Novel delivery methods to achieve immunomodulation.

Immunomodulation in infectious diseases, cancer, cardiovascular disease and autoimmunity can now be targeted by sophisticated protein design, altering cellular responses by increasing therapeutic cell numbers ex vivo and then reimplanting, or altering cell function by gene transfer of cells ex vivo. In the last year, vaccination has been applied to modulate responses to autoantigens, allergens, viral or cancer antigens. The application of these technologies has entered the clinical arena and is having a positive impact on the treatment and prevention of human diseases.

Introduction

Immunomodulation can impact many pathological processes including vaccination, autoimmunity, cancer and transplantation. Using naturally occurring compounds from steroids to antibodies and cytokines has been advantageous for disease management. Yet, their use is limited due to systemic effects on healthy tissue that can lead to unwanted side effects. In this review, we want to focus on new molecular and cellular technologies that are advancing the overall aim of targeting therapy to particular disease sites or mechanisms of disease and hence are more specific.

Intracellular delivery systems will be required for all molecules that have intracellular function. For example, nucleic acid molecules including encoding genes, oligonucleotides and RNA molecules must enter cells and target the nucleus when transcription is the target. Proteins with intracellular function require delivery into cells. Regardless of the molecules for delivery, a common requirement is the evasion of endosomal uptake that may cause degradation and denaturation. Several approaches are being developed that can be applied to the delivery of all these types of molecules at disease sites. For the goal to be fully achieved, cell-targeting strategies require still further development.

We concentrate this review on the engineering of molecules and cells that are helping the development of novel therapeutics (see Table 1). We are not going into the details of the methods of gene therapy as this is beyond the scope of this review.

Table 1

Delivery systems

Delivery system	Application	References
Antibody	Targeting cytokines and toxins to specific Ag expressing cells/tissues	[1,2•,3]
Latent molecules	Specific activity of proteins at sites of MMP activity	[4•,5]
Protein transduction domains	Intracellular delivery of proteins, DNA, RNA to any cell type	[8,10,11,13]
DNA vaccination	Recombinant vaccination through DNA expression of antigens	[26]
Replicating viruses	Specific proliferation of viruses in tumour cells	[17,21•]
DNA/RNA complexes	Delivery of RNA or DNA to cells avoiding endosomal degradation	[38,39]
T bodies	T cells engineered with chimeric receptors that target them to antigen bearing tumours or other cell types/tissues	[43,44,45•,46]

Protein delivery systems

Antibodies against tumour antigens have been used for both diagnostic and therapeutic purposes (see Figure 1a). The cellular immune response against cancers can be potentiated by delivering cytokines to sites of tumour growth. For this purpose, fusion proteins between anti-tumour antigen antibodies and cytokines (immunocytokines) have been developed [1]. A humanised anti-ganglioside GD2 antibody fused with IL-2 was investigated in phase I clinical trials in melanoma [2] and prostate cancer patients [3]. In both cases increased cellular immune responses were reported.

Figure 1

Schematic representation of the structure of antibodies, scFv and chimeric scFv signaling receptors. Panel a shows the structure of an antibody with its variable heavy (VH), variable light (VL) domains that comprise the antigen recognition/binding site with their respective constant regions (CH and CL). The whole structure is kept together by disulfide bonds. Panel b shows the structure of a scFv where the VH and VL are fused via a 15 aminoacid linker of the sequence (GGGGS)3. Panel c shows the structure of scFv chimeric receptor where the scFv is extracellular and is normally linked to the cytoplasmic signaling domain of a TCR chain such as the FcRγ or the ζ chain through a spacer or hinge region.

Immunocytokines provide increased half-life to cytokines, but the antibody moiety has a decreased half-life when compared with the non-fused protein. This is probably due to the interactions of immunocytokines with cytokine receptors that are widely expressed throughout the body. In this respect, engineering latent cytokines [4] overcome the interaction with cytokine receptors. Latent cytokines are fusion proteins between the latent-associated peptide (LAP) from transforming growth factor (TGF) β followed by a matrix metalloproteinase (MMP) cleavage site and the cytokine of interest. The LAP dimer encloses the cytokine in a shell-like structure conferring extended half-life to the cytokine that cannot interact with its receptor(s) until released from LAP at sites of high MMP activity found in pathologies with local inflammation and active tissue remodeling such as autoimmune disease, atherosclerosis and cancer.

Protein transduction domains (PTDs) are peptide sequences that can penetrate cell membranes independent of interaction with specific receptors or transporters. The first characterised PTD is the third α helix of Antennapedia homeodomain protein [6] whilst a region of HIV Tat containing a basic domain is probably the most widely applied PTD [7] (for review on PTDs see Reference [8]). The uptake mechanism for Tat is via a lipid raft-dependent macropinocytosis mechanism [9]. Following cell entry Tat targets molecules to the nucleus because of a strong nuclear localisation signal unless this signal is overridden by a nuclear export signal in the cargo molecule. In view of PTDs’ lack of immunogenicity, and versatility as carriers of a variety of molecules (proteins, peptides, siRNA, DNA–protein complexes and viruses) of unlimited size, their clinical application seems likely in the near future. The versatility of these molecules is illustrated in a recent study where PTDs were used to deliver pre-mRNA to correct aberrant splicing of a gene [10]. Also Bim, an antagonist of Bcl-2-mediated cell survival, was delivered to tumour cells with Tat that significantly slowed down tumour growth in murine models of pancreatic cancer and melanoma [11]. Another potential strategy for cancer treatment utilising a PTD is delivery of HSV-TK fused with Tat (TK-Tat), and in this in vitro study TK-Tat localised to the nucleus was stable and cells were sensitive to ganciclovir treatment with good bystander killing observed [12]. PTDs have been employed for immunomodulation in experimental models of inflammation; for example, the delivery of the Tat-IκBα superrepressor in a rat pleurisy model caused a reduction in both leucocyte recruitment and production of proinflammatory cytokines [13]. One interesting strategy is the use of a short amphipathic peptide carrier, Pep-1, which can be linked non-covalently to cargo molecules through Fmoc, which then dissociates immediately after it has crossed the cell membrane [14] and the liberated cargo can then distribute with normal tropism without influence from the attached PTD.

Human papilloma virus (HPV) has proven hard to eradicate because of ineffective means of in vitro culture. Cloning of the HPV capsid gene (L1) and its expression in heterologous systems [15] has allowed the development of effective vaccines. Hence, prophylactic vaccination to HPV types 6/11/16 and18 to prevent cervical cancer is becoming a reality [16].

New viral gene delivery systems

Somatic gene therapy was originally thought to be safer using non-replicating viruses. However, for cancer treatment scientists have now developed attenuated viruses that are capable of replicating more efficiently in cancer cells than in normal cells. These oncolytic viruses could in principle infect neighbouring cancer cells and lyse them sparing the normal tissue. A wide variety of RNA and DNA viruses are being investigated [17]. Replicating poxvirus can deliver antiangiogenic factors both as secreted factors or as shRNA [18,19] or cytokines such as GM-CSF [20]. Recently, the data of a phase I clinical trial in prostate cancer using intravenous injection of a replication-selective, prostate-specific antigen-targeted oncolytic adenovirus were published with some positive results [21].

The use of oncolytic viruses, the relevance of the animal models and the environmental safety issues involved raised some concerns [22]. The immune system appears to be dealing quite effectively with these oncolytic viruses because most humans are preimmune to these viruses either via natural exposure or by vaccination. Whether the antiviral immune response will allow for effective cancer therapy will need to be resolved in clinical trials.

Non-viral delivery systems

Plasmid DNA has no innate mechanism to enter cells, but direct injection in skeletal muscle achieves transfection and gene expression in most species from rodents [23] to humans [24]. This expression has led to the effective application of plasmid vectors in numerous vaccination studies conducted in rodents. DNA-based vaccines have several advantages including their simple preparation and stability. Vaccination is more flexible because the plasmid encoding the antigen can be combined with other genes that modify the immune response. Plasmid DNA also has immunostimulatory properties due to unmethylated CpG repeats that interact with TLR9 receptors expressed on antigen-presenting cells (APC), although this is not a prerequisite for vaccination as responses are also observed in TLR9-knockout mice [25]. Interestingly, the effectiveness of DNA vaccines in small animals has not translated to primates and humans and proof of activity in clinical trials has only recently been reported. In a phase I trial a DNA vaccine for bird flu was shown to be safe and achieved antibody responses at tested doses [26]. Transfection of plasmid DNA in skeletal muscle is efficiently enhanced in rodents by the use of electroporation that opens pores in the cells to permit direct entry of DNA [27]. Electroporation can enhance DNA vaccination [28], but there are few reports using electroporation in large animals. However, hydrodynamic delivery of plasmid DNA or siRNA into the vasculature of an occluded limb has proven equally effective in primates and rodents [29].

Nucleic acid molecules also need to be delivered intracellularly for function to the nucleus (for gene transcription, or for transcriptional decoy effects) or to the cytoplasm for inhibition of RNA translation or to target mRNA degradation (antisense RNA, ribozymes). RNA interference (RNAi) is an endogenous mechanism whereby double stranded RNA is processed by the RNAse III-like protein Dicer to produce short interfering RNA (siRNA) that are incorporated into the RNA-induced silencing complex (RISC) [30]. Synthetic siRNA are structurally related to endogenous microRNAs (miRNA) and can be used for sequence specific silencing. Bevasiranib, an siRNA targeting VEGF mRNA, is already in phase II clinical trials for the treatment of wet age-related macular degeneration [31]. These small molecules were applied in a variety of experimental models in saline, complexed with lipids or conjugated with molecules for improved pharmacokinetics or targeting [32]. Another interesting aspect of these molecules relevant to immunomodulation is their interaction with Toll-like receptor (TLR)-7 through which they can co-deliver an immunostimulatory response. In a recent study, immunostimulation was further enhanced with an siRNA targeting expression of IL-10 in combination with the TLR-7 stimulation by the molecule [33].

siRNA was used in arthritis models for targeting TNFα. siRNA was delivered to knee joints in combination with electroporation [34,35] or systemically complexed with cationic liposomes [36]. Both approaches required re-administration. This transient action is clearly a shortcoming for long-term effects in the treatment of chronic diseases such as rheumatoid arthritis. An alternative is the use of short hairpin RNA (shRNA) molecules expressed long term from gene delivery vectors. shRNA are additionally processed in the nucleus by an RNAse III protein (Drosha) producing pre-miRNA which are exported to the cytoplasm for further processing by Dicer to produce mature miRNA.

Ribozymes are catalytic RNA molecules that cleave mRNA sequences. Their design is more complex than siRNA. Recently, a hammerhead ribozyme targeting TNFα was shown to inhibit a model of arthritis when delivered intravenously before onset of the disease [37].

Naked DNA and RNA molecules can be combined with a variety of polycations in polyplexes for delivery to cells. The polycations used have different characteristics including ease of DNA unpackaging that is influenced by their molecular weight, toxicity, stability and ability to facilitate endosomal escape. Polyethylenimine (PEI), for example, is a polycation with strong endosome escape properties. To decrease the toxicity of polycations and retain their stability, reducible polycations are used that are specifically cleaved within cells to release the DNA. Synthetic vectors based on reducible (thiol-containing) polycations consisting of histidine and polylysine residues were shown to efficiently deliver DNA, mRNA and siRNA in a variety of different cells [38]. The reducible approach has also been used to coat PEI complexes [39].

To target liposomes, phage peptide libraries are used to screen for tissue-binding peptides in vivo, for example, in cancer [40] or synovial endothelium in arthritis [41]. These selected peptides can be conjugated to liposomes in order to target their cell of interest.

Cell delivery systems

The principle of immunosurveillance has always been central to immunotherapy approaches for the treatment of cancer. Despite the myriad of mechanisms by which cancer cells evade immunosurveillance, it has now been shown that infusion of autologous tumour antigen-specific T cells expanded ex vivo are therapeutic [42].

Another method for T cell therapy of cancer is the development of T bodies [43] in which chimeric receptors with extracellular scFv (see Figure 1b) of antibodies against tumour antigens are grafted onto the cytoplasmic and signaling domains of T cell receptor (TCR) subunits (see Figure 1c). The direct recognition of antigen by the scFv obviates the need for antigen processing and presentation by the major histocompatibility complex. In a clinical trial using T bodies, targeting of the tumour sites was poor and survival of the engineered cells very limited [44]. This may be due in part to the fact that the scFv used was of mouse origin. Similar work has also been done by cloning α and β chains of TCR from tumour infiltrating lymphocytes with some degree of success in melanoma [45]. It is of interest that the fate of the endogenous TCR chains of the transduced cells is unknown. Whether the endogenous TCR chains recognise autoantigens by themselves or by combination with the transduced TCR chains could have consequences for autoimmunity and needs a long-term follow-up of these patients.

Mesenchymal stromal cells (MSC) have immunosuppressive properties. Transplantation of MSC has been shown to inhibit steroid-resistant grade III–IV graft versus host disease (GvHD) in patients receiving allogeneic stem cell therapy [47]. MSC seem to affect the cytokine secretion profile of dendritic cells (DCs), naive and effector T helper cells, and natural killer (NK) cells to induce a more anti-inflammatory or tolerant phenotype [49].

GvHD is mediated by T cells from donor origin. T cells can be retrovirally transduced, after cell cycle activation ex vivo with anti-CD28/CD3 or phytohaemaglutinin, to express the ‘suicidal’ gene thymidine kinase from Herpes simplex virus (HSV tk) that renders dividing T cells sensitive to ganciclovir, and this approach has been used in clinical trials [50,51]. Recently, it has been shown that using IL-7-mediated proliferation of T cells preserves the effector function of T cells more effectively than TCR-mediated protocols for viral transduction [52].

Conclusions

The ingenuity of scientists and the exponential growth in the understanding of molecular mechanisms involved in different pathologies have increased substantially the possibilities for immunomodulation. The potential of novel approaches such as gene therapy and drug design enable targeting of therapeutics with better therapeutic index. The limitation of these therapies resides in our better understanding of the long-term outcome of delivering biologicals and manipulating the immune system. The immune system is finely tuned in health to react to foreign pathogens and oncogenic changes of cells in the body. Tilting the immune system with therapeutic agents to behave in a particular way is not without consequences. The challenge for immunomodulating therapies will be to obtain long-lived therapeutic outcomes with shorter therapeutic interventions. This may require the combination of immunotherapies alone or with stem cell therapies.

References and recommended reading

Papers of particular interest, published within the annual period of review, have been highlighted as:

• of special interest

•• of outstanding interest