Aquaporin-1 gene therapy as a novel method to treat radiation-induced salivary gland damage: the case for adeno-associated virus vectors

Abstract

Head and neck cancer imposes a substantial disease burden, with approximately 65,000 new cases annually in the United States and 500,000 cases worldwide. Radiation therapy remains a primary treatment modality for many patients with head and neck cancer. Despite advances aimed at limiting exposure of normal tissues within the field of radiation, salivary gland injury is a common and often unavoidable consequence of treatment. Although the mechanisms underlying radiation-induced salivary gland dysfunction remain incompletely understood, the resulting reduction in salivary flow is clinically evident and can cause substantial morbidity, including increased risk of oral infections, oral mucositis, dysphagia, and difficulty eating and speaking. Current therapeutic options are largely palliative, and durable, effective, and patient-acceptable treatments remain limited. Gene therapy offers a potential strategy for sustained functional improvement in irradiated salivary glands. The purpose of this work is to review the historical development of modified viral vectors for delivery of aquaporin-1, a water channel, to the salivary glands. We first summarize early studies using adenoviral vectors for gene delivery, which provided the basis for a Phase I clinical trial of adenoviral aquaporin-1 transfer for radiation-induced salivary gland dysfunction. We then describe the justification for our subsequent clinical trial using modified adeno-associated virus vectors for aquaporin-1 gene transfer. Finally, we outline the design of this trial, including subject recruitment, study procedures, and outcome measures.

Introduction

1. Head and neck radiation and salivary gland damage

Radiation therapy, alone and in combination with other modalities, is among the most common treatment options for head and neck cancer (1).This treatment involves the targeted delivery of high-energy, ionizing radiation (IR) to target tumors of the head and neck including: nasal cavity, nasopharynx and pharynx, tongue, base of tongue, tonsils and/or larynx. The salivary glands, particularly the parotid and submandibular glands, are often in or near the field of radiation. As radiation increases, damage becomes progressive and irreversible; functional parenchyma is lost and is replaced by fibrosis and inflammation (2). Clinical manifestations of this loss of secretory tissue include objective decrease in salivary flow (hyposalivation) and the subjective sensation of oral dryness (xerostomia) (d). Patients with a loss of saliva flow present with an increased risk for oral mucositis, oral infections and tooth decay. Loss of flow results in significant quality of life impacts, including challenges in mastication, swallowing, and speaking.

Several techniques attempt to minimize damage to healthy tissue during radiation therapy. Thermoplastic masks correctly position the patient and ensure accurate delivery of IR but are often poorly tolerated by patients (4, 5), even with open-face systems (6, 7). Systemic delivery of radioprotective agents, such as amifostine, is associated with significant side effects, is not tolerated by patients, and does not protect long-term salivary function (8, 9). Thus, there is still a high possibility of damage to salivary gland structures within the radiation field despite preventive strategies.

Current treatment options for patients with radiation-induced salivary gland damage remain short-term and palliative (10). Cholinergic agonists, such as pilocarpine and cevimeline, are common interventions which physiologically stimulate saliva flow, but effectiveness depends on continued use of the drug (11, 12). Further, the wide distribution of cholinergic receptors causes off-target effects including excessive sweating and urinary frequency leading to frequent medication discontinuation. Saliva substitutes, including lozenges, gels and mouthwashes, are options when these side effects are not tolerated by a patient (13). Though saliva substitutes offer temporary relief, they do not completely provide the benefits of normal salivation. There are currently no effective, well-tolerated, long-term treatment options to restore saliva flow following IR.

2. Salivary gland ion exchange and fluid secretion

Saliva production is initiated by water-permeable salivary gland acinar cells, which secrete an isotonic plasma-like filtrate. Unidirectional secretion and coordinated myoepithelial contractions push fluid from the acinar lumen through water-impermeable ductal cells which modify the fluid through ion channels. Na+/Cl- channels reabsorb sodium and chloride, while Na+/K+ and K+/H+ exchangers are postulated to secrete K+ and HCO3-into the ductal lumen to buffer and condition the saliva (Figure 1A). While the exact mechanism of K+ secretion is unclear, it is known that K+ concentration is significantly higher in the saliva than in the plasma and remains high in the lumen even in low flow conditions (14, 15).

Figure 1.

Changes in luminal ion concentration and saliva flow in health and disease (A) In a healthy salivary gland, isotonic fluid is secreted by the acinar cells via aquaporin-5 (AQP5) water channels, and sodium and chloride ions follow water flow. Fluid is modified in the ducts through the reabsorption of sodium and chloride and secretion of potassium, creating a hypotonic final saliva. (B) In a radiation-damaged salivary gland, acinar cells secrete minimal fluid while the ducts remain intact, minimal fluid movement leads to largely non-operative sodium and chloride channels. K+ concentration remains high. (C) In rAAV2hAQP1-transduced glands, AQP1 insertion in the ductal cells increases ductal permeability. Water flows down the osmotic gradient created by the K+ in the lumen, leading to an increase in saliva flow. Figure made in Biorender.

IR induces double-stranded DNA breaks causing cell death. Acinar cells are sensitive to radiation, resulting in cell death, reduced functional parenchyma and a decrease in primary fluid secretion (16, 17). Ductal cells (striated and excretory ducts) remain largely intact following radiotherapy. Since Na+ and Cl- ions follow water flow in the acinus, we reasoned that their low concentration would render the corresponding transporters essentially nonfunctional. (Figure 1B). However, the sustained high K+ concentration in the ductal lumen would create an osmotic gradient. The idea of utilizing this gradient to drive water flow was first proposed by Bruce Baum, DDS, PhD (18), who hypothesized that increasing water permeability of the residual ductal cells via an unregulated water channel (aquaporin-1) inserted into the membrane could permit water secretion into the lumen, increasing saliva flow (Figure 1C).

Human aquaporins (AQP1–5, AQP7–10) comprise a family of membrane water channels that move water following osmotic gradients. Human aquaporin-1 (hAQP1) is constitutively active and canonically present in red blood cells, endothelial cells, and renal proximal tubules (19, 20). While hAQP1 is not endogenously expressed by salivary gland acinar cells, ectopic expression of aquaporin-1 in salivary gland epithelial cells permits water movement along an osmotic gradient (21). Thus, a system to stably deliver hAQP1 to salivary epithelial cells could be a promising therapeutic option for IR-mediated salivary gland injury.

3. Gene therapy as a potential long-term solution to restore salivary gland function

Adenovirus vectors for AQP1 delivery

Adenoviruses are naturally occurring respiratory pathogens. The creation of replication-incompetent adenovirus vectors to package genes of interest has been fundamental in advancing gene therapy, and has led to clinical trials for the treatment of cystic fibrosis (22) and ornithine transcarbamylase deficiency (OTCD) (23).

The efficacy of adenovirus vector gene transfer in the salivary glands has been tested in several animal models (Table 1). A recombinant adenovirus carrying the hAQP1 gene (AdhAQP1) delivered to irradiated rat submandibular glands (18, 24), and minipig salivary glands (25) successfully increased saliva flow in a dose-dependent manner. Though recombinant adenovirus vectors are less immunogenic than wild-type viruses, there is still a risk of strong innate and adaptive immune responses. Toxicology and biodistribution studies in rats demonstrated that various vector dosages delivered to the submandibular glands resulted in no significant changes in serum lab chemistry or body weight. While vector was present in other body sites (e.g., brain, and liver) in a dose-dependent manner, the frequency of detection decreased over time, while vector detection in the submandibular glands remained highest (26, 27). Similarly, in the minipig model and in a nonhuman primate model (25, 28), there was minimal systemic immune response to vector administration, demonstrating that adenovirus vectors could be a safe and effective means for hAQP1 delivery.

Table 1.

Studies on virus vector-mediated AQP1 gene transfer to the salivary glands

AdhAQP1
Model	Max duration	Dose	Host Response	Saliva increase?	Reference
Rats	3 days	5×109 pfu	N/A (dexamethasone administered)	Yes	(18)
	9 days	5×109 pfu	-Elevated LDH, WBC, SGOT, serum P and K -Evident gross morphology changes in SMG -Neutrophil infiltration, focal necrosis, edema in glands	N/A	(24)
	92 days	2×108 to 2×1011 pfu	-Temporary SMG inflammation -Increase in WBC (females) -Dose-dependent increase in neutralizing antibodies -Vector present in gland at 90 days but not in other body tissues after 57 days	N/A	(27)
Minipigs	14 days	1×108 or 1×109 pfu	-Elevated globulins and WBCs	Yes, transient	(25)
Non-human primate	14 days	1×108 or 2×109 pfu	-Decrease in lymphocytes in either Ad control or AdhAQP1 -No changes in serum chemistry, liver values or -No histopathology	Inconsistent	(28)
Clinical trial	42 days-four years	4.8×107 to 4.4×1010 particles	-Elevated serum amylase, CRP, WBC - mild parotitis -Increased gallium uptake -Minimal neutralizing antibody increase -Elevated cytotoxic lymphocytes	-5/11 subjects	(29–31)
rAAV2hAQP1
Mice	13 weeks	1×108 to 4.4×1010 particles	-Mild inflammation due to manual manipulation -No changes in serum chemistry -Transient elevation in WBC, decrease in HGB -Vector distributed to treated SMGs and draining lymph nodes, not present in other tissues for 57 days	N/A	(40)
Minipigs	16 weeks	1×1010 to 3×1011 particles	-Elevation of GM-CSF in saliva -Hematocrit elevation -Increased neutralizing antibodies	Yes	(41)

AdhAQP1 clinical trial

In the first Phase I clinical trial to assess hAQP1 gene delivery to the salivary glands, an adenovirus vector was delivered to a single parotid gland in 11 subjects with IR-related salivary gland hypofunction (29). In this open-label, single-dose, dose-escalation trial, vector was targeted to the parotid gland via retroductal cannulation of Stensen’s duct to locally infuse an exact volume of vector. Low-dose vector administration led to an increase in saliva flow in 5 subjects at 42 days post-administration, which was sustained up to 2 years (30). Moderate inflammation was observed in this treatment-responsive group, including a systemic increase in cytotoxic lymphocytes and vector-related peripheral blood mononuclear cell (PBMC) proliferation (31). In contrast, high-dose delivery was more likely to trigger a significant adverse immune response without improvement in saliva production. Similar inflammatory effects were seen in non-responders but at higher levels. These findings prompted further investigation into a less immunogenic method to stably express hAQP1 in the salivary glands.

Adeno-associated virus (AAV) vectors for AQP1 delivery

AAVs, like adenoviruses, are naturally occurring viruses. Unlike adenoviruses, AAVs require helper-virus functions, classically from adenovirus, for productive replication. Of the dozens of serotypes, AAV2 and AAV5 are the most well-characterized as a delivery vehicles for gene transfer in humans and have approved drugs with suitable safety profiles (32). Compared to adenovirus vectors, AAV vectors are purposed to be less immunogenic (33). Though many individuals have pre-existing neutralizing antibodies to AAV due to natural infections (34), the influence of these antibodies on gene transfer efficacy is variable (35–37). Cell-mediated immune response to AAV vectors may also limit transgene expression at high doses (38, 39).

As with adenovirus, animal models have been used to assess AAV2 transduction of the salivary glands (Table 1). Delivery of recombinant AAV2 vector packaging human aquaporin-1 (rAAV2hAQP1) to murine salivary glands led to mild, transient inflammation at the site of delivery but no changes in overall body weight or serum chemistry. This response may partially be due to the limited vector biodistribution-by 57 days post-administration, vector DNA was only detected in mouse salivary glands and draining lymph nodes (40). When the same vector was delivered to minipig parotid glands, there were minimal serologic changes, and saliva flow increased in irradiated glands (41) without adverse health consequences. Finally, studies in non-human primates demonstrate similar, long-lasting expression of the vector transgene with negligible systemic effects and with nominal vector load in non-targeted tissues (42).

These findings, combined with the results from the adenovirus clinical trial, formed the foundation and rationale for our Phase I trial, “Safety of a Single Administration of AAV2hAQP1, an Adeno-Associated Viral Vector Encoding Human Aquaporin-1 to One Parotid Salivary Gland in People with Irradiation-Induced Parotid Salivary Hypofunction” (NCT02446249), discussed below.

Discussion

A novel therapeutic for IR-mediated salivary gland dysfunction

Given the lack of an effective, long-term treatment option for salivary hypofunction coupled with the systemic immune response to adenovirus vectors, we propose to investigate an AAV2 vector for delivery of hAQP1 to human salivary glands (rAAV2hAQP1). The specific aims for this study are as follows:

Primary outcome measure: assess the safety and tolerability of rAAV2hAQP1 delivery to human salivary glands.

Secondary outcome measures: measure changes in saliva flow, durability of transgene expression, and host response.

Participants

This study includes adults with a history of salivary gland damage due to radiation therapy for head and neck cancer. All individuals must have non-zero stimulated saliva flow in the targeted gland as an indicator of some remaining functional epithelial tissue. Five cohorts of three subjects each receive increasing vector doses (1×10¹⁰ to 6×10¹¹ vector particles/gland).

Pre-treatment procedures

Study procedures are summarized in Figure 2A. Pre-dosing visit 1 includes screening exams and collection of baseline information, including answers to xerostomia questionnaires and measurement of parotid and submandibular/sublingual saliva. If a subject meets screening requirements, they are further evaluated in pre-dosing visit 2, where a sialogram is performed to evaluate ductal architecture and a test infusion of buffer is administered to determine optimal infusion volume.

Figure 2.

Clinical trial procedures and timeline (A) Schematic showing the study procedures that are performed throughout the course of the study, including procedures performed prior to treatment, at every visit, periodically, and at the study endpoint. (B) Schematic of retrograde cannulation into Stensen’s duct of the parotid gland for delivery of rAAV2hAQP1 vector. (C) Timeline of the study through 36 months (“m36”). Subjects are seen at 90 days (“d-90”) and 14 days (“d-14”) prior to vector delivery, and at regular intervals thereafter. Following vector delivery, subjects stay at the in-patient hospital for two days. Phone call visits include administration of patient questionnaires and collection of health history from the subject. All other visits occur at the NIH Clinical Center. Figures made in Biorender.

Vector infusion

The rAAV2hAQP1 vector is replication-deficient and designed to transfer complementary DNA (cDNA) encoding hAQP1 to adult subjects. Vector is infused retrograde using positive pressure into the parotid gland via the Stensen’s duct (Figure 2B). Subjects are admitted to the NIH Clinical Center inpatient unit after infusion for 3 days and monitored for adverse events. Subjects return at post-infusion days 14, 28, 56, 90, 150 and 360 in the first year following vector administration and every 6 months thereafter until the end of the study period (Figure 2C).

Post-dosing assessments

Immune response is measured by serum anti-AAV2 antibodies, lymphocyte proliferation in response to vector and gallium-67 citrate scans of the glands (43). Unstimulated and citric acid-stimulated saliva from the parotid and submandibular/sublingual glands are collected using standardized protocols. Glandular uptake and secretion are measured using accumulation and release of technetium-99 pertechnetate (Tc-99m) (44). Vector DNA is measured in the serum and saliva throughout the duration of the study. The expression of the AQP1 gene in the targeted gland will be assessed by microscopic evaluation of core needle biopsy samples from treated parotid glands.

Conclusions

Gene therapy was proposed in the 1960s as an approach to treat diseases caused by genetic defects. Since then, the field has expanded beyond the correction of single genes in monogenic disorders to include strategies for genetically complex and acquired diseases. In this review, we discuss the scientific foundation for our clinical trial evaluating hAQP1 gene transfer for radiation-induced salivary gland dysfunction. The accumulated clinical experience with viral-vector gene delivery, together with our preclinical biodistribution and toxicology studies in mice and minipigs, provides a strong rationale for assessing the safety and feasibility of this approach in the proposed trial.

In addition to radiation injury, other causes of salivary gland dysfunction may also be amenable to hAQP1 gene therapy. For example, Sjögren’s disease (SjD), a common chronic systemic autoimmune disease, prominently reduces the function of the salivary and lacrimal glands. Like radiation-induced xerostomia, immune-mediated epithelial dysfunction and glandular injury can result in severe hyposalivation and xerostomia, as well as reduced water permeability of affected acinar cells (45). Current therapies for dryness symptoms in Sjögren disease are largely palliative and do not produce sustained restoration of salivary flow. As research on local AAV-mediated gene delivery advances, AAV2-hAQP1 therapy may prove to be a viable strategy for managing a wide range of diseases with sicca manifestations.