Red Blood Cell Hitchhiking: A Novel Approach for Vascular Delivery of Nanocarriers

Abstract

Red blood cell hitchhiking (RH) is a method of drug delivery that can increase drug concentration in target organs by orders of magnitude. In RH, drug-loaded nanoparticles (NPs) are adsorbed onto red blood cells and then injected intravascularly, which causes the NPs to transfer to cells of the capillaries in the downstream organ. RH has been demonstrated in multiple species and multiple organs. For example, RH NPs localized at unprecedented levels in the brain when using intra-arterial catheters like those in place immediately after mechanical thrombectomy for acute ischemic stroke. RH has been successfully employed in numerous preclinical models of disease, ranging from pulmonary embolism to cancer metastasis. In addition to summarizing the above versatility of RH, here we also describe studies into the surprisingly complex mechanisms of RH, as well as outline future studies to further improve RH’s clinical utility.

1. Introduction: advantages and challenges of drug delivery systems.

The modern pharmacopoeia now includes thousands of pharmacological agents. The spectrum of their complexity ranges from relatively simple “small molecule drugs” (e.g., aspirin) to biological macromolecules (e.g., nucleic acids and proteins) to, most recently, cells (e.g., transfusion of blood and modified lymphocytes in CAR-T therapy). Chemical drugs (small molecule drugs) are generally more stable, homogeneous, amenable to industrial development and permit administration via a variety of routes, including oral. Biologicals (macromolecules) offer more precise and generally more powerful effects. However, these complex, heterogeneous and labile agents are difficult to produce and use (1).

Both chemical and biological drugs require delivery from the administration site to the sit of action in the body. Generally, the oral route does not work for biologicals. These large and labile agents require injections. Furthermore, to exert activities, biologicals must get precisely to the target - the nucleus, or cytosol, or other cellular compartment, in the desired cells in the organ or tissue of interest(2). However, most biologicals, just like chemical agents, generally do not have such a natural homing. In order to improve drug delivery, diverse carriers have been devised by experts in chemistry, imaging and drug design, bioengineering, and material and pharmaceutical sciences, in collaboration with biomedical researchers(3).

These efforts evolved into a burgeoning multidisciplinary research enterprise. Countless academic and industrial labs around the world are busy devising various drug delivery systems including nanoparticles serving as nanocarriers for pharmacological agents (DDS, NP and NC, respectively)(4). Several of these carriers have been tremendously successful, as evidenced by the clinical approval and now common use of nanoparticles for gene therapies and imaging. However, DDSs still present major challenges associated with clearance from circulation, inactivation before reaching their targets, imprecision of delivery to the intended site, and too frequently, poor efficacy (5–7). Significant improvement of the precision and effectiveness of drug delivery remains one of the major challenges of pharmacotherapy, for both chemical and biological drugs.

Here we will briefly introduce two distinct types of carriers for DDSs, namely nanocarriers and red blood cells (RBCs), and will discuss an original approach combining NCs and RBCs into a novel drug delivery platform that we call “RBC hitchhiking” (RH). This technique, based on transient coupling of NCs to RBCs, dramatically changes the behavior of NCs in the body, providing a novel drug delivery paradigm, and a DDS platform with great potential for the treatment of a number of diseases.

2. Synthetic nanocarriers

During the last 50 years, the overlapping fields of drug delivery, targeting, and nanomedicine yielded battalions of DDSs aimed at improving the treatment of tumors, infectious diseases, stroke, inflammation, hematological, metabolic, cardiovascular, pulmonary, neurologic and other disease conditions. Benefits of DDSs include: A) enabling administration of poorly soluble and toxic drugs; B) optimizing pharmacokinetics (PK) via prolonging life-time in blood and limiting clearance and deposition in off-target tissues; c) minimizing undesirable interactions of a drug with the body via encapsulation into an “inert” carrier; and, D) optimization of spatiotemporal specificity, by directing drugs to the desired cells and subcellular compartments(8–10).

The rapidly growing roster of DDSs includes liposomes, liposome-like biodegradable polymersomes based on synthetic copolymers, a variety of dendrimers, nanogels, multilayered solid nanoparticles, carriers based on branched and linear molecules, multimolecular assemblies of natural and synthetic proteins, nucleic acids, carbohydrates and lipids, and their combinations. Liposomes are arguably the oldest, best characterized, and most widely clinically used nanocarriers (NC)(11). Hydrophobic and amphiphilic agents that can be loaded into liposome bilayer membrane and inner volume, respectively. These phospholipid-cholesterol-based vesicles provide a versatile DDS platform onto which innumerable variations have been based(12).

Coupling of antibodies, antibody fragments, and other affinity ligands provides targeting and enables intracellular delivery via engaging with cellular surface molecules involved in endocytosis(13). Coating the carrier surface with PEG inhibits opsonization by complement and uptake by phagocytes and other undesirable cells. Insertion of molecular triggers sensing the target microenvironment (e.g., via changes in pH, enzymatic activity, concentration of oxygen, or reducing agents), enables controlled carrier transformation leading to local drug release, fusion with or crossing the membranes, endosomal escape, activation of the drug cargo, and other responses elevating spatiotemporal control of DDS(14–18).

Many elements of DDS design are modular. This molecular Lego yields a rich and diversified palette of formulations with distinct materials and structures, geometries (size, shape), biomechanical characteristics (flexibility, stability), affinity/avidity to desired targets, responsiveness to the local microenvironment, interactions with blood components, binding to cells, uptake and intracellular trafficking and, ultimately, beneficial and unintended effects(19–22). Among these features, the surface characteristics of NCs control the behavior of a DDS in the body. They include hydrophobic vs hydrophilic or amphiphilic nature, charge, repulsive vs adhesive character, homogeneity and density of the molecular components that define those characteristics. Especially relevant in the context of this paper, surface features of NCs regulate its partition to cellular blood elements, either direct or mediated by plasma components(23, 24).

The very diversity of the roster of currently pursued DDS of different types hints that unlikely a specific DDS design or even a platform will become the universal champion. Each has a unique balance of advantages and disadvantages in the context of treatment of given disease. The advent of “personalized medicine” might even lead to further demand for a diversification of specific DDS designs, as they are engineered for individual patients, phases of the circadian cycle, age, and comorbidities (25). Notwithstanding this growing interest in the individualization of drug delivery, there are common challenging issues impeding utility of most DDSs. For example, fast blood clearance impedes bioavailability and delivery of most types of DDS. Even for DDSs featuring specific and high avidity to the target molecules, the percent accumulating in the target usually represents a relatively minor fraction of injected dose. General approaches that could help solve these issues are likely to make a significant impact (26, 27).

DDS biodistribution is ultimately determined by both DDS properties (e.g., size) and the physiology/pathophysiology of the body (28). For example, unlike the pulmonary vasculature that receives more than half of total heart blood ejection, all other organs and tissues share the leftovers. The first pass phenomenon is especially important for DDS uptake in organs that represent relatively modest fractions of the total blood volume and vascular surface. These organs will have no chance to collect a DDS rapidly cleared from the bloodstream, unless the DDS is injected directly upstream in the conduit artery. However, the high flow rate of blood passing through an organ does not favor effective accumulation (29, 30).

These considerations provided a major impetus to devise a strategy to combine the advantages of a NC with the unique features of a biological entity, the red blood cell (RBC). As we will explain below, employings RBCs as “super-carriers” of NCs can improve NC and drug delivery by orders of magnitude. But before explaining such complex NC-RBC systems, it is prudent to first provide an overview of the RBC as a carrier, and how it has been used with therapeutics outside the world of NCs.

3. Red blood cell (RBC): a drug delivery carrier.

RBC are easily available, enormously numerous and modestly expendable (to some degree). In comparison with other cells, RBCs are relatively simple and homogeneous (diameter and thickness of RBC biconcave discs are ~6–7 μm and ~2 μm). On average, each of 10₁₃ RBC (~4–5 million per microliter of blood) travels ~250 km in human body in the course of a life span of ~120 days. An RBC experiences millions of cycles of extremely fast flow in the arterial vasculature followed by squeezing through capillaries. Their unprecedented endurance, tensile strength and deformability is provided by the plasticity and sturdiness of the RBC membrane and cytoskeleton, lack of nuclei and organelles and molecular features of RBC surface. RBCs, which naturally evolved to deliver cargo (oxygen and carbon dioxide) in the bloodstream, seem well suited to be an ideal carrier for vascular drug delivery (31).

It was recognized fifty years ago that due to these features RBCs represents a nearly ideal carrier for drug delivery within the vascular system(32). Several groups embarked on the studies of RBC drug delivery including enzymes, genetic materials and other drugs(33–36). Alas, the emergence of HIV, hepatitis and other blood transmitting infections all but decimated this area of research in the eighties and nineties. Despite these setbacks, due to the steadfast persistence of a handful of labs and the general advent of cellular therapies, RBC drug delivery evolved into a blooming biomedical research enterprise (37–39). Several RBC-based DDSs are in clinical trials and a few are entering clinical practice (40–43).

Most drugs and NCs transfer from blood to the tissues via diffusion, and in some cases active transport, across and between the endothelial cells that form monolayer lining the lumen of blood vessels. In most cases, such extravasation is irreversible (44). In contrast, RBCs normally do not extravasate, except transient passage via the reticuloendothelial system (RES) comprised of specific compartments in the spleen, liver, and bone marrow, open for micron-sized objects circulating in blood. The RES phagocytic and immune cells survey passing objects --- white blood cells, RBCs, NCs --- and eliminate particles that appears abnormal, including senescent and damaged RBC, microorganisms, and, NCs (apparently, the latter objects are removed the most effectively). In pathology, RBCs extravasate in sites of tissue damage (hemorrhage) and to a lesser extent, inflammation. RBCs normally do not adhere to endothelium. Hydrodynamic forces drive RBCs to the center of the vascular lumen in large vessels, while in the microcirculation the repulsion of the negatively charged glycocalyx of the endothelium and RBCs minimizes their adhesive interaction. In pathology, RBCs adhere to the vascular lumen, due to inflammatory changes in the endothelium and the abnormal stiffness of some pathological RBCs, which has been reported in malaria, sickle cell disease (SCD) and, as emerging data imply, COVID.

There are two approaches to load drugs in RBC carriers (Figure 1). First is encapsulation into the inner volume of the RBC (45, 46). Such encapsulation is usually achieved by putting the RBCs into hypotonic solution to induce swelling and formation of openings (“pores”), with the solution containin the cargo drug of interest. The drug diffuses into the RBCs, and the RBCs are then put into normal osmotic conditions, which causes the pores to seal up (47, 48). Use of membrane-permeating agents and genetic modification of RBC precursors, reticulocytes, provide newer alternatives (49).

Figure 1: RBC-mediated drug delivery techniques that preceded RBC-hitchhiking:

A, Loading small molecule drugs and therapeutic proteins intracellularly in RBCs. RBCs in buffer solution are transferred into a hypotonic solution, which causes severe enough cellular swelling to create openings referred to as pores. Drugs in the external solution diffuse into the RBC cytosol. The RBCs are then shifted to isotonic solution, which causes the RBC to shrink back down to its original size, closing the pores. The drugs are thereby trapped inside the cytosol. B, Three different methods of coating RBCs with therapeutic proteins. Left panel, covalent attachment of a therapeutic protein (purple) to a random, intrinsic membrane protein. Covalent attachment can be mediated via a variety of conjugation chemistries, but the most common is to use an NHS-ester, which leaves an amide bond bridging the therapeutic protein and the RBC. Middle panel, Therapeutic proteins can be conjugated to monoclonal antibodies (blue) that bind a surface epitope on an RBC. Right panel, Therapeutic proteins can be genetically encoded and thereby fused genetically to a single chain of the variable fragment of a monoclonal antibody (scFv; blue), followed by expression as a single “fusion-protein.”

Approaches based on encapsulation of drugs into the RBC carrier represent the lion’s share of the RBC drug delivery research enterprise. In fact, all RBC-based DDSs that already entered the clinical domain involve loading of drugs into isolated RBC following the infusion into the patient. A variety of drugs have been loaded into RBC ghosts via reversible osmotic swelling. They include enzymes metabolizing diffusible pathological molecules in blood and other drugs that may benefit from prolonging their circulation time in the bloodstream. Loading into RBC ghosts drugs that slowly diffuse to plasma via concentration gradient creates the blood pool of agents that otherwise would quickly extravasate and clear (50, 51). In these applications RBC carriers help to improve the PK of the pharmacological cargo via prolonging its circulation, partitioning from plasma to the cellular fraction, and separating the cargo from the body (52, 53). The roster of cargoes loaded into carrier RBCs include biologicals (e.g., enzymes and antigens), but also chemical molecules (e.g., dexamethasone), imaging probes, contrast agents, and other compounds (54, 55).

As an alternative to encapsulation into the inner volume of RBCs, cargoes can be attached to the RBC surface. Several approaches can enable this attachment to the RBC surface, including: A) Direct adsorption onto the RBC membrane; B) Chemical conjugation, such as by an NHS-ester (56–58); C) Genetic modification of reticulocytes (59, 60); and, D) Targeting of drugs conjugated or fused with affinity ligands specifically binding to RBC determinants (61). Several labs established surface loading of cargoes onto RBCs by these methods (62, 63). Among other approaches, targeting drugs to RBCs using affinity ligands does not require RBC isolation, ex vivo loading and reinfusion of RBCs, or infusion of donor RBCs (64). Animal studies showed that drugs conjugated with antibodies that safely bind to RBC surface determinants immediately after injection bind to circulating naive RBCs, effectively becoming RBC-drug complexes. Replacing antibodies by their single chain fragments (scFv, lacking the Fc-fragment that may cause unintended effects) seems especially promising.

These methods for RBC surface loading have been successfully employed to attach to RBCs anti-thrombotic and anti-inflammatory recombinant proteins, antigens for modulation of the immune response, and glycoproteins protecting RBCs from complement (65–69). Therapeutic targets of these agents are accessible for RBC-drug complexes circulating in the bloodstream. Recombinant proteins represent a popular type of cargo for loading onto the RBC surface (70). Studies in animal models documented that coupling to RBCs retains these agents in the cellular fraction of blood and prolongs their circulation by orders of magnitude (68, 71, 72). Coupling to carrier RBC alters the drug pharmacokinetics and biodistribution (PK/BD), the repertoire of cellular and molecular targets encountered by the drug, and the spatiotemporal parameters of drug activity and effect(73, 74). In particular, this strategy enables vascular surveillance of biological pro-drugs that can be activated in the pathological sites (75, 76).

4. Nanoparticles hitchhiking on RBC (RH): transformative changes in PK/BD, vascular transfer, and potential medical utility.

Diverse types of NCs were first loaded into RBC ghosts, which are RBCs that have been emptied of their water soluble internal contents by exposing the RBCs to hypotonic solution for a prolonged time (77). However, the size of the transient pores induced in the RBC membrane using standard osmotic swelling does not exceed 50nm (78). Smuggling larger particles requires more invasive and damaging RBC approaches. In contrast, binding NCs to the RBC surface is less traumatic and more natural and permissive(79). Initial attempts of binding of immunoliposomes to RBCs were reported in the eighties(80). This line of investigation, however, was not actively followed, and, in general, research on the interaction of NCs with RBC remained dormant for several decades.

More recently, several groups have embarked on loading nanoparticles to RBC carriers, predominantly imaging probes, contrast agents and resonators for magnetic, isotope and ultrasound-based drug delivery and imaging approaches(38, 47). These new attempts, in part driven by recent successes in using cells for drug delivery and cellular therapies, revived the allbut-forgotten initial attempts to combine NCs with RBC to yield a hybrid DDS (in retrospect, this was somewhat naïve). In particular, the initial impetus for the collaborative studies of our labs came from testing this hypothesis that coupling to RBC may prolong circulation of nanocarriers beyond stealth technologies, including the somewhat problematic coating by polyethylene glycol (PEG) (81–83). These studies described below fortuitously evolved into the multifaceted concept of RBC hitchhiking by nanocarriers.

4.1. Effect of RBC hitchhiking on nanoparticle circulation and elimination by RES.

The initial studies testing adsorption on isolated rat RBC of polystyrene particles in the range of sizes 100–1,000 nm followed by intravenous injection in rats showed that RBC carriage prolongs nanoparticles circulation. In particular, while free nanoparticles within the size range of ~200–450 nm in diameter were cleared in minutes, RBC-bound nanoparticles remained in circulation for hours (~5% nanoparticles observed in blood even after 12 hours)(84). The reasons for this size optimum remain to be clarified. Perhaps, smaller particles did not exhibit sufficiently strong adhesion on RBC surface, while larger particles extended far from the RBC favoring dislodging(84).

Nanoparticle surface chemistry and charge modulated circulation of nanoparticles attached to RBCs(84). In vitro studies performed using viscometer showed that the in vitro rate of shear-induced detachment correlates with the in vivo circulation in rats(85). Scanning electron micrographs of RBCs have given clear indications of the dimples formed by the nanoparticles upon membrane adhesion indicating that the nanoparticle makes a firm contact with the RBC(86). Most likely, surface features and size of nanoparticles determine the adhesion strength on RBC.

RBC-hitchhiking nanoparticles, though remaining in circulation for prolonged times, were eventually cleared, thus leading to a question about the mechanisms behind prolongation of circulation as well as eventual clearance. Studies with dual labeling of RBCs and nanoparticles confirmed that while nanoparticles are removed from the blood, whereas the carrier RBCs to which nanoparticles were attached, remained in circulation(84). Studies(86) encompassing coating nanoparticles by inhibitors of phagocytosis, direct injection of these inhibitors and splenectomy suggested that the RBC-hitchhiking protects nanoparticles from uptake by RES macrophages in liver(85). Mechanical forces eventually remove nanoparticles from the RBC surface, which leads to their clearance by the liver and spleen (there is some component in spleen uptake that may be mediated by RBC delivery to this organ).

These studies demonstrated that adhesion of nanoparticles to RBCs is the foundation of hitchhiking and yielded hypotheses towards mechanisms of adhesion, which appears to be driven by a surface-wetting like process where a combination of hydrophobic and electrostatic interactions determine the extent of membrane spreading. A simple kinetic model of detachment of nanoparticles from RBC surface was consistent with the measured pharmacokinetics of RBC-hitchhiking nanoparticles(84). Nanoparticles that adhere well generally tend to perform with a higher amplitude in terms of extended circulation and changes in biodistributions.

4.2. Transfer of RBC hitchhiking nanoparticles to the vascular cells.

In addition to extension of half-life in circulation and reduction of RES uptake, RBC hitchhiking alters behavior of nanoparticles in a serendipitous and truly remarkable way (Figure 2). RBC-hitchhiking nanoparticles exhibited unusually high uptake in the lungs, maximal at 1 hour post injection and sustained at a high level for about 10 hours, after which, it decreased over a period of time up to 24 hours(86). This implies that hemodynamic factors and architecture of lung capillaries favor dislodgement of nanoparticles from RBCs. It seems logical in retrospect, given that the lungs represent a dense capillary network in the first major microvascular bed encountered by the RBC after IV injection. This is an example of the classical first pass phenomenon, described for several DDS targeted to endothelial cells(87).

Figure 2: RBC-hitchhiking efficiently transfers nanoparticles to cells in the target organ.

A, The original proposed mechanism of RBC-hitchhiking (RH). First, nanoscale drug carriers (nanocarriers; NCs) are mixed with washed RBCs, which causes the NCs to bind to the surface of the RBCs. The RBC-NC complexes are then injected into a blood vessel. The RBC-NC complex does not bump into the walls of large blood vessels, but when the RBC enters the capillaries, whose lumens are smaller than the diameter of RBCs, the NC is pressed against the capillary wall, leading to transfer of the NC to the endothelial cells of the capillary. B, Electronmicrograph of polystyrene-nanoparticles (PS-NPs) and lysozyme-dextran nanogels (nanogels) adsorbed onto RBCs. C, Experiments showed that in the lung, the best studied organ of uptake in RH, RH transfers nanoparticles (NPs) not just to endothelial cells, but also to marginated leukocytes. Marginated leukocytes are white blood cells that reside in the microvasculature of the lungs for prolonged periods. RH-mediated uptake of NPs into marginated leukocytes was shown in vivo and in vitro. D, Percent injected dose of a cargo drug that was found 30 minutes after injection in mice. A typical hydrophilic small molecule drug (DTPA) was only 0.08% of the injected dose, while liposomes targeted via surface conjugation of antibodies (anti-PECAM or -ICAM) and RH showed >300-fold higher levels of drug delivery.

Testing of a series of six the most widely used clinically-translatable NPs adsorbed onto RBCs, including 4 in frequent clinical use: liposomes, albumin-NPs, lipid nanoparticles, and AAV(88). All 6 NPs adsorbed onto RBCs. By comparison, free IgG molecules did not adsorb onto RBCs. This paper next optimized RH for lung delivery, by testing various NPs and by modifying NP surface properties. While RH with PS-NP only led to 5% of the injected dose of NPs localizing to the lungs, liposomes accumulated at >30% of the injected dose in the lungs. Further, RH improved lung-to-liver ratio for liposomes by almost 2 orders of magnitude. Thus, RBC-hitchhiking confers very high lung uptake, especially in some of the most clinically translatable NPs.

The very high lung uptake of some RBC-hitchhiking NPs is slightly higher even than NPs targeted via the much more common targeting method for NPs: coating NPs with affinity moieties such as antibodies. This begged the question of whether RBC-hitchhiking could be combined with antibody-targeting. This was addressed by coating NPs with antibodies targeting the endothelial epitope PECAM (88). While anti-PECAM-NPs accumulated in the lungs at 21% of the injected dose, anti-PECAM-NPs that were RBC-hitchhiking accumulated at 56%. The lung-to-liver ratio of anti-PECAM-NPs was even more impressively increased by 7.5-fold with the addition of RBC-hitchhiking. Thus, RBC-hitchhiking can be combined with traditional antibody-targeting to synergistically lead to unprecedented lung uptake.

From the standpoint of the mechanism, prototype PS-NPs had been assumed to transfer from RBC to the endothelial cells, but there was scant direct evidence. Furthermore, the PS-NP agglutination results and isotope tracing suggested that PS-NP’s might be deposited in the lungs as pulmonary emboli clogging capillary lumens and also retaining carrier RBC in the lung vessels (see below in section 5.1). Therefore, the question of cellular attribution has been revisited using safe, clinically employed liposomes and translatable nanogels free of these unintended effects exerted by PSNP(88). Collectively, the results of double-isotope tracing of NC and carrier RBC, flow cytometry and confocal microscopy asserted that: A) Carrier RBC leave these nanocarriers in the blood vessels and continue to circulate after unloading as normal RBC; B) Nanocarriers undamagingly delivered by RBC to the pulmonary microvasculature transfer to endothelial cells; and, C) Neither carrier RBC nor cargo nanocarriers clamp or stuck in the lung capillary lumens.

This promising result was coupled to a most unexpected result: a moderate fraction of the NPs were also in marginated leukocytes(88). Marginated leukocytes are white blood cells that reside (for minutes to hours) in the capillary lumen, with their greatest abundance by far in the lungs(89). Live cell imaging in vitro that stationary phagocytes would grab RBCs flowed past them, and remove the adsorbed NPs(88). Thus, RH with translatable NPs is produced by transfer from RBCs to endothelial cells and marginated leukocytes.

4.3. RBC-hitchhiking to the pulmonary vasculature: potential medical applications.

The pulmonary vasculature is naturally “designed” as one of the prominent targets for RH (perhaps, the most prominent). The lungs are the only organ that receives more than half of total cardiac blood output: all venous blood from the right chambers flows to the lungs via the pulmonary arteries, plus a fraction of arterial blood goes into the intercostal and bronchial arteries. All other organs including the brain, splanchnic organs and the heart itself share less than 50% of blood output going to the aorta. In addition, the flow rate in the lung vessels is slower than in the systemic vasculature, favoring interaction between vessel walls and blood elements squeezing through the alveolar capillaries.

RH provides clear advantages for applications that require lung targeting, including ARDS, pulmonary hypertension and pulmonary embolism, respiratory infections and many others. A number of studies in several animal models have confirmed the ability of RBC hitchhiking to deliver significant fractions of injected doses in lungs(86, 88, 90, 91). This opens opportunities for targeting various lung conditions arising from cancer or infections, among others. Indeed, RH of nanocarriers loaded with fibrinolytic plasminogen activators afford more effective dissolution of pulmonary thrombi in mouse model of fibrin clots lodged in the lung vasculature(88).

Studies have demonstrated efficacy of RBC-hitchhiking nanoparticles in treating lung metastasis in mice models(90, 91). Taking advantage of natural lung targeting, chemotherapy-loaded nanoparticles were delivered to the lungs to treat melanoma metastasis in the lungs. Compared to free nanoparticles, RBC-hitchhiking nanoparticles induced more than 15-fold improvement in lung deposition, thus leading to a high local concentration of encapsulated chemotherapeutics in the lungs. This provides a clear therapeutic advantage for the treatment of primary lung cancer or lung metastasis. Effective delivery of chemotherapeutics is often limited by off target effect which limit the dose that can be delivered to the patients. By creating high local concentrations at the disease organs, the therapeutic window of the drug can be substantially improved. Current approaches for drug targeting are based primarily on molecular targeting. Compared to the molecular approaches, RBC hitchhiking provides several advantages for lung targeting including high specificity and applicability to a broad variety of therapeutic cargos.

4.4. Universality of RBC-hitchhiking: animal species and target organs.

After initial tests of RBC-hitchhiking in mice, a major question remained about generalizability to other species. This was in question given that RH’s purported mechanism involved the size mismatch between RBCs and capillaries, and mice differ by humans in size by orders of magnitude(88). Fortunately, NP adsorption onto RBCs occurred in mice, rats, pigs, dogs (yet unpublished), and humans. Additionally, RH worked in vivo in pigs, and in fresh, ex vivo human lungs that were oxygenated and perfused(88). Thus, RH is likely universal among mammals.

Further, this drug delivery platform offers unique advantage generalizable translation of RH: targeting organs other than the lungs, which could be accomplished using a macro-scale targeting method: intra-arterial (IA) catheters. The idea is that RH deposits in the endothelial cells of the capillaries in whichever organ is immediately downstream of the intravascular injection site (Figure 3). After intravenous (IV) injection, the first capillary bed encountered is the lungs, and hence IV-RH targets the lungs.

Figure 3: RBC-hitchhiking can target any organ by selecting the catheter injection site.

In the center panel, we illustrate the two in-series circuits in the circulatory system: the pulmonary and systemic circuits. In the pulmonary circulation, venous blood (blue) returns to the right side of the heart (center), which pumps it via the pulmonary arteries to the capillary bed of the lungs. In the systemic circulation, blood returned from the lungs (red) is pumped via a large number of in-parallel arteries to capillary beds of organs, such as the brain. If RBC-hitchhiking (RH) nanoparticles (NPs) are injected via an intravenous (IV) catheter into a peripheral vein (as shown in the bottom left of the circuit diagram), the first capillary bed encountered is that of the lungs, leading to massive targeting of the NPs to the lungs (left panel). If RH-NPs are injected instead via an intra-arterial (IA) catheter (top right of the circuit diagram), the first capillary bed is that of whatever organ is immediately downstream of the catheter, here the brain. As in the right panel, this leads to unprecedented uptake in brain.

Using IA catheter upstream of, for example, the brain, IA-RH should target NPs to the brain. This indeed turned out to be the case for all 3 other organs we tried, including the brain. Indeed, IA-RH achieved higher brain uptake (12% of the injected dose) than the prior best technology (transferrin-receptor-targeted NPs, with 0.5 – 1% injected dose) by more than an order of magnitude. Importantly, IA delivery does not require a new clinical infrastructure or procedure. For both severe stroke and heart attack, the standard of care is to insert an IA catheter to remove the offending clot. While clot removal improves outcomes, the ensuing ischemia-reperfusion injury (IRI) produces much of the damage leading to still poor outcomes in strokes and some heart attacks. Therefore, we propose employing IA-RH immediately after IA-based clot removal using the IA catheters that are still in place, to deliver unprecedented levels of IRI drugs. These two indications, severe stroke and heart attack, are two of the biggest causes of morbidity and mortality in the US, and therefore should serve as excellent initial test cases for IA-RH.

5. Understanding and exploring RBC-hitchhiking: natural prototype(s), new applications, and advance towards specific molecular targeting.

A combination of carrier features of RBCs and NPs provided by hitchhiking affords a remarkable boost in drug delivery to the selected vascular areas in lungs and other organs of diverse animal species. Yet, this is a fairly new area of research and many aspects of the RH phenomenon remain unknown. Emerging results furthering our understanding of its mechanisms and its potential utility undoubtedly will expand its impact. Below we provide few specific examples of new tentative avenues for advancing RH.

5.1. “Immune adherence and trogacytosis”: natural transport of RBC-binding cargoes to host defense cells.

RBC-hitchhiking is an engineered technology, but it is likely exploiting natural mechanisms. In particular, RH bears striking similarities to the phenomenon of “immune adherence (IA)” (Figure 4). In IA, immune complexes formed in the bloodstream and various abnormal particles including microbes coated by immunoglobulins and complement (i.e., opsonized or appetized for phagocytic cells) adhere to specific complement receptors on RBCs, and these RBCs then shuttle the microbe-immune complexes to phagocytes without damage to RBC (reviewed in (92)). IA was first described in 1953(93), and has since been shown to occur with bacteria, parasites(94), and viruses(95, 96).

Figure 4: The immunological defense known as immune adherence bears striking similarities to RBC-hitchhiking.

In the left panel we illustrate the experimental protocol demonstrating immune adherence (IA), as well as it’s known mechanisms. Microbes are mixed with serum, which leads to the microbes being covalently opsonized C3b. The microbes are then mixed with RBCs, with the microbe-C3b complexes binding to clusters of CR1 on the RBCs. When the RBC-C3b-microbe complexes are injected intravenously, they accumulate primarily in the liver and spleen, with the best described cellular uptake being in the Kupffer cells. In the right panel, we diagram engineered RBC-hitchhiking (RH). RBCs are extensively washed in buffer such as saline or PBS, and then mixed with any of a diverse set of nanoparticles (NPs). In such serum-free conditions, the NPs adsorb onto the RBCs via unknown mechanisms. Upon IV injection, the NPs best target the lungs, where they are transferred to endothelial cells and marginated leukocytes.

In the classic initial IA experiments, bacteria mixed with serum bound to RBCs, but not if the serum had first undergone complement inactivation(92). Subsequent experiments elucidated the major steps and molecular players in IA, outlined briefly here. The first critical step in most forms of IA appears to be covalent opsonization of microbes by complement protein C3b(97, 98). C3b can be deposited on microbes through any of the 3 major pathways of complement activation: the classical pathway involving antibodies; the lectin pathway mediated by lectin proteins that bind molecular patterns such as polysaccharides; or the alternative pathway, which binds to any particle with surface nucleophiles, but at a low baseline rate. Once C3b is covalently attached to the microbe, clusters of C3b bind to complement receptor 1 (CR1) on RBCs(97, 98). Subsequently, the C3b-opsonized microbes on the RBCs are transferred to stationary phagocytes in the liver and spleen(92). This phenomenon provided a basis for approaches for clearing pathological agents from bloodstream using capturing antibodies conjugated with antibodies to CR1 devised by Ron Taylor (99, 100).

While the mechanistic details of IA have striking similarities to engineered RH, they also bear key differences in the two key steps of RH: the adsorption onto RBCs and the transfer to stationary cells and organs. Here we explain those differences, but also how those apparent differences might still be undergirded by the same underlying mechanisms.

First, while IA’s RBC-adsorption step is classically described as complement-dependent, engineered RH does not appear to be so. Engineered RH employs “washed RBCs”, meaning RBCs that have undergone multiple washes to removal all traces of serum, which also removes complement (note, “washing” is used clinically for transfusing RBCs into patients who have had allergic reactions to the serum in prior transfusions). Indeed, we have showed that the ex vivo adsorption of many, but not all, types of nanoparticles onto RBCs ex vivo is decreased dramatically in the presence of serum, which means such RH is decreased or eliminated in the presence of complement(88).

However, there are a few possible mechanisms of RBC-adsorption that could be shared by RH and IA. For example, IA has been shown to also work without C3, though less efficiently, via other opsonins such as mannose binding lectin (MBL)(98). Notably, MBL could remain bound to RBCs even after the washing steps of RH. Additionally, HIV virions engage in IA without opsonins, binding directly to Duffy antigen. Thus, the diversity of nanoparticle and microbe surface properties might explain how IA and RH have partial overlap in mechanisms of RBC-adsorption. To elucidate this, it is critical to test whether engineered RH works in settings of opsonin- and RBC-perturbations, such as genetic knockout or pharmacological inhibition of C3, MBL, and CR1.

The second major step of RH that differs from IA is the transfer of cargo to stationary cells and organs. IV-injected IA has been repeatedly shown to localize microbes to the liver and spleen, while IV-injected RH most efficiently deposits in the lungs. There are several possible explanations. First, it is possible that IA and RH simply have different mechanisms, with those of RH as yet barely uncovered. Second, it is possible they share largely the same mechanisms, but experimental differences have led to differences in apparent organ distribution of the cargo. In particular, the species of animals in the mechanistic studies were different, with mostly rats used in IA, but mice, pigs, and humans in RH. Also, IA experiments generally did not report on lung deposition, which could have been high but missed due to not investigating it after the first 1953 experiment did not initially find it. Finally, while RH massively increases the lung-to-liver ratio of nanoparticle deposition, there is still some liver deposition, and the liver deposition is especially high for the one microbe we tested in RH (adeno-associated virus).

Thus, it appears that IA and engineered RH likely share numerous overlapping mechanisms, but further experiments must prove which aspects are the same and which differ.

A much needed study would be a head-to-head comparison of IA and RH. Such a study should include: perturbations of C3, MBL, and CR1; investigation of destination cell types in lung, liver and spleen; and testing in multiple species. These mechanistic studies could then guide the rational design of improved RH. As discussed below, this may enable novel therapeutic interventions in host defense and immune system, while avoiding unintended alterations in these functions of the body.

5.2. Immunological aspects of RH.

Advancing RH to new immunological and host defense therapies is a breathtaking perspective. In theory, nanocarriers emulating the natural pathway(s) for elimination of invaders and pathological agents may supersede the potency and repertoire blood clearance afforded by via immune adherence and trogacytosis. Furthermore, enhanced uptake of modified RBC by defense cells, if it is achieved in a controlled and well understood mechanism, provides approaches to deliver cargoes to these cells(51, 101, 102).

For example, RBC filled with steroids and inhibitors of phagocytes provide alleviation of inflammatory processes, while RBC loaded with antigens modulate the immune response(103, 104). Amazingly, RBC-mediated delivery of antigens to diverse cells involved in the immunological response, may either stimulate or suppress this outcome(105–108). The factors controlling immunogenic vs tolerogenic effects of antigens delivered by RBC carrier are still emerging and their mechanisms are not sufficiently understood(109, 110). These factors include nature and extent of loading, character and extent of “non-specific” (i.e., common for many antigens) changes of RBC biocompatibility described above, molecular patterning of the antigen on the RBC surface, dose and route of administration of RBC-antigen and many other parameters of RBC-antigen complexes design(60, 111).

Most prior work on RBC hitchhiking has been performed with the goal of attaching nanoparticles to RBCs without inducing any modification to RBCs themselves. This is often controlled through controlling the density of attached nanoparticles on RBC surface. This is seen most clearly in case of polystyrene nanoparticles. Specifically, attachment of polystyrene nanoparticles induces concentration-dependent membrane alterations in RBCs, in particular, expression of phosphatidyl serine (PS) on the RBC surface. No significant PS expression was seen at low nanoparticle:RBC ratios, for example, 100:1. However, at nanoparticle:RBC ratio of 300:1 significant expression of PS could be seen. At the same time, the tissue deposition of nanoparticles was also significantly affected by the nanoparticle loading ratio. At lower ratios, as expected, the nanoparticles were deposited primarily in the lungs. However, at a high particle loading of 300:1, the biodistribution was shifted to spleen(112).

This was hypothesized to be driven by nanoparticle-induced changes in RBC membrane properties which potentially reduced the stretching of RBCs in lung capillaries, thus reducing particle deposition in the lungs, which in turn opens possibilities of depositing particles in other organs. This, in combination with nanoparticle-induced membrane alterations, led to deposition of nanoparticles in the spleen. Nanoparticles were captured in the spleen by the antigen presenting cells and led to a strong immune response against the antigen on the nanoparticle surface. Using ovalbumin-coated nanoparticles, this strategy was used to generate an immune response against ovalbumin and subsequently, an ovalbumin expressing lymphoma model.

Modulating immune response and host defense by RH may expand opportunities in this space provided by drug delivery systems such as design of vaccine delivery by nanocarriers providing an adjuvant effect, as well as DNA and RNA based vaccines(113–115). This is a formidable frontier of immense importance for human civilization (116, 117). The intricacies of instructive and innate branches of the immune systems rival the complexities of the brain. It is impossible to predict which types of immune-regulating nanocarriers may benefit from RH, and in which particular conditions. Further, the recipient’s status is an important factor. The same RBC-antigen formulation that confers tolerance in a healthy quiescent organism may induce a huge immune response in a similar organism in the state of inflammation.

5.3. Vascular and cellular addressing of NC.

Our knowledge of distribution of NCs transferred from RBC carrier to the vascular cells in the vasculature of addressee organs is lacking. It seems logical to postulate that the transfer occurs preferentially in the pre-capillary arterioles and especially in the capillaries, where the RBCs squeezing through the microvasculature come into intense direct contact with the vascular lumen. Uptake in the large arteries is expected to be less effective due to the unfavorable hemodynamics, whereas veins will receive just leftovers of RBC-bound NC.

It is also possible that some NC are more tightly bound to RBC membranes than NC swapped to vascular recipients in the first pass. Currently nanocarriers non-specifically attach to random components of RBC membrane, which may create a spectrum of affinities to RBC and degrees of spatial freedom for interactions with potential molecular partners in blood and vascular lumen. Understanding this aspect and design of selective coupling NC to carrier RBC may enable specific transfer to intended vascular areas and cells. In theory, endothelial cells covering the vascular lumen and likely intravascular leukocytes are the main addressees of RH.

Use of nanocarriers carrying affinity ligands towards vascular cells further expands the versatility of RH. For example, antibodies and single chain fragments to PECAM-1 and ICAM-1 are known to target cargoes to the vasculature, especially in the lungs after IV injection(118, 119). Adsorption of anti ICAM-1 antibody prolonged the residence of nanoparticles in the lungs. Specifically, the exposed part of the nanoparticles on the RBC surface was incubated with anti-ICAM-1 antibody to allow its adsorption. Anti-ICAM-1 antibody, as shown in prior studies(28), when coated on nanoparticles was thought to improve binding and internalization of nanoparticles in the lung endothelium which may explain their prolonged residence in lungs.

Engineering of nanocarrier shape in itself is known to impact nanoparticle circulation and targeting(120, 121). Specifically, rod-shaped particles have been shown to exhibit prolonged circulation and enhanced lung accumulation(120). These features of rod-shaped nanoparticles were further enhanced by their adsorption on RBCs. Specifically, RBC-hitchhiking anti-ICAM-1-coated rod-shaped nanoparticles exhibited a 6-fold improvement in lung/liver accumulation ratio compared to that by their spherical counterparts(122). Studies also showed a clear inverse correlation between the lung accumulation and spleen/liver accumulation.

Furthermore, covalent conjugation of monoclonal antibodies onto PS-NPs, both eliminated PS-NPs’ RBC agglutination, and offered the opportunity to combine RH with affinity-moiety-based targeting. RH anti-PECAM-PS-NPs accumulated in the lungs at ~70% injected dose, the highest reported for a nanocarrier of >100 nanometers (without pulmonary embolism), and improved lung-to-liver ratio >700-fold over free IgG-coated PS-NPs.

5. Safety aspects of RH.

RH is a relatively new approach for drug delivery. Its evolution, optimization and translation into the clinical domain will face challenges, some of which are difficult to anticipate at this early stage of development. Nevertheless, some scenarios and challenges can already be expected and deserve special attention(123). In particular, coupling to the carrier RBC of cargoes including nanocarriers may alter the cargo and carrier. In fact, the very hitchhiking is a manifestation of this postulate. However, some parameters of RBC and nanocarriers may change in unintended or undesirable ways(124).

Although the major fraction of the NC load swaps from RBC to the target vasculature, the residual fraction of RBC-bound NC will behave in the body differently from free NC. It is reasonable to anticipate that in comparison with free NC, the RBC-bound NC counterparts will feature more prolonged PK provided by carrier RBC. Further, RBC-bound NC partition in the cellular fraction of blood from plasma, inhabited by free NC. As result, hydrodynamic factors will separate RBC-bound NC from the vascular walls in the arteries. Biodistribution in the organs at the end of circulation phase will shift from dominant hepatic uptake typical of free NC, to the uptake shared between liver and spleen, typical of modified RBC.

Loading of the drug cargo inevitably alters RBC. The question is, whether changes in RBC biocompatibility caused by drug delivery are tolerable (or even detectable) in the context of the envisioned therapy. This is an important safety issue, somewhat common for many drug delivery strategies using RBC. For example, encapsulation of drugs into RBC leads to loss of hemoglobin content, and decrease in deformability, resilience and structural integrity of the membrane, as well as surface exposure of entities normally localized in the inner leaflet of the membrane (e.g., phosphatidyl serine, PS)(125–127). Uncontrolled coupling of cargoes results in inhibition of surface glycoproteins that protect RBC from complement and phagocytosis (e.g., Decay Acceleration Factor, DAF, and CD47, respectively). These abnormalities in many aspects are reminiscent of pathological changes in senescent, damaged and infected RBC(128–130) and RBC in disease conditions including malaria, sickle cell disease (SCD), paroxysmal nocturnal haematuria and sepsis surface(131, 132). Enhanced uptake by phagocytes in the RES and accelerated clearance impede RBC drug delivery function(133). In addition, enhanced uptake by RES phagocytes poses risk of inhibiting this host defense system due to overload, or, in opposite, its pathological activation leading to cytokine release syndrome(134, 135). Hemolysis pose risk of toxic effect of free hemoglobin on vascular and renal cells, whereas intravascular aggregation may cause RBC retention and occlusion of microvasculature, leading to pro-inflammatory changes in endothelium(136, 137).

Coupling rigid polystyrene NP affects RBC biocompatibility in a multifaceted fashion including sensitization to mechanical, oxidative and osmotic stresses, enhanced agglutination, and lysis in blood(138, 139). Fortunately, latex beads that cause these adversities do not have much of medical utility, whereas nanogels, liposomes and other prospective agents for RBC HH are remarkably more benign: coupling of 20–50 these nanocarriers per RBC caused no detectable changes of RBC biocompatibility parameters in vitro and in vivo in animal studies. Testing of these translational and clinically used nanocarriers including liposomes and nanogels caused no detectable agglutination, pulmonary embolism, inflammation and acute pulmonary arterial hypertension, reduced blood oxygenation or other adversities to RBC and the animals.

The issues discussed in this section require close attention and rigorous evaluation, in order to detect and defuse potential issues. As with any toxicological effect, the final benefit/risk ratio and appraisal of tolerability of potential issues depends on the dose (the Paracelsus Rule) and health state of the patient.

6. Conclusion.

Drug delivery systems utilizing RBC-hitchhiking represent the fruit of a fortuitous “collaboration” between two drastically different carriers. RBC are large, uniform, homogenous, long-circulating yet rapidly degradable biological carriers, naturally destined for uptake in the spleen. NCs are two orders of magnitude smaller, artificial, diverse, heterogeneous, rapidly eliminated yet degraded slower, with controlled degradation rate and mechanism, naturally destined to uptake in the liver.

Both carriers have no specific affinity including to each other (Table 1).

Table 1.

Features of red blood cells and nanocarriers for drug delivery systems utilizing red blood cell hitchhiking

Features	RBC	Nanocarriers
Nature	Natural cell	Artificial multimolecular assemblies
Size	6 micron	20–200nm
Homogeneity	Exceptionally high	Low
Diversity	None	High
Circulation time	Weeks to months	Minutes to hours
Degradation pathway	Phagocytosis and hemolysis	Phagocytosis and hydrolysis
Degradation rate	Minutes	Hours to weeks
Destination in body	Spleen	Liver
Affinity	None	Optional
Function in HH	Delivery NC to vasculature of interest	Delivery of drugs within vasculature of interest

Optimal combination of these distinct entities yields a novel drug delivery paradigm to the frontiers of biomedical research. No doubt, further studies will expand the horizons of this paradigm. For example, each medical application of RH will require meeting specific parameters of drug delivery, which can be modulated by optimizing size, shape, deformability, surface charge and other features regulating interaction of NC with carrier RBC and vascular counterparts. Further, delivery of diverse drugs will ask for different carriers based on formulating parameters, as well as rate and mechanism of drug release.

Cells other than RBC are explored in the context of NC hitchhiking, capitalizing on advances of cellular therapies involving natural and modified immune cells, stem cells, progenitor for specific cell types and fully developed tissue cells. In theory, many if not every cell type used for cellular therapies may find an additional utility for delivery of nanocarriers loaded intracellularly or on the cell surface via several methods including HH. For example, platelets and leukocytes offer features unattainable by RBC, including natural tropism to sites of desirable drug delivery. Thus, platelets can deliver cargoes to sites of blood clotting, whereas leukocytes to sites of inflammation. In addition to natural tropism, these cells possess natural mechanisms for activation and release of content from the inner granules in these sites of therapeutic interest.

On the other hand, the functional activity, diversity, sensitivity and complexity of cells other than RBCs also represent challenging aspects of co-opting them for NC HH. Our incomplete understanding and lack of control of the workings of these sophisticated objects inevitably poses concerns of unintended effects. In other words, these cells have their own agendas, which have nothing to do with drug delivery. The obvious fact that these agendas change under pathological conditions further complicates the drug delivery use of cells. The RBC, designed for transporting functions, generally does not present such challenges. Instead, the RBC offers great opportunities.