Evaluation of promoters for driving efficient transgene expression in neonatal porcine islets

Characterisation of transgenic pigs expressing a human T cell-depleting anti-CD2 monoclonal antibody

BACKGROUND

Pig islet xenotransplantation is a potential treatment for type 1 diabetes. We have shown that maintenance immunosuppression is required to protect genetically modified (GM) porcine islet xenografts from T cell-mediated rejection in baboons. Local expression of a depleting anti-CD2 monoclonal antibody (mAb) by the xenograft may provide an alternative solution. We have previously reported the generation of GGTA1 knock-in transgenic pigs expressing the chimeric anti-CD2 mAb diliximab under an MHC class I promoter (MHCIP). In this study, we generated GGTA1 knock-in pigs in which MHCIP was replaced by the β-cell-specific porcine insulin promoter (PIP), and compared the pattern of diliximab expression in the two lines.

METHODS

A PIP-diliximab knock-in construct was prepared and validated by transfection of NIT-1 mouse insulinoma cells. The construct was knocked into GGTA1 in wild type (WT) porcine fetal fibroblasts using CRISPR, and knock-in cells were used to generate pigs by somatic cell nuclear transfer (SCNT). Expression of the transgene in MHCIP-diliximab and PIP-diliximab knock-in pigs was characterised at the mRNA and protein levels using RT-qPCR, flow cytometry, ELISA and immunohistochemistry. Islets from MHCIP-diliximab and control GGTA1 KO neonatal pigs were transplanted under the kidney capsule of streptozotocin-diabetic SCID mice.

RESULTS

NIT-1 cells stably transfected with the PIP-diliximab knock-in construct secreted diliximab into the culture supernatant, confirming correct expression and processing of the mAb in β cells. PIP-diliximab knock-in pigs showed a precise integration of the transgene within GGTA1. Diliximab mRNA was detected in all tissues tested (spleen, kidney, heart, liver, lung, pancreas) in MHCIP-diliximab pigs, but was not detectable in PIP-diliximab pigs. Likewise, diliximab was present in the serum of MHCIP-diliximab pigs, at a mean concentration of 1.8 μg/mL, but was not detected in PIP-diliximab pig serum. An immunohistochemical survey revealed staining for diliximab in all organs of MHCIP-diliximab pigs but not of PIP-diliximab pigs. Whole genome sequencing (WGS) of a PIP-diliximab pig identified a missense mutation in the coding region for the dixilimab light chain. This mutation was also found to be present in the fibroblast knock-in clone used to generate the PIP-diliximab pigs. Islet xenografts from neonatal MHCIP-diliximab pigs restored normoglycemia in diabetic immunodeficient mice, indicating no overt effect of the transgene on islet function, and demonstrated expression of diliximab in situ.

CONCLUSION

Diliximab was widely expressed in MHCIP-diliximab pigs, including in islets, consistent with the endogenous expression pattern of MHC class I. Further investigation is required to determine whether the level of expression in islets from the MHCIP-diliximab pigs is sufficient to prevent T cell-mediated islet xenograft rejection. The unexpected absence of diliximab expression in the islets of PIP-diliximab pigs was probably due to a mutation in the transgene arising during the generation of the knock-in cells used for SCNT.

Sustained heterologous gene expression in pancreatic islet organoids using adeno-associated virus serotype 8

Genetic modification of pancreatic islet organoids, assembled in vitro prior to transplantation is an emerging alternative to direct in vivo genetic manipulations for a number of clinical and research applications. We have previously shown that dispersion of islet cells followed by re-aggregation into islet organoids, or pseudoislets, allows for efficient transduction with viral vectors, while maintaining physiological functions of native islets. Among viruses currently used for genetic manipulations, adeno-associated viruses (AAVs) have the most attractive safety profile making them suitable for gene therapy applications. Studies reporting on pseudoislet transduction with AAVs are, however, lacking. Here, we have characterized in detail the performance of AAV serotype 8 in transduction of islet cells during pseudoislet formation in comparison with human adenovirus type 5 (AdV5). We have assessed such parameters as transduction efficiency, expression kinetics, and endocrine cell tropism of AAV8 alone or in combination with AdV5. Data provided within our study may serve as a reference point for future functional studies using AAVs for gene transfer to islet cell organoids and will facilitate further development of engineered pseudoislets of superior quality suitable for clinical transplantation.

Cellular Immune Responses in Islet Xenograft Rejection

Porcine islets surviving the acute injury caused by humoral rejection and IBMIR will be subjected to cellular xenograft rejection, which is predominately mediated by CD4⁺ T cells and is characterised by significant infiltration of macrophages, B cells and T cells (CD4⁺ and CD8⁺). Overall, the response is different compared to the alloimmune response and more difficult to suppress. Activation of CD4⁺ T cells is both by direct and indirect antigen presentation. After activation they recruit macrophages and direct B cell responses. Although they are less important than CD4⁺ T cells in islet xenograft rejection, macrophages are believed to be a major effector cell in this response. Rodent studies have shown that xenoantigen-primed and CD4⁺ T cell-activated macrophages were capable of recognition and rejection of pancreatic islet xenografts, and they destroyed a graft via the secretion of various proinflammatory mediators, including TNF-α, reactive oxygen and nitrogen species, and complement factors. B cells are an important mediator of islet xenograft rejection via xenoantigen presentation, priming effector T cells and producing xenospecific antibodies. Depletion and/or inhibition of B cells combined with suppressing T cells has been suggested as a promising strategy for induction of xeno-donor-specific T- and B-cell tolerance in islet xenotransplantation. Thus, strategies that expand the influence of regulatory T cells and inhibit and/or reduce macrophage and B cell responses are required for use in combination with clinical applicable immunosuppressive agents to achieve effective suppression of the T cell-initiated xenograft response.

Increased expression of capsid protein in trans enhances production of single-round infectious particles by West Nile virus DNA vaccine candidate

West Nile virus (WNV; genus Flavivirus, family Flaviviridae) is an emerging pathogenic arbovirus responsible for outbreaks of encephalitis around the world. Whilst no vaccines are currently available to prevent WNV infection of humans, the use of cDNA copies of flavivirus RNA genomes with large internal deletions within the capsid (C) appears promising. C-deleted vaccines are able to replicate and secrete large amounts of non-infectious immunogenic subviral particles (SVPs) from transfected cells. We have previously generated a WNV DNA vaccine candidate pKUNdC/C where C-deleted WNV cDNA was placed under the control of one copy of the cytomegalovirus (CMV) promoter and the C gene was placed under the control of a second copy of the CMV promoter in the same plasmid DNA. This DNA was shown to generate single-round infectious particles (SRIPs) capable of delivering self-replicating C-deleted RNA producing SVPs to surrounding cells, thus enhancing the vaccine potential. However, the amounts of both SRIPs and SVPs produced from pKUNdC/C DNA were relatively low. In this investigation, we aimed at increasing SRIP production by optimizing trans-C expression via incorporating different forms of C and the use of a more powerful promoter. The construct containing an elongation factor EF1α promoter encoding an extended form of C was demonstrated to produce the highest titres of SRIPs and was immunogenic in mice. Additionally, SRIP and SVP titres were further improved via incorporation of a glycosylation motif in the envelope protein. The optimized DNA yielded ~100-fold greater titres of SRIPs than the original construct, thus providing a promising candidate for further vaccine evaluation.

Human embryonic stem cell derived islet progenitors mature inside an encapsulation device without evidence of increased biomass or cell escape

There are several challenges to successful implementation of a cell therapy for insulin dependent diabetes derived from human embryonic stem cells (hESC). Among these are development of functional insulin producing cells, a clinical delivery method that eliminates the need for chronic immunosuppression, and assurance that hESC derived tumors do not form in the patient. We and others have shown that encapsulation of cells in a bilaminar device (TheraCyte) provides immunoprotection in rodents and primates. Here we monitored human insulin secretion and employed bioluminescent imaging (BLI) to evaluate the maturation, growth, and containment of encapsulated islet progenitors derived from CyT49 hESC, transplanted into mice. Human insulin was detectable by 7 weeks post-transplant and increased 17-fold over the course of 8 weeks, yet during this period the biomass of encapsulated cells remained constant. Remarkably, by 20 weeks post-transplant encapsulated cells secreted sufficient levels of human insulin to ameliorate alloxan induced diabetes. Further, bioluminescent imaging revealed for the first time that hESCs remained fully contained in encapsulation devices for up to 150 days, the longest period tested. Collectively, the data suggest that encapsulated hESC derived islet progenitors hold great promise as an effective and safe cell replacement therapy for insulin dependent diabetes.

Nucleofection-mediated α1,3-galactosyltransferase gene inactivation and membrane cofactor protein expression for pig-to-primate xenotransplantation

Xenotransplantation of pig organs into primates leads to hyperacute rejection (HAR). Functional ablation of the pig α 1,3-galactosyltransferase (GalT) gene, which abrogates expression of the Gal α 1-3Gal β 1-4GlcNAc-R (Gal) antigen, which inhibits HAR. However, antigens other than Gal may induce immunological rejection by their cognate antibody responses. Ultimately, overexpression of complement regulatory proteins reduces acute humoral rejection by non-Gal antibodies when GalT is ablated. In this study, we developed a vector-based strategy for ablation of GalT function and concurrent expression of membrane cofactor protein (MCP, CD46). We constructed an MCP expression cassette (designated as MCP-IRESneo) and inserted between the left and the right homologous arms to target exon 9 of the GalT gene. Nucleofection of porcine ear skin fibroblasts using the U-023 and V-013 programs resulted in high transfection efficiency and cell survival. We identified 28 clones in which the MCP-IRESneo vector had been successfully targeted to exon 9 of the GalT gene. Two of those clones, with apparent morphologically mitotic fibroblast features were selected through long-term culture. GalT gene expression was downregulated in these 2 clones. Importantly, MCP was shown to be efficiently expressed at the cell surface and to efficiently protect cell lysis against normal human complement serum attack in vitro.

Genetic manipulation of the mouse developing hypothalamus through in utero electroporation

Genetic modification of specific regions of the developing mammalian brain is a very powerful experimental approach. However, generating novel mouse mutants is often frustratingly slow. It has been shown that access to the mouse brain developing in utero with reasonable post-operatory survival is possible. Still, results with this procedure have been reported almost exclusively for the most superficial and easily accessible part of the developing brain, i.e. the cortex. The thalamus, a narrower and more medial region, has proven more difficult to target. Transfection into deeper nuclei, especially those of the hypothalamus, is perhaps the most challenging and therefore very few results have been reported. Here we demonstrate a procedure to target the entire hypothalamic neuroepithelium or part of it (hypothalamic regions) for transfection through electroporation. The keys to our approach are longer narcosis times, injection in the third ventricle, and appropriate kind and positioning of the electrodes. Additionally, we show results of targeting and subsequent histological analysis of the most recessed hypothalamic nucleus, the mammillary body.

Anti-CD2 producing pig xenografts effect localized depletion of human T cells in a huSCID model

BACKGROUND

We investigated whether graft produced anti-human CD2, mediated by adenovirus (Adv) transduction of pig neonatal islet cell clusters (pNICC), would protect xenografts in a humanized mouse model from immune attack and whether such immunosuppression would remain local.

METHODS

A mouse anti-human CD2 Ab (CD2hb11) previously generated by us was genetically engineered to produce chimeric and humanized versions. The three forms of CD2hb11 were named dilimomab (mouse), diliximab (chimeric) and dilizumab (humanized). All 3 forms of CD2hb11 Ab were tested for their ability to bind CD3(+) human T cells and to inhibit a human anti-pig xenogeneic mixed lymphocyte reaction (MLR). They were administered systemically in a humanized mouse model in order to test their ability to deplete human CD3(+) T cells and whether they induced a cytokine storm. An adenoviral vector expressing diliximab was generated for transduction of pNICC. Humanized mice were transplanted with either control-transduced pNICC or diliximab-transduced pNICC and human T cells within grafts and spleens were enumerated by flow cytometry.

RESULTS

Dilimomab and diliximab inhibited a human anti-pig xenogeneic response but dilizumab did not. All 3 forms of CD2hb11 Ab bound human T cells in vitro though dilimomab and diliximab exhibited 300-fold higher avidity than dilizumab. All 3 anti-CD2 Abs could deplete human CD3(+) T cells in vivo in a humanized mouse model without inducing upregulation of activation markers or significant release of cytokines. Humanized mice transplanted with diliximab-transduced pNICC afforded depletion of CD3(+) T cells at the graft site leaving the peripheral immune system intact.

CONCLUSIONS

Local production of a single Ab against T cells can reduce graft infiltration at the xenograft site and may reduce the need for conventional, systemic immunosuppression.

Human thyroxine binding globulin (TBG) promoter directs efficient and sustaining transgene expression in liver-specific pattern

The liver performs a vital role in metabolic process, which makes it an attractive target organ for gene therapy. To improve the effects of gene therapy in disorders caused by metabolic disturbance, we quantitatively evaluated six promoters, CMV, EF1α, PGK, apoE, thyroxine binding globulin (TBG), and cytochrome P450 2E1 (CYP2E1) by measuring the expression of α1-antitrypsin, which is controlled by these promoters and introduced via a lentivirus-mediated delivery system in the liver. The results showed that the TBG promoter presents as highly active though in general it is slightly lower than the ubiquitous CMV and EF1α. The expression of exogenous genes driven by the TBG promoter demonstrates to be much higher than by PGK, apoE, and CYP2E1 promoters, and the fragment of -435bp to -26bp from transcription start site (TSS) in the TBG promoter region is identified as the optimum region to direct transgene expression at a higher level. In addition, we further confirmed that the TBG promoter confers transgene persistent and specific expression within the liver up to several months after integration. The data suggests that the TBG promoter is a valuable tool and will greatly facilitate the optimization of vector design in hepatic gene therapy.

Nucleic-acid based gene therapeutics: delivery challenges and modular design of nonviral gene carriers and expression cassettes to overcome intracellular barriers for sustained targeted expression

The delivery of nucleic acid molecules into cells to alter physiological functions at the genetic level is a powerful approach to treat a wide range of inherited and acquired disorders. Biocompatible materials such as cationic polymers, lipids, and peptides are being explored as safer alternatives to viral gene carriers. However, the comparatively low efficiency of nonviral carriers currently hampers their translation into clinical settings. Controlling the size and stability of carrier/nucleic acid complexes is one of the primary hurdles as the physicochemical properties of the complexes can define the uptake pathways, which dictate intracellular routing, endosomal processing, and nucleocytoplasmic transport. In addition to nuclear import, subnuclear trafficking, posttranscriptional events, and immune responses can further limit transfection efficiency. Chemical moieties, reactive linkers or signal peptide have been conjugated to carriers to prevent aggregation, induce membrane destabilization and localize to subcellular compartments. Genetic elements can be inserted into the expression cassette to facilitate nuclear targeting, delimit expression to targeted tissue, and modulate transgene expression. The modular option afforded by both gene carriers and expression cassettes provides a two-tier multicomponent delivery system that can be optimized for targeted gene delivery in a variety of settings.

Physiological and tissue-specific vectors for treatment of inherited diseases

After more than 1500 gene therapy clinical trials in the past two decades, the overall conclusion is that for gene therapy (GT) to be successful, the vector systems must still be improved in terms of delivery, expression and safety. The recent development of more efficient and stable vector systems has created great expectations for the future of GT. Impressive results were obtained in three primary immunodeficiencies and other inherited diseases such as congenital blindness, adrenoleukodystrophy or junctional epidermolysis bullosa. However, the development of leukemia in five children included in the GT clinical trials for X-linked severe combined immunodeficiency and the silencing of the therapeutic gene in the chronic granulomatous disease clearly showed the importance of improving safety and efficiency. In this review, we focus on the main strategies available to achieve physiological or tissue-specific expression of therapeutic transgenes and discuss the importance of controlling transgene expression to improve safety. We propose that tissue-specific and/or physiological viral vectors offer the best balance between efficiency and safety and will be the tools of choice for future clinical trials in GT of inherited diseases.

Specific transgene expression in mouse pancreatic beta-cells under the control of the porcine insulin promoter

The availability of regulatory sequences directing tissue-specific expression of transgenes in genetically modified mice and large animals is a prerequisite for the development of adequate models for human diseases. The rat insulin 2 gene (Ins2) promoter, widely used to achieve transgene expression in pancreatic beta-cells of mice, also directs expression to extrapancreatic tissues and performs poorly in isolated pancreatic islets of human, mouse, and pig. To evaluate whether the full 5' untranslated region (UTR) of the porcine insulin gene (INS) confers robust and specific expression in beta-cells we generated an expression cassette containing 1500bp of the porcine INS 5' UTR and the 3' UTR of the bovine growth hormone gene (GH). The cassette was designed to allow easy exchange of the sequences to be expressed and easy removal of the vector backbone from the expression cassette. To evaluate the properties of the cassette, we initially inserted a cDNA encoding human betacellulin, a growth factor known to affect structural and functional parameters of beta-cells. After confirming the functionality and specificity of the construct in vitro, transgenic mouse lines were generated by pronuclear DNA microinjection. Using RT-PCR, immunohistochemistry and immunofluorescence, we show that transgenic mice expressed human betacellulin exclusively in beta-cells. Confirming the proposed insulinotropic effect of betacellulin, transgenic mice showed improved glucose tolerance. We conclude that the newly designed expression cassette containing 1500bp of the porcine insulin promoter 5' UTR confers robust and specific transgene expression to beta-cells in vitro and in transgenic mice.