Targeting nanoparticles to malignant tissues for improved diagnosis and therapy is a popular concept. However, after surveying the literature from the past 10 years, only 0.7% (median) of the administered nanoparticle dose is found to be delivered to a solid tumour. This has negative consequences on the translation of nanotechnology for human use with respect to manufacturing, cost, toxicity, and imaging and therapeutic efficacy. In this article, we conduct a multivariate analysis on the compiled data to reveal the contributions of nanoparticle physicochemical parameters, tumour models and cancer types on the low delivery efficiency. We explore the potential causes of the poor delivery efficiency from the perspectives of tumour biology (intercellular versus transcellular transport, enhanced permeability and retention effect, and physicochemical-dependent nanoparticle transport through the tumour stroma) as well as competing organs (mononuclear phagocytic and renal systems) and present a 30-year research strategy to overcome this fundamental limitation. Solving the nanoparticle delivery problem will accelerate the clinical translation of nanomedicine.
The possibility of engineering nanoparticles that selectively detect and destroy cancer cells in the body remains an exciting concept1,
In this article, we explore the influence of delivery on the efficacy of nanoparticles for cancer-targeting applications. Delivery is important because systemically administered nanoparticle carriers cannot function as designed if they do not access the diseased cells and tissues at a sufficiently high dosage19,
Delivery efficiency and consequences
How many nanoparticles accumulate in a tumour? Upon systemic administration, the mononuclear phagocytic system (MPS) and the renal clearance pathway compete with the tumour for nanoparticles. The MPS is a network of organs (most notably the liver and spleen) that contains phagocytic cells that take up nanoparticles, while the renal (kidney) system excretes nanoparticles smaller than 5.5 nm in hydrodynamic diameter26,
To determine the current delivery efficiency to solid tumours, we used SciFinder and Google Scholar databases and the search term ‘nanoparticle delivery’, and identified 224 manuscripts (Fig. 1a,b). We started by analysing the most recent publications and did not observe any significant variance in the median of the delivery efficiency over the past 10 years. Thus, we conclude that these data represent the current stage of development. From this compilation of studies, the data from 117 reports were tabulated (Supplementary information S1 (table)) and standardized to calculate the delivery efficiency based on a non-compartmental linear trapezoidal analysis model31. This method is commonly used in pharmacokinetic studies to determine area-under-the-curve (AUC) values, and it enabled the quantification of the AUC for a given nanoparticle–tumour interaction, providing temporal and spatial information with a minimum of three time points (Fig. 1c). The remaining 107 papers were excluded from our analysis because of insufficient time points (Fig. 1b; Supplementary information S2 (table)). More time points might have allowed a more precise calculation of the AUC, but very few researchers presented data with more than three time points (Fig. 1b). In many cases, authors were contacted to provide additional information to accurately calculate the AUC. Our analysis revealed that a median of 0.7% of the injected dose (ID) of the nanoparticles reached the tumour (this value was derived from 232 data sets) and that the median delivery efficiency has not improved in the past 10 years (Fig. 1d). This suggests that only 7 out of 1,000 administered nanoparticles actually entered a solid tumour in a mouse model. We used the median rather than the arithmetic mean to describe the delivery efficiency because the variability of data sets was high and the median is less affected by extreme values. We also provide alternative ways to interpret the delivery efficiency (for example, %ID per gram of tissue) in the Supplementary information.
Several trends may be observed in the data. First, nanoparticles composed of inorganic materials tend to provide a higher delivery efficiency than those made from organic materials (0.8% and 0.6%ID, respectively). Second, particles with hydrodynamic diameters smaller than 100 nm tend to show a higher delivery efficiency than larger particles (0.7% and 0.6%ID, respectively). Third, the delivery efficiency of nanoparticles exhibiting neutral zeta potentials, as defined from −10 to +10 mV, tends to be higher than the delivery efficiency of nanoparticles with positive (>10 mV) or negative (<−10 mV) zeta potentials (0.7%, 0.6% and 0.5%ID, respectively). Fourth, active-targeting strategies tend to outperform passive-targeting strategies, affording delivery efficiencies of 0.9% and 0.6%ID, respectively. Fifth, rod-shaped nanoparticles tend to exhibit a higher delivery efficiency compared with spherical, plate or flake, and other shapes, with values of 1.1%, 0.7%, 0.6% and 0.9%ID, respectively. Last, orthotopic tumour models tend to display higher delivery efficiencies than heterotopic tumour models (Fig. 1e–m; Table 1).
In 4 out of the 117 publications, delivery efficiencies higher than 5%ID were reported; the only physicochemical commonalities of these studies were the use of nanoparticles with hydrodynamic diameters less than 100 nm and near-neutral zeta potentials (Supplementary information S1 (table)). However, it is difficult to draw more specific conclusions from the analysis of these four papers32,
A multivariate analysis was further conducted to determine the contribution of a single parameter or combination of biological, physical and chemical parameters to the 0.7%ID. This analysis was undertaken to tease apart the effect of multiple interacting factors on the delivery efficiency. The data showed that the cancer type and the tumour model were major factors; however, they influenced the delivery efficiency of organic and inorganic nanoparticles in different manners as shown in Table 2 and Supplementary information S3–S5 (multivariate analysis) (a lower P value suggests a higher probability for the contribution of a specific parameter on the delivery efficiency).
Consequences of low nanoparticle delivery efficiency. We used a simple dosimetry analysis to estimate the consequences of using nanoparticles with 0.7%ID delivery efficiency for treating patients with cancer. Although an accurate determination of the human dose on the basis of animal studies is inherently complex, our analysis provides a simple, quantitative and general approximation that can be refined in future analyses.
We are evaluating two drug-loading strategies for nanoparticles: drug encapsulation in the nanoparticle core and drug loading on the nanoparticle surface. Provided that the efficiency of drug loading is the same for both strategies, the amount of drug is proportional to the cube and square of the nanoparticle radius, for each respective strategy (Fig. 2; Supplementary information S6,S7 (table, estimation)). For our estimation, we used a nanoparticle (60 nm in diameter) carrying anticancer drugs with a molecular weight of 500 g mol−1, a half maximal inhibitory concentration (IC50) value of 1.0 μM and a delivery efficiency of 1.0%ID as an example. To achieve the IC50 in a tumour (volume of 0.5 cm3) of a mouse (body weight of 20 g), it would be necessary to inject 1.2 × 1012 nanoparticles or a dose of 6.5 mg kg−1 provided that the nanoparticles encapsulated 20 wt% of the drug, which is a typical nanoparticle drug-loading capacity. If drugs were loaded on the particle surface with a density of one drug molecule per nanometer squared, the required dose would increase to 2.8 × 1012 nanoparticles or 15.7 mg kg−1. Such doses should be feasible in mice, because a typical injection dose for a mouse study is approximately 10 mg kg−1. However, if the dose is scaled from a mouse to a human, it becomes more difficult to achieve the required quantity; that is, one would need 2.7 × 1014 drug-encapsulated nanoparticles (19 mg m−2) or 6.4 × 1014 surface-loaded nanoparticles (45 mg m−2) for one person using a body-surface-area-based dosing strategy36. This would result in a systemic administration of 90 ml (drug encapsulation strategy) or 213 ml (drug surface loading strategy) of colloidal nanoparticle dispersions (assuming a nanoparticle concentration of 5.0 nM) into the patient to elicit a therapeutic effect similar to that in a mouse model (Supplementary information S7 (estimation)). Calculations for other nanoparticle sizes are presented in Supplementary information S8 (figure).
Our rough estimations using a delivery efficiency of 1.0%ID suggest that there are serious problems for the translation of cancer targeting nanoparticles for human use. First, the required amount of nanoparticles that need to be injected into humans would be high and would require the ability to scale up nanoparticle production, which is not a trivial task. Colloidally stable nanoparticles are commonly synthesized on a small scale for use in academic settings and for small animal (for example, mouse) models. This scale of synthesis is easily achieved for most nanoparticle types. When synthesis is scaled to prepare a large amount of nanoparticles in one batch, the nanoparticle quality and function may be compromised (for example, through the formation of aggregates, a polydisperse size distribution or shape irregularities). Second, the cost of nanoparticles for conventional use in patients may be prohibitively high because of the large number of nanoparticles required per injection dose. Third, the relatively high injection volume provides a technical limitation for using nanoparticles in patients; further concentration of nanoparticles to reduce the dispersion volume can compromise their colloidal stability and cause irreversible aggregation and/or shorten the shelf life. Fourth, because approximately 99%ID of all administered nanoparticles will interact with off-target tissues, high nanoparticle doses increase the risk of toxicity. Last, for engineering nanoparticles for cancer cell and subcellular targeting, the dosage requirement and cost may be even higher. It is possible that the amount of nanoparticles reaching cancer cells and their subcellular compartments in vivo is much less than 0.7%ID because nanoparticles need to cross the tumour extracellular matrix to reach the cancer cells. These nanoparticles may interact nonspecifically with non-malignant cells and other structural tissue components. A central strategy for addressing all of these issues is by increasing the delivery efficiency. For example, if delivery efficiencies improve from 1% to 10%ID, the injection volume for nanoparticles that follow a drug encapsulation strategy would decrease from 90 to 9 ml.
Elucidating the specific mechanisms involved in nanoparticle–tumour interactions will help to develop more effective strategies to increase the delivery efficiency beyond 0.7%ID. The delivery of nanoparticles to tumours is based on specific and nonspecific cellular interactions. Specific (or active) targeting relies on functionalizing the surface of nanoparticles with ligands that are complementary to the target sites, which may be peritumoural and intratumoural blood vessels, the extracellular matrix, tumour cells or intracellular targets. In nonspecific (or passive) targeting, the nanoparticle surface is coated only with anti-fouling and/or stabilizing agents. The current view is that nanoparticles cross the tumour vascular barrier through intercellular gaps and are retained in the tumour owing to pressure created by poor lymphatic drainage — a process termed ‘enhanced permeability and retention’ (EPR)37. Nanoparticles could be further retained in the tumour by active targeting32,38. In this article, we focus on intratumoural nanoparticle targeting.
Once nanoparticles reside in tumour blood vessels, they can extravasate into the tumour microenvironment. This section describes where and how nanoparticles extravasate from the tumour vasculature, and compares and contrasts the intercellular versus transendothelial transport mechanisms.
Tumour blood vessels. A basic understanding of tumour blood vessels is important to explain nanoparticle–tumour interactions and transportation. When nanoparticles are injected into tumour-bearing animals, they rapidly pass from the systemic circulation to chaotic vessels that supply the tumour with oxygen and nutrients. The tumour vasculature is highly abnormal, exhibiting an uneven distribution with zones of both increased and sparse vascular density, hierarchical disorganization, serpentine structure and irregular branching, and includes vascular malformations that form arteriovenous shunts39,40. Tumour vascular density is generally highest at the tumour/host interface; by contrast, central portions of tumours tend to be less well vascularized and often exhibit zones of necrosis owing to insufficient blood supply41,42. There are at least six distinctly different types of tumour blood vessels: feeding arteries, mother vessels, glomeruloid microvascular proliferations (GMP), vascular malformations, capillaries and draining veins39,43,
Nanoparticles are likely to cross the vascular barrier and enter the tumour compartment through mother vessels. These vessels are derived from normal venules and capillaries and are the first angiogenic type of blood vessel to form. They exhibit abnormal hyperpermeability to plasma proteins because they are lined by a single, thin layer of flattened endothelial cells, have disrupted basement membranes and have little or no pericyte coverage47,
Intercellular extravasation. The predominant belief in the nanotechnology community is that nanoparticles extravasate from the tumour blood vessels into the tumour microenvironment through gaps between adjacent endothelial cells. This is one of the key fundamental principles for the EPR effect37,54,
Transcellular extravasation. An alternative hypothesis is that nanoparticles can extravasate into tumours via a transendothelial cell pathway. Normal venule and capillary endothelial cells contain a structure termed vesiculo–vacuolar organelles (VVOs), which span the cytoplasm from the lumen to the ablumen (Fig. 3b). VVOs are composed of an extensive plexus of interconnected cytoplasmic vesicles and vacuoles, the membranes of which collectively constitute a surface area several times that of the venular plasma membrane59. They provide a pathway for the extravasation of plasma during acute vascular hyperpermeability induced by vascular endothelial growth factors (VEGFs) and other vascular permeabilizing agents60. VVOs contribute to mother vessel formation by providing the necessary membrane to the cell surface, which allows a four- to fivefold vessel enlargement. As a result, mother vessels contain fewer VVO vesicles than the normal venular endothelium from which they originate. Despite the reduced number of vesicles available for transendothelial transport, the distance VVOs have to traverse across thinned mother vessel endothelium is also reduced41 (Fig. 3b). Both horseradish peroxidase and ferritin have been shown to cross solid tumour blood vessels via the VVO pathway59,60.
Which extravasation mechanism is more important? It is unclear which pathways dominate the extravasation of macromolecules and nanoparticles in tumours. Macromolecules can extravasate from mother vessels and GMP by way of intercellular gaps; however, these gaps are exceedingly rare, as was determined in mother vessels induced by an adenoviral vector expressing murine VEGFA164. In addition, some or many of the reported interendothelial cell gaps may actually represent open transcellular VVOs. VVOs are located close to and interact with lateral endothelial cell junctions and could thus give the appearance of interendothelial gaps. It was shown using serial section electron microscopy that many of these openings are in fact transendothelial cell pores57. Distinguishing between intercellular gaps and transendothelial cell pores is difficult, and intercellular gaps can only be demonstrated definitively by labelling the margins with selective junctional markers such as vascular endothelial cadherin; to our knowledge, this has never been attempted. However, the mechanisms and pathways that mediate nanoparticle transport to the tumour are important. If the extravasation is mediated primarily by the transcellular route, nanoparticles that will actively target this transport pathway can be designed. At present, the nanotechnology community has not investigated this transport mechanism. Instead, a heavy emphasis has been placed on studying nanoparticle transport through intercellular gaps via the EPR mechanism — an approach that has thus far yielded poor delivery efficiency. Therefore, there is a need to probe the tumour endothelium transport mechanism to guide the future design of nanoparticles.
Intratumoural nanoparticle interactions
Once the nanoparticles have crossed the vascular barrier, they have to navigate through the tumour microenvironment to reach and target cells or specific intratumoural structures. This section provides a description of these tumour biological barriers and explains the current view of how the physicochemical properties of nanoparticles dictate the transport process in the tumour microenvironment.
Components of the tumour microenvironment. Nanoparticles interact with heterogeneous components of the tumour stroma (that is, non-malignant cells such as fibroblasts, pericytes and immune cells) and parenchyma (that is, malignant cells) after extravasation from the tumour vasculature. In solid tumours, the structural rigidity is supported by extracellular matrix components such as collagen, fibronectin, hyaluronan, fibrin and proteoglycans synthesized by fibroblasts61. Tumours are highly heterogeneous because the composition of parenchyma and stroma cells, and the type and ratio of the support and secreted proteins vary depending on tumour type and stage62,63. The interstitial fluid pressure inside the tumour can be elevated by a factor of 10–40 compared with normal tissues64,65, creating pressure gradients and heterogeneous flow in the interstitium. This pressure can influence the transport and distribution of chemotherapeutics, imaging agents, macromolecules and nanoparticles in the tumour24. Furthermore, the high tumour interstitial fluid pressure increases the outward interstitial flow to the stroma and lymphatics with increased fluid transport to the nearest ‘sentinel’ lymph node; coupled with lymphangiogenesis, which is found in some but by no means all solid tumours, these characteristics strongly correlate with tissue invasiveness and metastasis65.
Intratumoural targeting. After extravasation, the physicochemical-dependent transport and penetration depth of nanoparticles in the tumour matrix are likely to be influenced by the tumour type, size and stage because of the specific tumour microenvironment properties (as described in the previous section)66,67. The general consensus is that smaller nanoparticles penetrate more deeply into the tumour extracellular matrix away from the blood vessel, whereas their larger counterparts are restricted to the immediate vicinity of the vascular extravasation point66,68,69. The physicochemical properties may determine whether nanoparticles coated with tumour-cell-targeting aptamers, antibodies, peptides and/or small molecules are able to interact with malignant cells. A head-to-head comparison of active and passive targeting reveals a lack of significant difference in tumour accumulation32,70,71. To fully elucidate the nanoparticle fate in the tumour tissue, we have to move beyond 2D images72 and employ 3D imaging and quantify nanoparticle–cell interactions in vivo to describe, confirm and evaluate specific intratumoural nanoparticle targeting.
Organs competing with tumours
The MPS and renal system compete with the tumour for circulating nanoparticles26,
Mononuclear phagocytic system. Nanoparticles are identified by the MPS as foreign substances that need to be sequestered, degraded and eliminated. The MPS consists mainly of the liver, spleen, lymph nodes, bone marrow, skin and other organs containing resident phagocytic cells such as macrophages. Resident macrophages in the MPS organs are derived from circulating monocytes and can be phenotypically diverse and heterogeneous in an organ73. Most administered nanoparticles are sequestered by the liver and the spleen, and thus the discussion will focus on these two organs.
Macrophages in the liver include Kupffer cells and motile macrophages, and reside in liver sinusoids along with fenestrated endothelial cells (Fig. 4a). While nanoparticles flow through the liver sinusoidal capillaries, it has been suggested that scavenger receptors on Kupffer cells recognize the opsonins adsorbed on the nanoparticle surface and engulf them. The way in which the cells process the nanoparticles varies based on the nanoparticle type and chemistry; for example, large inorganic nanoparticles can reside in these cells for long periods of time, whereas organic nanoparticles are more rapidly degraded and eliminated74.
The specificity and role of spleen macrophages in terms of nanoparticle uptake has not been fully investigated. The spleen is compartmentalized into the red and white pulp that are separated by a marginal zone, and sinuses with pores of approximately 3 μm in diameter75. The red pulp contains macrophages that are involved in erythrocyte degradation73, whereas the marginal zone macrophages, B cells and dendritic cells provide defence against pathogens. Metallophilic macrophages and tingible body macrophages reside in the white pulp and are responsible for innate immunity and clearing of apoptotic cells, respectively76. Macrophages in either the red or white pulp can potentially take up circulating nanoparticles, but the specifics remain unclear.
Macrophage uptake mechanisms. The predominant mechanism of nanoparticle sequestration by macrophages is phagocytosis, although nanoparticles can also be taken up by macropinocytosis, clathrin-mediated and caveolin-mediated endocytosis, and by other endocytic pathways77. It is well known that low-density lipoproteins and bacteria are taken up through scavenger receptors, and this mechanism has been adapted to explain how nanoparticles are sequestered. In some studies, nanoparticles were shown to be taken up through scavenger receptors78,79; however, more receptors may be involved, because engineered nanoparticles are more chemically diverse than pathogens and low-density lipoproteins. Thus, the mechanism of nanoparticle interaction with MPS organs and macrophages is likely to be more complex.
When nanoparticles are administered into the body, their physicochemical properties change. The nanoparticles interact with serum proteins to form a ‘protein corona’ and may aggregate. The corona is diverse and consists of ions, opsonins and other serum proteins80,81. This process has been described as a change in the molecular identity of the nanoparticle, from a ‘synthetic identity’ that describes the measured size, shape and surface chemistry post-synthesis to a ‘biological identity’ that describes the physicochemical properties after interaction with the biological fluid82. It is the biological identity that determines the interaction with the cells. Analysis of the protein corona of over 70 formulations of gold nanoparticles revealed unique serum protein coatings for each formulation, and thus it is unclear how a single protein or a combination of these proteins mediate their interactions with macrophages83,84.
Researchers have observed interesting trends in nanoparticle–macrophage interactions. Larger nanoparticles are cleared more rapidly than smaller nanoparticles, and particles with cationic surfaces show the strongest macrophage uptake, followed by those with anionic surface charge and then by those with zero net surface charge. For example, poly(ethylene glycol) (PEG)-coated gold nanoparticles that are 90 nm in diameter have a 4-fold higher uptake by J774A.1 macrophages than their 15- and 30-nm counterparts with similar surface charge83. Comparable trends were observed in vivo for gold nanoparticles between 10 and 250 nm: the 10-nm gold nanoparticles showed an increased distribution throughout the body, whereas the 250-nm gold nanoparticles were retained to a greater extent in MPS organs85. To escape macrophage uptake and extend the circulation lifetime of nanoparticles in vivo, researchers typically coat the nanoparticles with neutral polymers, such as PEG or erythrocyte membranes, to camouflage the particles from these cells. A longer circulation lifetime increases the chance for nanoparticles to accumulate in the tumour. There is a direct correlation between the length and density of PEG ligands and the blood half-life of nanoparticles66. However, these ‘stealth’ nanoparticle formulations have not fully solved the sequestration and targeting problem. Thus, it is important to obtain a detailed understanding of how the MPS organs and specific macrophage cells interact with nanoparticles.
Renal clearance. The kidney glomeruli have three layers: a fenestrated endothelium, a glomerular basement membrane and a lining epithelium86 (Fig. 4b). The charge- and size-selective permeation are the functions of the latter two layers, respectively87. The basement membrane is a structural gel composed of a heterogeneous network of type IV collagen, fibronectin and heparan sulfate proteoglycan, among other components, and has a net polyanionic charge86. The epithelium lining contains filtration slits of 4–6 nm in width87 and allows the passage of nanoparticles smaller than this size cut-off. Nanoparticles filtered out of the blood in glomeruli pass into the proximal convoluted tubule and are eventually eliminated in urine. Nanoparticles filtered from glomeruli can also be endocytosed by cells that line the proximal tubules before they reach the bladder88.
The single key word that will continue to stimulate nanomedicine research in the next 30 years is ‘translation’. Nanoparticles are designed to alter the biodistribution and pharmacokinetics of a small molecule drug or contrast agent in patients and enable the delivery of a larger dose to the targeted diseased tissue in an effort to improve the therapeutic index, reduce systemic toxicity and/or offer better imaging signals. These advantages have been dampened by the lack of translation to patient care, despite the large investment (more than US$1 billion in North America in the past 10 years) and success in imaging and treating tumours in mouse models. As a result, nanomedicine has acquired a reputation of being ‘hype’ that cannot deliver and has not transformed patient cancer care as it promised 15 years ago.
Nanoparticles can be approved for human use by health agencies if they offer an improvement in diagnosis, therapeutic index and/or a reduction of toxicity. At present, only a few non-targeted nanoparticle formulations (for example, Abraxane and Doxil) have been clinically approved30,89,
Researchers are starting to develop non-conventional strategies to improve the delivery efficiency, including: normalization of tumour vasculature94,
We need to consider a more rational and longer-term strategic approach to overcome the nanoparticle delivery problem. The community must take a step back and evaluate the principles that have guided the field and assess whether targeting is possible. Many researchers regard a delivery efficiency of 0.7%ID to represent nonspecific interaction rather than specific targeting. The current underlying principles of nanoparticle targeting have not produced the desired clinical outcomes, and if these principles remain unquestioned, the next 30 years will only provide more unsatisfactory results. If nanoparticles cannot accumulate in solid tumours and reach target tumour cells, how can they effectively work as designed?
We suggest a 30-year systematic and coordinated strategy to enable the community to make decisions that move the field forward (Fig. 5). The key question that needs to be addressed is: can nanoparticles target diseased tissues? This can only be answered by conducting careful and systematic studies. The next 10 years should focus on probing the fundamental interactions of nanoparticles with organs and tissues that accumulate, sequester or eliminate nanoparticles (such as the liver, spleen and kidney), as well as the interactions between nanoparticles and tumours, with respect to the physicochemical properties of the nanoparticles. These studies must challenge the current paradigm of why the MPS sequesters nanoparticles, whether the nanoparticles are taken into tumours solely by the EPR effect, whether nanoparticles can be transported effectively into and through tumours, and whether the pressure inside tumours can be measured and altered to improve nanoparticle transport. These are merely starting points, and there are many other questions that need to be answered. These studies should drill down to the details, from the synthetic and biological identities of the nanoparticles to the biological systems from host, organ, tissue, cell and molecular perspectives. There is a need to develop techniques and tools that allow quantitative analysis of these fundamental questions. These studies need to be multi-parametric (multiple particle types with multiple cell and tissue systems) and may require computational tools to deconvolve the data and present models of interactions. Computational tools and mathematical modelling can be used to analyse tumour interactions and response to drugs and nanoparticles in both preclinical and clinical stages117,
The second 10 years could focus on using these fundamental studies to design optimal nanoparticle systems for targeting tumours and to test these formulations in various animal models. By analysing diverse animal models in the same study, the delivery efficiency can be determined and the mechanisms involved can be identified. Achieving the desired research goals will require multidisciplinary teams with a detailed plan of action. Again, the focus should be on measuring the delivery efficiency and therapeutic indexes using designs that were optimized in the first 10 years. If the delivery efficiency cannot be increased, the community will need to change their way of thinking about nanomedicine. Because we know that nanoparticles are heavily sequestered by macrophages, perhaps a better scenario may be to combine nanoparticles with macrophages for use in immunotherapy or to assess two- or multistage strategies (for example, first removing macrophages followed by the injection of therapeutic nanoparticles)120.
The final 10-year period will need to focus on clinical trials and making use of the optimal nanoparticle formulations. The results need to be fed back to the foundational studies and database so that the data can be mined to assess the fundamental interactions of nanoparticles with biological systems from the mouse model to the human level. An example of an early cross-animal analysis was recently conducted for the assessment of CALAA-01, a nanoparticle formulation for small intefering RNA delivery121. Despite the fact that our proposed 30-year approach is slower and more controlled than might be hoped for in advancing nanomedicine, it has the potential to generate a foundation that will serve the field and facilitate nanomedicine translation in the latter part of this century. Furthermore, this approach will allow researchers and clinicians to truly assess the potential of nanotechnology for cancer applications and enable the community to make rational research decisions.
We must admit that our current approach is broken, and that is why we have not observed significant clinical translation of cancer nanomedicines. Many academic studies focused on the potential of nanoparticles for in vivo medical applications and showed that nanoparticles may be delivered into tumours by the EPR effect. However, publishing ‘proof-of-concept’ studies will only lead to curing mice and will unlikely translate to patient care, irrespective of the number of nanoparticle design permutations used for cancer-targeting studies. Despite the lack of translational progress, nanotechnology remains an exciting topic of research and does have promise for improving patient care. The main advantages of using nanoparticles are that they can be engineered with precise functional properties and can access biological systems and compartments. However, our inability to control the nanoparticle transport inside the body presents a major limitation for using nanotechnology to diagnose and treat cancer, as well as diabetes, cardiovascular disorders and other diseases for which nanotechnology is actively researched.
Cancer Nanomedicine Repository (CNR): http://inbs.med.utoronto.ca/cnr/ (June 2016)
The authors thank all authors of the studies that were surveyed for providing additional information on request to calculate the nanoparticle delivery efficiencies (Supplementary information S1 (table)).The authors thank M. Samarakoon and A. Göpferich for fruitful discussions. The authors also thank S. Patel, M. Chalsev, A. Mahmood, J. D. Mora, and Y. Y. Chen for creating the database system. W.C.W.C. acknowledges the Canadian Institutes of Health Research (CIHR), Natural Sciences and Engineering Research Council (NSERC), and Prostate Cancer Canada for supporting his research program. H.F.D. acknowledges financial support from the US National Institutes of Health (NIH) grant P01 CA92644 and by a contract from the National Foundation for Cancer Research. S.O. acknowledges the Japan Society for the Promotion of Science (JSPS) for a Research Fellowship (PD, No. 5621) and the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan for a Grant-in-Aid for Young Scientists (B) (No. 26820356).