The Transduction of Rat Submandibular Glands by an Adenoviral Vector Carrying the Human Growth Hormone Gene is Associated with Limited and Reversible Changes at the Infusion Site

Abstract

Adenoviral vectors have been shown to efficiently deliver exogenous genes to salivary glands and have therefore been investigated as tools for the treatment of human disease. The purpose of this study was to evaluate the response of F344 rats to intraductal infusion of the right submandibular salivary gland with an adenoviral vector encoding the gene for human growth hormone (AdCMVhGH). Co-administration of hydroxychloroquine (HCQ) was used to redirect the secretion of human growth hormone (hGH) from saliva into serum. This paper documents the findings of the pathology evaluation of this National Toxicology Program study. The right submandibular salivary gland (infusion site) was the primary target organ, with microscopic lesions characteristic of a mild to moderate insult observed at 3 days post infusion in vector exposed animals. These lesions were characterized by variable degrees of acute glandular inflammation, degeneration and necrosis, with more severe lesions in the higher dose groups. Rats at 28 days post infusion had milder inflammation, degeneration and necrosis compared to day 3 rats, with variable degrees of regeneration. In conclusion, the effects on the salivary glands are reversible as indicated by the milder inflammation and degeneration in the day 28 rats concomitant with mild to moderate regeneration. Therefore, the vector appears relatively innocuous with limited tissue toxicity.

[The supplemental data referenced in this paper is not printed in this issue of Toxicologic Pathology. It is available as a downloadable file in the online edition of Toxicologic Pathology, 34(4). In order to access the full article online, you must have either an individual subscription or a member subscription accessed through www.toxpath.org.]

Introduction

Gene therapy offers the opportunity for correcting protein deficiencies through the expression of a transgene coding for the required protein, with an ultimate goal of providing therapeutic and sustained levels of transgene-encoded proteins in the bloodstream (Verma and Somia, 1997). The application of gene transfer methods requires a safe and effective tissue site for introduction of the transgene as well as a vector capable of stable transduction of this tissue.

The gene transfer management of models of inherited and acquired single protein deficiency disorders is most often achieved by the administration of the gene transfer vector to either a critical-for-life organ (e.g., liver or lung) (Snyder et al., 1997; Auricchio et al., 2002) or a tissue not physiologically intended for secretion (e.g., injection of insulin into muscle for the treatment of diabetes) (Bohl et al., 2000; Johnston et al., 2003). Neither of these approaches is ideal due to the requirement for relatively high vector doses to achieve effective levels of the therapeutic protein. Furthermore, there are potentially serious side effects of targeting a critical organ (Marshall, 2000).

Salivary glands are recognized as novel target sites for gene therapeutics due to several inherent advantages, such as easy access, noninvasive excretory duct cannulation for vector delivery and encapsulation that limits the spread of vector (Voutetakis et al., 2004). Moreover, the salivary gland is not crucial for life, the cells are well differentiated and slow dividing, and they normally produce large amounts of protein for export, in both exocrine and endocrine directions (Isenman et al., 1999).

Clinical data on recombinant adenoviral vectors has demonstrated vector safety, efficiency and efficacy in humans following delivery to lung, sinus, skeletal muscle, brain and liver (Snyder and Flotte, 2002). Moreover, recombinant adenoviral vectors have been shown to correct transient disease in humans treated for hemophilia B and cystic fibrosis (Snyder and Francis, 2005). In the rat, previous studies have demonstrated the utility of adenoviral vectors for transduction of submandibular salivary glands resulting in secretion of proteins encoded by the vector transgene into saliva and serum (Baum et al., 1999, 2004; Delporte et al., 1997; He et al., 1998). When human growth hormone (hGH) is expressed from an adenoviral vector delivered to salivary glands, most of the hGH follows the regulated, tissue-specific, exocrine secretory pathway into the saliva where it is not therapeutically useful. However, the direction of protein secretion can be altered by modification of amino acid-based sorting signals. A weak amine, such as chlorquine, or its hydroxy derivative, hydroxychloroquine (HCQ), will alkalinize intracellular vesicles involved in secretory protein sorting (Caplan et al., 1987; Davis and Mecham, 1998; Moore et al., 1983). Previous studies have shown that co-administration of HCQ via intraperitoneal (IP) injection enhances the endocrine secretion of adenovirus-directed hGH from rat submandibular glands in vivo, a more useful and desirable outcome for delivery of therapeutic proteins (Hoque et al., 2001).

The recombinant adenovirus AdCMVhGH used in this study is a conventional first generation, adenoviral serotype 5 (Ad5) vector encoding hGH (Deleport et al., 1997; Wang et al., 2005). The construction of the recombinant adenoviral vector for this study and the biodistribution in F344 rats following submandibular duct infusion was previously reported (Zheng et al., 2005). The distribution and availability of the AdCMVhGH transgene were evaluated using real time quantitative PCR analysis in multiple tissues (testis/ovary, spleen, liver, lungs, heart, draining lymph nodes and the treated and contralateral untreated salivary gland) and by the presence of hGH in saliva and serum. At three days post-infusion, the transgene was distributed to all tissues analyzed with the exception of the gonads and heart. By day 29, most organs, other than the targeted and contralateral submandibular glands, were negative for the presence of the transgene, providing evidence of a trend towards vector clearance over time.

The purpose of this study was to evaluate the potential local and systemic tissue toxicity of a single intraductal infusion of the right submandibular salivary gland with an adenoviral vector encoding the gene for hGH (AdCMVhGH), with co-administration of HCQ by IP injection.

Material and Methods

This was a National Toxicology Program study, performed at BioReliance in Rockville Maryland, and was designed and conducted to conform to the United States Food and Drug Administration Good Laboratory Practice regulations. Animal treatment was performed according to NIH guidelines and had IACUC review. The histopathological evaluation was subjected to a pathology working group peer review at the National Institute of Environmental Health Sciences in Research Triangle Park, North Carolina.

Recombinant Adenovirus Construction and Viral Titer Determination

The recombinant adenovirus construction (He et al., 1998) and determination of viral titer (Zheng et al., 2005) were previously reported.

Animals

Male and female F344 rats were obtained from Taconic Farms (Germantown, NY) at 6 weeks of age. The rats were quarantined and acclimated to laboratory conditions for 28 to 30 days (males) or 35 to 37 days (females) before dosing. All rats survived the quarantine/acclimation period and appeared normal prior to study initiation. The rats were then randomized based on stratified body weights.

Animal Husbandry

During the testing period, the animals were housed individually in compound- and dose-specific polycarbonate shoe-box cages (Lab Products, Inc., Seaford, DE). All cages were bedded with Irradiated Heat-Treated Sani-Chip Hardwood (P.J. Murphy Forest Products, Montville, NJ). All animals were provided irradiated NTP-2000 Open Formula Diet, meal form (Zeigler Brothers, Inc., Gardners, PA) ad libitum, except for overnight fasting periods prior to the scheduled and terminal sacrifices. Drinking water, which met US EPA drinking water standards, was provided ad libitum during quarantine and by water bottle for the week prior to and during the study.

Study Design

A total of 72 rats/sex were randomly distributed by weight (12/sex/group) into treatment Groups 1–6. Groups 1 and 2 received submandibular duct infusion of saline while groups 3–6 received submandibular duct infusions of AdCMVhGH (described in the next section). Groups 1 and 4 also received intraperitoneal injections of saline while Groups 2, 3, 5, and 6 received intraperitoneal injections of HCQ (also described in the next section). For each treatment group, there were a total of 4 male rats and 4 female rats sacrificed at 3 days post-treatment and a total of 8 males and 8 females sacrificed at 28 days post treatment (Table 1).

AdCMVhGH and HCQ Administration

AdCMVhGH administration was staggered so that animals from each group were treated over several days. Twenty-four hours before the administration of AdCMVhGH, animals in groups 2, 3, 5 and 6 received HCQ (10 mg/kg) while animals in groups 1 and 4 received an equal volume of saline by intraperitoneal injection. On study day 1, anesthesia was induced with an intramuscular injection of a ketamine (60 mg/kg) and xylazine (8 mg/kg) cocktail. Atropine (0.5 mg/kg) was administered subcutaneously or via intraperitoneal injection. Once the animal was anesthetized, a PE 10 polyethylene cannula was inserted into the right submandibular duct and the cannula was fixed into position with glue. A single injection of 200ul of either AdCMVhGH (groups 3, 4, 5, 6) or saline (groups 1, 2) was infused into the duct via an insulin syringe connected to the cannula. The cannula was left in place for approximately 10 minutes after dosing and then removed. Immediately after dosing with AdCMVhGH or saline, animals in Groups 2, 3, 5 and 6 were given another IP injection of HCQ (10 mg/kg) while animals in Groups 1 and 4 were given an IP injection of an equal volume of saline. The HCQ and saline IP injections were repeated on study day 2.

Clinical Assessments

All rats were observed in their cages twice daily for clinical evidence of toxicity, morbidity and mortality. A detailed hands-on clinical observation was performed on test days 1, 3, 8, 15, 22, and 29. Body weights, food consumption and water intake for this study were measured and previously reported (Zheng et al., 2005).

Blood Collection

Rats scheduled for sacrifice on days 3 and 28 were fasted overnight (approximately 16 to 18 hours) prior to blood collection and then anesthetized with CO₂/O₂. Blood was collected from the retro-orbital sinus for clinical pathology studies on the morning of each scheduled sacrifice date.

Necropsy and Tissue Handling

An extensive, systematic necropsy was performed on all animals after scheduled sacrifice. In order to identify both the target organ and any lesions indicative of potential systemic toxicity, a full standard list of tissues were collected and processed for histological evaluation (Chhabra et al., 1990). All tissues, except the eyes, were fixed in 10% neutral buffered formalin (NBF), trimmed, embedded, sectioned (5–6 um) and stained with hematoxylin and eosin. The eyes were fixed for a minimum of 24 hours and a maximum of 72 hours in Davidson’s solution and then transferred to 10% NBF and processed with the other tissues. Salivary glands, both eyes and Harderian glands were kept separate as to left and right.

PCNA Immunohistochemistry

Antigen retrieval was performed by incubating deparaffinized tissue sections in deionized water for 2 intervals of 5 minutes each, in a microwave oven at 50% power. There was a 5 minute cool-down between heating intervals. A 20 minute cool-down in a running water bath was done after the second heating interval. Endogenous peroxidase was inhibited by incubation with freshly prepared 3% hydrogen peroxide for 10 minutes at room temperature. According to manufacturer’s protocol, no protein blocking for nonspecific staining was required for this assay. The sections were then incubated with mouse monoclonal anti-PCNA (Chemicon) at a 1:1000 dilution for 30 minutes at room temperature and a secondary antibody (Jackson ImmunoResearch) at a dilution of 1:400 for 30 minutes at room temperature. Antibodies were diluted in a combined mixture of 1% BSA and 1% non-fat powdered milk solution. This was followed by a Ready-to-Use Elite ABC-peroxidase label (ABC kit, Vector Laboratories, Burlingame CA) according to manufacturer’s protocol. Immune complexes were visualized with 3,3′-diaminobenzidine tetrahydrochloride (DAB) as a chromagen (4 minutes at room temperature). As a negative control, normal serum was used instead of primary antibodies. The sections were counter-stained with Modified Harris hematoxylin for 30 seconds at room temperature, dehydrated through graded alcohols then coverslipped.

Results

Mortality and Clinical Observations

All male and female rats survived until the scheduled sacrifices. There were no significant clinical signs noted in either sex over the course of the study.

Clinical Pathology

All clinical pathology data for this study was previously reported by Zheng et al. (2005).

Gross Lesions

Gross necropsy findings on day 3 and day 29 did not include any abnormal findings that were considered test article-related.

Complete Histopathological Evaluation

A complete histopathological evaluation of all tissues revealed that there were no significant treatment-related lesions in organs other than the salivary glands at 3 and 28 days post-infusion. Since the incidence and severity of lesions in the right submandibular gland were similar between the male and females at both 3 and 28 days post-injection, only the male data is provided in Figure 1. Supplementary data with a more extensive summary of findings for both male and female 3-day and 28-day right submandibular gland lesions and parotid gland lesions are available in the online edition of Toxicologic Pathology, 34(4). Please see the abstract of this article for more information.

Histopathology Findings from the 3 Day Scheduled Sacrifice Male and Female Rats Infused with AdCMVhGH

Right Submandibular Salivary Glands

Infusion of Ad-CMVhGH with vehicle saline or vehicle HCQ (Groups 3–6) into the right submandibular gland via intraductal cannulation was associated with variable degrees of necrosis, degeneration and inflammation, depending on the dose group (Figure 1). The low- and mid-dose groups had lesions that consisted of primarily mild to moderate acute inflammation, moderate degeneration and mild to moderate necrosis (Groups 3–5) (Figures 2B–2D). The lesions were more severe in the high-dose groups (Group 6) with predominately marked acute inflammation, degeneration and necrosis.

The necrotic areas were characterized by either single cell necrosis that includes cell shrinkage, condensation of nuclear chromatin and cytoplasm, convolution of the cell and the presence of apoptotic bodies (Figure 2C), or as locally extensive areas of necrosis with tissue that is paler than the surrounding viable tissue, consisting of swollen cells with variable degrees of eosinophilia, hyalinized cytoplasm, vacuolated cytoplasm, nuclear pyknosis, karyolysis, and/or karyorrhexis with associated cellular debris (Figure 3A). Necrosis was typically patchy throughout the gland and contiguous groups of acini were affected (Figure 2B). The regions of cellular degeneration were characterized by cellular swelling, cytoplasmic vacuolation and chromatin clumping with acinar cell loss and replacement by macrophages and fibroblasts (Figures 2C and 2D).

The inflammatory response was subacute and consisted of primarily neutrophils admixed with lymphocytes, plasma cells and macrophages. There was locally extensive infarction diagnosed in 2 female 3 day animals, 1 in the high-dose group (Group 6) and one in the mid dose saline group (Group 4) characterized by a locally extensive region of coagulative necrosis with associated vascular thrombosis, hemorrhage, and fibrin deposition.

Parotid Salivary Glands

Lesions in the 3-day male and female rat right parotid glands consisted of minimal to moderate degeneration with single cell necrosis, usually more severe along the outer edges of the glands (Figure 3B). Similar changes were present in the left parotid gland, although fewer left parotid glands were affected and the affected glands had minimal to mild degeneration and single cell necrosis.

Histopathology Findings from the 28 Day Scheduled Sacrifice Male and Female Rats infused with AdCMVhGH

Right Submandibular Salivary Glands

The lesions in the 28 day sacrifice animals were indicative of a previous insult that was in an advanced stage of resolution. The cellular changes were primarily in the treatment Groups 3–6 and were similar across groups (Figure 1). These lesions consisted of predominately minimal to moderate acinar and ductular regeneration with minimal to mild inflammation, minimal degeneration and minimal single cell necrosis (Figures 3C and 3D). Most of the above lesions had a multifocal distribution. Regeneration was characterized by acini and ductules lined by increased numbers of cuboidal cells that were jumbled and “piled up” with decreased amounts of cytoplasm and large nuclei containing vesicular chromatin. Regeneration was demonstrated with strong nuclear PCNA immunoreactivity (Figure 3E). These regenerative acini and tubules were located both within the parenchyma and within areas of interstitial fibrosis. The inflammatory response in these groups was chronic, consisting of mostly lymphocytes with few plasma cells.

There were no lesions in the male (Figure 1) and female vehicle saline groups (Group 1) and in the female vehicle HCQ group (Group 2). In the male vehicle HCQ group, 6 of the 8 animals had normal submandibular gland morphology. Of the remaining 2, 1 animal had moderate regeneration and minimal degeneration, necrosis and chronic inflammation while the other animal had lesions consistent with diffuse septic inflammation characterized by marked neutrophilic inflammation, moderate necrosis, moderate degeneration, and mild regeneration (Figure 1).

Parotid Salivary Glands

By 28 days post-infusion, 98% and 100% of the male right and left parotid glands were normal and 100% and 98% of the female right and left parotid glands were normal, respectively.

Histopathology Findings Associated with Salivary Gland Saline Infusion

Infusion of saline (Groups 1 and 2) into the submandibular salivary gland was associated with predominately minimal inflammation, minimal-to-moderate degeneration and minimal scattered single cell necrosis at 3 days post-infusion (Figure 1). The exception is a female rat in Group 1 (vehicle saline) with a more severe neutrophilic inflammatory response and moderate degeneration and mild necrosis, suggestive of infection. By 28 days, the submandibular lesions resolved in all Group 1 male and female rats.

Histopathology Findings Associated with IP Injection of Saline or HCQ

IP injection of HCQ (Group 2) resulted in more animals affected by inflammation, degeneration and necrosis at 3 days post-injection, compared to IP injection of saline (Group 1), however, the response was predominately minimal-to-mild (Figure 1). By 28 days, the submandibular lesions resolved in all Group 2 female rats and in 6 out of 8 Group 2 male rats (Figure 1). One of the 28 day Group 2 male rats had lesions consistent with septic inflammation characterized by marked neutrophilic inflammation, degeneration and necrosis.

Discussion

There is scant literature about the adverse effects following local administration of compounds to the rat salivary gland. Most available references about adverse effects are in reference to drugs administered by other routes such as beta-sympathomimetic agonists that result in hypertrophy and hyperplasia of acinar cells in the submandibular and sublingual salivary glands (Nicolau and Ferreira, 1989) or the administration of high doses of furosemide which can result in the atrophy of acinar cells in the submandibular and parotid salivary glands (Scarlett et al., 1988). Long-term administration of Cyclosporin A has been shown to result in a variety of severe functional and morphological alterations in the submandibular glands of pilocarpine treated rats (Dehpour et al., 1996) and the administration of certain antihistamines can cause cellular apoptosis with foci of acinar and ductular atrophy in both the parotid and submandibular salivary glands (Liu and Linn, 1970). The IP administration of theophylline, a phosphodiesterase inhibitor, will induce a transient decrease in weight associated with the depletion of secretory vesicles in the acinar cells of the submandibular and parotid glands (Kajikawa et al., 2005). However, the co-administration of propranolol will partially abolish this theophylline-induced glandular reduction, suggesting that this transient reduction in size and depletion of secretory vesicles is the result of both phosphodiesterase inhibition and β-adrenergic receptor activation. Oral dosing of rats with a high dose of a novel synthetic steroid with combined estrogenic and progestagenic properties resulted in multifocal hyperplastic lesions within submandibular salivary glands that originated from intercalated ducts. Additional studies suggested that the specific ratio of estrogenic and progestagenic activity of the steroid compound played an important role in the hyperplasia of intercalated ducts (de Rijk et al., 2003). Non-drug-related lesions include salivary gland atrophy that can result from duct ligation (Takahashi et al., 2005), liquid diets (Hall and Schneyer, 1964) or denervation (Wells and Peronace, 1967) or enlargement of the submandibular and sublingual salivary glands following amputation of the lower incisor teeth (Takeda et al., 1986).

The only adverse effect of adenoviral vector infusion into salivary glands appears to be from local tissue injury to the submandibular gland. Adesanya et al. (1996) reported an acute neutrophilic inflammatory response and a transient decrease in salivary function that resulted from adenovirus-mediated gene transfer of beta-galactosidase in rat salivary glands. The time point of salivary gland examination was 3 days post-infusion. Treatment with dexamethasone reduced the inflammatory response, preserved the salivary function and extended the transgene expression. Ewens et al. (2005) have demonstrated that the intravenous injection of fluorouracil plus leucovorin in rats results in facial edema, submandibular gland enlargement and decreased salivary secretion not typically observed in humans. Histologically, there was bacterial infection with inflammation and parenchymal necrosis. The authors suggest that facial edema can result in blockage of the glandular ducts and that the decreased salivary secretion in turn can predispose to gland infection with resident opportunistic pathogens. The mechanical administration of adenoviral vector in our study may have provoked local tissue damage, edema and deranged salivary outflow leading to secondary inflammation and occasional secondary bacterial infection (from resident oral bacteria) or, perhaps, the decreased salivary outflow may have resulted in interstitial leakage of salivary proteolytic enzymes, causing local tissue damage. Dexamethasone or another anti-inflammatory may therefore serve as a useful adjunct therapy for gene transfer using adenovirus vectors.

There are reports of salivary gland neoplasms in rats, such as squamous cell carcinoma, that result from ionizing radiation (Glucksmann and Cherry, 1976) or by compounds such as potassium iodide (Takegawa et al., 1998) and 7,12-dimethlbenz[a]anthracene (DMBA) (Alam and Alam, 1987). Squamous metaplasia of the ductular epithelium, considered a preneoplastic condition, typically precedes salivary gland neoplasia. In our study, evaluation of all 28-day submandibular salivary glands did not reveal any changes that may be considered as preneoplastic (i.e., metaplasia and/or hyperplasia), and staining for PCNA did not reveal any preneoplastic or neoplastic treatment-related changes.

Although infusion of the AdCMVhGH vector into the right submandibular salivary gland did not result in gross lesions at either time point, histological lesions indicate that the right submandibular salivary gland is the primary target organ. The absence of significant lesions in other organs indicates the lack of systemic toxicity. Although there was initial tissue damage to the right submandibular salivary gland due to the adenoviral vector infusion at day 3 post-infusion, there was significant tissue repair by day 28 in both sexes and in all dose groups indicating that the initial tissue damage is reversible.

The right and left parotid glands were also affected in the 3-day sacrifice male and female rats however these tissue changes were minimal and usually confined to the outer edges of the glands. Rat salivary glands are less well encapsulated than their human counterparts therefore, during and subsequent to the infusion process, there may have been leakage of infusion material from the submandibular salivary gland into the surrounding tissue. The parotid glands, due to their close proximity, may have subsequently been in contact with the infusion material. Another likely scenario is the local release of inflammatory mediators, such as cytokines, from an adjacent inflamed submandibular gland that then had an affect on the surrounding tissue. Although the left submandibular gland was not affected, the inflammatory mediators or infusion material released from the right submandibular gland may have been responsible for the minimal edge effect seen in the left parotid gland, which is located approximately 1 centimeter from the right submandibular gland. The sublingual glands are also in close proximity to the submandibular glands, however, they were not similarly affected. Their lack of histological lesions may be due to the nature of the acini. The parotid glands consist of serous acini, similar to the exocrine pancreas, whereas the sublingual glands are mucinous. The mucous content may provide protection for this tissue. The majority (98–100%) of male and female right and left parotid glands were histologically normal by 28 days post-infusion.

The persistence of AdCMVhGH in the salivary glands and the clearance of this vector from most other organs in the majority (~85%) of the animals at 28 days post-infusion were previously reported (Zheng et al., 2005). The limited tissue toxicity and evidence of tissue repair reported here, combined with the successful and sustained expression of the vector in the salivary glands, indicate the efficacy of this method of in vivo gene transfer and expression. These findings also provide supporting evidence that the salivary gland may be considered a potentially useful gene target site for the correction of a variety of human single protein deficiency diseases with a low risk of adverse reactions.