Nanotechnology has opened new avenues for tackling unmet challenges in medicine . In the field of oncology, a notable application involves the use of engineered nanoparticles (NPs) designed to transport therapeutic agents with precise delivery to tumor sites. This approach aims to mitigate toxic effects associated with off-target drug delivery and optimize therapeutic efficacy.

For decades, the enhanced permeability and retention (EPR) effect has stood as the central mechanism for driving passive NP delivery to tumors . In this model, leaky blood vessels and a compromised lymphatic drainage system contribute to the preferential NP extravasation and accumulation within solid tumors. However, recent evidence challenges this paradigm, suggesting that NP extravasation into tumors primarily occurs via transendothelial transport pathways . Regardless of the mode of NP extravasation, active targeting strategies have been widely explored to further enhance NP accumulation in tumors and NP internalization by cancer cells . Active targeting involves the modification of NPs with targeting ligands (i.e., small molecules, peptides, or antibodies) that bind to overexpressed receptors within the tumor microenvironment.

Despite the promise of nanomedicine, neither passive nor active delivery strategies have significantly improved clinical therapeutic outcomes for solid tumors . Reasons for the poor clinical performance of passive tumor targeting are the considerable heterogeneity of the EPR effect in humans, alongside the restricted diffusion of NPs across the dense tumor stroma . Reasons for the limited performance of active targeting include its reliance on passive targeting, the more complex designs of targeted NPs, the potential for attached functional ligands to increase phagocytic capture and shorten blood circulation time, and the formation of a protein corona that may block the targeting ligand on the particle surface .

Important classes of inorganic usNPs under investigation for cancer nanomedicine include metallic usNPs (gold, silver), oxide and sulfide usNPs (silica, iron oxide, copper sulfide), and rare earth-based usNPs (cerium oxide, gadolinium oxide) . Ultrasmall NPs have dimensions comparable to those of a typical globular protein of 3 to 6 nm in diameter , although the precise size criteria can vary among researchers. For the purpose of this discussion, usNPs are defined as being small enough to undergo renal clearance. While this usually entails usNPs smaller than the kidney filtration barrier of 5–6 nm , slightly larger particles have also been found to undergo renal excretion in some cases.

Ultrasmall NPs are situated at the interface between small molecules and conventional NPs, and so they provide a unique opportunity to leverage distinctive properties inherent to both domains . On one hand, usNPs and their conjugates can behave as biomolecules in terms of biomolecular interactions and physiological behavior . Additionally, certain types of usNPs, especially gold nanoclusters (AuNCs), manifest molecule-like physical and chemical properties, such as luminescence . Simultaneously, usNPs – whether used alone or conjugated to drugs, diagnostic probes, and targeting ligands – can function as a more conventional NP platform in nanomedicine applications . In diagnostic applications, usNPs have been employed in diverse imaging modalities, including optical imaging , X-ray computer tomography , photoacoustic imaging , magnetic resonance imaging , and positron emission tomography . In therapeutic applications, usNPs have been used for drug delivery as well as served as phototherapeutic agents and radiosensitizers .

The incorporation of active targeting strategies is expected to further enhance the selectivity and performance of usNPs for cancer treatment. By designing usNPs to target surface receptors on cancer cells, tumor retention can be improved by minimizing particle intravasation back to tumor blood vessels. Active targeting can also promote usNP transport to the cell interior, potentially leading to more effective drug delivery and chemotherapy. It must be noted that the success of these strategies relies on efficient passive targeting in the first place . Nevertheless, cumulative evidence suggests that actively targeted usNPs can enhance tumor accumulation compared to non-targeted particles (Section 5). Furthermore, usNPs containing tumor homing and penetrating peptides can target the more accessible tumor vasculature, potentially aiding in particle accumulation within the tumor site .

A direct comparison of the impact of NP size on the tumor accumulation and retention of actively targeted particles was undertaken by Xu and colleagues . The authors synthesized transferrin-coated iron oxide NPs with core sizes of 3 and 30 nm and assessed their binding to transferrin receptors overexpressed in a 4T1 xenograft breast cancer model. Their findings revealed that actively targeted 3 nm NPs produced a sixfold higher level of tumor retention compared to non-targeted counterparts. In contrast, the corresponding improvement in tumor retention was only 1.15-fold in the case of the larger NPs. This difference was attributed to easier tumor clearance (tumor intravasation back to blood vessels) of off-targeted 3 nm NPs compared to 30 nm ones.

Common functional ligands employed in actively targeted usNPs encompass small molecules such as folate, aptamers, peptides, full antibodies, and antibody fragments. These ligands can be covalently attached to the underlying surface coat through standard bioconjugation chemistry or utilizing bioorthogonal bioconjugation strategies such as click-chemistry . Additionally, in certain situations, the targeting ligand can be directly conjugated to the NP inorganic core, exemplified by the S–Au bond formed between cysteine-containing molecules and gold NPs . Actively targeted AuNCs can also be prepared using bioactive peptides or proteins via a one-step biomineralization process, in which case the peptide or protein serves the purpose of both surface stabilization and functionalization . Importantly, the use of antibodies and other proteins as targeting agents may increase the HD of usNPs beyond the threshold for renal filtration, and so careful consideration is needed in the design of such constructs.

To ensure effective interaction with cell surface receptors on cancer cells, the incorporation of targeting ligands onto usNPs must optimize the exposure, orientation, and conformation of the functional portion. For small molecules and peptides in particular, the functional moiety must circumvent both steric hindrance from the underlying surface coat and undesired intermolecular interactions on the ligand shell. In this regard, computer simulations emerge as a powerful tool for optimizing the size and composition of usNPs designed for receptor targeting. For example, Häkkinen and colleagues designed a series of 1.7 nm AuNCs functionalized with RGD peptides as targeting ligands along with chemotherapy drugs and inhibitors of signaling pathways . Their simulations revealed that the system composition and the peptide/drug ratio critically influenced the targeting ability of the particles. In addition to computer simulations, a detailed experimental characterization of the surface properties and interactions of targeted usNPs is indispensable for elucidating their biological behavior and optimizing their performance.

Table 1: Receptor/ligand combinations employed in active targeting strategies involving usNPs for cancer diagnosis and treatment.

aApparent binding affinities estimated for the free ligand binding to corresponding receptor. The binding affinity could differ when the ligand is immobilized on an usNP. Some entries report direct KD measurements, while others report IC50 or Ki values determined from competition assays. bKapp et al. performed a comprehensive evaluation of the binding affinity (IC50 values) of different RGD peptide ligands to various integrin receptors . While short linear peptides demonstrated binding to the αvβ3 integrin with affinities ranging from 12 to 89 nM, short cyclic peptides displayed stronger affinities in the range of 1.5 to 6 nM. cBasilion and colleagues developed a peptide-based high-affinity ligand for PMSA, referred to as PSMA-1 . Moreover, a recent review highlights the latest developments in PSMA-targeted therapy for prostate cancer . dAffinity values for five distinct antibodies were reported by Münz et al. .

Targeted usNPs with weak binding to cancer cell surface receptors may not provide any additional value over non-targeted particles. As stressed by Ruoslahti and colleagues , many peptide ligands bind their receptors with weak affinities in the high-nanomolar to low-micromolar range. This implies that delivering a substantial excess of targeted usNPs locally would be needed for receptor saturation ([usNP] = 9 × KD for 90% saturation), but challenges with insufficient NP tumor penetration and diffusion make this unlikely. Even if delivering a high local concentration of targeted usNPs were possible, the contribution of active targeting would not be distinguishable from the nonspecific background in this case. To address the challenge of weak ligand–receptor affinity, one can opt for a more suitable high-affinity ligand, or design usNPs to leverage avidity effects.

Another layer of complexity arises from the in vivo system operating in an open, non-equilibrium state, where concentrations constantly change and biological processes are dynamically regulated . Consequently, it becomes important to extend the characterization beyond binding affinity and include the examination of binding kinetics between targeted usNPs and their receptors . For a simple one-step binding model, KD = koff/kon and tr = 1/koff, where kon and koff are the association and dissociation rate constants of the binding reaction, respectively, and tr is the residence time of the complex. The value of koff (or tr) is determined by short-range non-covalent interactions at the binding interface, reflecting the stability of the bound complex. For instance, with a KD of 1 nM and a characteristic kon of 1 × 106 M−1·s−1, the residence time would be approximately 17 min. In chemically related compounds, kon generally remains more or less invariant, and relative changes in KD follow corresponding changes in koff. Avidity effects also manifest through a reduction in koff while kon remains unaffected. Importantly, a prolonged residence time may prove beneficial when targeting the tumor vasculature, as the targeted usNPs would remain bound to their receptors even as most of the circulating particles are cleared from the body. Furthermore, a prolonged residence time could be advantageous for retaining usNPs within the tumor, particularly for the smallest particles (e.g., few-atom AuNCs) that may experience not only efficient renal clearance but also rapid efflux from the tumor. It is noteworthy that this concept has been experimentally demonstrated in a mouse xenograft model using very small DARPin proteins (14.5 kDa) with a range of affinities (0.09 to 270 nM through differences in koff) for the HER2 receptor . It was found that the highest-affinity DARPin reached 8% ID/g tumor accumulation, whereas the lowest-affinity DARPin reached only 0.6% ID/g. These results were consistent with a modeling analysis of the effects of molecular size and binding affinity on tumor accumulation, developed by Schmidt and Wittrup .

Table 2: Quantitative assessment of tumor uptake for actively targeted usNPs compared to control groups.

aCuNCs, copper nanoclusters; usAuNPs, ultrasmall gold NPs; AuNCs, gold nanoclusters; AGuIX, Gd-chelated polysiloxane NPs; C’ dots, Cornell dots (core–shell silica NPs), b% ID/g was determined from the information provided in the manuscript, cTumor uptake reported as % ID, dUptake detected in the spine.