the liposomes spontaneously reorganized from simple single-bilayer spheres into complex multilamellar and multivesicular (ML/MV) architectures—onions of nested membranes and vesicles-within-vesicles. A quantitative membrane area fraction, measured from fluorescently stained images, correlated almost perfectly (r = −0.99) with reduced one-hour release. A dithionite-quenching assay on fluorescently labeled liposomes independently confirmed the greater lamellarity of DOPC formulations, and liposomes made with negatively charged DOPG, which repel adjacent bilayers electrostatically and remain unilamellar, released drug nearly tenfold faster despite having the same degree of unsaturation.

The benefits extended broadly across hydrophilic molecules. Compared with saturated DSPC liposomes, DOPC formulations showed 1.6- to 3.3-fold higher encapsulation and loading for metformin, moroxydine, neosaxitoxin, the STING agonist 2′3′-cGAMP, the peptides insulin and teriparatide, and the protein interleukin-2, with one-hour release reduced 5- to 17-fold and seven-day release 2- to 9-fold slower. Notably, the more hydrophilic the molecule, the greater the advantage—consistent with the idea that each additional bilayer in the diffusion path presents a formidable barrier to water-soluble drugs. In contrast, the amphiphilic anesthetic bupivacaine, which can partition into and permeate lipid membranes, showed the opposite trend, releasing faster from more unsaturated liposomes.

The in vivo proof of principle came in a rat sciatic nerve block model, a demanding test because TTX is lethal at low doses if it escapes systemically. Free TTX at 5 micrograms was uniformly fatal, and saturated-liposome formulations at 110 micrograms produced only hours of block with signs of systemic spread. The DOPC liposomes, by contrast, delivered sensory nerve blockade lasting from 6.5 days at 80 micrograms of TTX to 13 days at 260 micrograms—50 times the lethal free dose—with no fatalities and essentially no contralateral deficits indicating systemic toxicity. When bupivacaine was co-loaded to accelerate onset, block was established within 30 minutes and lasted over eight days. Fluorescence imaging showed that both liposome types persisted at the injection site for weeks, confirming that the prolonged effect stemmed from slower release rather than slower clearance, and histology revealed inflammation and myotoxicity scores comparable to or better than the commercial liposomal bupivacaine product Exparel.

The implications reach well beyond anesthesia. A simple, injectable platform that passively achieves high loading and ultra-slow release of hydrophilic small molecules, peptides, and proteins—without covalent chemistry, aptamers, or device implantation—could transform sustained delivery of cancer immunotherapies like interleukin-2 and 2′3′-cGAMP, antiviral agents, and drugs for metabolic disease. The authors note that nanoscale extruded DOPC liposomes retained their ML/MV structure and sustained release, opening a path toward systemic applications. By reframing lipid unsaturation from a liability into an asset, the study suggests that some of drug delivery’s most stubborn constraints may yield not to more elaborate engineering, but to a reconsideration of the physics already hidden inside the lipid bilayer.

Subject of Research: Liposome-based sustained local drug delivery using unsaturated phospholipids

Article Title: Ultra-slow release of hydrophilic drugs via multilamellar–multivesicular liposomes formed by unsaturated phospholipids

Article References: Wang, Y., Xue, T., Torre, M., Han, Y., Shao, R., & Kohane, D. S. (2026). Ultra-slow release of hydrophilic drugs via multilamellar–multivesicular liposomes formed by unsaturated phospholipids. Nature Biomedical Engineering. https://doi.org/10.1038/s41551-026-01793-6

Image Credits: AI Generated

DOI: 10.1038/s41551-026-01793-6

Keywords: liposomes, drug delivery, unsaturated phospholipids, DOPC, tetrodotoxin, multilamellar vesicles, local anesthesia, controlled release, hydrophilic drugs, nerve block, biomedical engineering, Nature Biomedical Engineering

The trial, led by Chunmei Zhao and Fang Luo of Beijing Tiantan Hospital at Capital Medical University, together with colleagues at multiple centers in China and Nepal, enrolled 80 patients whose episodic cluster headache had resisted oral medications. To be eligible, participants had to present to a pain clinic within five days of the onset of a new cluster period, a narrow window that allowed the researchers to capture the natural arc of an episode from its earliest days. Patients were randomly assigned in equal numbers to one of two procedures targeting the sphenopalatine ganglion: pulsed radiofrequency treatment or a conventional nerve block. Everyone in both groups then took oral verapamil at a dose of 360 milligrams per day, the standard preventive drug for cluster headache, ensuring that the comparison isolated the effect of the procedure itself.

The sphenopalatine ganglion, also called the pterygopalatine ganglion, has long fascinated headache specialists because it acts as a major relay station for the facial pain pathways and parasympathetic autonomic signals that surge during cluster attacks. Attacks of cluster headache involve intense pain around one eye or temple, typically accompanied by tearing, nasal congestion, eyelid drooping, and restlessness, all hallmarks of trigeminal-autonomic activation in which the sphenopalatine ganglion plays a central role. Blocking this ganglion with local anesthetics has been a recognized short-term rescue strategy for years, but the effects of a nerve block fade within days. Pulsed radiofrequency, by contrast, applies brief bursts of an alternating electrical field to the ganglion through a needle electrode, generating controlled thermal and electromagnetic effects that are thought to modulate pain signaling without destroying nerve tissue, potentially offering relief that lasts for months.

The technical execution of the procedure demands precision. Under imaging guidance, clinicians advance a specialized electrode through the face to reach the pterygopalatine fossa, the small anatomical pocket where the ganglion sits, and confirm correct placement before delivering the pulsed radiofrequency cycles. In this trial the technique was performed as an add-on to verapamil rather than a replacement for it, reflecting the reality that most patients with cluster headache need layered treatment. The researchers registered the study prospectively on ClinicalTrials.gov as NCT03567590, obtained ethics approval from the Institutional Review Board of Beijing Tiantan Hospital, and followed Consolidated Standards of Reporting Trials guidelines, with a blinded-endpoint design and a Data and Safety Monitoring Board overseeing the study, features that strengthen confidence in the findings.

The primary endpoint was the duration of the cluster period, and the difference between the groups was dramatic. Patients treated with pulsed radiofrequency had a median cluster-period duration of 15.0 days, with an interquartile range of 6.2 to 30.0 days, while those treated with nerve block had a median duration of 61.0 days, with an interquartile range of 17.0 to 90.0 days. The median difference between groups was minus 34.0 days, with a 95 percent confidence interval of minus 56.0 to minus 17.0 days and a P value below 0.001, a result that is both statistically robust and clinically profound. In practical terms, patients who received the pulsed radiofrequency treatment were spared roughly an additional month of recurrent attacks compared with those receiving the standard block.

Secondary endpoints told a consistent story across nearly every measure of attack burden. Pain intensity during attacks, scored on the numerical rating scale, was significantly lower in the pulsed radiofrequency group at day 3, week 1, week 2, and month 1, with median between-group differences of minus 1.0, minus 2.0, minus 3.0, and minus 4.0 points respectively, the last representing a large shift on a scale where every point matters to a patient in agony. Attack frequency and attack duration were also significantly reduced in the pulsed radiofrequency group at week 1, week 2, month 1, and month 3, and patients used less rizatriptan, a fast-acting triptan medication taken to abort individual attacks, over the same follow-up points. By month 6 and year 1 the between-group differences on these secondary measures were no longer statistically significant, which is consistent with the episodic nature of the disease, since many patients in both groups had exited their cluster periods by then.

Safety is always a central question when a procedure involves threading an electrode near critical structures at the skull base, and the trial reported few adverse effects with pulsed radiofrequency treatment. The authors concluded that in patients with episodic cluster headache resistant to oral medications, sphenopalatine ganglion pulsed radiofrequency combined with verapamil provided greater pain relief than sphenopalatine ganglion nerve block combined with verapamil, with few adverse effects. The study was supported by the National Key Research and Development Program of China and by Capital’s Funds for Health Improvement and Research, among other funders, which had no role in the design, analysis, or decision to publish. The authors declared no competing interests, and all participants provided written informed consent before enrollment.

The implications reach beyond the immediate result. Cluster headache affects roughly one in a thousand people, and while abortive options such as subcutaneous sumatriptan, high-flow oxygen, and newer CGRP-targeting therapies exist, preventive options remain limited and verapamil, the mainstay, requires slow titration and cardiac monitoring. For patients in the midst of a cluster period, every day of shortened episode duration translates into dozens of avoided attacks, restored sleep, and reduced reliance on acute medications. If pulsed radiofrequency of the sphenopalatine ganglion can reliably compress a cluster period from two months to two weeks, it could reshape the treatment algorithm for medication-refractory episodic disease, positioning a one-time procedural intervention alongside pharmacology rather than as a last resort.

Important caveats temper the enthusiasm. The trial focused specifically on episodic cluster headache, the form in which attacks occur in bounded bouts separated by remission periods, and its results cannot be assumed to extend to chronic cluster headache, in which attacks persist for a year or more without relief. The comparison arm was an active control, a nerve block rather than a sham procedure, which is methodologically respectable but means the study cannot fully separate the specific neuromodulatory effects of pulsed radiofrequency from the general effect of any ganglion-directed intervention combined with verapamil. The sample of 80 patients, while adequate for the primary endpoint, is modest, and the open nature of the treatment delivery, with only the endpoint assessment blinded, leaves some room for expectation effects on the self-reported secondary outcomes.

Even with those limitations, the trial stands as one of the most convincing demonstrations to date that targeted neuromodulation of the sphenopalatine ganglion can alter the natural history of a cluster period rather than merely blunting individual attacks. The prospective registration, blinded endpoint assessment, multicenter recruitment, and consistent signal across primary and secondary outcomes give the finding unusual weight for a procedural pain study. For the patients who spend their cluster seasons pacing dark rooms waiting for the storm to pass, the message from this research is concrete and hopeful: a precisely delivered field of pulsed electrical energy, aimed at a ganglion the size of a grain of rice hidden behind the cheek, may cut weeks of suffering from the cruelest headache disorder medicine knows. Larger studies and replication in other populations will be needed before the procedure becomes standard practice, but the direction of the evidence now points firmly toward the sphenopalatine ganglion as the right target at the right time.

Subject of Research: Pulsed radiofrequency neuromodulation of the sphenopalatine ganglion for treating episodic cluster headache

Article Title: Efficacy and safety of sphenopalatine ganglion pulsed radiofrequency in episodic cluster headache: a randomized controlled trial

Article References: Zhao, C., Bian, J., Ren, H., Wu, D., Meng, L., Shen, Y., Wang, B., Shrestha, N., & Luo, F. (2026). Efficacy and safety of sphenopalatine ganglion pulsed radiofrequency in episodic cluster headache: a randomized controlled trial. BMC Medicine. https://doi.org/10.1186/s12916-026-05242-8

Image Credits: AI Generated

DOI: 10.1186/s12916-026-05242-8

Keywords: cluster headache, sphenopalatine ganglion, pulsed radiofrequency, randomized controlled trial, verapamil, nerve block, neuromodulation, pain management, neurology, headache, trigeminal-autonomic, BMC Medicine

This architectural surprise matters because it solves one of the most stubborn problems in drug delivery: the sustained release of hydrophilic, or water-loving, drugs. Molecules that dissolve readily in water tend to leak out of conventional liposomes rapidly, producing a burst release that drains most of the payload within hours or days. For applications that demand drug levels to be maintained over weeks—long-acting analgesia, psychiatric medications, or depot formulations for chronic disease—that burst behavior has been a fundamental obstacle. The new multicompartment liposomes suppress burst release almost entirely and enable what the authors describe as ultra-slow release kinetics.

The practical payoff was demonstrated in a rat model of regional anesthesia. When the multilamellar–multivesicular liposomes were loaded with tetrodotoxin, a potent hydrophilic sodium-channel blocker, they produced a sciatic nerve block lasting several weeks. Nerve blocks with conventional formulations of local anesthetics typically last hours to at most a day or two. A weeks-long block from a single injection, if it can be translated safely to patients, could transform postoperative pain management, reducing or eliminating the need for opioid analgesics after major surgery.

To understand why the finding is unexpected, it helps to consider lipid phase behavior. Saturated phospholipids, with straight hydrocarbon tails, pack tightly into ordered gel-phase membranes and are the traditional building blocks of multilamellar vesicles used for slow release. Unsaturated phospholipids, by contrast, carry kinks in their tails where double bonds interrupt the chain. Those kinks frustrate tight packing, keeping membranes fluid and permeable—which is precisely why unsaturated lipids dominate the composition of many clinically approved liposomal drugs designed for efficient release and circulation. The intuition, therefore, has been that unsaturated lipids are poorly suited to retention-based, slow-release formulations.

Wang and colleagues found the opposite. When formulated under their conditions, unsaturated phospholipids self-organized into complex, hierarchical vesicles in which multiple bilayers are wrapped around, or packed alongside, multiple aqueous compartments. A hydrophilic drug loaded into these internal aqueous spaces must diffuse across several lipid barriers and traverse tortuous pathways between compartments before it can reach the surrounding tissue. Each bilayer acts as a diffusion barrier, and the compartmentalized geometry multiplies the effective path length. The result is a release profile that unfolds over days and weeks rather than minutes and hours, without the initial burst that plagues single-compartment vesicles.

Efficient encapsulation is the other half of the achievement. Hydrophilic drugs partition into the aqueous volume of liposomes, but in unilamellar vesicles that volume is limited and the enclosing membrane is thin, so leakage is rapid. In the multivesicular architecture, the internal aqueous volume is substantially larger and subdivided into many isolated reservoirs. Even if some compartments rupture or empty early, others remain sealed, smoothing the release curve. The study reports that these structures encapsulate hydrophilic drugs efficiently and retain them until the lipid matrix itself degrades or remodels slowly in tissue.

The broader context is a field increasingly focused on long-acting formulations. Reviews by Li and colleagues in Nature Reviews Materials and by Gao, Karp, Langer and Joshi in Chemistry of Materials have argued that extending the duration of action of drugs is one of the central challenges of pharmaceutical science, particularly for molecules that the body clears quickly. Injectable depots, implants and in situ gels all attempt to hold a drug near its target and meter it out slowly, but each carries trade-offs in invasiveness, reproducibility and manufacturability. Liposomes have long been attractive because they are biocompatible, biodegradable and produced at scale—yet their classic limitation with hydrophilic payloads has kept them out of most long-acting applications.

Previous efforts to slow liposomal release have included manipulating membrane cholesterol content, crosslinking bilayers, polymer-coating vesicles, and, most recently, engineering the lipid chemistry itself. A 2023 study in Nature Communications by Li and colleagues reported aromatized liposomes in which modified phospholipids increased drug loading, slowed release and improved in vivo efficacy, illustrating the strategy of tailoring lipid structure to retention goals. The new work is notable because it achieves the same ends not by designing exotic molecules but by exploiting an unexpected self-assembly behavior of ordinary unsaturated phospholipids—the workhorses already used across the liposome industry. If a standard, clinically familiar lipid class can be coaxed into a depot architecture through formulation rather than synthesis, the path to translation could be considerably shorter.

Caution is warranted, as with any preclinical result. The weeks-long nerve block was demonstrated in rats, and the dose, tolerability, reversibility and tissue response of a multiweek anesthetic block in humans remain open questions. A block that lasts weeks is only desirable if it can be timed and, when necessary, reversed; prolonged numbness or weakness could itself cause harm. The authors’ tetrodotoxin payload is a neurotoxin whose safety margin in clinical use would demand exceptionally precise control of release rates—precisely the control the multivesicular architecture appears to provide, but at a stringency that regulatory science has rarely been asked to certify. Questions about batch-to-batch consistency of such hierarchical structures, and about how they behave in tissues other than peripheral nerve, will shape the next stages of development.

Even so, the conceptual contribution is likely to endure. The study reframes a class of lipids long considered unsuitable for depot formation into a platform for ultra-slow release, and it suggests that the phase behavior of unsaturated membranes is richer than the field assumed. Multilamellar–multivesicular liposomes add a distinct mechanism to the slow-release toolbox described in recent reviews: retention by nested, compartmentalized barriers rather than by membrane rigidity or polymer matrices. For the growing list of hydrophilic drugs— biologics, neurotransmitter modulators, ion-channel blockers—that would benefit from weeks of controlled exposure, the finding opens a formulation route built from molecules the pharmaceutical industry already knows how to make, purify and approve. The distance between a rat’s sciatic nerve and a patient’s postoperative recovery remains considerable, but the first demonstration that a simple lipid recipe can hold water-soluble medicine for weeks is the kind of result from which translational stories often begin.

Subject of Research: Multilamellar–multivesicular liposomes formed by unsaturated phospholipids for ultra-slow release of hydrophilic drugs

Article Title: Unsaturated phospholipids form multicompartment liposomes that extend release of hydrophilic drugs

Article References: Unsaturated phospholipids form multicompartment liposomes that extend release of hydrophilic drugs. (2026). Nature Biomedical Engineering. https://doi.org/10.1038/s41551-026-01791-8

Image Credits: AI Generated

DOI: 10.1038/s41551-026-01791-8

Keywords: liposomes, drug delivery, phospholipids, hydrophilic drugs, sustained release, multilamellar vesicles, tetrodotoxin, nerve block, long-acting formulations, biomedical engineering, self-assembly, anesthesia