Towards Protein-Based Viral Mimetics for Cancer Therapies

Towards Protein-Based Viral Mimetics for Cancer Therapies Ugutz Unzueta 1, 2, 3 , María Virtudes Céspedes 3, 4 , Esther Vazquez 1, 2, , Neus Ferrer-Miralles 1, 2, 3 , Ramón Mangues 3, 4 * and Antonio Villaverde 1, 2, 3 * 1 Institut de Biotecnologia i de Biomedicina, Universitat Autònoma de Barcelona, Bellaterra, 08193 Barcelona, Spain 2 Departament de Genètica i de Microbiologia, Universitat Autònoma de Barcelona, Bellaterra, 08193 Barcelona, Spain 3 CIBER de Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN), Bellaterra, 08193 Barcelona, Spain 4 Oncogenesis and Antitumor Drug Group, Biomedical Research Institute Sant Pau (IIB-SantPau), Hospital de la Santa Creu i Sant Pau, C/ Sant Antoni Maria Claret, 167, 08025 Barcelona, Spain


Drug-based cancer therapies
Since age is a main factor of risk, the high prevalence of cancer in high-income countries places this disease as a second highest cause of death (around 1 in 4 deaths), after cardiovascular diseases [1]. Despite possible compensatory effects of early detection, the high mortality among cancer patients stresses the limitation of current treatments, many of which are essentially based on surgery and adjuvant chemotherapy [2]. Low molecular weight cytotoxic chemicals, such as 5-Fluorouracil, Cisplatin or Doxorubicin have been developed and used for decades and they represent the current basis of treatment for most cancers [3].
These drugs induce DNA damage, leading to tumor cell death, and are administered at maximum tolerated doses. The resulting high systemic drug levels cause severe toxicities related to DNA damage in highly proliferative healthy tissues (e.g. bone marrow), which worsens patients' quality of life [4].
Poor drug penetration due to abnormal tumor architecture and composition [5], and clearance through hepatic metabolism [6] or renal clearance (with a cut-off around 7 nm; see Glossary) [7] are additional factors that hamper a desired dose reduction to safer, less toxic values.
Renal filtration can be largely minimized by increasing the molecular size of the drug, through conjugation to large molecules such as proteins, which act as carriers. In addition to allowing longer circulation times in the bloodstream, drugprotein conjugation reduces hepatic clearance and increases drug concentration in tumors, compared to free-drug administration. This is because its nanometric size promotes higher nanoconjugate accumulation in tumor tissue because of the enhanced permeability retention (EPR) effect; that is, a form of passive targeting [8;9] (Box 1). In this regard, nab-paclitaxel has been incorporated into treatment regimens for advanced breast, lung or pancreatic cancer. In nab-paclitaxel, the bound albumin stabilizes paclitaxel and in effect increases the size of the drug.
Because of the many possible benefits of having drugs that are larger than small molecules, nanoparticles (usually ranging from 10 to 100 nm) are promising agents in the development of cancer therapies [10]. Most nanoparticles currently used in the clinic exhibit passive targeting (e.g. liposomal doxorubicin, nabpaclitaxel) [11]. In this context, only about 5 % of the injected therapeutic reaches the tumor because the high accumulation (50-80 % of the dose) of nanoparticles in the mononuclear phagocytic system (MPS) especially in the liver [12][13][14]. This process could be attenuated through the covalent attachment of polyethylene glycol (PEGylation) to the nanoparticle [15] (Box 1). However, the penetration of nab-paclitaxel into tumors might also benefit from indirect effects.
Thus, the albumin component of the nanoparticle may bind to SPARC, a protein secreted by stromal fibroblasts to the tumor extracellular space, or to the gp60 receptor, facilitating nab-paclitaxel endothelial transcytosis [16;17].

Cell targeting in cancer treatments
A relevant and distinctive property of cancer tissues is that the proteins that drive tumor progression, such as cytokine, hormone or grow factor receptors are differentially overexpressed in cancer stem cells (CSC), as compared to healthy tissues [18]. Such differential expression can enable the molecular tagging of cancer cells for the delivery of next generation drugs. Molecular tags are already implemented in combination with conventional therapies to inhibit signalling from a specific target protein (eg, VEGF, EGFR, HER-2 or B-Raf) [19]. Although less aggressive than in chemotherapy, toxicity can also arise if target activity is inhibited in normal tissues, and resistance can develop through target or pathway mutation (e.g. EGFR amplification) or the activation of alternative or compensatory pathways [18].
Learning from these lessons, cell targeting in cancer treatment should be primarily exploited to engineer the biodistribution of conventional, well-known drugs as cargos in long-circulating nanoconjugates, aimed to increase the effective drug concentration in tumor cells. In this regard, if the administered drugs would be introduced in such a way that they only (or preferentially) penetrate tumor cells, doses could be largely reduced and toxicity issues essentially minimized. CSCs are responsible for tumor and metastasis initiation and maintenance and closely associated with aggressiveness. Active drug targeting aimed at eliminating CSCs is then a promising anticancer strategy. This therapeutic approach takes advantage of the differential expression of membrane receptors between CSCs and the bulk of the tumor, mainly composed of differentiated cells [20].

Proteins, virus-like functions and artificial viruses
In nature, animal viruses, which are nanoscale in size, exhibit exquisite specificity for cell surface receptors displayed on target cells. The specific interactions that trigger infection are mediated by cross-molecular interactions between peptide motifs in capsid proteins that act as ligands, and target surface cell proteins that act as receptors for the virus. The multivalency of ligandreceptor binding based on the repetitive and regular architecture of viruses ensures a high degree of tissue and cell penetrability, and increases the likelihood of interaction. In parallel, an increasing number of peptides and protein domains have been described as tumor-homing peptides. They exhibit the ability to specifically bind cell-surface protein markers in CSCs or in more differentiated cells [21][22][23], with an important degree of discrimination between specific tumor types [24]. Alternatively, nonspecific cell-penetrating peptides have been engineered to be activated by local stimuli, such as low pH, or by metaloproteases, which are present in tumor tissues [25].
All these categories of peptides are valuable tools in enabling the targeting of drugs to specific tumors or tumor cell sub-populations, provided they functionalize nano-sized vehicles in a multivalent and regular distribution. The 'artificial virus' concept was proposed to define any manmade biocompatible nanomaterial exhibiting virus-like characteristics and size, with the potential to be cell-targeted carriers in molecular therapies [26]. Metals, polymers, carbon nanotubes or lipids may be suitable for nanoparticle fabrication [27]. However, proteins are likely the most convenient materials for the construction of effective viral mimetics in therapy, since they are the ultimate supporters of biological functions and specificity in molecular interactions. Being fully biocompatible, proteins have been produced since decades in cell factories by cost-effective scalable bioproduction (or by chemical synthesis if short peptides), to be used, among other applications, as pharmaceuticals [28;29]. In this regard, the regulatory issues linked to the administration of proteins to humans have been already well addressed, and the number of endotoxin-free and generically recognized as safe (GRAS) microorganisms available for biological production of proteins is lately expanding [30]. In addition, precise protein engineering by conventional genetic approaches allows the modulation of their functional and structural properties in a very versatile way.
Furthermore, cost-effective large-scale production of difficult-to-express proteins and nanostructured protein materials is now becoming feasible due to accumulating advances in genetics and systems biotechnology [31;32] and the increasing availability of cell factories adapted to complex protein production challenges [30;33;34]. The multiple virus-like functions necessary for molecular transport and intracellular delivery can only be achieved by proteins, and different functions can be assumed by protein complexes of by the construction of single chain modular polypeptides that recruit diverse functional domains from independent origins [35]. The unique functional and structural plasticity of proteins is ideal for the generation of multifunctional vehicles adapted to the targeted transportation of specific drug types, including nucleic acids (in nonviral gene therapy) and chemicals (in chemotherapy). Although protein-based viral mimetics have great potential for use in cancer therapy [36], rapid development of therapeutic artificial viruses has been unfortunately impaired by still limited structural comprehension of protein-protein interactions and by the lack of universal tools to predict and engineer precise contacts between designed polypeptides. The ability to arrange building blocks in regular patterns to generate multivalent constructs of defined nanoscale size, is an unavoidable requirement for the de novo generation of virus-like assemblies. Although control over particle size has been more easily reached in the design of liposomes and related polymer-based vehicles, the issue is much more challenging in the case of protein vehicles. Some recent successes in the computing-assisted design of complex protein nanostructures [37;38] permit to envisage, however, the feasibility of tailoring multimeric protein nanomaterials.

Emerging nanoarchitectonic principles, viral mimetics and antitumoral drug delivery
In this context, protein science has benefited from multiple approaches to engineering protein self-assembly [36;39], which have resulted in the generation of a wide range of nanoparticles and nanostructured materials [40]. The most promising routes to reach functional protein nanoparticles include: exploitation of the amphiphillic character of peptides and proteins, the adaptation of natural oligorimerization domains and the manipulation of charge distribution to modulate electrostatic protein-protein interactions (Table 1) (Table 1). Some multifunctional proteins of this kind have already entered clinical trials [42].
In a paradigmatic example of viral mimetics, multifunctional single chain proteins were developed based on the linear fusion of three main cassettes: an amino terminal cationic peptide, a core scaffold protein and a carboxy terminal polyhistidine [43]. Such an engineering scheme is extremely efficient in promoting the self-organization of the whole chimera under aqueous physiological conditions [41], as nanoparticles of regulatable size between 10 and 80 nm [44]. This is irrespective of the particular scaffold protein used as building block core, and the particular amino acid sequence of the amino It is already possible to load artificial viruses with expressible DNA for gene delivery [48]. Coupling artificial viruses to anti-tumor compounds would be a logical next step.

Concluding remarks and future perspectives.
In summary, the versatility of protein engineering regarding structure and