Rapamycin control of exocrine protein levels in saliva after adenoviral vector-mediated gene transfer

Assessment of the safety and biodistribution of a regulated AAV2 gene transfer vector after delivery to murine submandibular glands

Clinical gene transfer holds promise for the treatment of many inherited and acquired disorders. A key consideration for all clinical gene transfer applications is the tight control of transgene expression. We have examined the safety and biodistribution of a serotype 2, recombinant adeno-associated viral (AAV2) vector that encodes a rapamycin-responsive chimeric transcription factor, which regulates the expression of a therapeutic transgene (human erythropoietin [hEpo]). The vector, AAV2-TF2.3w-hEpo (2.5 × 10(7)-2.5 × 10(10) particles), was administered once to a single submandibular gland of male and female mice and mediated hEpo expression in vivo following a rapamycin injection but not in its absence. Control (saline treated) and vector-treated animals maintained their weight, and consumed food and water, similarly. Vector delivery led to no significant toxicological effects as judged by hematology, clinical chemistry, and gross and microscopic pathology evaluations. On day 3 after vector delivery, vector copies were not only abundant in the targeted right submandibular gland but also detected in multiple other tissues. Vector was cleared from the targeted gland much more rapidly in female mice than in male mice. Overall, our results are consistent with the notion that administration of the AAV2-TF2.3w-hEpo vector to salivary glands posed no significant risk in mice.

Prevention of diabetic neuropathy by regulatable expression of HSV-mediated erythropoietin

Previous studies have demonstrated that gene transfer of genes coding for neurotrophic factors to the dorsal root ganglion (DRG) using nonreplicating herpes simplex virus (HSV)-based vectors injected subcutaneously can prevent the progression of diabetic neuropathy. Because prolonged expression of neurotrophic factors could potentially have unwanted adverse effects, we constructed a nonreplicating HSV vector, vHrtEPO, to express erythropoietin (EPO) under the control of a tetracycline response element (TRE)-minimal cytomegalovirus (CMV) fusion promoter. Primary DRG neurons in culture infected with vHrtEPO express and release EPO in response to exposure to doxycycline (DOX). Animals infected with vHrtEPO by footpad inoculation demonstrated regulated expression of EPO in DRG under the control of DOX administered by gavage. Mice rendered diabetic by injection of streptozotocin (STZ), inoculated with vHrtEPO, and treated with DOX 4 days out of 7 each week for 4 weeks were protected against the development of diabetic neuropathy as assessed by electrophysiologic and behavioral measures. These studies indicate that intermittent expression of EPO in DRG achieved from a regulatable vector is sufficient to protect against the progression of neuropathy in diabetic animals, and provides proof-of-principle preclinical evidence for the development of such vectors for clinical trial.

Systemic delivery of bioactive glucagon-like peptide 1 after adenoviral-mediated gene transfer in the murine salivary gland

An adenoviral (Ad) vector that expresses bioactive glucagon-like peptide 1 (GLP-1) was generated, and its effectiveness at modulating glucose homeostasis was evaluated after transduction of murine salivary glands. The construct was engineered with the signal sequence of mouse GH to direct the peptide into the secretory pathway, followed by a furin cleavage site and the GLP-1(7-37) sequence encoding an Ala to Gly substitution at position 8 to achieve resistance to degradation. When expressed in Neuro2A and COS7 cells, an active form of GLP-1 was specifically detected by RIA in the conditioned medium of transduced cells, showed resistance to degradation by dipeptidyl-peptidase IV, and induced the secretion of insulin from NIT1 pancreatic beta-cells in vitro. In vivo studies demonstrated that healthy mice transduced with Ad-GLP-1 in both submandibular glands had serum GLP-1 levels approximately 3 times higher than mice transduced with the control Ad-luciferase vector. In fasted animals, serum glucose levels were similar between Ad-GLP-1 and Ad-luciferase transduced mice in keeping with GLP-1's glucose-dependent action. However, when challenged with glucose, Ad-GLP-1 transduced mice cleared the glucose significantly faster than control mice. In an animal model of diabetes induced by alloxan, progression of hyperglycemia was significantly attenuated in mice given the Ad-GLP-1 vector compared with control mice. These studies demonstrate that the bioactive peptide hormone, GLP-1, normally secreted from endocrine cells in the gut through the regulated secretory pathway, can be engineered for secretion into the circulatory system from exocrine cells of the salivary gland to affect glucose homeostasis.

Evaluation of a rapamycin-regulated serotype 2 adeno-associated viral vector in macaque parotid glands

OBJECTIVES

Salivary glands are useful target organs for local and systemic gene therapeutics. For such applications, the regulation of transgene expression is important. Previous studies by us in murine submandibular glands showed that a rapamycin transcriptional regulation system in a single serotype 2, adeno-associated viral (AAV2) vector was effective for this purpose. This study evaluated if such a vector was similarly useful in rhesus macaque parotid glands.

METHODS

A recombinant AAV2 vector (AAV-TF-RhEpo-2.3w), encoding rhesus erythropoietin (RhEpo) and a rapamycin-inducible promoter, was constructed. The vector was administered to macaques at either of two doses [1.5 x 10(11) (low dose) or 1.5 x 10(12) (high dose) vector genomes] via cannulation of Stensen's duct. Animals were followed up for 12-14 weeks and treated at intervals with rapamycin (0.1 or 0.5 mg kg(-1)) to induce gene expression. Serum chemistry, hematology, and RhEpo levels were measured at interval.

RESULTS

AAV-TF-RhEpo-2.3w administration led to low levels of rapamycin-inducible RhEpo expression in the serum of most macaques. In five animals, no significant changes were seen in serum chemistry and hematology values over the study. One macaque, however, developed pneumonia, became anemic and subsequently required euthanasia. After the onset of anemia, a single administration of rapamycin led to significant RhEpo production in this animal.

CONCLUSION

Administration of AAV-TF-RhEpo-2.3w to macaque parotid glands was generally safe, but led only to low levels of serum RhEpo in healthy animals following rapamycin treatment.

Ultrasound-assisted non-viral gene transfer to the salivary glands

We report a non-viral gene transfer method using ultrasound induced microbubble destruction to allow the uptake of plasmid gene transfer vectors to the cells of the mouse salivary gland. The Luciferase (Luc) reporter gene, driven by a cytomegalovirus (CMV) promoter, was delivered unilaterally to the submandibular salivary gland via retroductal cannulation and Luc expression was monitored with in vivo imaging. The CMV-Luc plasmid was delivered to the salivary gland in a carrier solution containing microbubbles composed of lipid-encased perfluoropropane gas, with two different concentrations of microbubbles used (100 and 15% volume/volume). An Adenoviral (Ad) vector using an identical CMV-Luc expression cassette was used as a positive control at two different dosages. Whereas ultrasound-assisted gene transfer (UAGT) with 100% microbubbles was weak and rapidly extinguished, UAGT with the 15% microbubble solution was robust and stable for 28 days. UAGT seems to be a practicable and promising method for non-viral gene delivery to the salivary glands.

AAV5-mediated gene transfer to the parotid glands of non-human primates

Salivary glands are potentially useful target sites for multiple clinical applications of gene transfer. Previously, we have shown that serotype 2 adeno-associated viral (AAV2) vectors lead to stable gene transfer in the parotid glands of rhesus macaques. As AAV5 vectors result in considerably greater transgene expression in murine salivary glands than do AAV2 vectors, herein we have examined the use of AAV5 vectors in macaques at two different doses (n = 3 per group; 10(10) or 3 x 10(11) particles per gland). AAV5 vector delivery, as with AAV2 vectors, led to no untoward clinical, hematological or serum chemistry responses in macaques. The extent of AAV5-mediated expression of rhesus erythropoietin (RhEpo) was dose-dependent and similar to that seen with an AAV2 vector. However, unlike results with the AAV2 vector, AAV5 vector-mediated RhEpo expression was transient. Maximal expression peaked at day 56, was reduced by approximately 80% on day 84 and thereafter remained near background levels until day 182 (end of experiment). Quantitative PCR studies of high-dose vector biodistribution at this last time point showed much lower AAV5 copy numbers in the targeted parotid gland (approximately 1.7%) than found with the same AAV2 vector dose. Molecular analysis of the conformation of vector DNA indicated a markedly lower level of concatamerization for the AAV5 vector compared with that of a similar AAV2 vector. In addition, cellular immunological studies suggest that host response differences may occur with AAV2 and AAV5 vector delivery at this mucosal site. The aggregate data indicate that results with AAV5 vectors in murine salivary glands apparently do not extend to macaque glands.

Long-term transduction of miniature pig parotid glands using serotype 2 adeno-associated viral vectors

BACKGROUND

Previously, using an adenoviral vector, we showed that miniature pigs could provide a valuable and affordable large animal model for pre-clinical gene therapy studies to correct parotid gland radiation damage. However, adenoviral vectors lead to short-term transgene expression and, ideally, a more stable correction is required. In the present study, we examined the suitability of using a serotype 2 adeno-associated viral (AAV2) vector to mediate more stable gene transfer in the parotid glands of these animals.

METHODS

Heparan sulfate proteoglycan was detected by immunohistochemistry. beta-galactosidase expression was determined histochemically. An AAV2 vector encoding human erythropoietin (hEpo) was administered via Stensen's duct. Salivary and serum hEpo levels were measured using an enzyme-linked immunosorbent assay. Serum chemistry and hematological analyses were performed and serum antibodies to hEpo were measured throughout the study. Vector distribution was determined by a quantitative polymerase chain reaction.

RESULTS

Transgene expression was vector dose-dependent, with high levels of hEpo being detected for up to 32 weeks (i.e. the longest time studied). hEpo reached maximal levels during weeks 4-8, but declined to approximately 25% of these values by week 32. Haematocrits were elevated from week 2. Transduced animals exhibited low serum anti-hEpo antibodies (1 : 8-1 : 16). Vector biodistribution at animal sacrifice revealed that most copies were in the targeted parotid gland, with few being detected elsewhere. No consistent adverse changes in serum chemistry or hematology parameters were seen.

CONCLUSIONS

AAV2 vectors mediate extended gene transfer to miniature pig parotid glands and should be useful for testing pre-clinical gene therapy strategies aiming to correct salivary gland radiation damage.

Novel gene switches

Controlling gene activity in space and time represents a cornerstone technology in gene and cell therapeutic applications, bioengineering, drug discovery as well as fundamental and applied research. This chapter provides a comprehensive overview of the different approaches for regulating gene activity and product protein formation at different biosynthetic levels, from genomic rearrangements over transcription and translation control to strategies for engineering inducible secretion and protein activity with a focus on the development during the past 2 years. Recent advances in designing second-generation gene switches, based on novel inducer administration routes (gas phase) as well as on the combination of heterologous switches with endogenous signals, will be complemented by an overview of the emerging field of mammalian synthetic biology, which enables the design of complex synthetic and semisynthetic gene networks. This article will conclude with an overview of how the different gene switches have been applied in gene therapy studies, bioengineering and drug discovery.

Enhanced transduction of mouse salivary glands with AAV5-based vectors

We previously demonstrated that recombinant adeno-associated virus vectors based on serotype 2 (rAAV2) can direct transgene expression in salivary gland cells in vitro and in vivo. However, it is not known how other rAAV serotypes perform when infused into salivary glands. The capsids of serotypes 4 and 5 are distinct from rAAV2 and from each other, suggesting that they may direct binding and entry into different cell types. In the present study, we investigated the tropisms, transduction efficiencies, and antibody response to AAV vectors based on AAV serotypes 2, 4, and 5. Administration of rAAV2beta-galactosidase (betagal), rAAV4betagal, or rAAV5betagal to murine submandibular salivary glands by retrograde ductal instillation resulted in efficient transduction of salivary epithelial cells, with AAV4 and AAV5 producing 2.3 and 7.3 times more betagal activity compared with AAV2. Improved transduction with AAV5 was confirmed by QPCR of DNA extracted from glands and immunohistochemical staining for transgene expression. Like AAV2, AAV5 primarily transduced striated and intercalated ductal cells. AAV4 transduction was evident in striated, intercalated, and excretory ductal cells, as well as in convoluted granular tubules. In keeping with the encapsulated nature of the salivary gland, the majority of persistent viral genomes were found in the gland and not in other tissues. Neutralizing antibodies (NABs) found in the serum of virus-infused animals were serotype specific and there was no crossreactivity between serotypes. No NABs were detected in saliva but sialic acid conjugates present in saliva could neutralize AAV4 at low dilutions. Together our data suggest that because of differences in receptor binding and transduction pathways, other serotypes may have improved utility as gene transfer vectors in the salivary gland and these differences could be exploited in gene therapy applications.

Effect of serotype 5 adenoviral and serotype 2 adeno- associated viral vector-mediated gene transfer to salivary glands on the composition of saliva

Key to the development of a useful clinical therapy is the minimization of side effects. Routine safety testing, however, does not provide information about the physiological status of many potentially useful gene transfer target sites. In this study, we evaluated the longitudinal effects of intrasalivary duct delivery of recombinant serotype 5 adenoviral (rAd5; 10(9)-10(10) particles/gland in rats) and recombinant serotype 2 adeno-associated viral (rAAV2; 10(8)-10(9) particles/gland in mice) vectors on salivary composition. Both vectors led to modest, transient alterations in several salivary components that thereafter returned to normal. The changes suggested two initial specific consequences of rAd5 and rAAV2 vector administration: (1) a modest breach of the mucosal barrier in the targeted glands, indicated by elevations in salivary albumin, total protein, and Na+ levels, and (2) an innate host response, indicated by transient elevations in either salivary lactoferrin and IgA levels (rAd5) or mucin (rAAV2). These studies are consistent with the notion that administration of modest doses of rAd5 and rAAV2 vectors to salivary glands for a therapeutic purpose can be accomplished without severe or permanent injury to the target tissue, or compromise to its essential exocrine physiological function.

Salivary gland gene therapy

Salivary glands have proven to be unusual but valuable target sites for multiple clinical gene transfer applications. Access to salivary glands for gene transfer is easy. Multiple studies in animal models have yielded proofs of concept for novel treatments for damaged salivary glands following therapeutic irraditation, in Sjögren's syndrome, and for gene therapeutics systemically by way of the blood-stream and locally in the oral cavity and upper gastrointestinal tract.

Rapamycin control of transgene expression from a single AAV vector in mouse salivary glands

Salivary glands (SGs) appear to be a useful target site for gene therapeutics. The ability to control transgene expression is essential for clinical application. Previously, in a proof-of-concept study, we have shown that the rapamycin-inducible transcriptional regulation system can regulate protein expression after adenoviral-mediated gene transfer to SGs. To evaluate the potential ability to utilize this regulatory system for long-term control of transgene expression in this tissue, we employed a 'third generation', single adenoassociated serotype 2 viral (AAV2) vector encoding human erythropoietin (hEPO) under the control of a rapamycin-inducible promoter. The vector, rAAV-TF2.3-hEPO (10(10) particles/animal), was delivered to mouse SGs. No detectable increase in serum hEPO or hematocrit levels was observed in the absence of rapamycin administration. However, rapamycin induced elevation of serum hEPO levels, as well as concomitant hematocrit changes, that were dose-dependent, completely reversible, and relatively stable over the course of this study (6 months), with no appreciable change in rapamycin responsiveness. Our results suggest that the rapamycin transcriptional regulation system delivered in a single AAV2 vector to SGs may be valuable for systemic protein replacement applications.

Pharmacologic transgene control systems for gene therapy

Pharmacologic transgene-expression dosing is considered essential for future gene therapy scenarios. Genetic interventions require precise transcription or translation fine-tuning of therapeutic transgenes to enable their titration into the therapeutic window, to adapt them to daily changing dosing regimes of the patient, to integrate them seamlessly into the patient's transcriptome orchestra, and to terminate their expression after successful therapy. In recent years, decisive progress has been achieved in designing high-precision trigger-inducible mammalian transgene control modalities responsive to clinically licensed and inert heterologous molecules or to endogenous physiologic signals. Availability of a portfolio of compatible transcription control systems has enabled assembly of higher-order control circuitries providing simultaneous or independent control of several transgenes and the design of (semi-)synthetic gene networks, which emulate digital expression switches, regulatory transcription cascades, epigenetic expression imprinting, and cellular transcription memories. This review provides an overview of cutting-edge developments in transgene control systems, of the design of synthetic gene networks, and of the delivery of such systems for the prototype treatment of prominent human disease phenotypes.

Partial redirection of transgenic human growth hormone secretion from rat salivary glands

Regulated secretory pathway proteins, when delivered as transgenes to salivary glands, are secreted predominantly into saliva. This is not useful for those proteins whose therapeutic function is required systemically, for example, human growth hormone (hGH). One strategy to improve the efficiency of hGH secretion into the bloodstream involves manipulation of existing sorting signals. The C terminus of hGH is highly conserved and contains a domain similar to the regulated pathway sorting domain of pro-opiomelanocortin (POMC). We hypothesized that, similar to POMC, mutation of this domain would divert hGH secretion from the regulated to the constitutive pathway, which in salivary glands leads to the bloodstream. Several mutations were made in the C terminus of the hGH cDNA and tested in vitro. One biologically active mutant containing E174A and E186A substitutions, and with an included C-terminal extension, was studied in greater detail. Compared with wild-type hGH, we found that this mutant hGH accumulated in the Golgi/trans-Golgi network and showed increased basal secretion in AtT20 cells, a model endocrine cell line. Importantly, in vivo, the mutant hGH displayed a relative increase in the proportion of constitutive pathway secretion seen from rat salivary glands, with a significantly lower saliva-versus-serum secretion ratio (p=0.03). Although this mutant is unlikely to be therapeutically beneficial, these results suggest that the final destination of a transgenic secretory protein may be controlled by reengineering its sorting determinants.

Salivary glands: novel target sites for gene therapeutics

Directing the local or systemic expression of therapeutic proteins is a potentially important clinical application of gene transfer technology. Gene-based therapeutics theoretically offer many advantages over protein therapeutics. Numerous tissues have been evaluated for this purpose in animal models, most commonly the liver and skeletal muscle. Based on pre-clinical studies, we suggest that salivary glands are a valuable, yet under-appreciated, target tissue for both systemic and upper gastrointestinal tract gene therapeutic applications.