Targeted Intracellular Delivery of Antibodies: The State of the Art

1First established in the 1950s, the drug delivery field has since grown substantially and evolved through several generations, from simple formulations for oral and transdermal delivery to complex delivery systems with the ability to overcome biological barriers. A simple search for “drug delivery” in PubMed reveals a slow growth in the number of results until the early 2000s, followed by rapid expansion up to 2015, and a subsequent plateau of approximately 15,000 publications per year. Nevertheless, a tremendous number of in-demand drugs exert effects within cells, and an increasing number of new drugs intended to influence protein regulatory pathways are realized through protein–protein interactions (PPI). Pharmaceuticals acting on intracellular macromolecules must be delivered intracellularly into the target cells. Small molecules, which include most drugs, often cross biological membranes readily owing to their size, amphiphilic nature, and the availability of transmembrane carriers. Meanwhile, despite the generally successful transmembrane penetration of many small molecules, some potential intracellular targets are excluded from their modulation. Among such elusive targets, are certain PPIs that regulate essential cellular functions and contribute to the signaling pathways involved in cancer pathogenesis. Small molecules are often less effective than macromolecules for the interruption of PPIs, except for proteins with small-molecule-binding pockets such as enzymes, ion channels, and cell receptors, because of the large flat contact areas and lack of poorly defined deep pockets and grooves in which small molecules could insert and subsequently perturb its activity. PPIs often involve large surfaces, which typically range from 1500 to 3000 Å2, compared with 300–1000 Å2 in protein interactions with small molecules. A typical example of PPIs is the interaction of antibodies with protein antigens. In recent years, trends have revealed the growing impact of approved polypeptide macromolecular drugs, particular antibodies to cell-surface and soluble targets.

2The size of the human interactome size was estimated to include 650,000 or more distinct interactions, but this estimation has since increased. Recent analysis has revealed that only a small part of the human proteome (less than 700 proteins) is pharmaceutically accessible by approved drugs, which has led to the concept of “undruggable” targets that are not amenable to small-molecule interventions. This list includes certain infamous and notorious cancer-causing molecules, such as c-Myc, Ras, and NF-κB. Intracellular proteins that act at immune checkpoints are other possible targets for which inhibition by antibodies might reactivate the host’s immune response against cancer cells. In this regard, a huge field of applications has been opened for intracellular treatment and diagnostics using antibodies. One mechanism of antibody action is the alteration of signal transduction through a physical block of the interaction between two proteins that are components of a cellular pathway. This is exemplified by antibodies binding receptors, such as those of the EGFR family, on cell membranes, and sterically preventing them from transmitting a signal. The tumor signaling perturbation through antibody interference has been shown to be a successful approach for the treatment of some diseases. The feasibility of antibodies to downregulate intracellular targets has been proven in experiments with microinjections and transfections with a corresponding gene (intrabody approach;; ). However, the microinjection technique is limited to a small number of cells in vitro. The intrabody format has the potential for development, but still exhibits the same risks as gene therapy: the possible genome integration of viral vectors, low efficiency, and possible toxicity of non-viral vehicles, which has necessitated the development of direct antibody introduction methods. Currently, the routine introduction of antibodies into cells to visualize the cellular targets requires membrane permeabilization, which prevents the practical implementation of antibodies against intracellular targets for therapeutic purposes. Thus, the development of convenient technologies for the targeted delivery of antibodies, their fragments, or antibody-like molecules inside living cells without harming the cells is urgently needed. This overall task contains several stages, such as target-cell recognition, internalization, and transport to a relevant intracellular compartment. To date, the relevant literature details different approaches to the delivery of antibodies to their intracellular targets: transfection, cell-penetrating peptides, fragments of bacterial toxins, lipid-based molecules, physical methods (electroporation and microinjection). None of these methods have provided a complete solution for all the necessary stages of intracellular transport. In this review, we have analyzed different approaches intended for recombinant antibody delivery and discussed their limitations and pitfalls.

3One possible way to engage in the therapy of intracellular “undruggable” protein targets is the use of antibodies. Antibodies could be raised against different protein antigens, including those with flat contact surfaces that are not druggable via small-molecule approaches. Since the first therapeutic antibody entered the market in 1986, various antibodies against different targets have been developed and more than 60 antibody therapeutics have been approved. The major advantages of antibodies are supreme specificity and affinity for a target, which make them invaluable basic experimental tools, as well as the standard of care, for indications, such as rheumatoid arthritis, psoriasis, solid tumors, and blood cancers. The considerable success of antibody therapy in the clinic has sparked a major effort into antibody research, including the development of engineered humanized molecules with improved physicochemical properties and safety. Similarly, recent large-scale antibody generation projects established a pipeline for the high-throughput development and validation of binder proteins for numerous possible targets.

4As a potent tool with wide applications, the antibodies can capture splice variants and proteins subjected to post-translational modifications, as well as different protein isoforms and conformational variants. For fundamental purposes, this approach might allow better clarification of the activity of intracellular proteins in their natural environment for better elucidation and validation of disease mechanisms. Furthermore, antibodies that can interact with a particular domain of a target protein are more specific than approaches such as gene knockout, transfection, and changes in gene expression. To date, most antibodies are limited to cell membrane targets, such as receptors or secreted proteins, such as cytokines, growth factors, and hormones. The number of transmembrane and secreted proteins is estimated to be approximately 8000–9000 proteins, which implies that two thirds of the proteome could be unlocked for antibody modulation in vivo. To this end, the antibody must be delivered inside the living cell.

5The features of antibodies, such as large size, complex architecture and molecular composition, and costly manufacturing, have catalyzed a significant research effort to finding alternative binder molecules based on smaller antibody fragments or alternative protein scaffolds. Thus, 150-kDa multidomain antibodies produced in eukaryotic expression systems have inspired the search for smaller antibody fragments amenable for specific target recognition and suitable for less expensive production in Escherichia coli (Figure 1). The most common, 26-kDa ScFv and 45-kDa Fab, are characterized by lower expression costs but also by rapid clearance than the parent molecules. However, they still contain intradomain disulfide bonds, which hamper correct folding upon expression in E. coli. Some natural antibodies (for example, those discovered in camels) contain 15-kDa heavy-chain only antibodies lacking a disulfide bond (dsAb or VHH, also known as Nanobodies®). VHH antibodies are possible alternatives to conventional antibodies, which have good stability, solubility, and expression yields. In addition, dozens of antibody-mimic types have been described through advances in combinatorial selection techniques. The most advanced clinical stage scaffolds are a 10-kDa monobody (Adnectin) based on the human fibronectin FN3 domain, a 6-kDa affibody based on the Z-domain of staphylococcal protein A, 20-kDa anticalin derived from human lipocalins, and the 14- to 21-kDa designed-ankyrin-repeat proteins (DARPins) engineered consensus sequence based on ankyrin repeat proteins. Alternative antibody scaffolds are much smaller in size, and combine a disulfide-free single domain with sufficient affinity and easy production in different hosts. A feature of small-size alternative binders and antibody fragments is their short circulation time in the blood, which may be advantageous for diagnostic functionality. Additional opportunities are presented by a class of bispecific antibodies able to recognize simultaneously two different epitopes on the same or different antigens. This concept was proposed for classical immunoglobulins G (IgGs) in 1983 but has been challenging owing to the complexities of their production. Expansion of antibody fragments could push this field forward, which has become a “hot topic” in research since 2010, with more than 60 molecules in development.

6The interior part of a cell is not a single cavity inside the cell plasma membrane, but a sophisticated diversity of compartments separated by intracellular membranes. Therefore, the detection of a substance within a cell can mean quite different things, and the terms “intracellular transport” or “transport into a cell” for an antibody, as well as for any macromolecules, may also have different meanings. In many cases, this transport is limited to entrapment within the compartments of the endocytosis/transcytosis pathway when molecules entering a cell remain within closed membrane formations, from which there were three common pathways: a recirculation path back to the external environment; transcytosis to the other side of the cell; and a degradation path through the lysosomes. However, the most in-demand approach is the transport of antibodies into the cytosol, which may or may not be followed by their subsequent transport to other compartments (e.g., the nucleus), where the majority of potential therapeutic targets are concentrated.

7Although it is generally believed that antibodies do not penetrate the plasma membrane, there is a considerable data pool that illustrates the presence of various antibodies within cells as a result of antigen-driven or antibody-driven cell entry of the free antibody or antibody–antigen complex.

8Antigen-driven cell entry can be observed for antibodies bound to antigens of several viruses and bacterial toxins that naturally penetrate into cells, mainly through endocytosis followed by subsequent endosomal escape to the cytosol, where they are destroyed by the intracellular Fc-receptor TRIM21 and direct intracellular pathogens into proteasomes. TRIM21 is a widely expressed E3 ligase recognizing the antibody Fc domain with high affinity (600 pM after TRIM21 dimerization). The function of this system is to protect the cell against intracellular pathogens, such as viruses and bacteria, internalized with attached host antibodies. The delivery of an antibody against a particular cell protein could mobilize the TRIM21 system to act against the endogenous proteins. Nevertheless, to take advantage of the TRIM21 system for the disruption of a particular protein in target cells seems to be a tall order, requiring the delivery of full-sized-antibodies inside living cells.

9The intracellular localization of antibodies and their complexes with antigens that enter the cell by antibody-driven mechanisms [e.g., via several types of Fc-receptors] is initially normally limited to endocytic vesicles, such as endosomes or phagosomes. Subsequent antibody recycling through neonatal Fc receptors FcRn, and transcytosis or degradation in lysosomes, also prevent antibodies from crossing the membrane and entering the cytosol. However, a recently elucidated mechanism of antigen presentation in dendritic cells, including Sec61-mediated antigen transport from endosomes in the cytosol for cross-presentation, might allow the antibody, if already bound to antigen when internalized by the dendritic cell, to enter the cytosol.

10As already noted above, the property of most antibodies that limits their applications to targeting extracellular proteins is their inability to cross cellular membranes to reach the cytosol and compartments such as the nucleus. One exception is the small subset of naturally occurring mouse and human autoantibodies that target host self-antigens during autoimmune diseases, such as multiple sclerosis, systemic lupus erythematosus, and Sjogren syndrome and artificial ones based on them. It has been reported that these autoantibodies possess an intrinsic ability to penetrate living cells and capture intracellular and intranuclear antigens, such as DNA, histones, ribosomal protein P, nuclear ribonucleoproteins, and others. Penetration into different types of cells has been described, but it was not antibody isotype-specific. In some cases, autoantibody internalization causes cell apoptosis or cytokine release, which contributes to the inflammation associated with autoimmunity. Several mechanisms of internalization have been suggested, including the interactions of basic residues in the complementarity-determining region of autoantibodies with a negatively charged cell surface, Fc receptor-mediated entry, or heparin sulfate proteoglycan-mediated clathrin-dependent endocytosis, or via caveolae/raft-dependent endocytosis of nucleoside transporter ENT2. The ability of some autoantibodies to cross the cell membrane was explored with 3D8 and 3E10 anti-DNA autoantibodies as delivery vectors for attached payloads such as nanoparticles and proteins. However, a deep understanding of the penetration mechanisms of autoantibody and the extent to which they may be modified is still lacking.

11Except for the above-described, relatively limited group of antibodies that can penetrate cell membranes, most antibodies targeted to promising intracellular targets do not demonstrate cell-penetrating ability, making their delivery inside a cell a key bottleneck. Various technologies have been developed and tested to accomplish this task; some success has occurred. Those strategies can be grouped into several major classes: direct physical delivery, direct intracellular expression, fusion with the part of internalizing autoantibodies responsible for their intrinsic ability to enter the cells, the use of protein-transduction domains or their mimics, and the use of various nanoparticle carriers (including inorganic nanoparticles, liposomes, polymersomes, and viral envelopes). These main approaches are schematically presented in Figure 2.

12One of the most straightforward and oldest strategies of intracellular antibody delivery is their direct physical transfer into the target cell; the predominantly exploited methods include electroporation and microinjection. These methods are successfully used in protein functional studies, although a rising number of studies aimed at establishing an in vivo electroporation technique, primarily for nucleic acid delivery, including encouraging clinical trials (NCT01440816) makes their practical implementation theoretically conceivable for some local applications. Recently, several new physical delivery approaches have been recruited for intracellular antibody delivery. For example, anti-E6 HPV 16 oncoprotein antibody delivered by sonoporation, which is based on the simultaneous use of high intensity focused ultrasound with microbubbles, was shown to restore p53 expression, whose degradation is promoted by the E6 oncoprotein. A new sophisticated method called the biophotonic laser-assisted surgery tool, relying on laser-induced cavitation bubbles for the creation of transient plasma membrane pores, was demonstrated to deliver different cargoes, including antibodies, into cells. In addition, an approach utilizing rapid cellular membrane deformation (microfluidics) was used for the delivery of a range of different macromolecules, including anti-tubulin antibodies, into live HeLa cells. All these physical delivery methods offer the advantage of direct antibody delivery inside the cytosol with the ability to target other compartments (e.g., the nucleus) upon the attachment of the specific localization signal, as shown for the nuclear delivery of anti-PCNA antibody conjugates with NLS delivered into the cells via electroporation. Possible limitations of these methods include a lack of cell-specificity, in addition to rather difficult in vivo, and difficulties in clinical translation. Examples of different approaches to intracellular physical delivery of antibodies are presented in Table 1.

13In addition to the physical delivery described above, a well-established straightforward method for intracellular antibody delivery is the so-called intrabody technique. An intrabody is an antibody or its derivative produced within the same cell where the antigen is located. Intrabody expression can be achieved by cell transfection with a plasmid or virus carrying the gene encoding the antibody or its fragment. This technology has become more attractive since one of the pioneering proof-of-principle works demonstrating the inhibition of alcohol dehydrogenase I in Saccharomyces transfected with cDNA-coding antibody was published in 1988. The advantages of this approach are the direct expression within the cell and relatively easy direction of intrabody to the desired cell compartment where the specific antigen should be bound (e.g., membrane or secreted protein in the ER or nuclear protein in the nucleus). The latter is usually achieved by the attachment of a targeting sequence. The targeting of intrabodies to be retained in the ER appears to be the most straightforward approach, enabling their maturation and folding in the native environment.

14A secretory signal peptide targets intrabody to the ER lumen, and ER retention by KDEL (Lys-Asp-Glu-Leu) (or similar) sequence introduced into the C-terminus of the intrabody traps the intrabody–antigen complex within the ER–Golgi complex. These types of intrabodies do not need to be strictly neutralizing to exert their knockdown functions, as the mere act of retention of secretory or membrane proteins within the ER knock down their functionality. In contrast, intrabodies aimed at cytosolic, nuclear, or mitochondrial antigens require much more precise tuning to provide for their appropriate folding within the cytosol. To overcome this limitation, many different strategies for the generation and/or selection of suitable antibodies for cytosolic expression have been developed. These approaches include: use of single-domain antibody variants [camelid VHH, shark-derived variable new antigen receptors, or variable domain of the heavy chain (VH) or light chain (VL) selected from human antibodies]; fusion with different protein domains [maltose binding protein, Fc-domain, proteasome-targeting PEST motif, and others; grafting of complementary determining regions; construction and selection of intrabodies lacking S–S bonds; numerous eukaryotic in vitro selection-based strategies aimed at isolating a functional and soluble antibody fragment and even electrostatic manipulation via the introduction of negative charges. In addition to direct interaction with the antigen located in the cytosol, simple fusion with an appropriate signal peptide can redirect intrabodies expressed in the cytosol to bind the antigens within the cell nucleus or mitochondrion.

15As phage display technology allows rapid and relatively easy generation of antibodies to almost any antigen of interest, intrabodies aimed at many different proteins, including their splice variants and specific post-translational modification sites, have been developed and shown to be rather efficient in vitro and sometimes in vivo. The most widely utilized targets include cancer-linked antigens, neurodegenerative-disease-related antigens, toxins, viral infection (e.g., HIV or HCV) – related antigens, and miscellaneous antigens, mostly for protein function studies, within the cell. The therapeutic potential of intrabodies against cancer, Huntington’s disease, and Alzheimer’s disease has been already demonstrated in mouse models of these diseases. Moreover, in a phase I clinical trial conducted in 2000, the feasibility of adenoviral-mediated gene therapy using an anti-erbB-2-directed intrabody in the context of human ovarian cancer was shown: the authors reported anti-erbB-2 scFv-encoding gene expression measured by polymerase chain reaction and real-time polymerase chain reaction in most patients with very limited toxicity. However, contrary to the effects demonstrated by an anti-erbB-2-directed intrabody in animal studies, no dramatic clinical benefit was observed in this trial, leaving much room for the improvement of this strategy.

16Thus, the attractive potential therapeutic use of intrabodies still has unresolved issues; the major issues are finding an efficient and safe DNA transfection method in vivo and insufficient tissue specificity. The other issues include the delayed onset of the effect, uncertain duration and level of expression, and whether the transformation is transient or permanent (especially when using such vectors as polyplexes, liposomes, adenoviruses). However, extensive research into the development of promising new nucleic acid delivery systems, including tissue-specific variants for various gene therapy applications have led to a number of clinical trials and a few already clinically approved approaches favor the introduction of intrabody-based therapy to the clinic.

17Although the intrabody approach relies on straightforward intracellular expression within the same antigen-expressing cells, another strategy is based on the use of the intrinsic ability of several naturally occurring autoantibodies (see Section “Internalization of Antibodies Into Cells Via Normal and Pathological Processes”) to enter the cells. The most straightforward pathway utilizes the recently discovered potential of a subset of cell and nuclear penetrating lupus erythematosus anti-DNA autoantibodies to serve as therapeutic agents targeted toward DNA repair-deficient malignancies. Specifically, the lupus erythematosus anti-DNA autoantibodies 3E10 and their more potent divalent mutants, including humanized and re-engineered ones, as well as the nucleolytic autoantibody 5C6 were shown to bind DNA and either inhibit key steps in DNA repair or damage single-stranded DNA in a manner, making them selectively lethal to cancer cells with defective homology-directed repair of DNA double-strand breaks.

18A more flexible and universal approach recruits the fragments responsible for cell entry of those cell penetrating autoantibodies for fusion with antigen-recognizing parts of the antibody of choice. A series of articles have reported proof-of-principle of the feasibility of this approach to the demonstration of significant antitumor effects in vivo. In particular, humanized VL of lupus erythematosus autoantibody m3D8, with its ability to penetrate the cell and localize in the cytosol, has been engineered into different human IgGs to give so-called cytotransmabs capable of targeting an intracellular antigen of choice within the cytosol. Based on this strategy, the human IgG1 format antibody named iMab RT11-i containing m3D8 humanized VL targeted to the cytosolic activated GTP-bound form of oncogenic Ras mutants has been developed. Decorated with tumor-homing integrin binding RGD10 cyclic peptide, this iMab inhibited the growth of oncogenic Ras-mutated tumor xenografts in mice, but not wild-type Ras-harboring tumors, demonstrating the feasibility and potential of this approach for the cytosolic delivery of antibodies. Nuclear delivery of antibodies has been demonstrated by using a bivalent scFv derivative of cell and nuclear penetrating autoantibody 3E10 fused to an anti-MDM2 antibody, which impaired the growth of melanoma cells and melanoma xenografts sensitive to MDM2 inhibition.

19Contrary to the limited number of studies utilizing the part of internalizing autoantibodies responsible for their intrinsic ability to enter the cells, a somewhat similar approach relying on fusion to protein transduction domains, or CPPs, is a widely utilized technique to shuttle various cargos (proteins, plasmid DNA, RNA, oligonucleotides, liposomes, imaging agents, and anti-cancer drugs) into living cells. Since the discovery of the first and prototypical HIV 1 trans-activating (TAT) protein, capable of cell membrane penetration in 1988, a wide variety of CPPs and some of their mimics was derived and recruited for the intracellular delivery of various agents for research studies or potential therapeutic applications, with a few reaching clinical trials to date (NCT01975116, NCT00914914). Unsurprisingly, this promising strategy has also been attempted to enhance intracellular antibody delivery, which has led to cell penetrating TransMabs or transbodies. An early study, revealing enhanced cell retention and internalization of Fab fragments decorated with a TAT protein-derived peptide, was followed more than a decade later by a set of articles demonstrating that antibody-based fusion proteins containing a short membrane transport-facilitating peptide were transported into the cells and bound to intracellular structures and that anti-caspase-3 antibody based TransMabs were able to inhibit apoptosis in cells. Subsequently, this approach gained popularity and the number and use of CPP-conjugated whole antibodies, and their derivatives, started to increase; see Table 2 for examples. Recently, a transbody based on mAbs against human HBcAg coupled with the TAT protein transduction domain was demonstrated to be effective in vivo.

20However, there are two main limitations to be addressed when using this strategy. The first stems from the pool of data that demonstrate entrapment of the internalized CPPs and CPP-conjugates inside endocytic vesicles, which has raised the question as to whether CPPs and CPP-conjugated macromolecules, including transbodies, are able to be released from the endosomes efficiently, making efficient endosomal escape one of the limitations of this approach. The insertion of an additional endosomolytic moiety was suggested to improve the situation. A recent study reported an efficient endosomal escape of an antibody co-delivered with an endosomolytic peptide derived from the cationic and membrane-lytic spider venom peptide M-lycotoxin. An approach utilizing light-controlled endosomal photosensitizer-driven escape of the construct based on an anti-Ki67 mAb decorated with CPP Pep-1 co-delivered with photosensitizer was revealed as an effective strategy to facilitate the endosomal escape of antibodies.

21Another limitation of CPP-based technology, including transbodies, may be the absence of tissue and cellular specificity. Although, in the case of transbodies, cell-specific action can be reached on the antigen level, the problem of unnecessary unspecific delivery to cells remains, but may be solved by using some cell-homing CPPs or the addition of a cell-targeting ligand to a transbody. For example, a fusion of anti-mutated K-ras scFv and cancer-cell specific CPP BR2 demonstrated significant and cancer-cell-selective effects in vitro.

22Nanocarrier-driven targeted intracellular delivery of different payloads is a widespread strategy that has the following advantages: allows systemic delivery of a relatively large amount of cargo with reduced cargo-related side effects; lowers off-target organ toxicity; enables the possibility of sustained release; and shields the cargo from premature enzyme degradation. Moreover, nanocarriers permit relatively easy decoration with functional (e.g., cell-targeting) moieties, which, in contrast to direct antibody modification, cannot hamper their reactivity. Thermosensitive, enzyme-sensitive, pH-sensitive, light-responsive, and combo-stimuli-responsive nanocarriers can be easily designed to provide a controlled payload release in response to thermal (body, increased tumor temperature, or localized heating), specific enzymatic (e.g., tumor microenvironment or intracellular enzymes) and specific pH microenvironments (e.g., acidic tumor microenvironment or acidifying endocytic vesicles) as well as to external light exposure or any combination of these.

23Various types of nanocarriers (see Table 3 for examples), including inorganic nanoparticles, polymersomes, liposome-based nanocarriers, as well as modified viral and virus-like particles have all been recruited for antibody intracellular delivery with reasonable success.

24Lipid-based nanoparticles are extensively used for the intracellular delivery of therapeutic proteins, including antibodies. Several cationic-lipid based nanocarriers have been proposed for the intracellular delivery of antibodies. For example, Courtete et al demonstrated that a lipid reagent for siRNA transfection could be successfully adjusted to give efficient anti-HPV16 E6 antibody delivery and resulted in the downregulation of this oncoprotein activity in vivo.

25Polymer based nanoparticles are another widely used approach for intracellular protein delivery. Several years ago, polymer-based liposomes, called polymersomes were shown to be effective for the cytosolic delivery of antibodies aimed at various intracellular antigens inside the cells, resulting in the respective epitope binding and subsequent biological effect in different cell types, as no cell-specific moiety was required in this proof-of-principle study. A few years later, a sophisticated polymersome-based approach for cell-specific antibody delivery across the blood–brain barrier was developed. Polymersomes were decorated with Angiopep-2 peptide targeting low density lipoprotein receptor-related protein 1 receptor to yield dual functionality, as this receptor is associated with both endothelial transcytosis and endocytosis into the cells of the central nervous system. As a result, Angiopep-2-decorated polymersomes delivered IgGs into the central nervous system cell in vitro, efficiently crossed the blood–brain barrier, and co-localized with astrocytes, neurons, and glial cells in vivo; thus, this approach warrants further development.

26Inorganic nanoparticles composed of several inorganic materials have been used for the intracellular delivery of antibodies. For example, mesoporous silica nanoparticles functionalized with nitrilotriacetic acid–metal ion complexes for pH-sensitive binding and release of the delivered His6-tagged chromobodies (anti-GFP fluorophore labeled nanobodies) exhibited the ability to deliver these chromobodies inside the cells. A noticeable release of chromobodies to cytosol required the use of endosomal release triggers, of which chloroquine and dimethyl sulfoxide were the most effective ones, which may be unsuitable for practical use. Another group developed mAbs to anti-influenza A and anti-Dengue viral proteins biomineralized with calcium phosphate, yielding sphere-like 130–150 nm nanoparticles, which, in contrast with native unmodified mAbs were able to enter the cells and co-localize with their antigens within the endosomal compartment. Moreover, the biomineralized mAb led to a significant reduction in the mortality of the infected mice compared with the corresponding naked mAb. Both these nanoparticle-based delivery systems possessed difficulties in endosomal escape of the delivered antibody that necessitated the use of endosomal release triggers or were limited to antigens that are known to localize within the endosomes. An easy way to reach the cytosol is to enter the cell by a non-endocytosis-involving pathway, as recently demonstrated for oxidized carbon black nanoparticles, which are supposed to cause temporary disruption of the lipid bilayer and thus provide antibody delivery inside the cells via an electroporation-like mechanism. Another solution to facilitate the delivery of antibody-bearing nanoparticles into the cytosol been proposed by, who decorated gold-coated iron oxide nanoparticles with IgG-binding streptococcal bacterial protein G fused with TAT peptide for endosomal escape. The resulting functionalized magnetic nanoparticles loaded with anti-mitochondria mAb were able to enter the cell and reach mitochondria within HeLa cells. However, the efficiency of TAT peptide-mediated endosomal escape is supposed to be rather low; thus, in order to improve endosomal release, another group developed an approach that recruited gold nanoparticles decorated with TAT peptide fused with pH-sensitive influenza virus hemagglutinin protein HA2; in this, the former moiety was responsible for cell-entry and the latter for the introduction of efficient endosomal release of the delivered anti-actin antibody.

27The main limitation of all these approaches is their non-specific action; however, this may be solved by additional inclusion of nanoparticles with a cell-homing moiety.

28Being extensively used for gene delivery, viral-based vehicles have been recruited successfully for intracellular antibody delivery. Efficient entrapment of the antibody within these viral-based nanoparticles is often provided by the inclusion of Staphylococcus aureus protein A or its sub-domains capable of IgG binding. A decade ago, a system based on a protein A Z-subdomain dimer fused with nucleocapsid protein incorporated into the envelope of hemagglutinating virus of Japan revealed the ability to efficiently incorporate IgGs and target them to the respective intracellular antigens (nuclear pore complex and α-tubulin). Owing to the possible toxic and immunogenic issues connected with viral vectors of animal origin, intracellular antibody-delivering plant virus-like nanoparticles, which are considered to be nonpathogenic in humans, were recently developed. Based on Sesbania mosaic virus coat protein modified with IgG binding protein A B-domain self-assembling virus-like particles, this platform for intracellular delivery of antibodies proved to be effective for intracellular delivery of three different mAbs, which exerted their biological effects owing to respective epitope binding. Generally, owing to the high efficiency of antibody delivery across the membrane, combined with the high loading efficiency compared with most nanovehicles (e.g., inorganic nanoparticles or polymersomes), viral-based particle approaches will encourage new horizons of study in this field, especially if an additional ability to target only the specified cells can be introduced into this type of nanoparticle.

29A few self-assembled protein nanocarriers, polyion complex micelles, and pharmacokinetic improving erythrocyte membrane-coated mAb nanoparticles have been developed, with some exhibiting encouraging results. For example, a self-assembled protein nanocarrier based on an α-helical peptide self-assembling into a hexameric coiled-coil bundle fused with an Fc-binding Protein A fragment was recently developed. It was further decorated with an endosomolytic Aurein 1.2 sequence. This nanocarrier enabled a high antibody loading and efficient delivery into the cytosol. The authors proposed possible future modifications of this nanocarrier with cell specific targeting ligands to dramatically increase the attractiveness of this already promising nanovehicle.

30Targeted intracellular delivery of antibodies can be aimed at the antigens localized in various intracellular compartments, predominantly the ER, cytosol, and nucleus. Examples of efficient antibody intracellular delivery approaches aimed at specified intracellular compartments are presented in Table 4, and the advantages and limitations of different approaches are summarized in Table 5. The ER is very easily targeted by intrabodies, as natural antibodies already contain a signal for secretion; thus, only the additional ER retention motif (KDEL, usually) should be encoded within the antibody gene to achieve delivery into the cell, where an intrabody is expressed. Cytosolic and nuclear antigens require more effort to achieve intrabody targeting, but are still vastly applicable. The main problem with regard to the use of intrabodies is the requirement of a safe, efficient, and cell-specific gene delivery method for their in vivo use. The targeted delivery of the antibodies themselves inside the cell from the outside seems to lack many of the safety parameters associated with gene delivery, but narrows the variety of intracellular compartments that have been reached to date by the delivered antibody; to the best of our knowledge, no approaches for targeting the ER by an externally delivered antibody have yet been published. An efficient cell-specific targeted delivery of antibody for the intracellular antigen of interest from the cell outside requires a stepwise delivery, starting from the specific recognition of the target cells, followed by effective internalization of the construct, with the subsequent endosomal escape to the cytosol and – if necessary – further transport to the required cell compartment (e.g., the nucleus or mitochondria). Unfortunately, just a few published delivery strategies integrate all these crucial steps together in one construct. For example, a recent approach for the construction of modular transport systems, including DARPins, for cell specific receptor recognition, bacterial toxin-derived component for endosomal escape, and a different DARPin for intracellular antigen recognition was published. This system was demonstrated to deliver DARPins into the cytosol efficiently and cell-specifically. When an antigen of interest is localized within another compartment (e.g., the nucleus or mitochondria), additional signal peptides for specific compartment delivery are required. The delivery system that can serve an appropriate candidate to accomplish this step-by-step targeted cell-specific delivery of antibody (or its derivative/mimic) into the designated intracellular compartment (e.g., the nucleus) is the modular nanotransporters platform. Modular nanotransporters are recombinant polypeptide-based delivery vehicles, consisting of a ligand module for cell-specific recognition and subsequent internalization, the translocation domain of Diphtheria toxin as an endosomolytic module for successful endosomal escape, and the optimized SV-40 large T-antigen nuclear localization sequence for nuclear import. The feasibility of this platform has already been demonstrated in vitro and in vivo for the delivery of locally acting anti-cancer drugs, such as photosensitizers, several types of radionuclides [α-emitters, and Auger electron emitters] into various types of cancer cells. The cancer cell specificity was easily modulated by the choice of appropriate ligand module [e.g., epidermal growth factor, α-melanocyte stimulating hormone, and folic acid] for the characteristic internalizable receptor overexpressed on the target cells. Recently, the principal feasibility of the construction of modular nanotransporters bearing anti-cMyc scFv (as an example) was demonstrated, which may be a promising new step toward the efficient targeted delivery of antibodies inside the target cells.

31Another theme that emerges from the intracellular antibody delivery concept is the ability to relocate protein-of-interest within the cells or deplete it through a cellular protein elimination system. Protein relocation is an approach to interfere with target protein function through trapping it with binder molecule and thereby preventing the target protein transport to a defined cellular compartment. The feasibility of such strategy was demonstrated with intracellular antibodies, which inhibited heterochromatin protein 1β traffic to the nucleus, resulting in altered nuclear morphology and apoptosis. Another group reported intracellular antibody binding to Sec61 and prevention of its transport from the ER toward endosomes. To interfere with the target protein function, the intracellular antibody must be: (1) delivered into cells at a stoichiometric concentration; (2) bound to the target in a way that blocks its function, and (3) ensured for high affinity binding, otherwise the equilibrium will shift toward target escape from antibody complex before reaching desirable effect. These are strict conditions that limit the possible applications of intracellular antibodies against relatively low-abundant targets and/or molecules with a very high affinity toward the target.

32Notwithstanding the elimination of the target, once bound to the antibody, it could circumvent this equilibrium conundrum. For specific degradation of proteins, cells contain proteasome systems, which are highly efficient and elaborate machinery for the regulation of protein turnover and protection from misfolded and damaged proteins. A ubiquitin-dependent system is the major pathway for specific protein degradation, for which protein ubiquitination is a prerequisite step. Ubiquitination is a multistep enzymatic process, that involves three types of enzyme: E1 ubiquitin activating enzyme; E2 ubiquitin conjugating enzyme; and E3 ubiquitin protein ligase. The substrate specificity is conferred by hundreds of E3 ubiquitin ligases, belonging to HECT, RING, and U-box families, according to their mode of action. RING and U-box E3 ubiquitin ligases function as scaffolds for E2 ubiquitin conjugating enzymes, whereas HECT proteins form a thiol-ester bond with ubiquitin before transferring it to the substrates. Harnessing an endogenous protein control system is a viable approach to downregulate the target proteins. For this purpose, the binder moiety should be fused with a module targeting the protein to the proteasomes. Several examples of such reprogrammed E3 ubiquitin ligases have been described, based on multi-subunit RING E3 enzymes, targeting β-catenin or mono-molecule U-box protein CHIP E3 ligase, degrading KRAS, c-Myc, epidermal growth factor receptor mutants. Fusion with CHIP has several advantages over RING E3 ligases, such as broad substrate diversity and a lack of dependence on other subunits. By virtue of the catalytic nature of the ubiquitin system, enabling multiple rounds of activity, depletion of abundant targets, even exceeding the amount of delivered antibody may be possible.

33Three decades of progress in strategies for efficient targeted intracellular delivery of antibodies has led to the development of various approaches, from direct physical methods and gene delivery to sophisticated synthetic delivery vehicles. Although widely utilized for intracellular protein function studies, the therapeutic applications of this promising precise approach remains in the in vitro stages and rarely progresses to in vivo studies; thus, there is much room for improvement. We hope that both the elucidation of normal behavior of antibodies within cells, including their relocation and degradation pathways, as well as the development of specific, effective, and clinically applicable systems of targeted intracellular protein delivery will allow this approach to be used as a clinical therapy in future.

34TS, AU analyzed the literature, wrote the article, and designed the figures and tables in the article. AR and AS analyzed the literature, wrote the article, and critically reviewed the article.

35The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.