The use of gene transfer for the protection and repair of salivary glands

The concept of transferring genes to tissues for clinical applications has been discussed for nearly half a century (Wolff and Lederberg, 1994; Schecter, 2002). The exponential increase in our ability to manipulate the genetic material of a cell via recombinant DNA technology has brought this goal closer to realization. The original perception that gene therapy should be considered only for a few major organs as a means of treating life-threatening disorders that are refractory to conventional treatment has changed. There are many non-life-threatening conditions that adversely affect a patient's quality of life for which there are no effective treatments. The lack of suitable treatment has permitted morbidity to become a rational basis for extending the scope of gene therapy. Implicit in this is the realization that what constitutes acceptable target tissues has changed. Much evidence supports the idea that the major salivary glands (parotid and submandibular) are useful targets for gene therapy (Baum, Wellner & Zheng, 2002). Cellular transduction can be accomplished without requiring surgically invasive techniques. The acinar and ductal cells of human submandibular and parotid glands are easily accessed via retrograde instillation through their respective ducts (see below). The approach has been shown to be both reliable and reproducible (Mastrangeli et al, 1994; Baccaglini et al, 2001), and proofs of concept using intraductal cannulation of salivary glands for transgene delivery has been achieved by many others (Barka and Van der Noen 1996; Goldfine et al, 1997; Baum et al, 2002). Additionally, salivary glands are secretory organs capable of producing large amounts of protein for both endocrine and exocrine functions (Baum et al, 1999). This review will focus on two conditions that affect the oral cavity and are associated with high morbidity. Sjögren's syndrome (SS; an autoimmune exocrinopathy) and radiation induced salivary gland damage together provided the impetus for our early work on gene transfer to salivary glands. Both conditions result in a significant reduction of salivary output via irreversible damage to the secretory parenchyma. Unfortunately for the affected patients, no effective conventional treatments exist. Primary SS affects about one million people (>90% females; prevalence ∼0.5%) in the USA alone (Fox et al, 1998; Pillemer et al, 2001). Radiation induced damage of the salivary glands is also not uncommon. There are in excess of 40 000 new cases of head and neck cancer diagnosed each year in the USA, almost all treated in part with ionizing radiation (Greenlee et al, 2001). The purpose of this review is to examine the progress made in addressing gene transfer strategies for preventing damage to, as well as correcting already damaged, salivary glands. There are two general methods for transferring genes to mammalian cells: viral and non-viral. Viral vectors typically are highly efficient for gene transfer in vivo, but they can pose a significant safety risk. Non-viral methods, while much safer for the host, at present are relatively inefficient for gene transfer in vivo. Both viral and non-viral methods have been used successfully to transfer genes to salivary glands in animal models. For the purposes of this presentation both general methods have application, and thus are relevant. It is important for the reader to recognize that at present there is no such thing as a perfect vector for transferring genes to any tissue. All currently available vectors, both viral and non-viral, have certain shortcomings as well as specific benefits. However, it seems quite likely that the quality and utility of both types of vectors will considerably improve in the near future. Access for delivering vectors and transferring genes to salivary gland cells in vivo is excellent via direct, intraoral cannulation of the main excretory ducts and retrograde infusion (Mastrangeli et al, 1994; Baum et al, 2002). This delivery strategy is analogous to that used clinically to infuse contrast media into the salivary glands for routine contrast radiographs (sialographs). Salivary gland epithelial cells for the most part exist as a cellular monolayer along the ductal tree. Thus, from a cannulated ductal orifice the apical membranes of almost all parenchymal cells are available as targets for vectors. Additionally, salivary glands are generally well encapsulated, a circumstance that is likely to limit the undesirable spread of vector beyond the targeted gland. As salivary epithelial cells in vivo divide very slowly, the choice of which viral vector to use is limited to those capable of infecting non- or slowly dividing cells. Accordingly, retroviral vectors are not especially useful for salivary gland gene transfer because for them to be effective the target cell must be actively dividing. Conversely, as adenovirus (Ad) and adenoassociated virus (AAV) vectors can infect non-dividing cells (Table 1), both are useful for gene transfer to mammalian salivary glands (Baum et al, 2002). Serotype 5 Ad (Ad5) vectors have been used extensively by our laboratory for salivary gland gene transfer studies, as well as by others for studies in numerous other epithelial tissues. These double-stranded DNA viruses readily infect both rat and mouse salivary cells in vivo. Previously, our laboratory reported that quite high levels of cellular transduction occur after Ad5-mediated gene transfer to rat salivary glands in vivo (e.g. ∼20–35%) (Delporte et al, 1997a). All cell types in rodent salivary glands appear to be good targets for Ad5 vectors, with the exception of cells in the main excretory duct (Delporte, Redman & Baum, 1997b). However, adenoviral-mediated gene transfer is by definition transient, as these vectors can only rarely integrate into the host cell genome (Harui et al, 1999). Typically, Ad vectors exist in an extra-chromosomal location in the nucleus and are `diluted' with each cell division. Of more practical concern, Ad vectors elicit a potent immune response in salivary glands as in other tissues (Adesanya et al, 1996; Kagami et al, 1998). This leads to the generation of neutralizing antibodies that blunt the effect of subsequent vector delivery, as well as considerable cellular immune-mediated target cell destruction. As a consequence, Ad5-mediated transgene expression in rodent salivary glands is usually detected for no more than 7–10 days. The net result is that unless quite low doses of vector can be used therapeutically, enabling the virus to escape host immune surveillance (Adesanya et al, 1996), Ad5-mediated gene transfer is best considered for proof of concept experimental studies in salivary tissue. The vast majority of gene transfer studies historically conducted in our laboratory have used Ad5 vectors for obtaining such proofs of concept. AAV is a member of the parvovirus family and is naturally replication-incompetent. Importantly, there are no known pathologies linked to AAV infections in humans. AAV is a very small, single-stranded DNA virus, approximately one-eighth the size of an Ad virus. Because of its small size (<5kb), only limited space is available for the transgene cassette (i.e. minimally the promoter, transgene, and polyadenylation sequence). Consequently, AAV vectors are not useful for transferring large genes. For several years now, we have used serotype 2 AAV (AAV2) vectors to transfer genes into murine salivary glands (Braddon et al, 1998; Yamano et al, 2002). In contrast to the wide cellular targeting seen with Ad5 vectors in salivary glands, AAV2 vectors are quite restricted, primarily infecting intralobular duct cells. In our experience, AAV2 vectors rarely, if ever, transduce acinar cells, presumably because these cells lack an appropriate receptor on their apical membrane surface. Interestingly, we also have not seen reproducible transduction of any rat salivary epithelial cell type with AAV2 vectors, likely for the same reason. Wild type AAV2 has the ability to integrate into human cells at a specific, non-mutagenic locus in human chromosome 19 (Kotin et al, 1991), however, this property is not retained by recombinant AAV2 (rAAV2) vectors. The rAAV2 vectors can integrate into the genome randomly but at low frequency, and more typically exist in an extra-chromosomal location in the nucleus, like Ad vectors, but as stable concatamers. Importantly, rAAV2 vectors appear to elicit only a modest immune response, much less than that generated by Ad vectors (Sarukhan et al, 2001). Accordingly, in slowly dividing cells transgene expression from rAAV2 vectors can be quite long-lived. In fact, we routinely follow rAAV2-mediated transgene expression in murine salivary glands for 2–3 months. At present, rAAV2 vectors show considerable promise for use in novel treatments of salivary gland disorders, e.g. with SS (Baum et al, 2002; see below), where long-term, stable transgene expression is required. There are several ways for transferring genes by non-viral methods, including electroporation, and physical penetration by either a so-called gene gun (using DNA adsorbed to tiny particles), or by direct delivery of plasmid DNA. The latter can involve using the transgene alone (often referred to as a `naked plasmid') or the plasmid complexed to a poly-cation (typically a cationic lipid; referred to as a `lipoplex'). Both types of plasmid delivery have been used successfully with salivary glands following intraductal cannulation (O'Connell et al, 1995; Goldfine et al, 1997; Baccaglini et al, 2001). However, the levels of transgene expression obtained with plasmid delivery are much lower than those achievable with a recombinant viral vector. For example, when human growth hormone is used as a transgene, expression from a relatively modest dose (108 plaque forming units) of an adenoviral vector is several hundred-fold greater than seen with an optimized lipoplex preparation (Baccaglini et al, 2001). Furthermore, transgene expression is quite transient following plasmid-mediated gene transfer, with peak levels of the encoded protein rapidly declining to background within about 48 h after lipoplex delivery (Baccaglini et al, 2001). Nonetheless, as described below, there are clearly circumstances when plasmid delivery to salivary glands may be sufficient to result in therapeutic benefits. As previously mentioned, there are approximately 40 000 new cases of head and neck cancer diagnosed in USA every year (Greenlee et al, 2001). Most of these patients will undergo head and neck radiation treatment as part of their therapy. Unfortunately, the fluid secreting acinar cells of the salivary gland are exquisitely sensitive to ionizing radiation (see earlier review in this series by Nagler) (Stephens et al, 1986; Vissink et al, 1991; Franzen et al, 1992; Johnson et al, 1993). Acinar cells are the only fluid producing cells in these glands (see earlier review in this series by Turner and Sugiya). The morbidity associated with radiation-induced damage to the salivary glands is thus very significant, for while ductal cells are less sensitive to radiation, they are relatively impermeable to water and are not capable of generating salivary flow. Therefore, loss of the acinar cells in the gland will have profound consequences for the patient (see earlier reviews in this series by Ship and by Nagler). Unfortunately, there are no effective conventional treatment approaches available for these patients (Baum et al, 2002). Initially, our work focused on increasing salivary secretion in damaged salivary glands postirradiation. Specifically, we examined rats whose submandibular glands were subjected to a single radiation dose of 17.5 or 21 Gy. We hypothesized that by increasing water permeability in surviving cells (presumably mostly ductal) increased fluid secretion would occur. We constructed a recombinant adenovirus (AdhAQP1) encoding the human water channel protein aquaporin-1 (hAQP1), driven by the cytomegalovirus (CMV) promoter/enhancer (Figure 1) (Delporte et al, 1997a). Three to 4 months postirradiation rats were administered a single dose of AdhAQP1, or a control virus, via retrograde ductal instillation. Three days later, stimulated saliva was collected from all rats. The control virus had no effect on salivary flow and the irradiated rats exhibited marked salivary flow of Conversely, rats a to increase in salivary salivary flow for treated with the control virus (Delporte et al, 1997a). we examined the utility of for damage in (O'Connell et al, 1999). In this a single parotid gland of was irradiated with a single dose of Gy. This dose salivary flow in all (O'Connell et al, 1999). was administered at 19 postirradiation and salivary secretion examined and days The however, were and only two of the increased salivary flow with the animal administered an virus (O'Connell et al, 1999). for the from the rat studies an of the virus into the glands, these animal in the of viral receptor on the of gland cells (e.g. et al, as well as in the target cells. At present it is not the strategy of gene transfer will be useful for of radiation damage in human salivary glands, and we are our animal of hypothesized by which fluid secretion from irradiated salivary glands. duct cells are in a with only and The is to the of the cell and the is to the The cell is shown as water of this cell with the water channel aquaporin-1 is into the apical and membranes a by which water can flow in response to an We have hypothesized that this would be generated by of and into the is on the in et As an of our studies on the of radiation we to an important safety of salivary specific expression of the transgene in vivo. mammalian genes that cell type specific for and This circumstance has been successfully by several to target transgene expression to tissues in vivo et al, 1986; and 1998). Specifically, we to a of specific for either acinar or ductal cells, to our recombinant vectors. this we constructed two novel recombinant each a relatively specific The human salivary is relatively acinar while the is relatively ductal et al, 2001). is by acinar cells in salivary glands, is in salivary glands by ductal cells and We the cellular expression of adenoviral each of which the transgene and only in the and et al, 2001). We hypothesized that expression with these vectors would be either relatively in acinar cells or ductal cells, or was to rat salivary glands there was a expression of in acinar cells ductal cells. use we detected a of in ductal cells. of the promoter/enhancer to in the levels of expression in acinar and ductal cells et al, 2001). In we a therapeutic of this We the cell type expression and of two adenoviral vectors. which exhibited primarily ductal cell as efficient at increasing salivary flow in irradiated rats as et al, 2001). In these studies that it should be to use in the of gene transfer vectors. vectors should limit transgene expression to salivary cells, preventing any undesirable ionizing radiation with a the transfer of the of a of known as These and highly have an on the of and not these can on numerous cellular including and DNA. is also in the it is not a it have The of in a cell is to control as it can to in the of such as and (see earlier review in this series by Nagler). In can cell membranes a means of increasing the target for In have a limited ability to membranes and The in damage is the In this the transfer of that are and those that are The of most damage in tissues is the & of the most the can of and almost any There are several that can to the of in vivo, including direct from of water by ionizing As the can be via a of The as described in by this can in vivo, a known as the is more likely to occur as is usually present and such as such as and in a to their in the of these et al, the there are no that Therefore, a gene therapeutic strategy at to be to focus on the levels of both and from the of a single to The cellular is the There are of this in cells (Figure a a and an et al, et al, and These the following the of the most are membrane associated also and amounts of and have also been shown to exist in the nucleus, and of in from at cells from ionizing radiation have focused on the levels of In a series of studies, et that treatment of mouse to with encoding as a plasmid complexed with cationic or within a recombinant from both the and of ionizing radiation, including radiation and et al, In they also that of mouse with a plasmid encoding human in a of several in radiation-induced damage et al, that also radiation-induced in both human and in in vivo et al, et al, are likely the of damage to cells, these studies show evidence for the of the levels of the for of damage et al, 2002). on the of and we are currently applications to salivary glands. Specifically, we are to limit radiation damage of salivary gland cells by the levels of both and via a cationic We are using the for human and We have hypothesized that of these transgene alone or in will increased of salivary cells from In we that a approach should the levels of and thus an effect on It is important to recognize that the of by as a Therefore, we are extending the work of et to the within a cell in to cellular and can to levels (Figure the and both the following There are known of that specific expression as well as in The family of can the reduction of and Of these have been shown to be important in cell membrane and 1986; et al, on the other is to be limited to the (Figure where it from and 2001). These studies are in Interestingly, it may be that a focus on of salivary cells that can be may studies radiation induced damage in that the of the in fact, may be the in syndrome et al, 2001). The of this is that damage to the of salivary glands to postirradiation studies are to The single most salivary glands is SS (see by et earlier in this SS is an autoimmune that targets salivary and glands, either in the or of autoimmune (e.g. As the prevalence of SS in the USA is about one million and the affects et al, 2001). The of the salivary of the is the of a in the salivary glands by a conventional with a gland This is of cells. At present, the of SS is not well Thus, with this there are no therapeutic targets for gene transfer approaches or for more conventional treatments. however, a general has to clinically addressing the certain autoimmune including This is mostly from of experimental studies of 2001). this that the present in an autoimmune like or SS is a result of an of in the target tissue. In for example, of gland from patients show high levels of the & 2001). In gland from SS patients high levels of other like and that are by the salivary epithelial cells et al, 2001). Accordingly, it is that either the expression of or blunt their will to an within the gland. This would of the target and This has been by several in animal of several autoimmune & 1994; et al, 1997; et al, 1998). Thus, when recombinant a that can cells, is administered to these murine the is following the of recombinant protein to a murine of the profound that is in these is & In there has been a considerable to the of with recombinant for treating types of have been used in clinical a and a receptor Both recombinant show a significant clinical and are for use to in USA et al, et al, Additionally, a using that this approach may be in SS patients et al, 2001). It is generally well that while the treatment of many via the of recombinant has patient e.g. in to growth hormone it is not Accordingly, many have using gene transfer technology to stable levels of therapeutic and the for the of (Table and several have shown that a single delivery of the or can to of in et al, et al, et al, 2001). et have reported that use of a rAAV2 vector to transfer the also can experimental in et al, there are several that a single of a gene transfer vector encoding can the in for a of to several months et al, 1998; et al, et al, 2001). the of gene transfer used to the in a of SS with murine was described et al, 2001). The approach used is a of what has been described as these are in and their and constructed a recombinant Ad vector encoding the gene and administered it by direct into mouse salivary glands with in with a control Ad encoding the treated with the virus encoding exhibited a significant reduction in the of seen in the salivary glands, as well as a reduction in destruction. we a series of studies to the that transfer of a gene encoding a may be in correcting the salivary seen in murine of the mouse et al, 1999). These an and reduction in salivary flow that with in their salivary glands et al, 1994; Yamano et al, 1999). For these studies we to use rAAV2 vectors, because of their relatively lower with Ad vectors We constructed a rAAV2 vector encoding human and its in vivo in a of et al, 2001). an or of to to the dose of and considerably less than not treated with this vector. We that is to direct the expression of human in mouse salivary glands when using only modest levels of the vector et al, 2002). Most we have administered this vector to the salivary glands of and its on stimulated salivary flow were the vector either at salivary or at treatment after salivary was Salivary were made when the were In both the salivary flow than with their We are our studies to the of this general strategy for SS At the of this we the reader that the concept of using genes to is not and has been considered for at the 40 years (Wolff and The using gene transfer clinically is that many and disorders, such as salivary gland damage and have no conventional such as gene therapy to patients with such The progress in recombinant DNA which genetic readily has gene therapy The clinical gene therapy was conducted in with much and considerable there has been and as well as levels of from an there has been very There are at high quality focused on gene and many other high quality with to, or gene therapy or numerous animal studies, such as those described for salivary glands, proofs of concept for clinical Additionally, major have been made in vector as well as in vector and & 2001). In the only known from a gene therapy was reported 2001). within a year there the of clinical following a clinical gene transfer et al, This latter two from a which had them to life in an from both were treated using a gene transfer vivo, and the cells and to both have been to their and an this it is to that in gene therapy is at a In this in a in the of In our this has not However, also that gene therapy was well in and that its clinical promise was We also this The methods and strategies described are and will be seen as such a from However, the approaches are on We are most with increased and gene transfer will become a part of the for treating salivary gland disorders within We and Wellner for their on this