Physiological principles underlying the kidney targeting of renal nanomedicines

Abstract

Kidney disease affects more than 10% of the global population and is associated with considerable morbidity and mortality, highlighting a need for new therapeutic options. Engineered nanoparticles for the treatment of kidney diseases (renal nanomedicines) represent one such option, enabling the delivery of targeted therapeutics to specific regions of the kidney. Although they are underdeveloped compared with nanomedicines for diseases such as cancer, findings from preclinical studies suggest that renal nanomedicines may hold promise. However, the physiological principles that govern the in vivo transport and interactions of renal nanomedicines differ from those of cancer nanomedicines, and thus a comprehensive understanding of these principles is needed to design nanomedicines that effectively and specifically target the kidney while ensuring biosafety in their future clinical translation. Herein, we summarize the current understanding of factors that influence the glomerular filtration, tubular uptake, tubular secretion and extrusion of nanoparticles, including size and charge dependency, and the role of specific transporters and processes such as endocytosis. We also describe how the transport and uptake of nanoparticles is altered by kidney disease and discuss strategic approaches by which nanoparticles may be harnessed for the detection and treatment of a variety of kidney diseases.

Introduction

A variety of physiological principles distinguish nanoparticles from small molecules with respect to their transport and interactions in the body^1–5. These differences underlie the potential for engineered nanoparticles to address challenges in disease diagnosis and treatment that cannot be readily achieved with small molecules. Nanoparticles that are engineered to have medical functions — also known as nanomedicines — generally consist of a nano-sized core coated with surface ligands that confer stability, reduce their non-specific accumulation in healthy tissues and organs, and/or enhance selective targeting to the diseased tissues. Medical functions are achieved by loading therapeutic or imaging agents into the core or conjugating them to the nanoparticle surface (Fig. 1a). Currently, nanomedicines are predominantly used in the context of cancer diagnosis and treatment. Of the 31 nanomedicines^6–8 approved by the FDA and EMA, nine (~29%) have been approved for cancer treatment and six (~19%) are for the imaging of cancer and other diseases. The successful translation of cancer nanomedicines into the clinic has deepened our understanding of the physiological principles involved in the transport and interactions of engineered nanoparticles in normal and cancerous tissues^9–13, providing a foundation for the design of innovative nanomedicines for diseases beyond cancer.

Fig. 1 |. Glomerular filtration of nanoparticles.

a, Nanomedicines typically consist of a nano-sized core coated with versatile ligands. Therapeutic and imaging drug molecules can be loaded into the core or conjugated to the nanoparticle surface. Surface ligands such as polyethylene glycol are often used as ‘antifouling’ agents to stabilize nanomedicines in the physiological environment and reduce their non-specific accumulation in off-target organs and normal tissues. Active targeting ligands that can target specific receptors at disease sites are also used to enhance targeting to the diseased tissues. The in vivo transport and interactions of nanomedicines can be further modulated by engineering diverse structural properties including the size, shape, surface chemistry, density and stiffness of the nanoparticles. b, Nanoparticles can be filtered by the glomerulus to enter the renal tubules. To enter the Bowman’s space, nanoparticles in the glomerular capillaries must filter through the three layers that comprise the glomerular filtration barrier. The first layer comprises glomerular endothelial cells, which are covered by a dense surface glycocalyx and are perforated by fenestrae of 60–80 nm. The middle layer is the glomerular basement membrane (GBM) — a non-cellular meshwork composed of negatively charged and porous extracellular matrix with an average pore size of ~10 nm. The outer layer comprises podocytes with long foot processes separated by slit pores with an average size of ~12 nm. c, The glomerular filtration barrier behaves as a ‘band-pass’ size filter and thus the glomerular filtration of nanoparticles is size-dependent. In the nanometre range, filtration efficiency increases with reductions in nanoparticle hydrodynamic diameter (HD). When dispersed in solutions, nanoparticles can interact with the solvent molecules, forming a hydrodynamic layer on the surface. HD takes into consideration the core and the surface coating of nanoparticles and the hydrodynamic layer attached to the particle surface. However, filtration efficiency decreases for nanoparticles in the sub-nanometre (<1 nm) range, due to the stronger trapping effect of the endothelial glycocalyx. The negatively charged GBM has traditionally been thought to facilitate the filtration of positively charged dextran nanoparticles with a radius in the range 1.8–4.4 nm compared with the filtration of neutral and negatively charged nanoparticles. However, contradictory findings suggest that a deeper understanding of the charge dependency of glomerular filtration is needed. ID, injected dose.

Compared with cancer nanomedicines, engineered nanoparticles for kidney disease management (renal nanomedicines) are notably underdeveloped, perhaps in part due to the misconception that kidney disease is less life-threatening than cancer and because of the general availability of kidney dialysis or transplantation for the management of kidney failure in high-income regions. However, kidney disease is associated with substantial morbidity and mortality, and earlier and better treatments are needed. Most current medications for kidney diseases are small-molecule drugs. For instance, small-molecule drugs with supportive functions, such as loop diuretics and renin–angiotensin system inhibitors, are intended to delay the decline in kidney function^14–22. Other small-molecule drugs, such as antioxidants (for example, N-acetylcysteine) and antifibrotic drugs (for example, pirfenidone), are also under investigation for their ability to reduce levels of reactive oxygen species (ROS) and treat fibrosis, respectively, but low and non-specific targeting to damaged kidney tissues limit their efficacy^23–29. Nucleic acid-based drugs, such as small interfering RNAs (siRNAs) and antisense oligonucleotides, are emerging as a novel type of therapeutic for kidney disease³⁰; however, their utility is hampered by a lack of specific delivery vectors to target these agents to the kidneys³¹. By contrast, the successes of nanomedicines in oncology naturally inspire applications of engineered nanoparticles to address the therapeutic challenges of kidney diseases. To date, only three iron-based colloidal nanoparticles (Ferrlecit, Venofer and Ferumoxytol) have been approved for use in patients with kidney disease — specifically for the treatment of iron deficiency anaemia⁸.

Although knowledge derived from cancer nanomedicines will accelerate the development of nanomedicines for other diseases^32–37, including kidney diseases^38,39, important differences between nanoparticle transport and interactions between cancer and kidney disease should be considered. For instance, solid tumours are in general highly permeable to nanoparticles — even those >100 nm in size — due to the presence of leaky vasculature^12,40,41 through which cancer nanomedicines can passively extravasate. By contrast, kidney-targeted nanoparticles often need to be small enough to be filtered through the glomerular capillary wall, and thus the kidney filtration threshold for nanoparticles is only ~5.5 nm⁴². In addition, the high metabolic rate of cancer cells may enhance the accumulation of nanoparticles⁴³; by contrast, the cellular uptake of nanoparticles by injured proximal tubular epithelial cells (PTECs) is lower than that by healthy PTECs⁴⁴. Moreover, unlike the distinct boundaries between tumour and adjacent healthy tissue, normal and injured kidney cells are often mixed in diseased kidneys, posing additional challenges for the selective targeting of injured kidney tissue. These differences suggest that a deep understanding of the physiological principles of the intra-kidney transport and interactions of nanoparticles in healthy and diseased kidneys will be required for the development of renal nanomedicines with optimal efficacy and minimal toxicity.

Over the past 15 years, we and other groups have worked on the development of renal-clearable nanoparticles — a class of engineered nanoparticles that are eliminated through the kidneys without notable accumulation in the liver⁴⁵ and other organs. In this Review, we summarize our latest understanding of the physiological principles regulating the transport and interactions of nanoparticles in the kidneys as well as advances in the development of renal nanomedicines for the treatment of kidney diseases. We also discuss a mechanism by which healthy PTECs reduce their long-term retention of nanoparticles and thereby the potential nephrotoxicity of nanoparticles. We hope this Review will serve as a bridge to connect the nanomedicine and nephrology communities to accelerate the development and clinical translation of renal nanomedicines.

Physiological principles

The selective delivery of engineered nanoparticles into the kidneys is the first step in applying nanomedicines for the diagnosis and treatment of kidney diseases. As one of the major organs involved in the elimination of substances from the body, the kidneys receive and filter more than 20% of the cardiac blood output every minute through its basic functional unit — the nephron — which is composed of the glomerulus, the proximal tubule, the loop of Henle, the distal tubule, the collecting duct and the peritubular capillaries. Once introduced into the circulation, nanoparticles are transported into the kidneys through the glomerular and peritubular capillaries. Thus, the targeting efficiency of nanoparticles to the kidneys depends on the competing elimination of the blood-circulating nanoparticles by other organs; the efficiency of their glomerular filtration from glomerular capillaries; and the efficiency of their tubular secretion in peritubular capillaries. Over the past few decades, studies have revealed key physiological principles that regulate the entrance of nanoparticles from the bloodstream into the kidneys and their subsequent interaction with cells of the kidney.

Size-dependency and glomerular filtration

The glomerulus contains a cluster of branching and anastomosing glomerular capillaries that are structurally supported by the mesangial cells and surrounded by the Bowman’s space. The unique multi-layer structure of the glomerular capillary wall behaves as a band-pass filter that regulates the filtration of nanoparticles from the blood into the Bowman’s space and the tubular cavities (Fig. 1b). The glomerular endothelial cell layer is covered by a dense surface glycocalyx and perforated by fenestrae that are much larger than those of other capillaries (60–80 nm versus ~5 nm, respectively^46–49). The glycocalyx is composed of glycoproteins including proteoglycans, and is an important part of the glomerular filtration membrane that regulates the glomerular filtration of substances^50,51. Surrounding the glomerular endothelial cell layer is the glomerular basement membrane (GBM), a non-cellular meshwork composed of negatively charged and porous extracellular matrix macromolecules with a pore size up to ~10 nm^52–58. The outer surface of the GBM is covered by a layer of podocytes with long foot processes separated by slit pores with an average size of ~12 nm⁵⁹. Mesangial cells are separated from the blood only by the glomerular endothelial cell layer. Therefore, nanoparticles that are larger than the pore size of the GBM but comparable to the size of glomerular endothelial fenestrae can extravasate the glomerular endothelial layer and be taken up by mesangial cells⁶⁰ or deposited on the GBM^61,62, but cannot be filtered through the GBM to reach the podocytes and tubules.

One early study used quantum dots of different sizes to investigate the size-dependency of filtration through glomerular capillaries⁴² (Fig. 2a). The first key finding of that work was that the quantum dots needed to be highly resistant to serum protein binding to retain their original size, since the binding of serum protein could increase the size of the quantum dots by >15 nm. The second key discovery was that increasing the size of the quantum dots from 4.36 nm to 8.65 nm decreased their renal clearance from ~80% of the injected dose (%ID) to ~20 %ID. Based on these analyses, the researchers determined the kidney filtration threshold for quantum dots to be ~5.5 nm. This observed filtration trend is consistent with the size-dependent filtration of neutral dextran nanoparticles for which the fractional clearance increases with a decrease in size from 8.8 nm to 3.6 nm (ref. 63). The identification of the glomerular filtration threshold of quantum dots greatly facilitated the development of renal-clearable engineered nanoparticles, including gold nanoparticles (AuNPs)^64–66 (Fig. 2b), copper nanoparticles⁶⁷, silver nanoparticles^68,69, iron oxide nanoparticles^70–72, silica nanoparticles (such as Cornell dots)⁷³ (Fig. 2c), polysiloxane nanoparticles (such as AGuIX)^74,75 (Fig. 2d) and polymeric cyclodextrin nanoparticles (such as H dots)^76,77 (Fig. 2e). Among these, Cornell dots^78,79 and AGuIX⁸⁰ nanoparticles are currently in clinical trials for the imaging and treatment of cancer. Renal-clearable nanoparticles composed of novel metal oxides^81,82 and dextran polymer^83–85 have also now been reported. In addition to nanoparticles, renal-clearable macromolecules, such as the PEGylated peptide–dye conjugates (for example, LUM015)⁸⁶, and small molecules such as 3,6-diamino-2,5-bis(N-[(1R)-1-carboxy-2 hydroxyethyl]carbamoyl)pyrazine (MB-102)⁸⁷, have been developed for the imaging ofvarious diseases. LUM015 (ref. 88) and MB-102 (ref. 89) are currently in clinical trials for cancer imaging and the transdermal measurement of kidney function, respectively. In 2017, we further advanced our fundamental understanding of the size-dependency of glomerular filtration by investigating the elimination of even smaller engineered nanoparticles — gold nanoclusters with sizes below 1 nm — by the kidney⁹⁰ (Fig. 1c). Surprisingly, we found that the 24-h renal clearance efficiencies of sub-nanometre gold nanoclusters decreased from 51.57 %ID to 19.07 %ID as the size decreased from 25 gold atoms (Au₂₅SG₁₈) to 10–11 gold atoms (Au_10–11SG_10–11). Further electron microscopic studies led to the discovery that the dense glycocalyx layer of the glomerular endothelial cells contribute to the filtration of sub-nanometre gold nanoclusters by forming sub-nanometre cavities that temporarily trap smaller gold nanoclusters and slow their filtration rate. However, continued renal elimination of the sub-nanometre gold nanoclusters was evident 1 month after their intravenous administration. This inverse size-dependency in the glomerular filtration of sub-nanometre nanoclusters was also observed for soft nanoparticles⁹¹, highlighting the important role of the glycocalyx layer in the glomerular filtration of engineered, sub-nanometre nanoparticles.

Fig. 2 |. Representative nanoparticles eliminated by the kidneys through glomerular filtration.

a, Quantum dots with a cadmium selenide (CdSe) core and a zinc sulfide (ZnS) shell with a core–shell size of 2.8–4.3 nm and a hydrodynamic diameter (HD) of 4.4–8.7 nm. Coating of the surface with cysteine prevents binding of the quantum dots to serum proteins in the circulation. b, Gold nanoparticles coated with glutathione or polyethylene glycol (PEG). The core sizes are less than 3 nm. The HDs are less than 7 nm. c, Cornell dots are ~6 nm silica nanoparticles coated with PEG and embedded with IRDye 800CW for the fluorescence imaging of cancer and lymph nodes. d, AGuIX nanoparticles are ~4 nm polysiloxane nanoparticles coated with gadolinium-dodecane tetraacetic acid (DOTA) for MRI and photodynamic therapy of cancer. e, H dots are ~6 nm polymeric nanoparticles with a polylysine backbone conjugated with β-cyclodextrin molecules and ZW800 (a zwitterionic dye molecule with a near-infrared fluorescence emission at ~800 nm). MW, molecular weight.

Charge-dependency and glomerular filtration

The negative charge of the GBM^54,92,93 led to the suggestion that glomerular filtration is charge-dependent, with electrostatic interactions resulting in the more efficient filtration of positively charged dextran nanoparticles compared with that of neutral and negatively charged dextran^94–97 (Fig. 1c). However, later studies showed that the charge-dependent glomerular filtration of nanoparticles is much more complicated than initially believed. The amount of positively charged dextran nanoparticles retained in the glomerulus was shown to be more than ten times higher than that of negatively charged dextran⁹⁸. In addition, negatively charged Ficoll (a polysaccharide nanoparticle) is more easily filtered through the glomerulus than neutral Ficoll⁹⁹. These findings suggest that simple coulombic interaction-mediated charge selectivity might not be the only contributor to the charge-selectivity of glomerular filtration. More thorough investigations are therefore needed to advance our understanding of the charge-selectivity in the glomerular filtration of engineered nanoparticles.

Particle density and glomerular filtration

In addition to size and charge, other factors such as particle density can affect the glomerular filtration of nanoparticles. Our group investigated the effect of particle density on the renal clearance of nanoparticles by systematically comparing the renal clearance efficiencies of gold, silver and gold–silver alloy nanoparticles⁶⁹. We found that the renal clearance efficiency of silver nanoparticles (~45 %ID; density ~10.5 g cm⁻³) at 2 h after injection is higher than that of AuNPs (~30 %ID; density ~19.3 g cm⁻³) due to the weaker margination effect of low-density silver nanoparticles towards the blood vessel wall, which favours their glomerular filtration.

Transporter-mediated tubular secretion of nanoparticles

Although glomerular filtration is generally considered to be the main pathway through which engineered nanoparticles enter the kidneys, tubular secretion can also result in the selective targeting of nanoparticles to the kidneys (Fig. 3a). In contrast to the passive process of glomerular filtration, tubular secretion is an active process whereby small-molecule and low molecular weight (MW) metabolites¹⁰⁰, toxins¹⁰¹ and drugs¹⁰² are transported directly from the blood into the tubular cavities, bypassing glomerular filtration. Once these small molecules reach the peritubular capillaries, in which the fenestrae are ~5 nm size⁴⁹, they can extravasate into the interstitial space where they bind to influx transporters expressed on the basal membrane of PTECs, enter the PTEC cytoplasm and are subsequently excreted into the tubular cavities by efflux transporters on the luminal membrane of the PTECs. Although the prevailing understanding is that tubular secretion mainly applies to small molecules such as mercaptoacetyltriglycine^103,104 and para-aminohippuric acid¹⁰⁵, we have shown that this active elimination pathway is also used to enhance the kidney targeting of ultrasmall organic nanoparticles. For example, we have shown that an ultrasmall organic nanoparticle, ICG–PEG45 (ref. 106) (MW ~ 3,000 Da) — which we generated by conjugating a polyethylene glycol (PEG) 45 chain (that is, a PEG with 45 repeating ethylene glycol units) with the clinically approved, near-infrared-emitting dye, indocyanine green (ICG¹⁰⁷) — is mainly eliminated through the kidneys, unlike free ICG dye, which is eliminated through the liver¹⁰⁸ (Fig. 4a). At 24 h after intravenous injection, 92.9 %ID of ICG–PEG45 was found in the urine whereas free ICG was undetectable in the urine. Furthermore, administration of the organic anion transporter inhibitor, probenecid¹⁰⁹ — to inhibit the basolateral uptake of ICG–PEG45 into PTECs — reduced the renal clearance efficiency of ICG–PEG45 by 50% 30 min after injection, whereas clearance of the glomerular-filtrable control probe IRdye 800CW conjugated with PEG45 (800CW–PEG45 (refs. 91,106)) remained unchanged. We have also shown that large PEGs with molecular weights around 11,000 Da can also be secreted directly from the blood into the urine through the proximal tubules¹⁰⁶. These findings clearly indicate that the tubular secretion pathway is an important pathway for nanoparticles to target the kidneys through the actions of both influx and efflux transporters.

Fig. 3 |. Tubular secretion and luminal tubular uptake of nanoparticles.

a, Tubular secretion describes the process by which nanoparticles cross the peritubular capillary wall to reach the basolateral side of proximal tubule epithelial cells (PTECs), followed by their intracellular transport to reach the luminal side of PTECs from where they are eventually secreted into the tubular lumen for elimination in urine. Nanoparticles can extravasate the peritubular capillary wall through paracellular or intracellular pathways. The uptake of nanoparticles by PTECs can be either transporter-mediated (whereby nanoparticles enter the basolateral side of the PTECs through membrane influx transporters (IT) and are eliminated from the luminal side of PTECs through efflux transporters (ET)) or via transcytosis (whereby nanoparticles enter the basolateral side of the PTECs through endocytosis and are released into the tubular lumen through exocytosis). b, Nanoparticles can also be taken up by PTECs following their glomerular filtration. This uptake is via endocytosis and is dependent on the surface physicochemical properties of the nanoparticles. Endocytosed nanoparticles are not necessarily reabsorbed back into the peritubular blood circulation.

Fig. 4 |. Representative nanoparticles that are processed by the kidney tubules.

Representative nanoparticles eliminated by the kidneys through tubular secretion (parts a–d). Indocyanine green (ICG) conjugated with a polyethylene glycol (PEG) polymer with 45 repeating units (PEG45; molecular weight (MW) ~2,800 Da) (part a). The tubular secretion of ICG-PEG45 is mediated by organic anion transporters. Iron oxide nanoparticles coated with 6-nitrodopamine–PEG as surface ligands (hydrodynamic diameter (HD) ~140 nm) (part b). Poly(lactic-co-glycolic acid) (PLGA, MW 24–38 kDa) nanoparticles coated with different glycosaminoglycans (heparin, MW 16–17 kDa; chondroitin sulfate, MW 300 kDa; hyaluronic acid, MW 900 kDa; dermatan sulfate, MW 43 kDa) with HDs of 130–180 nm (part c). Mesoscale PLGA–PEG polymeric nanoparticles (HD ~400 nm) (part d). The tubular secretion of the large nanoparticles in parts b, c and d is mediated through transcytosis. Representative nanoparticles taken up by proximal tubule epithelial cells (PTECs) from the lumen after glomerular filtration (parts e–g). Positively charged polyamidoamine dendrimers have a high affinity for the negatively charged microvilli on the luminal surface of PTECs (part e). 1,2-Distearoyl-sn-glycero-3-phosphoethanolamine-poly(ethylene glycol) (DSPE–PEG) nanoparticles conjugated with a kidney targeting peptide (KTP) (part f) enable efficient luminal tubular uptake after glomerular filtration through megalin-mediated endocytosis. Polyvinylpyrrolidone-co-dimethyl maleic anhydride [poly(VP-co-DMMAn)] polymeric nanoparticles (MW 10 kDa) (part g) also demonstrate high luminal tubular uptake after glomerular filtration.

Transcytosis-mediated tubular secretion of nanoparticles

As described above, glomerular filtration of nanoparticles is limited by the filtration size threshold (~5.5 nm) whereas transporter-mediated tubular secretion of nanoparticles requires specific interaction of the nanoparticles with relevant transporters; thus, selective targeting of large nanoparticles (>10 nm) to the kidneys remains a challenge. Using intravital microscopic imaging, a 2019 study¹¹⁰ showed that pegylated iron oxide nanoparticles of ~140 nm can be secreted directly from blood into the urine, although with much lower efficiency than that of ICG–PEG45 nanoparticles (Fig. 4b). Another study also showed the efficient renal clearance of glycosaminoglycan-modified poly(lactic-co-glycolic acid) (PLGA) nanoparticles with sizes ranging from 130 nm to 180 nm (56–74 ID% 2 h after intravenous injection)¹¹¹ (Fig. 4c). Mesoscale PLGA–PEG nanoparticles of ~400 nm also selectively accumulated in peritubular capillary endothelial cells and PTECs^112,113 (Fig. 4d). The kidney targeting of these large nanoparticles was attributed to transcytosis, first through their uptake by endothelial cells of the peritubular capillaries via endocytosis and their release into the interstitial space by exocytosis, followed by their uptake by PTECs via endocytosis or micropinocytosis and secretion into the tubular lumen through exocytosis (Fig. 3a). However, more detailed studies are needed to validate this process at the molecular level.

Luminal uptake of nanoparticles by proximal tubules after glomerular filtration

Once nanoparticles enter the proximal tubular cavities after glomerular filtration, their transport and interactions strongly depend on the surface properties of the nanoparticles. Depending on their physicochemical properties, nanoparticles can be efficiently taken up by PTECs from the luminal side after glomerular filtration (Fig. 3b) without the need for reabsorption back to the bloodstream, as is the case for some small molecules¹¹⁴ and albumin^115,116. The most commonly used ligands for renal-clearable nanoparticles, such as PEG, inulin, sinistrin, cyclodextrin and dextran, are classic ligands that reduce the protein and cellular affinity of nanoparticles¹¹⁷ and can therefore contribute to the glomerular filtration of nanoparticles by minimizing protein binding; however, these ligands might also reduce interaction of the nanoparticles with PTECs after their glomerular filtration. By contrast, ligands that can efficiently interact with the cell membrane may facilitate the luminal uptake of nanoparticles by PTECs after glomerular filtration. For instance, since the luminal surface of PTECs is covered by densely packed and negatively charged microvilli, positively charged small nanoparticles, such as AuNPs⁴⁴, polysiloxane nanoparticles¹¹⁸, low-molecular-weight chitosan^119,120, low-generation dendrimers^121,122 (Fig. 4e), polyethyleneimine¹²³ and lysozymes^124,125, are more easily retained in the proximal tubules and taken up by PTECs than negatively charged small nanoparticles. Moreover, due to the high expression of membrane receptors on the luminal surface of PTECs, nanoparticles coated with active targeting ligands, such as peptides^126–128 (Fig. 4f), light chain proteins^129,130 and low-molecular-weight polymers^131–133 (Fig. 4g), are specifically taken up by PTECs through receptor binding¹³⁴. We have demonstrated that AuNPs are specifically taken up by PTECs through endocytosis following their glomerular filtration and that a positive surface charge facilitates their uptake through clathrin-mediated endocytosis⁴⁴.

Targeting diseased kidneys

Kidney diseases can be caused by a variety of factors and can involve injury to different components of the kidney, inevitably altering the transport and interactions of therapeutically administered renal nanomedicines. For example, injury to physiological barriers, such as glomerular size and charge barriers^135–137, may affect the transport and interaction of nanoparticles in the kidneys. However, different kidney diseases can affect physiological barriers in different ways, and therefore different strategies will be needed to target renal nanomedicines to different components of the nephron (see Supplementary Table 1).

Targeting glomerular diseases

The glomerulus is the key compartment involved in the filtration of fluid and is also one of the major sites of injury in kidney diseases. Glomerular damage can result from genetic mutations, drugs, toxins and the deposition of immune complexes, or can be a consequence of other chronic diseases^138,139. However, the exact sites of injury within the glomeruli may vary, and therefore the strategies for targeting different disease sites within the glomerulus may also differ. For instance, in addition to providing structural support for the glomerular tuft, mesangial cells also have critical roles in the initiation and progression of glomerular diseases through the chemotaxis of inflammatory cells, and proliferation and excessive production of extracellular matrix^140,141. As mentioned earlier, mesangial cells are separated from blood only by the glomerular endothelial cell layer and they can therefore be targeted by large and non-renal-clearable nanoparticles that are smaller than the glomerular endothelial fenestrations (~60–80 nm). A study that used pegylated AuNPs to test the size-dependency of nanoparticles in targeting mesangial cells found that pegylated AuNPs of ~75 nm were most effective⁶⁰ (Fig. 5a). In another study, albumin nanoparticles of ~100 nm were used as a vehicle to deliver the antioxidant celastrol to mesangial cells in a rat model of mesangioproliferative glomerulonephritis¹⁴² (Fig. 5b). The celastrol–albumin nanoparticles demonstrated greater efficacy than free celastrol in alleviating proteinuria, inflammation, glomerular hypercellularity, and the excessive deposition of extracellular matrix. In addition, the celastrol–albumin nanoparticles showed minimal systemic cardiotoxicity, hepatotoxicity and neurotoxicity compared with the toxicity of free celastrol due to their lower accumulation in off-target organs. Other studies have targeted membrane receptors expressed by mesangial cells. For example, large liposomes (~170 nm) conjugated with OX7 antibody fragments (directed against Thy1.1 antigen)¹⁴³ (Fig. 5c) and siRNA-loaded polycation cyclodextrin nanoparticles conjugated with mannose ligand of ~70 nm (specific for the mannose receptor of mesangial cells)¹⁴⁴ (Fig. 5d) demonstrated enhanced targeting to mesangial cells compared with that of non-conjugated nanoparticles.

Fig. 5 |. Representative nanomedicines that target the glomerulus.

a, Pegylated gold nanoparticles (~75 nm) efficiently target mesangial cells within the glomerular tuft. b, Targeting of albumin nanoparticles (~100 nm) loaded with celastrol to mesangial cells showed therapeutic efficacy and minimal systemic toxicity in a model of mesangioproliferative glomerulonephritis. c, OX7 antibody fragment-conjugated pegylated liposomes (~170 nm) target the Thy1.1 antigen on mesangial cells. d, Mannose-coated polycationic cyclodextrin nanoparticles (~70 nm) target mannose receptors on the mesangial cells. Adamantane–polyethylene glycol (PEG) and adamantane–PEG–mannose ligands are conjugated onto the polycationic cyclodextrin backbone through cyclodextrin-adamantane host–guest interaction for stabilization in the physiological environment and active targeting, respectively. Negatively charged small interfering RNA (siRNA) is encapsulated in the polycationic backbone through electrostatic interaction as a therapeutic agent. e, Positively charged cationized ferritin (~13 nm) can target the negatively charged glomerular basement membrane (GBM). f, Similarly, positively charged polycationic cyclodextrin nanoparticles (~70 nm) can also target the negatively charged GBM. g, The ‘shamporter’ system targets podocytes through monovalent IgG for the delivery of siRNA. Active targeting monovalent IgG is conjugated with positively charged protamine through a biotin–streptavidin interaction. Negatively charged siRNA is attached to positively charged protamine through an electrostatic interaction. MW, molecular weight.

The GBM is an important component of the glomerular filtration barrier, and defects in the GBM are associated with a wide spectrum of kidney diseases including Alport syndrome, Pierson syndrome and anti-GBM disease^52,145. The GBM also serves as a unique reservoir for nanomedicines, facilitating the slow release and delivery of therapeutics. The GBM is covered by the glomerular endothelial cell layer and thus can only be reached by nanoparticles that are smaller than the endothelial fenestrae (~60–80 nm). Of note, the GBM was originally identified as a negatively charged barrier following the observation that cationic ferritin was preferentially trapped in the GBM (Fig. 5e) compared with anionic and neutral compounds¹⁴⁶. The preferential accumulation of 13 nm cationic ferritin in the negatively charged GBM has been used to achieve non-invasive, in vivo MRI of the GBM in normal mice and in a mouse model of focal segmental glomerulosclerosis^147,148. A further study demonstrated that polycation cyclodextrin nanoparticles of ~70 nm⁶¹ loaded with siRNA (Fig. 5f) accumulate and disassemble in the negatively charged GBM, acting as a reservoir for the slow release the siRNA in the kidneys.

Podocytes are located on the outer surface of the GBM and represent a therapeutic target for diseases such as congenital nephrotic syndrome and diabetic kidney disease. Nanoparticles need to cross the GBM layer to reach podocytes. Hence, podocyte-targeting nanoparticles should be smaller than the pore size of the GBM (~10 nm) and are often conjugated with ligands that target specific receptors on the podocyte membrane¹⁴⁹. For example, albumin–drug conjugates¹⁵⁰ and quantum dots coated with cyclo(RGDfC) peptide (a cyclic pentapeptide that contains the αVβ3 integrin-binding sequence arginine–glycine–aspartate)¹⁵¹ have been developed to specifically target the neonatal Fc receptor and the ανβ3 integrin receptor, respectively, on podocytes. Another study investigated the development of a ‘shamporter’ system¹⁵², comprising a monovalent fragment of an anti-mouse podocyte antibody and a polycationic nuclear protein protamine for siRNA loading, to selectively deliver nephrin siRNA and TRPC6 siRNA to podocytes for the in vivo inhibition of nephrin and TRPC6, respectively (Fig. 5g).

Targeting proximal tubule injuries

The proximal tubules constitutes over 50% of the kidney mass¹⁵³. Their critical roles in regulating the secretion, reabsorption and uptake of metabolic wastes, toxins and drugs makes them a key site of kidney injury and a therapeutic target for renal nanomedicines. By taking advantage of the altered physiological barriers in diseased kidney (Fig. 6), different approaches have been used to deliver different types of nanomedicines to injured proximal tubules (Fig. 7).

Fig. 6 |. Targeting of nanoparticles to proximal tubules in the healthy and injured kidney.

a, The proximal tubules of healthy kidneys can be targeted by ultrasmall glomerular-filtered nanoparticles that demonstrate high luminal uptake efficiency and by tubular-secreted nanoparticles. However, non-renal-clearable nanoparticles that are neither filtered by glomeruli nor secreted by tubules cannot target healthy proximal tubules. b, Following injury, the luminal uptake efficiency of glomerular-filtered nanoparticles by injured proximal tubule epithelial cells (PTECs) decreases as a consequence of cell damage (for example, through apoptosis or necrosis). However, the glomerular filtration of non-renal-clearable nanoparticles that cannot be filtered by healthy glomeruli may increase due to impairment of the glomerular filtration membrane, thus allowing them to reach the proximal tubular lumen where — depending on the uptake efficiency of damaged PTECs — they may be taken up by PTECs. Tubular-secreted nanoparticles can still target PTECs from the basolateral side although the efficiency may be affected by PTEC damage. However, enlargement of the fenestrations of peritubular capillaries may increase the interstitial infiltration of tubular-secreted nanoparticles and non-renal-clearable nanoparticles, which might promote their targeting to the interstitial spaces and the therapeutic treatment of renal interstitial inflammation and fibrosis. In addition, retention of nanoparticles in the tubular lumen can be prolonged as a result of protein cast-mediated tubular obstruction, which may potentially prolong the therapeutic window of nanomedicines. GBM, glomerular basement membrane.

Fig. 7 |. Representative nanomedicines that target the renal tubules and interstitium.

a, Antioxidant DNA origami nanostructures (rectangular, triangular and tubular) have been used to reduce renal tubule levels of reactive oxygen species (ROS) in a model of rhabdomyolysis-induced acute kidney injury (AKI). b, Similarly, antioxidant ultrasmall copper oxide (Cu_5.4O) nanoparticles (~4.5 nm) coated with ascorbic acid and dehydroascorbic acid also target renal tubules to reduce ROS in rhabdomyolysis-induced AKI. c, A folic acid-conjugated dendrimer (~6 nm) targets renal tubules to deliver a small interfering RNA (siRNA) against prolyl hydroxylase domain protein 2 (PHD2) for the treatment of ischaemia–reperfusion injury (IRI)-associated AKI. d, Oltipraz-loaded poly(lactic-co-glycolic acid) (PLGA) nanoparticles (~100 nm) target renal tubules following their filtration via the impaired glomerular filtration barrier for the treatment of IRI. e, A celastrol-loaded albumin nanocomplex (~120 nm) targets proximal tubules via impaired glomerular filtration through megalin–clathrin-mediated endocytosis for the treatment of IRI. f, IL-10-loaded exosomes (~130 nm) also target the renal tubules via impaired glomerular filtration for the treatment of IRI. g, Neutrophil membrane-coated coenzyme Q10-loaded PLGA–polyethylene glycol (PEG) nanoparticles (~100 nm) target the distal tubules via impaired glomerular filtration for the treatment of IRI. h, Tubular-secreted mesoscale PLGA–PEG nanoparticles (~400 nm) loaded with different types of drugs (triptolide, formoterol, edaravone, oligonucleotide antagonist and a renalase agonist peptide) target the proximal tubules for the treatment of different kidney diseases including cisplatin-induced AKI, IRI and cisplatin-induced chronic kidney disease. i, Tubular-secreted mesoscale PLGA–PEG nanoparticles (~400 nm) loaded with emodin can target the interstitium for the treatment of renal fibrosis following unilateral ureteral obstruction (UUO). j, Gold nanoparticles (~60 nm) pegylated with thiolated PEG (HS-PEG) and coated with anti-collagen I antibody target the interstitium following their extravasation via the enlarged fenestrae of damaged peritubular capillaries for CT imaging of renal fibrosis in renal artery stenosis. k, Fibronectin-binding peptide-conjugated pegylated liposomes (~110 nm) loaded with celastrol also target the interstitium following their extravasation via the enlarged fenestrae of damaged peritubular capillaries for the treatment of renal fibrosis in a UUO model of renal fibrosis. l, Similarly, fibrotic kidney-homing peptide-conjugated pegylated liposomes with a PLGA core (~70 nm) loaded with sorafenib target the interstitium following their extravasation via the enlarged fenestrae of damaged peritubular capillaries for the treatment of renal fibrosis in a UUO mouse model.

Injury to PTECs (for example, through the loss of microvilli, cell apoptosis or necrosis or obstruction of the tubular lumen as a consequence of protein casts) can significantly alter the transport and interactions of injected nanoparticles (Fig. 6). We found that the accumulation of renal-clearable, luminescent, ultrasmall AuNPs¹⁵⁴ was enhanced in a model of mild unilateral ureteral obstruction (UUO)¹⁵⁵, whereas the intra-kidney transport of AuNPs from the medulla to the pelvis was slower as a consequence of tubular obstruction¹⁵⁶. By contrast, in a model of severe UUO, the accumulation of AuNPs was decreased due to a reduction in blood perfusion¹⁵⁵. We have also shown that the uptake of positively charged AuNPs by injured PTECs is dramatically decreased in a cisplatin-induced model of acute kidney injury (AKI), regardless of whether PTEC injury was mild or severe⁴⁴. Thus, the high luminal uptake of renal-clearable, positively charged nanoparticles, such as chitosan^119,120, dendrimer^121,122 (Fig. 7c), lysozyme¹²⁵ and low-molecular-weight polymer¹³², by normal PTECs after their glomerular filtration might not be equally demonstratable in the context of PTEC injury. For example, the uptake of ~4 nm serine-modified dendrimer¹²¹ decreased from ~80 %ID to ~55 %ID following mercury chloride-induced injury to the proximal tubules.

The decline in nanoparticle uptake by injured PTECs differs from that by cancer cells, which often exhibit more efficient cellular uptake of nanoparticles than normal cells⁴³. The decline in nanoparticle uptake also complicates the ability to selectively target injured proximal tubules while sparing healthy tubules. One strategy to address this challenge is to develop therapeutic nanoparticles that can be filtered by the glomerulus and can elicit their therapeutic functions extracellularly in proximal tubules without the need for PTEC uptake. A series of antioxidant nanoparticles that can be readily eliminated by the kidneys have been developed to quench the expression of ROS in injured renal tubules in the context of AKI^157–161. These therapeutic nanoparticles do not require the additional loading of external drugs but show greater antioxidant capacity than small-molecule antioxidants in models of AKI. One approach used a series of antioxidant DNA origami nanostructures¹⁶² with rectangular (60 × 90 nm), triangular (120 nm) and tubular (400 nm) shapes (Fig. 7a). Surprisingly, these large nanostructures were readily eliminated by the kidneys into the bladder in healthy mice, although the clearance pathway is not yet clear. The rectangular DNA origami exhibited comparable efficacy to that of high-dose N-acetyl cysteine (a clinically approved antioxidant small molecule) in a mouse model of rhabdomyolysis-induced AKI. Antioxidant metal nanoparticles have also been used in models of AKI. One study showed that ultrasmall antioxidant copper nanoparticles¹⁶³ with an average size of ~4.5 nm can be readily filtered by the glomerulus (Fig. 7b) and demonstrate a scavenging capacity for ROS about tenfold greater than that of DNA origami nanostructures, which enabled a ~250-fold lower injection dose when used to treat rhabdomyolysis-induced AKI.

Enhanced glomerular filtration of large nanoparticles resulting from impairment of the glomerular filtration barrier can also be used to target the tubules (Fig. 6). One study that assessed PLGA nanoparticles of different sizes determined that nanoparticles of 100 nm selectively accumulate in ischaemic mouse kidneys¹⁶⁴. Subsequent delivery of the antioxidant, oltipraz via 100 nm PLGA nanoparticles (Fig. 7d) led to significant decreases in serum creatinine, blood urea nitrogen, tubular necrosis and renal fibrosis compared with that achieved with free oltipraz. Large celastrol-loaded albumin nanocomplexes of ~120 nm¹⁶⁵ (Fig. 7e) and selenium-loaded albumin nanoparticles of ~80 nm¹⁶⁶ also targeted injured renal tubules in a model of ischaemia–reperfusion injury (IRI) through megalin–cubilin-mediated endocytosis of albumin. The celastrol-loaded albumin nanocomplex accumulated mainly in the glomeruli of healthy mice but higher levels of the nanocomplex were observed in the renal tubules of mice with IRI due to impairment of the glomerular filtration barrier. Exosomes of ~130 nm¹⁶⁷ (Fig. 7f) and membrane-coated nanoparticles of ~100 nm¹⁶⁸ (Fig. 7g) have been used to deliver IL-10 and coenzyme Q10 to proximal tubules and distal tubules, respectively, for the treatment of IRI-induced AKI.

Nanoparticles that can directly enter PTECs from peritubular capillaries via the basolateral membrane could also be used to deliver therapeutics to injured proximal tubules (Fig. 6). For example, mesoscale PLGA–PEG nanoparticles can efficiently target PTECs from the basolateral side and have been widely used to deliver a variety of therapeutic agents in different models of kidney disease (Fig. 7h). In one study that used these nanoparticles to deliver anti-inflammatory triptolide for the treatment of IRI¹⁶⁹, the amount of triptolide delivered to the kidneys by the mesoscale nanoparticles was ~50 times higher than that of free triptolide. Moreover, administration of triptolide-loaded nanoparticles to mice with IRI-induced AKI reduced serum creatinine, blood urea nitrogen, complement component C3 and phosphate extracellular signal-regulated kinase to a significantly greater extent than free triptolide. Mesoscale nanoparticles have also been used to deliver formoterol^170,171 and Toll-like receptor 9 (TLR9) antagonists¹⁷² for the induction of mitochondrial biogenesis and to deactivate inflammatory signalling, respectively, in the context of IRI. The delivery of formoterol to the renal cortex via formoterol-loaded nanoparticles was about ten times higher than that achieved with free formoterol, and was associated with significantly greater reductions in levels of serum creatinine and the injury marker KIM1 (also known as HAVCR1), and in tubular necrosis in a model of IRI-induced AKI¹⁷⁰. Moreover, formoterol-loaded nanoparticles ameliorated the tachycardic and hypotensive effects of formoterol, and unlike free formoterol, did not induce changes in heart rate or blood pressure¹⁷¹. Nanoparticles loaded with TLR9 antagonists demonstrated protection against IRI when injected 6 h before renal ischaemia, during reperfusion or 1.5 h after reperfusion, through attenuation of tubular necrosis, apoptosis and inflammation¹⁷². Nanoparticles loaded with a renalase agonist¹⁷³ and the ROS scavenger, edaravone¹⁷⁴, have also been delivered to proximal tubules for the prevention of cisplatin-induced chronic kidney disease (CKD) and the treatment of cisplatin-induced AKI, respectively. The attenuation of cisplatin-induced CKD by renalase agonist-loaded nanoparticles was associated with reduced levels of inflammatory cytokines in plasma, inhibition of the regulated necrosis of kidney cells, preservation of tubular epithelial cells and vascular cells, and suppression of inflammatory macrophages and myofibroblasts¹⁷³. Use of mesoscale nanoparticles substantially increased the delivery of edaravone to the kidneys of mice with cisplatin-induced AKI and significantly improved its therapeutic efficacy compared with that of free edaravone¹⁷⁴.

Targeting the interstitium

The interstitium was long considered to be a passive tissue that provided structural support to tubular epithelial cells; however, we now know that interstitial components have essential roles in the endocrine functions of the normal kidneys and in the progression of kidney diseases, such as acute interstitial nephritis and renal fibrosis^175,176. Interstitial injury is characterized by the infiltration of inflammatory cells, interstitial expansion and the development of fibrosis^177,178. Most nanoparticles cannot extravasate the peritubular capillaries of healthy kidneys to reach the interstitium due to the small size of the peritubular capillary endothelial fenestrae (~5 nm). However, nanoparticles that are actively eliminated by the kidneys through tubular secretion can still cross the peritubular capillary wall to target the interstitial space (Fig. 6). For example, mesoscale PLGA–PEG nanoparticles have been used to deliver the antifibrotic agent, emodin, to the interstitial space of mice with UUO-induced renal fibrosis¹⁷⁹ (Fig. 7i). Moreover, the fenestrae of peritubular capillaries become enlarged following injury, enabling the transport of larger nanoparticles into the interstitial space (Fig. 6). For example, anti-collagen 1 antibody-conjugated pegylated AuNPs¹⁸⁰ (62.5 ± 6.8 nm) have been synthesized to target collagen fibres for CT imaging of fibrosis in a model of renal artery stenosis (Fig. 7j). Moreover, active targeting pentapeptide Cys–Arg–Glu–Lys–Ala (CREKA)-conjugated liposomes of ~110 nm¹⁸¹ (Fig. 7k) and phage-display fibrotic kidney-homing peptide-conjugated liposomes of ~70 nm¹⁸² (Fig. 7l) loaded with celastrol and sorafenib, respectively, have been reported to specifically target interstitial myofibroblasts following their delivery via injured peritubular capillaries in models of renal fibrosis. The accumulation of CREKA–liposomes in the kidney was about two times higher and their in vivo binding to myofibroblasts was about five times greater than that of non-conjugated liposomes. Consequently, celastrol-loaded CREKA–liposomes decreased renal fibrosis, injury and inflammation to a greater extent than celastrol-loaded non-conjugated liposomes and free celastrol in a mouse model of UUO-induced fibrosis¹⁸¹. The accumulation of phage-display fibrotic kidney-homing peptide-conjugated liposomes was comparable to that of non-conjugated liposomes in the contralateral normal kidney but two to three times higher than that of non-conjugated liposomes in the fibrotic kidney. Compared with free sorafenib and sorafenib-loaded non-conjugated liposomes, sorafenib-loaded peptide-conjugated liposomes significantly reduced the infiltration of α-smooth muscle actin-expressing myofibroblasts and the deposition of collagen I and enhanced renal plasma flow in a mouse model of UUO-induced renal fibrosis¹⁸².

Re-elimination of endocytosed nanomedicines

Efforts to develop nanoparticles for the treatment of kidney diseases will inevitably encounter the long-standing challenge of nanotoxicity. Although glomerular filtration is often considered to be a passive and rapid elimination pathway, filtered nanoparticles can be actively taken up by PTECs. As described above, the uptake efficiency of nanoparticles by PTECs depends on the physicochemical properties of the nanoparticles, such as size and surface chemistry. Moreover, nanoparticles may be designed to be taken up by the kidney cells for specific diseases; thus, the re-elimination of nanoparticles from kidney cells following completion of their medical tasks is critical to their future clinical translation. An organelle-mediated mechanism of extrusion⁴⁴ that enables PTECs to re-eliminate endocytosed nanoparticles, might provide a foundation for reducing the potential nephrotoxicity of nanomedicines.

Proximal tubular extrusion

The re-elimination of endocytosed nanoparticles is often mediated by the membrane fusion of intracellular vesicles with the cell membrane^183–186. However, in 2023 we described a previously unknown pathway — which we termed proximal tubular extrusion — that involves membrane extrusion with no need for membrane fusion, and is used by PTECs to selectively remove their intracellular contents⁴⁴ (Fig. 8). Specifically, we found that following glomerular filtration and their endocytic uptake by PTECs, ultrasmall renal-clearable AuNPs can undergo a process of ROS-dependent biotransformation to form large, flower-like gold nanoassemblies in endosomes and lysosomes. Biotransformation of nanoparticles has previously been shown to reduce their exocytosis from human breast cancer cells¹⁸⁷ and human fibroblasts¹⁸⁸. However, we found that both endocytosed AuNPs and their nanoassemblies can be eliminated from the luminal side of PTECs through a process of cell membrane extrusion. By spontaneously squeezing a ~5 μ m balloon-like extrusion through their dense microvilli, PTECs can jettison endosomes and/or lysosomes (with or without AuNPs) into the membrane extrusion along with other intracellular organelles, such as mitochondria and endoplasmic reticulum. Pinching of the extrusion off the luminal membrane into the lumen eliminates the endocytosed AuNPs as extruded vesicles. We also observed this process of proximal tubular extrusion in healthy proximal tubules without administration of AuNPs, suggesting that this proximal tubular extrusion pathway is not activated by the injected AuNPs but is an intrinsic process of PTECs to remove components of their intracellular contents without the need for cell division. That is, this process is an intrinsic self-renewal mechanism of mitotically quiescent PTECs to maintain their cellular homeostasis, and explains the observation that gold nanoclusters⁹⁰ and polysiloxane nanoparticles¹¹⁸ are efficiently eliminated from the kidneys within days to weeks despite being taken up PTECs. More importantly, this mechanism provides a foundation for the clinical translation of renal nanomedicines with minimal nephrotoxicity.

Fig. 8 |. Proximal tubular extrusion of endocytosed gold nanoparticles by proximal tubule epithelial cells.

Proximal tubule epithelial cells (PTECs) can re-eliminate endocytosed gold nanoparticles (AuNPs) into the tubular lumen by spontaneously squeezing ~5 μm balloon-like extrusions through their dense microvilli. The endocytosed ultrasmall AuNPs undergo a process of biotransformation to form large, flower-like gold nanoassemblies in endosomes and lysosomes. Endosomes and/or lysosomes (with or without AuNPs) are jettisoned into the membrane extrusion along with other intracellular organelles such as mitochondria and endoplasmic reticulum (ER); the extrusion is then pinched off the luminal membrane into the lumen as extruded vesicles. The proximal tubular extrusion process has also been observed in healthy proximal tubules that had not been treated with AuNPs, suggesting that this process is not activated by nanoparticles but is an intrinsic process of PTECs to remove intracellular components without cell division. We propose that tubular extrusion is an intrinsic self-renewal mechanism by which mitotically quiescent PTECs maintain cellular homeostasis.

Conclusions

Although the field of renal nanomedicines is still in its nascent stage compared with the field of cancer nanomedicines, its development is undoubtedly accelerating with increased awareness of the potential therapeutic importance of nanomedicines for the treatment of kidney disease. A deepened fundamental understanding of nanoparticle transport and interactions in the healthy and diseased kidney is critical to enhancing the selectivity of renal nanomedicines to diseased kidney tissues and minimizing potential toxicities to healthy kidney tissues. The unique physiological principles that regulate nanoparticle transport and interactions in the kidneys highlight required differences in design strategies between renal and cancer nanomedicines; however, insights obtained from each field can benefit the other. The ability to develop nanoparticles that are renally clearable has been recognized as an important strategy to expedite the clinical translation of cancer nanomedicines^78–80, whereas progress in the development of strategies for cancer nanomedicines — including advances in the delivery of genetic modifiers and immunological drugs using nanoparticles^189–191 — may contribute to the development of renal nanomedicines for the treatment of genetic and autoimmune kidney diseases. Although substantial progress has been made in our understanding of the physiological principles involved in the intra-kidney targeting and elimination of renal nanomedicines, further research is needed to unravel exactly how nanoparticle transport and interactions are changed by disease to aid the translation of renal nanomedicines and safely improve outcomes in patients with kidney disease.

Key points.

• Despite considerable advances in cancer nanomedicines, renal nanomedicines for the treatment of kidney diseases are markedly underdeveloped.

• The physiological principles that regulate the glomerular filtration, tubular secretion, luminal tubular uptake and re-elimination of nanoparticles in the kidneys may facilitate the selective targeting of nanoparticles to specific segments of the nephron.

• Targeting of nanoparticles to different cell types in the glomerulus or to the glomerular basement membrane can be achieved through fine-tuning of their physicochemical properties based on our understanding of glomerular filtration and the glomerular filtration barrier.

• Different transport pathways can be used to deliver nanoparticles to different components of the renal tubules, including the luminal and the basolateral sides of tubular epithelial cells and the interstitium.

• Differences in nanoparticle transport and interactions between healthy and diseased kidney tissues offer opportunities for the design of nanoparticles that can selectively target specific kidney diseases.

• A newly discovered organelle extrusion mechanism facilitates the elimination of endocytosed nanoparticles from the kidneys and may minimize the nephrotoxicity of future nanomedicines.