Nanoparticle Delivery to Tumours: From EPR and ATR Mechanisms to Clinical Impact

Standfirst

New insights into active versus passive nanoparticle tumour entry and exit mechanisms are enriching the understanding of tumour-targeted drug delivery. Here, we align the principles of the enhanced permeability and retention (EPR) and active transport and retention (ATR), and outline how their mechanistic features may be employed to improve the performance and clinical impact of cancer nanomedicines.

Since its inception, in 1986, the enhanced permeability and retention (EPR) effect has served as the central dogma for nanoparticle delivery to tumours¹. The EPR principle states that nanoparticles enter tumours through inter-endothelial gaps and are retained there owing to dysfunctional lymphatics. However, recent reports have questioned the validity of EPR-based passive targeting, demonstrating that active (that is, cell-dependent) mechanisms are also involved in nanoparticle transport in and out of tumours^2–5. Using ‘zombie’ mice, it was shown that active transcytosis through tumour endothelial cells plays an important role in nanoparticle entry, at least in case of gold nanoparticles²; here, a subset of tumour endothelial cells, termed ‘nanoparticle-transport endothelial cells’ (N-TECs), overexpress transport pathway-related genes and are responsible for gold nanoparticle entry³. In addition, it was found that tumor-associated macrophages (TAMs) migrate towards extravasated gold nanoparticles, internalize them and transport them away from the vasculature into the tumour interstitium⁴. Finally, it was shown that lymphatic drainage in tumours is not dysfunctional, and that small and large nanoparticles exit tumours via peri- and intratumoral lymphatics, respectively⁵.

These mechanisms have been summarized as the active transport and retention (ATR) principle, stating that nanoparticles enter tumours predominantly through active transport processes, that is, by endothelial transcytosis, vesiculo-vacuolar organelles and/or immune cell migration, rather than through inter-endothelial gaps; and that nanoparticles are retained in tumours owing to interactions with TAMs, fibroblasts, cancer cells and/or extracellular matrix components, and not owing to dysfunctional lymphatics⁶ (Fig. 1a). Collectively, the ATR principle suggests that nanomedicine tumour targeting may be more active than passive, thereby replacing the EPR effect⁶.

Figure 1. Mechanisms and strategies in nanoparticle delivery to tumours.

(a) The enhanced permeability and retention (EPR) effect posits that nanoparticles enter tumours through inter-endothelial gaps, including via vascular bursts, and are unable to exit owing to impaired lymphatic drainage. The active transport and retention (ATR) principle posits that nanoparticles enter tumours primarily through transcytosis by nanoparticle-transport endothelial cells (N-TECs), vesiculo-vacuolar organelles and/or engagement with circulating immune cells, followed by nanoparticle trafficking and retention in tumours by tumour-associated macrophages (TAMs), and partial nanoparticle clearance from tumours through intra- and peritumoral lymphatics. (b) Nanoparticle tumour delivery may result from a mixture of both EPR and ATR mechanisms. In high-accumulating tumours, passive transvascular transport mechanisms may dominate, whereas in low-accumulating tumours, the relative absence of inter-endothelial gaps may favor active transport. Although not yet fully elucidated, soft nanoparticles appear to be mostly associated with passive transport mechanisms, whereas hard nanoparticles may be more inclined to exploit active endothelial transcytosis. (c) Nanoparticle delivery and translation may be improved by promoting transcytosis by N-TECs, for example, by active targeting using ligand-coated nanoparticles. (d) Regular endothelial cells could be transformed into N-TECs by gene or RNA therapy. (e) TAMs could be targeted and modulated towards enhanced anti-tumour efficacy. (f) Immune cells may be engaged for active migration to and into tumors. (g) Strategies for patient stratification should address the heterogeneity in delivery mechanisms and in the extent of nanoparticle tumour targeting, thereby improving the clinical translation and performance of cancer nanomedicines.

Delivery mechanisms depend on tumour and nanoparticle type

The ATR concept, and specifically active endothelial transcytosis as the mode of nanoparticle entry, has been a subject of debate in the drug delivery community. The dominant mode of nanoparticle entry was validated using zombie mice, where fixatives are employed to shut down transcellular transport; however, fixation may also affect the number and size of inter-endothelial gaps, and the morphology and density of the tumour extracellular matrix, thereby also potentially reducing passive entry. Furthermore, it needs to be taken into account that the dominant entry mechanism(s) strongly depend(s) on the tumour type. For example, by investigating the accumulation of ferritin nanoparticles in 32 mouse tumour models⁷, it was found that 70% (23/32) of tumour models had low and 30% (9/32) of tumour models had high ferritin nanoparticle accumulation. Importantly, only in former, tumour accumulation was found to be higher in normal mice as compared to zombie mice, indicating that active transvascular transport pathways dominate in tumours with low levels of target site accumulation. By contrast, in the 30% of tumour types with high levels of nanoparticle accumulation, ferritin nanoparticle concentrations were comparable in control and zombie mice, illustrating that in these tumours, transvascular transport mainly takes place through EPR-based passive mechanisms. Thus, passive and active delivery pathways are not mutually exclusive, but complement each other to different extents in different tumour types (Fig. 1b). It is important to note in this regard that in clinical settings, particularly tumours with high levels of (mostly passive) nanoparticle accumulation are of interest, as patients demonstrating efficient tumour targeting are more likely to respond to nanotherapy than patients with low levels of target site localisation.

The above studies provide key new insights towards a broad(er) mechanistic understanding of tumour targeting. U87-MG and 4T1 tumour models are used in both the active transcytosis paper² and the 32-model comparative analysis⁷. In the latter, they are identified as low-accumulation models, in which, indeed, active transvascular transport pathways dominate. This suggests that tumour model selection may have unintentionally been biased towards low-accumulation models in the active nanoparticle entry experiments. In addition, only relatively ‘hard’ gold and ferritin nanoparticles were used in these two studies. This could also lead to bias, with regard to interpretation and generalization. Systematic follow-up studies are needed to elucidate the balance between passive and active extravasation pathways for soft and clinically more-widely used nanocarriers, such as liposomes, micelles and antibody-drug conjugates, which may behave differently in the bloodstream and at the endothelial interface as compared to hard nanoparticles.

Alternative transvascular transport pathways should also be kept in mind. For example, macro- and micropinocytosis, which may play prominent roles in the transport of soft nanomaterials, have not yet been investigated in depth. In addition, ‘nanoparticle-induced endothelial leakiness’ (NanoEL) may be relevant, positing that certain types of hard nanoparticles bind to the inter-endothelial cell adhesion molecule vascular endothelial-cadherin, triggering actin remodeling and degradation, and thereby actively creating inter-endothelial gaps through which nanoparticles then passively leak into tumours⁸. Moreover, vascular bursts may contribute to nanoparticle accumulation. These bursts are characterized by dynamic temporary eruptions that open and close the endothelium⁹, presumably because of changes in tumour blood vessel perfusion and/or pressure, enabling nanoparticles to enter the tumour likely through passive transvascular transport, which is in line with the notion that they have only been reported for soft nanoparticles to date.

Exploiting mechanisms to improve cancer nanomedicine performance

The question remains how insights into passive and active nanoparticle transport mechanisms can inform the design of cancer nanomedicines with improved performance. To enhance nanoparticle delivery and efficacy, N-TECs could be exploited, if their role in transporting soft nanomedicines and their presence in human tumours can be confirmed. For instance, N-TECs could be targeted by ligand-modified nanomedicines, or normal endothelial cells could be transformed into N-TECs through gene transfection (Fig.1 c,d). Moreover, TAMs and myeloid cells can be targeted, in order to repolarize them towards an antitumour phenotype and to assist in actively transporting nanomedicines into tumours, respectively (Fig. 1e,f).

Reflecting on clinical translation and patient benefit, it is crucial to consider that both the mechanisms and extent of nanoparticle tumour targeting are heterogeneous. A pragmatic way forward towards creating clinical impact is via the use of biomarkers. Biomarkers are commonly employed for patient stratification in oncology; for example, to select the right patients for treatment with antibodies and kinase inhibitors, receptor (over)expression and driver mutations are assessed by histopathology and genetic analysis. By contrast, identifying suitable biomarkers for cancer nanomedicines is not straightforward. To visualize and quantify the extent of nanoparticle tumour targeting, non-invasive imaging could be employed, together with companion nanodiagnostics or nanotheranostics, to select patients accordingly. Moreover, histopathological biomarkers can be used for cancer nanomedicine patient stratification; for instance, by staining blood vessels and TAMs in tumour tissue sections or biopsies, the accumulation of nanoformulations in tumours can be predicted, to guide patient selection in clinical trials¹⁰.

If the presence and prominence of N-TECs can be confirmed, and if specific cell surface markers for N-TECs can be identified, then N-TEC-based histopathology may aid in refining the prediction of nanoparticle tumour targeting (Fig. 1g). To further increase predictive accuracy, this could be combined with histopathological assessment of TAM density. Such biomarker strategies are crucial to deal with heterogeneity in nanoparticle tumour targeting, and they are needed to help promote cancer nanomedicine translation and treatment.