Nanobody-based Cancer Therapy of Solid Tumors

Three Steps Leading to Therapy

For each of these platforms, the distribution through the body and the delivery into the tumor tissue is different, and the efficiency of this process contributes strongly to the efficacy of the treatment. A therapeutic formulation that is very effective in a 2D cell culture set-up is not necessarily effective in an in vivo preclinical model. In this section of the article, we will discuss the steps that the nanobody-based therapeutic molecules go through before reaching their therapeutic target in vivo.

Delivery of therapeutic agents can be conducted in different ways: oral, intravenous, intraperitoneal or intratumoral. Nanobody technology is applicable to all of the aforementioned administration routes, although the first three have been used previously (e.g.,^[46,48,53]). Each pathway has different demands for the nanobody-based formulation. For instance, with the oral application or the intraperitoneal injection, the nanobody requires resistance to extreme conditions (i.e., proteases and/or acidic pH). Nanobodies can be made resistant to proteases by adaptation of the sequence or by the introduction of an additional disulfide bond in order to improve resistance to pepsin and chymotrypsin.^[54] For intravenous injection, stability in serum is essential. Although most nanobodies have been described as very stable, when combined with effector domains or nanoparticles, the stability of these systems might be different. Instability of nanobody-based formulations may give rise to an early release of the drug before reaching the cancerous mass, which could result in severe side effects and decreased therapeutic benefits. As nanobodies are usually sufficiently stable for intravenous injection, this method of administration is currently the most frequently used method for in vivo nanobody-based cancer therapy studies.

Systemic Circulation Upon Intravenous Injection

Intravenous administration of therapeutics is not always performed in close proximity to the tumor mass.^[53] Consequently, the injected material needs to travel for a sufficient period of time along the circulatory system in order to reach the tumor. Sufficient tumor accumulation therefore requires a sufficient residence time of the nanobody in the blood stream, which differs for the type of nanobody platform. Naked nanobodies are rapidly cleared from the bloodstream, which reduces the time interval to bind to their target molecule. On the other hand, efficient clearance also decreases the risk of unwanted toxic side effects. Therefore, an appropriate balance between these factors might be essential for successful therapy. In case of larger nanobody drug formats, opsonization and subsequent recognition and uptake by the reticuloendothelial system (RES) may occur, leading to hepatic clearance of these therapeutic compounds.

Renal Clearance. Renal clearance is a multifaceted process involving glomerular filtration, which depends on the size of the molecule.^[55] Molecules with an in vivo hydrodynamic diameter (HD) <6 nm in size are filtered by fenestrations in the endothelial cell layer, in contrast to molecules with a HD >8 nm. In general, the average weight cutoff for renal clearance is approximately 60 kDa.^[56,57] For intermediate-sized molecules, the filtration is further dependent on their charge. Positively charged molecules are more likely to be filtered due to the negative charge of the globular membrane.^[55,58] Besides this, the charge of a molecule may provide interactions with plasma proteins, increasing the HD and preventing renal clearance.^[59] However, in the case of nanobodies, there is no general rule for the net charge in vivo. Due to the small size of these molecules, their isoelectric point is mainly determined by the different amino acid composition of the complementary determining region (CDR) regions. It is important to note that, in general, an extremely high or low isoelectric point will render some of the nanobodies unsuitable for in vivo use.^[60] The size (2.5 nm diameter and 4 nm height) and prolate shape of nanobodies predict rapid renal clearance.^[61] This prediction has already been confirmed by several in vivo studies.^[18,62] Importantly, the renal clearance and/or retention of the nanobody–toxin conjugates in the kidney may lead to renal toxicity. Whether these compounds are retained in the kidneys depends on the added size, change in charge and/or overall HD due to the coupled effector domain. For instance, a PS such as IRDye700DX results in the addition of only 2 kDa,^[41] which, in this case, will result in clearance of the nanobody–PS construct through the kidneys. Nevertheless, in this particular context, nephrotoxicity is minimized due to the fact that the PS only leads to toxicity when it is activated through specific illumination.

Although no nanobody–effector domain platform for cancer therapy has been characterized in vivo so far, pharmacokinetic toxicology studies of other immunotoxins, such as ref-43-pokeweed antiviral immunotoxin, have already showed dose-dependent kidney toxicity due to renal retention.^[63,64] To avoid toxicity, renal retention of nanobody–drug conjugates should be minimized. As the renal retention relies on the endocytic pathway, the coinfusion of gelofusine and/or lysine in order to compete with megalin may lower the retention.^[65] In addition to this, substitution of negative or positive residues of the nanobodies could affect renal retention.^[17,66–67] Since the nanobody scaffold can be engineered to a certain extent, it can be designed to ensure lower renal retention.^[17,56] A different method for reducing renal accumulation and retention is to lower the renal filtration rate,^[57] resulting in an increase in half-life and the chance of improved tumor uptake.^[68] Increasing the size by, for instance, glycosylation,^[69] PEGylation^[70] or noncovalent binding to circulating serum proteins (albumin), such as the fusion with an albumin-binding nanobody,^[19] can prolong the half-life and thus lower renal retention.^[68]

In contrast to the described kidney clearance for antagonistic nanobodies and targeting nanobodies with effector domains, nanoparticles decorated with nanobodies are, due to their size, not eliminated through the kidneys. These larger types of nanobody platforms (nanoparticles) are cleared by the liver.

Hepatic Clearance. The hepatobiliary system is the primary route of excretion for drugs that are too large for renal filtration.^[55] Compounds and particles that undergo hepatic clearance are catabolized by hepatocytes.^[71] Kupffer cells and hepatocytes are parts of the biliary system and particles endocytosed by these cells are excreted into the bile. Kupffer cells have a much higher phagocytotic capacity than hepatocytes and form the RES or mononuclear phagocyte system. Particles taken up by Kupffer cells rely exclusively on intracellular degradation; however, particles that are not broken down will be retained inside the cells. Hepatic clearance has a preference for the removal of particles with a HD of 10–20 nm, as their primary task is filtration of, for instance, viruses.^[59] In addition to the liver, phagocytic cells of the RES also reside in the spleen, making this organ another target of the clearance of nonglomerular-cleared compounds.

Nanobody-based drug delivery systems, such as albumin nanoparticles, liposomes or micelles, are spherical with a large HD.^[55] A biodistribution study of untargeted liposomes indeed showed an accumulation in liver and spleen.^[72] Of interest is that saturation of the liver accumulation results in a shift of the liposome distribution to the spleen,^[73] appointing the liver as the main clearance organ of liposomes. Moreover, accumulation of liposomes into the tumor might also occur after saturation of the liver. Modifications of liposomes can be carried out in order to avoid 'first-pass' hepatic clearance, at least to a certain extent. For instance, PEGylation of liposomes lowers opsonization by plasma proteins and increases the circulation time by avoiding phagocytosis by the RES components.^[74] Although accumulation at the tumor depends on the circulation time of liposomes, an increase in circulation time does not directly translate into increased tumor uptake.^[75]

Nanobody Extravasation & Tumor Penetration

The second phase of drug delivery into a solid tumor, once the drug has reached the tumor blood circulation or the nearby blood supply, is the extravasation from the bloodstream and retention at the tumor site in order to allow interaction with cancer cells, resulting in accumulation in the tumor. Transport of nanobody-based therapeutics across the vessel wall is mediated by diffusion and fluid transport. In normal tissue, a net negative pressure between blood vessels and the interstitial space exists, resulting in fluid movement towards the interstitial space and further to lymphatic ducts. In tumors, the interstitial fluid pressure (IFP) is higher than that of surrounding tissue. This elevated IFP limits the transport of large molecules (e.g., mAbs) and particles into the interstitial matrix, and the tumor penetration becomes more dependent upon diffusion.^[76] Targeted therapeutics are aimed at binding to receptors present on tumor cells. However, the binding site barrier effect, first suggested by Fujimori et al.,^[77] was described as a limiting factor of high-affinity binding mAbs that, due to their large size, hampers the diffusion of other mAb molecules into the tumor tissue.^[76] Importantly, this effect was not observed with small molecules such as affibodies and nanobodies, which are able to distribute throughout tumors in a more homogenous manner.^[12,68] In a study by Oliveira and coworkers, the tumor distribution of 15-kDa nanobodies was compared with a 150-kDa mAb after conjugation to the fluorophore IRDye800CW (IR).^[12] The EGFR-specific 7D12-IR nanobody showed a homogeneous distribution of the probe in A431 human tumor xenografts at 30 min to 2 h postinjection, which led to a relatively high tumor uptake, whereas the negative control R2-IR did not accumulate in the tumors. An irregular distribution of cetuximab-IR in the tumor stroma was observed, possibly due to the binding site barrier effect.^[78] Similar results were obtained for the anti-EGFR affibody, a small binding scaffold based on protein A, conjugated to IR, in contrast to cetuximab-IR, which was confined to the center of the tumor.^[79] Homogenous distribution of the drug throughout the tumor mass is essential for successful treatment. If only part of the tumor mass os exposed to the drug, complete tumor eradication will not be achieved, leading to eventual tumor regrowth.^[69] In this respect, nanobodies are expected to outperform mAbs.

As molecular size is an important factor for diffusion, the diffusion capacity of nanobody-targeted nanoparticles (platform C) is dependent on their size, in that the smallest will have a better chance of diffusing into the tumor (Figure 3). Importantly, the blood and lymphatic vasculature differ substantially in tumors and normal tissues. Blood vessels of healthy tissues are normally well sealed and continuous, which prevents extravasation of therapeutic compounds. By contrast, immature, dilated tumor vessels are leaky due to the presence of much larger pores in postcapillary venules, often exceeding 100 nm in size.^[71] This hyperpermeability of the tumor vasculature allows leakage of macromolecules and nanoparticles into the tumor. This phenomenon is referred to as the enhanced permeability and retention effect, and was first described by Matsumura and Maeda in 1986.^[80] High levels of doxorubicin were delivered to the tumor site through cross-linked polymeric micelles (with a diameter of ˜70 nm) decorated with EGFR-targeting nanobodies (EGa1), due to the enhanced permeability and retention effect.^[39] On the other hand, no significant effect on tumor growth inhibition of a 14C tumor xenograft model was observed with liposomes encapsulating the small tyrosine kinase inhibitor AG538 and decorated with anti-EGFR nanobody (EGa1),^[42] despite the fact that a clear inhibitory effect on cell proliferation was observed in vitro. This lack of toxic effect can be attributed to the electrostatic interactions between this cationic liposome formulation with serum proteins, thereby affecting the circulation time and subsequent accumulation at the tumor site. On the other hand, active targeting by the surface-bound nanobodies does not contribute to significant accumulation of nanoparticles in solid tumors, but has a vital contribution to the subsequent step.

Nanobody Interaction With Targets

For all of the mentioned nanobody platforms, the final step before the therapeutic mechanism of action, is the actual binding to the target molecule, which is mediated by the nanobody. The specificity of this last stage is essential for the therapy to occur with minimal side effects. For antagonistic nanobodies, a high binding affinity is essential, as these nanobodies are expected to compete with the natural ligands, which normally bind with high affinity themselves to their receptor. Phage display selections can specifically be aimed at the retrieval of high-affinity binders. In addition, the improvement of binding affinity can be obtained by preparing a family library based upon the CDR3 sequence of an already-selected nanobody.^[81] A disadvantage of the nanobody technology is that the conjugation to an effector domain might have a severe effect on the binding properties of the nanobody. Crystal structures of nanobodies have shown that the N-terminus is positioned close to the site of the CDR sequences and conjugation to this site of the protein might affect antigen binding.

Although not in all cases, random conjugation to the primary amines (lysines and the N-terminus of the protein) was found to affect the binding properties of the nanobody.^[18] In this study, the conjugation of the fluorophore IR was shown to completely prevent the binding of a HER2-directed nanobody to its target in vivo. Importantly, affinity was retained after conjugation of this fluorophore to a C-terminal cysteine.^[18] Thus, the best solution for the conjugation of effector domains to the nanobody appears to be via the C-terminus. Two nanobodies fused to effector domains have been described and both have been fused to the C-terminus: the anti-VEGFR2 has been fused to PE38^[46] and an anti-EGFR nanobody to soluble TRAIL.^[26] In the latter case, a drop in affinity was observed, but in this particular set-up, it did not hamper the efficacy in killing cancer cells. In addition, other examples have documented site-directed conjugations using a C-terminal cysteine.^[33,82–83] Alternatively, click chemistry and intein- and sortase-based conjugation systems are in development and may contribute to further functionalization of the nanobodies.^[84–88]

Recently, we have randomly conjugated EGFR-targeted nanobodies (named 7D12 and 7D12-9G8) to a traceable PS for photodynamic therapy (PDT).^[41] The binding affinities of these EGFR-targeted nanobody–PS conjugates remained in the low-nanomolar range and these conjugates are expected to behave in vivo very similarly to what has been observed in molecular imaging studies.^[12] After the preclinical testing, more will be known of the feasibility of the approach in which the fluorescent nanobody–PS conjugate can be detected through optical imaging, enabling guidance of the actual treatment (i.e., PDT). After binding of the nanobodies to their target receptor (e.g., EGFR), the nanobodies undergo a very slow internalization (one round of internalization is completed after 24 h). For a more rapid internalization of the cargo, the use of biparatopic nanobodies (e.g., 7D12–9G8) was recently introduced.^[51] These biparatopic nanobodies consist of two different nanobodies binding to the same target protein (EGFR), but on different, nonoverlapping sites. As a result, these nanobodies stimulate receptor clustering, which induces receptor internalization and subsequent degradation in the lysosomes. Similar results were shown for antibody constructs.^[89] This method allows specific binding to the target cells, followed by internalization, enabling the reversible conjugation of drugs that are sensitive to intracellular proteases, such as cathepsin B, to then be released for their action.

When nanobodies are employed as targeting moieties of long-circulating nanoparticles, such as PEGylated liposomes or branched gold nanoparticles, affinity becomes less critical. This is mainly because the affinity will be sufficient as a result of avidity, as several nanobodies are present on the same particle. Mamot et al. have shown that the targeting moiety has a function in the cellular uptake of the particles.^[75,90–91] This has been demonstrated with nanobodies binding to cell membrane proteins, such as anti-EGFR nanobodies (EGa1) conjugated to liposomes,^[39,39,42] polymeric micelles^[38,43] or to albumin nanoparticles.^[44] Another example is the binding of the anti-c-Met nanobodies conjugated to albumin nanoparticles to the human ovarian carcinoma cell line TOV, stably expressing c-Met.^[30] In this case, the presence multiple nanobodies on the surface of those particles also results in clustering of their target receptor at the membrane, causing their internalization. The c-Met-targeted albumin nanoparticles clearly entered by the route of early endosomes, late endosomes and lysosomes, where degradation of both nanoparticle and c-Met took place.^[30] These nanoparticles were also able to induce phosphorylation of c-Met. However, this activation of c-Met was insufficient to induce complete c-Met signaling, as cell migration was not induced. Similar observations were previously reported for agonistic antibodies against c-Met as well as EGFR.^[92,93] The most prominent effect of the anti-c-Met nanoby-albumin nanoparticles (NANAPs) seems to be the degradation (i.e., downregulation) of the c-Met receptor. By contrast, EGFR-targeted liposomes were able to inhibit EGF-induced activation and also induced EGFR receptor downregulation.^[39] Interestingly, scFv liposomes also targeting EGFR were unable to induce the same EGFR downregulation effect, which is believed to be related to the fact that nanobodies dissociate from their targets only at very acidic pH levels (below the pH of late endosomes).^[39] Thus, by remaining attached to the nanobody, EGFR is unable to recycle to the cell membrane, and therefore the EGFR–nanobody–liposome is directed to lysosomes for degradation.

Lysosomal routing and subsequent degradation is very valuable as it leads to the downregulation of receptors that play an active role in tumor proliferation. This routing also opens up the possibility of incorporating sensitive linkers (pH and protease), enabling the release of cargos from the nanoparticle in the endolysosomal system. For instance, the EGFR-targeted nanobody liposomes containing IGF-1R-targeted kinase inhibitors were clearly able to release the kinase inhibitor, which then reached the target site of action (i.e., cytoplasm).^[40,42] Similarly, the EGFR-targeted nanobody albumin nanoparticles that trafficked to the lysosomes were also able to release the multikinase inhibitor from its linker, as this could carry out its mechanism of action (in an in vitro study).^[44] In case of the EGFR-targeted nanobody–micelles containing doxorubicin as the payload,^[38,43] doxorubicin was coupled with the polymer via a hydrolytically degradable linker (at pH 5), enabling the release of doxorubicin from the polymer upon trafficking of the nanoparticles to the late endosomes/lysosomes.^[94] In these examples, cargos were very-low-molecular-weight drugs that could diffuse out of the late endosome/lysosome compartments. More complex will be the situation in which the cargo cannot cross the cell membrane, unless binding of the nanobody to the target protein is sufficient for the mechanism of action. This seems to be the case for the HER2-targeted branched gold nanoparticles for photothermal therapy^[23] and the HER2-targeted nanobodies for radionuclide therapy,^[45] as well as for EGFR-targeted nanobodies (7D12 and 7D12-9G8) conjugated to PSs for PDT.^[41] Nevertheless, in the latter case, enhanced toxicities were documented upon increased internalization of the conjugates.^[41]

Nanomedicine. 2015;10(1):161-174. © 2015  Future Medicine Ltd.