Protein nano-cages: Novel carriers for optimized targeted remedy

Ferritin

This protein belongs to a family of iron storage proteins and consists of 24 subunits that are assembled to form a hollow sphere, globular protein⁴. In vertebrates, ferritin subunits are composed of both heavy (H) and a light (L) chains, which are spontaneously assembled together and surround an aqueous cavity within the internal and external diameters of about 8 and 12 nm, respectively. In 1937, ferritin was isolated from horse spleen, but its structural features remained unidentified until 1991, when its crystal structure was resolved and determined⁷. The Douglas research group was the first to report the characteristics of this protein family. They showed that any changes in the ferritin external surface would not lead to a significant change on its inner surface, which is another fascinating applied feature of this protein⁸.

Proteins of the ferritin family are thermodynamically stable and can tolerate high temperature (up to 85°C) and pH (8.5–9) stress. Hence, they can remain unchanged for a long time even in rigorous ionic situations. In addition, biodegradability and a non-immunogenic manner can be mentioned as the most important characteristics of these proteins, which can be applied widely in medical theranostics. Therefore, this protein family can be broadly used as mineral scaffolds and contrast-agents in MRI, and recently they have been applied for the construction of nano-protein cargoes⁸.

The ability of ferritin for binding to a variety of tumor markers makes it a key determinant for a variety of medical applications, especially tumor diagnosis. The H chain of ferritin (Hfn) consists of an arginylglycylaspartic acid (RGD) domain at its N-terminus, by the sequence of -Cys, Phe, Cys, Asp, Gly, Arg, Cys, Asp, Cys-. The RGD domain, also displayed by “C4” symbol (denoting the 4 cysteine residue present in the domain), is the main core of a ferritin protein and participates in its attachment to other structures. Additionally, the RGD domain shows some genetic variations in different species and can be adopted for different applications⁹. It has been shown that Hfn-c4-RGD (Hfn integrated with RGD or C4 domain) has the ability to target THP-1, a human monocyte cell line that is derived from acute monocyte leukemia¹⁰. There are also some other reports that show that the mutant type of Hfn may have a higher performance than its wild type¹⁰.

Another well-known target for the RGD domain of Hfn is epidermal growth factor receptors (EGFR). EGFR is a famous biomarker that shows a high level of expression in some cancer cells. T cell-replacing factor receptor, another highly expressed receptor in cancer cells, is the other target for the RGD domain of ferritin nano-cages⁸.

Overall, the ability of ferritin nano-particles to co-bind to tumor biomarkers and anti-tumor agents gives them a notable therapeutic property. As an example, the co-binding of the RGD domain of Hfn to both doxorubicin (Dox) and transferrin receptor 1 facilitates the entrance of the drug to tumor cells. However, this targeted and facilitated drug delivery will lead to reduced drug dosage, as well as unwanted multidrug-resistant effect which occurs as the result of reduced peripheral exposure and decreased systemic toxicity via a more effective bio-distribution^10,11.

Small heat shock proteins

Heat shock proteins (HSP) are a family of proteins that are widely expressed in living cells. They respond to stressful situations, such as heat shock, after which they were first described. The key mission of the HSP family is to maintain the correct 3D structure of proteins that are faced with undesirable denaturing conditions. Different types of HSPs exist in nature, which are named according to their molecular weight - small HSPs (sHSPs) are the category that is placed at the lower end of this classification. One of the most well-established types of sHSPs are those that are derived from Methanococcus jannaschii) MjHSP. MjHSP can form a protein nano-cage by self-assembling of 24 identical subunits into an octahedral structure¹². This structure, which is characterized by 16.5 kDa weight and 12 and 6.5 nm size in the outer and inner dimensions, respectively, resembles a typical ferritin nano-cage; but it has a number of large 3 nm pores on its surface, which makes the transport of small molecules across the cage more feasible^13,14. In addition, the structure has the benefit of dramatic thermodynamic stability, similarly to ferritin nano-structures¹⁵.

MjHSP was first introduced in 2003 by Flenniken et al.¹⁶ as a multifunctional nano-scale platform. They demonstrated that this protein nano-cage has variable internal and external surfaces with a variety of reactive amine groups on their structures. The authors also engineered special selectivity for these structures by adding functional thiol groups via polymerase chain reaction-mediated site-directed mutagenesis to enhance the functional reactivity of these nano-structures. Consequently, these genetically modified structures would be able to attach to peptides, various ligand molecules and antibodies, and show versatile therapeutic applications¹⁶.

In addition, the Flenniken research group used a modified thiol group of sHSP to form a hydrogen bond with the maleimide group of Dox, which led to a successful encapsulation and efficient delivery of this chemotherapy agent. Since 24 thiol groups are located in the internal surface of sHSP, it is possible to connect 24 Dox molecules to a single sHSP nano-cage, which can then be gradually released within 24 to 25 hours of incubation in an acidic pH of about 4.5 to 5.5. It has been shown that 50% of these Dox molecules can be released after the first 1.5 hours at pH=5. Effective detachment of Dox from sHSP in lysosomes has been also reported under the same conditions¹⁶.

β-Lactoglobulin

β-Lactoglobulin (BLG) is a well-known protein that is present in many mammal species. It is the major constituent of the whey obtained from sheep and cow’s milk. It is a relatively small globular protein, and its regular structure comprises of an 8-stranded antiparallel β-sheet, 3-turn structure and an α-helix on the outer surface. Different studies have reported different ligand binding sites on this cone-like protein structure, and have proposed that a variety of hydrophobic ligands can bind successfully to BLG at pH values between 6 and 8.1¹⁷. For example, retinol and fatty acids can bind to BLG’s calyx binding site. This property of BLG, besides its stability in the acidic pH, makes the structure a noteworthy choice for oral administration of drugs¹⁸. Like many other nano-structures, many studies have been conducted to improve the efficacy of nano-BLG particles to use in medical applications. For instance, solvent accessibility of ligands that bind to BLG is the main drawback of this protein as a drug carrier - this problem has been partially fixed by different approaches, such as applying a second coat, such as pectin, to protect the ligand during the delivery process¹⁹. Assembly of BLG into nano-structures has been also improved to enhance the capability of these drug delivery systems^20,21 .

Albumin

Albumins belong to a family of highly water-soluble, globular proteins. Ovalbumin, the main protein of egg white, bovine serum albumin, the albumin derived from cow’s serum, and human serum albumin, the most abundant protein of human serum, can be mentioned as the three major types of this protein family²². The albumin family has been widely used as drug carriers, since they are easily available, have a low cost, have a feasible purification procedure, are stable in a large pH range (4–9), show unusual ligand-binding properties, due to their highly charge nature, and most importantly represent a preferential uptake in inflamed tissue and, surprisingly, tumors. These properties, besides their nontoxic and biodegradable nature, makes them a favorable choice for drug delivery. Assembly of albumins into well-defined nano-structures gives them additional therapeutic advantages, including high tolerance, lack of undesirable interactions with serum, high capacity for drug loading²³.

Plant viruses

Cowpea chlorotic mottle virus (CCMV). The use of CCMV nanoparticles as drugs and chemical carriers, besides the manipulation of their functional groups to enhance their capabilities, was first made by Douglas et al. in 2002. These authors also tried to mimic the functionality of ferritin, as a well-established iron reservoir in living organisms⁸. Today, the structural features of these viral nano-particles have been well defined. We know that CCMV’s protective coat consists of 180 identical protein subunits, each composed of 190 amino acids. These identical subunits are self-assembled together to form an icosahedral protein cage, which represents a T=3 symmetry. This complex structure forms in a way so that the amine terminals of the protein chains locate in the interior side of the cavity, making it a favorable option to carry negatively charged components²⁴. This nano-particle, which is 18 nm wide in its interior cavity and 28 nm wide as a whole (twice the diameter of ferritin), shows a very dynamic structure. For example, reassembly, disassembly and swelling of this structure due to pH alteration is a feature of its dynamicity. This is an important point because researchers can utilize the reversible swelling property of CCMV nano-cages, hence the process can lead to the formation of 60 openings of 2 nm diameters, suitable for an appropriate pH-dependent load and release of the medical cargo²⁵.

CCMVs can be easily obtained from the infected leaves of the virus corresponding host, Vigna unguiculata. It can also be obtained by the means of genetic engineering and the use of E. coli or yeast expression systems. When genetic engineering is the method of choice, it would be possible to make some pseudo particles with different functionality. For example, by truncating the N-terminus of the capsid’s subunits, the number of the subunits, present in the resulting particles, will be changed, along with the order of symmetry and shape, rather than icosahedral structures²⁶.

Cowpea mosaic virus (CPMV). CPMV is another plant virus with an icosahedral structure. CPMV particles show interesting properties for drug delivery systems. For example, the recovery yield of these particles from their infected host, Vigna unguiculata, is relatively high and cost-effective. The amino acid composition and arrangement of these nano-structures is another outstanding feature. CPMV’s capsid is composed of 60 protein subunits, mainly of two types, small 24 kDa and large 41 kDa units²⁷. These are self-assembled in a 28 nm structure, in which lysine and cysteine residues are located on the outer surface and make them a suitable carrier for many chemicals and drugs. They can also be used to improve the functionality of the structure. For example, the formation of polyethylene glycol (PEG)-maleimide ligands, which aid in reducing the particle’s hydrophilicity and enhance their biocompatibility. Different ligands, such as carbohydrates, antibodies, oligonucleotides, organic dyes, biotin, peptides and proteins, react with the cysteine residues on the outer surface of the nano-cage²⁸. In addition to these characteristics features, CPMV particles have shown good stability against harsh chemical modifications (e.g., when conjugated with PEG), and a wide range of temperature alteration, solvent and pH fluctuations, which are all vital for drug delivery systems.

Bacteriophages

MS2 bacteriophage. MS2, the pathogen of the Enterobacteriaceae family, is an icosahedral bacteriophage. Its capsid is formed by self-assembly of 180 identical protein subunits into a 27 nm porous structure (32 pores of 1.8 nm in diameter). This protein nano-cage is stable against harsh environmental changes, such as wide variations in pH range (3–10) and temperature fluctuations²⁹. The porous structure of the MS2 protein nano-cage has good stability against harsh environmental changes, making the particle a very good platform for chemical modifications and functional alterations.

The Wu research group (1995) was the first to try targeted drug delivery using the MS2 bacteriophage³⁰. The RNA genome of this bacteriophage consists of a specific region known as translational repression RNA stem-loop (TR), which naturally has an affinity to the virus coat protein. The affinity of TR for the viral coat not only prevents the binding of ribosomes and inhibits the translation of the replicase cistron, but also plays an important role in the assembly of viral particles. The Wu research group used this affinity to encapsulate drugs into the MS2 viral cage. They successfully attached the ricin A chain (a highly fatal herbal toxin) to the TR region of the bacteriophage and showed that multiple copies of the TR-RAC conjugate can be incorporated into a single viral shell. They also made some genetic and chemical manipulations on the MS2’s protein coat to add human transferrin to the surface of the shell, in order to establish receptor-mediated endocytosis in target cells³⁰. Later, this work was followed by other researchers worldwide. For example, in 2002 a research group from the University of Leeds, manipulated the MS2 viral shell via a site-directed mutagenesis and made chimeric VLPs, which were strongly immunogenic when carrying either B or T cell epitopes. This group not only made use of a directed immunogenic response of these manipulated particles, but could also protect and effectively deliver nucleic acid-based drug cargoes^31,32.

Qβ bacteriophage. Qβ bacteriophage is another icosahedral virus with a diameter of 28 nm and a T=3 symmetry. The protein coat of this virus consists of 180 identical proteins, of 14 kDa molecular weight. Therefore, the structure is very similar to the MS2 bacteriophage³³. Like other protein nano-particles, many studies have been conducted to make use of the Qβ bacteriophage as an efficient drug vehicle. The onset of these efforts goes back to the Strable et al. (2008) who successfully incorporated unnatural amino acids into the Qβ VLPs and made some notable structural changes from the natural coat, and investigated the stability and properties of the new structure³⁴. Many other researchers have been published since then, which report successful functionalization and/or targeting of Qβ VLPs. For instance, attachment of human transferrin to the bacteriophage protein coat³⁵, modification of Qβ coat so that viral particles can successfully target CD-22 for the treatment of certain autoimmune disorders and cancers³⁶. The most recently published work is by Chen et al. (2016) who reduced the cytotoxicity of Dox by dual functionalization of Qβ VLP by PEG and this chemotherapy agent³⁷.

Human viruses

Simian vacuolating virus 40 (SV40). SV40 is a DNA onco-virus that is found in both monkeys and humans, and belongs to the family of Polyomaviridae that all have an icosahedral structure. Viral proteins that participate in the structure of SV40 varies depending on the early or late translation of the virus genome. The main coat protein of SV40 is viral protein 1 (VP1), which is produced by early transcription of this viral genome with a Svedberg constant of 16. Late transcription of the SV40 genome not only yields VP1, but also two other smaller (19s) viral proteins, VP2 and VP3. A SV40 virus that only consists of VP1 yields a T=7 d (triangulation number 7 dextro) in a 45 nm icosahedral structure composed of 72 pentamers assembled together^38,39.

Like the other VLPs, SV40 VLPs can also be used as a drug delivery system. The variable structure of these VLPs, influenced by the types of VPs, provides a specific flexibility in design and applications. Many studies have been focused on specific targeting, as well as functionalization of SV40 VLPs. For instance, genetic modifications of the core proteins for directed targeting^40,41, gene delivery and transduction of multiple cell type and so on^42–44.

Murine polyomavirus. Like SV40, murine polyomavirus belongs to the family of Polyomaviridae. The protein coat of this virus is composed of an outer layer, which resembles SV40’s structure, and forms via self-assembly of 360 VP1 units, ordered in 72 pentamers and a T=7 icosahedral structure. However, it differs from SV40 due to the presence of an additional inner layer, which is composed of both VP2 and VP3 subunits⁴⁵.

In 2005 Brinkman and his colleagues surveyed the beneficial therapeutic effects with different particulate structures of murine polyomavirus VP1-coat protein. They showed that polyomavirus-like-particles (PLPs) represent a highly potent antigen-delivery system for inducing cytotoxic T lymphocyte responses, which can be applied for the treatment of malignant diseases via immunotherapeutic approaches⁴⁶. Many other notable studies have focused on the genetic manipulation of VP1 subunits of the PLP outer layer in order to render some specific functionalities to these VLPs. For instance, Schmidt et al. added a special protein domain named WW (because of its two conserved tryptophan residues) to PLP’s VP1s⁴⁶. Via this modification, PLPs gained a specific affinity for proline-rich ligands. The authors showed that this affinity leads to the efficient encapsulation of a variety of ligands to the PLPs⁴⁷. Attachment of charged groups to the interior layer of the PLPs⁴⁸, and efficient encapsulation of the foreign cargoes via an anchoring technique for VP1s⁴⁹ can also be mentioned as some other examples.