Indomethacin Treatment Post-irradiation Improves Mouse Parotid Salivary Gland Function via Modulation of Prostaglandin E2 Signaling

Annually, >600,000 new cases of head and neck cancer (HNC) are diagnosed worldwide with primary treatment being surgery and radiotherapy. During ionizing radiation (IR) treatment of HNC, healthy salivary glands are collaterally damaged, leading to loss of function that severely diminishes the quality of life for patients due to increased health complications, including oral infections and sores, cavities, and malnutrition, among others. Therapies for salivary hypofunction are ineffective and largely palliative, indicating a need for further research to uncover effective approaches to prevent or restore loss of salivary gland function following radiotherapy. Previous work in our lab implicated prostaglandin E2 (PGE2) as an inflammatory mediator whose release from radiation-exposed cells promotes salivary gland damage and loss of function. Deletion of the P2X7 purinergic receptor for extracellular ATP reduces PGE2 secretion in irradiated primary parotid gland cells, and salivary gland function is enhanced in irradiated P2X7R–/– mice compared to wild-type mice. However, the role of PGE2 signaling in irradiated salivary glands is unclear and understanding the mechanism of PGE2 action is a goal of this study. Results show that treatment of irradiated mice with the non-steroidal anti-inflammatory drug (NSAID) indomethacin, which reduces PGE2 production via inhibition of cyclooxygenase-1 (COX-1), improves salivary gland function compared to irradiated vehicle-treated mice. To define the signaling pathway whereby PGE2 induces salivary gland dysfunction, primary parotid gland cells treated with PGE2 have increased c-Jun N-terminal Kinase (JNK) activation and cell proliferation and reduced amylase levels and store-operated calcium entry (SOCE). The in vivo effects of blocking PGE2 production were also examined and irradiated mice receiving indomethacin injections have reduced JNK activity at 8 days post-irradiation and reduced proliferation and increased amylase levels at day 30, as compared to irradiated mice without indomethacin. Combined, these data suggest a mechanism whereby irradiation-induced PGE2 signaling to JNK blocks critical steps in saliva secretion manifested by a decrease in the quality (diminished amylase) and quantity (loss of calcium channel activity) of saliva, that can be restored with indomethacin. These findings encourage further attempts evaluating indomethacin as a viable therapeutic option to prevent damage to salivary glands caused by irradiation of HNC in humans.

Annually, >600,000 new cases of head and neck cancer (HNC) are diagnosed worldwide with primary treatment being surgery and radiotherapy. During ionizing radiation (IR) treatment of HNC, healthy salivary glands are collaterally damaged, leading to loss of function that severely diminishes the quality of life for patients due to increased health complications, including oral infections and sores, cavities, and malnutrition, among others. Therapies for salivary hypofunction are ineffective and largely palliative, indicating a need for further research to uncover effective approaches to prevent or restore loss of salivary gland function following radiotherapy. Previous work in our lab implicated prostaglandin E 2 (PGE 2 ) as an inflammatory mediator whose release from radiation-exposed cells promotes salivary gland damage and loss of function. Deletion of the P2X7 purinergic receptor for extracellular ATP reduces PGE 2 secretion in irradiated primary parotid gland cells, and salivary gland function is enhanced in irradiated P2X7R −/− mice compared to wild-type mice. However, the role of PGE 2 signaling in irradiated salivary glands is unclear and understanding the mechanism of PGE 2 action is a goal of this study. Results show that treatment of irradiated mice with the non-steroidal anti-inflammatory drug (NSAID) indomethacin, which reduces PGE 2 production via inhibition of cyclooxygenase-1 (COX-1), improves salivary gland function compared to irradiated vehicle-treated mice. To define the signaling pathway whereby PGE 2 induces salivary gland dysfunction, primary parotid gland cells treated with PGE 2 have increased c-Jun N-terminal Kinase (JNK) activation and cell proliferation and reduced amylase levels and store-operated calcium entry (SOCE). The in vivo effects of blocking PGE 2 production were also examined and irradiated mice receiving indomethacin injections have reduced JNK activity at 8 days post-irradiation and reduced proliferation and increased amylase levels at day 30, as compared to irradiated mice without indomethacin. Combined, these data suggest a mechanism whereby irradiation-induced PGE 2 signaling to JNK blocks critical steps in saliva secretion

INTRODUCTION
Each year, >600,000 new cases of head and neck cancer (HNC) are diagnosed across the world (Johnson et al., 2020). Effective approaches to treat HNC include surgical excision of the tumor followed by IR, with or without chemotherapy (Cramer et al., 2019). During radiation treatment, salivary glands, located proximal to tumors, are collaterally damaged leading to reduced salivary gland function. Reduced saliva output causes numerous health complications, including increased rates of oral infections, cavities, and malnutrition and an overall poorer quality of life (Grundmann et al., 2009). Current treatment options for salivary hypofunction are palliative, relatively ineffective and costly to patients, indicating a need for further research to improve the quality of life for HNC patients (Jensen et al., 2019).
Wound healing is a complex regenerative process that occurs in three sequential but overlapping phases following tissue damage: (i) hemostasis/inflammation, (ii) proliferation/reepithelialization, and (iii) remodeling. While each phase of the response is essential for adequate wound repair, dysregulation at any stage can lead to insufficient repair or chronic inflammation and excessive scarring. Hemostasis/inflammatory responses begin immediately following a wounding event and typically last for about 3 days (Reinke and Sorg, 2012). Radiationinduced inflammatory responses have been studied in salivary glands but present conflicting results (Jasmer et al., 2020). Research suggests that interleukin (IL)-6 mediates induction of cellular senescence in irradiated (13 Gy) submandibular glands (SMGs), with elevated levels seen at 3 h and 14 days post-IR. However, both IL-6 knockdown and IL-6 treatment prior to radiation exposure unexpectedly protected SMGs from senescence at 8 weeks post-IR, leaving the role of IL-6 during the inflammatory response to radiation difficult to understand (Marmary et al., 2016). Interestingly, treatment of SMGs with an adenovirus containing the neurotrophic factor, neurturin, prevents hypofunction in irradiated (6 Gy fractions × 5 days) salivary glands, when given pre-but not post-IR (Lombaert et al., 2020). Radiation treatment also resulted in a significant increase in inflammation-associated gene expression in irradiated mice (300 days post-IR) and minipigs (16 weeks post-IR) that was reduced by neurturin-expressing adenovirus administration prior to radiation exposure and was associated with normalized morphology and increased size and function of the salivary gland (Lombaert et al., 2020). In contrast, another study demonstrated that there is a decrease in immune-related gene expression and reduced macrophage numbers in irradiated (15 Gy) SMGs at days 7-28 post-IR, whereas adenoviral-induced activation of the Sonic hedgehog (Shh) pathway at day 3 post-IR increased immune gene expression and macrophage numbers (Zhao et al., 2020). These studies suggest that an effective therapy to prevent salivary gland dysfunction due to radiation should target the hemostasis/inflammation phase of tissue damage, i.e., 0-3 days post-IR.
The second phase of the wound healing response, cell proliferation, is necessary to replace cells lost following damage; however, the homeostatic regulation of proliferation and differentiation is necessary to promote functional tissue repair. It has previously been proposed that the proliferative phase encompasses days 3-21 of the wound healing process, with transition to the remodeling phase occurring from day 21 through 1 year post-damage (Reinke and Sorg, 2012). In irradiated salivary glands, it has been shown that compensatory proliferation begins at day 5 post-IR and is mediated by activation of c-Jun N-terminal kinase (JNK) signaling (Wong et al., 2019). However, cell proliferation rates in irradiated salivary glands remain elevated compared to non-irradiated glands at chronic timepoints, days 30, 60, and 90 post-IR (Grundmann et al., 2010). Despite the increase in cell proliferation rates in irradiated salivary glands, the cells remain undifferentiated, as indicated by decreased amylase levels (Grundmann et al., 2010;Hill et al., 2014;Morgan-Bathke et al., 2014). Various pharmacological agents have been evaluated for restoration of irradiated parotid glands, where reducing the proliferative response correlates with increased salivary gland function in vivo (Grundmann et al., 2010;Hill et al., 2014;Morgan-Bathke et al., 2014). These data suggest that dysregulated signaling during the transition from the proliferative to remodeling phases of the wound healing process should be targeted to enhance salivary gland function following radiation damage.
The P2X7 receptor (P2X7R) for extracellular ATP released from damaged cells is a component of the innate immune system. Previous work from our lab showed that deletion or pharmacological antagonism of the P2X7R prevents salivary gland dysfunction in mice caused by radiation exposure (Gilman et al., 2019). Interestingly, mouse primary parotid gland cells lacking the P2X7R secrete significantly lower levels of the biologically active lipid, prostaglandin E 2 (PGE 2 ), basally and following radiation exposure, which suggests that P2X7Rmediated PGE 2 release may lead to salivary gland dysfunction that could be reversed by blocking the P2X7R (Gilman et al., 2019). PGE 2 is produced from plasma membrane phospholipidderived arachidonic acid that is first converted to prostaglandin H 2 by cyclooxygenases (COXs), COX-1 and COX-2, and then to PGE 2 by microsomal PGE synthase-1 (mPGES-1), mPGES-2 or cytosolic PGE synthase (cPGES) . PGE 2 acts by binding to four different E-prostanoid receptors (EPRs), EP1-4R, to induce the activation of multiple G proteins . EPRs, typically EP2R and EP4R, also can transactivate the epidermal growth factor receptor (EGFR) (Jiang et al., 2017). Based on their ability to activate G protein and EGFR signaling, PGE 2 -bound EPRs regulate multiple physiological processes, including, proliferation, differentiation, survival, cytokine production, immune cell migration and vasodilation/vasoconstriction (Dennis and Norris, 2015;Gilman and Limesand, 2020). PGE 2 also induces the phosphorylation of intracellular c-Jun terminal kinase (JNK) and it's downstream target c-Jun (Zeng et al., 2015;Zhong et al., 2015), which we have previously shown mediates the induction of compensatory proliferation in irradiated salivary glands (Wong et al., 2019). Indomethacin is a non-steroidal anti-inflammatory drug (NSAID) that functions via nonselective and reversible inhibition of COXs to reduce the synthesis of eicosanoids, including PGE 2 (Lucas, 2016). Due to the postulated role of PGE 2 in the induction of radiation-induced salivary gland dysfunction, we evaluated the hypothesis that indomethacin treatment would restore irradiated salivary gland function by blocking PGE 2 production and subsequent activation of JNK-mediated compensatory proliferation.

Mice
Animals were maintained according to the University of Arizona Institutional Animal Care and Use Committee (IACUC) regulations with protocols approved by the IACUC. Four to eight week-old C57BL/6J (stock no. 000664) or FVB/NJ mice (stock no. 001800) were purchased from Jackson Labs (Bar Harbor, ME, United States). Age-and sex-matched mice of the same genotype were randomly assigned to treatment groups. Mice were on 12 h light/dark cycles and housed in vented cages with food and water ad libitum. Where indicated, mice received intraperitoneal (IP) injections of vehicle (sterile saline with 10% ethanol) or indomethacin (1 mg/kg body weight, Sigma, no. I7378, St. Louis, MO, United States) prepared from a 10 mg/mL stock solution in 100% ethanol, warmed at 55 • C for 5 min, diluted to 1 mg/mL in saline and sterilized by passage through a 0.22 µm Polyvinylidene difluoride (PVDF) filter.

Radiation Treatment
Mice were sedated via an IP injection of a mixture of ketamine/xylazine (50-10 mg/kg), constrained in 50 mL tubes and shielded with >6 mm lead, leaving only the head and neck region exposed. Mice or cells were placed in the radiation field and received a single 5 Gy dose of radiation from a 60 Co Teletherapy unit at 80 cm distance from the source and ∼0.3-0.4 Gy/min (Theatron-80, Atomic Energy of Canada, Ottawa, ON, Canada) or a 225 kV X-ray unit at 48 cm distance from the source and 1.4 Gy/min (RS2000, Rad Source Technologies, Buford, GA, United States).

Saliva Collection
Salivary flow rates were evaluated on days 3, 10, or 30 following radiation. Saliva production was stimulated with an IP injection of carbachol (0.25 mg/kg body weight) and whole saliva was collected for 5 min via vacuum aspiration into pre-weighed tubes and then snap-frozen. Salivary flow was calculated by taking the difference in tube weight (post-collection minus pre-collection) and dividing by 5 min to express saliva secretion in milligrams per minute. Saliva flow rates were normalized to the average of the non-irradiated, vehicle-injected group on each day of collection.

Cyclooxygenase Activity Assay
Primary cells were prepared as described above. On day 2 of culture, cells were treated with indomethacin (25 µM in saline from the 100% ethanol stock solution) or vehicle (saline with 0.1% ethanol) and cells were collected 24 h later. Protein was extracted as described below and COX activity was measured via a COX activity assay (Abcam, no. ab204699, Cambridge, United Kingdom) following the manufacturer's instructions.

Prostaglandin E 2 Enzyme-Linked Immunosorbent Assay
Primary cells were prepared as described above. On day 2 of culture, media was replaced and cells were treated with indomethacin (25 µM in saline) or vehicle (saline with 0.1% ethanol) 1 h prior to receiving a 5 Gy dose of radiation. Cellfree supernatants were collected at indicated timepoints following radiation and the PGE 2 concentrations were determined with an enzyme-linked immunosorbent assay (PGE 2 ELISA, R&D Systems, no. KGE004B, Minneapolis, MN) following the manufacturer's instructions, where PGE 2 was normalized to the protein concentration of the corresponding cell culture dish.
United States) were diluted in serum-free primary cell media and used to treat dispersed parotid cell aggregates at the indicated timepoints for dose-response and kinetic analyses. For amylase staining, PGE 2 was diluted to 10 µM in serum-complete primary cell media and cells were treated on culture plates for 24 h prior to collection. Cells were centrifuged and resuspended in lysis buffer to extract proteins as described below.

Protein Isolation and Quantification
Parotid glands were harvested and snap-frozen from untreated and irradiated mice at day 8 or 30 post-IR. Tissues were homogenized in radioimmunoprecipitation assay (RIPA) buffer, with protease inhibitor cocktail (30 µL/mL; Sigma, no. P8340), sodium orthovanadate (5 mM) and phenylmethylsulfonyl fluoride (PMSF, 1 mM). Tissue homogenates were incubated on ice for 30 min, sonicated for 1-2 min and centrifuged at 12,000 RPM for 10 min at 4 • C to remove cell debris. Cells or aggregates were collected by scraping, centrifuged, resuspended in tissue protein extraction reagent (T-PER, Thermo no. 78510) with protease inhibitor cocktail (30 µL/mL), sodium orthovanadate (5 mM), and PMSF (1 mM), incubated on ice for 10-15 min and centrifuged at 12,000 rpm for 10 min at 4 • C to remove cell debris. Protein content was measured with the Pierce Coomassie Plus Bradford assay (Thermo, no. 23236, Waltham, MA, United States) or the Bicinchoninic acid (BCA) assay (Thermo, no. 23225).

5-Ethynyl-2 -Deoxyuridine Incorporation Assay
Primary parotid gland cells were prepared as described above and were cultured on 18 mm collagen-coated glass coverslips (Neuvitro, no. H-18-collagen, Vancouver, WA, United States). On day 4 of culture, cells were treated with vehicle (DMSO), PGE 2 (Cayman Chemical, no. 14010), or EP receptor-selective (EP1R-EP4R) agonists at indicated doses: 17-phenyl trinor PGE 2 (EP1R agonist; Cayman Chemical, no. 14810), AH13205 (EP2R agonist; Santa Cruz Biotechnology, no. sc-214513), sulprostone (EP3R agonist; Cayman Chemical, no. 14765) and CAY10598 (EP4R agonist; Cayman Chemical, no. 13281). Cells were serum-starved for 2 h prior to treatment. All compounds were solubilized in DMSO, mixed into serum-free primary cell culture media (described above) and cells were treated for 24 h. During the last hour of EPR agonist treatment, cells were incubated with 5-ethynyl-2'-deoxyuridine (EdU) from the Click iT EdU cell proliferation kit (Invitrogen, no. 10337). EdU was fluorescently labeled following the manufacturer's instructions. Coverslips were mounted on glass slides with Prolong Gold Antifade Mountant (Thermo Fisher Scientific, no. 36934). Images were captured on a Leica DM5500 fluorescence microscope (Leica Microsystems, Wetzlar, Germany) with 4-megapixel Pursuit camera (Diagnostic Instruments, Inc.) at 200× magnification. EdU positive cells and total cells were manually counted. Groups were quantified by averaging the number of positive cells out of the total number of cells from 5 fields of view/slide and 3 slides per treatment. Graphs depict the average number of positive cells/total number of cells. Each symbol represents an independent sample.

Intracellular Calcium Quantification
Intracellular free Ca 2+ concentration ([Ca 2+ ] i ) in isolated parotid epithelial cells was quantified as previously described (Woods et al., 2018). Briefly, primary parotid gland cells untreated or treated with 10 µM PGE 2 for 24 h in serumfree DMEM/F12 media containing gentamycin (50 µg/ml) were washed with assay buffer (120 mM NaCl, 4 mM KCl, 1.2 mM KH 2 PO 4 , 1.2 mM MgSO 4 , 1 mM CaCl 2 , 10 mM glucose, 15 mM HEPES, 1% (w/v) BSA, pH 7.4) and then adhered to chambered coverslips using Cell-Tak cell adhesive (Corning Inc., Corning, NY, United States) and loaded with 2 µM of the calcium indicating dye, fura-2-AM (Life Technologies, Carlsbad, CA, United States) in assay buffer for 30 min at 37 • C, washed and incubated in dye-free assay buffer for 30 min. Prior to use, cells were washed again and placed in calcium-free assay buffer containing 0.2 mM EDTA and baseline fluorescence values were collected for 60 s prior to stimulation with carbachol (100 µM) for 210 s to evaluate muscarinic type 3 receptor functionality. Then, 3 mM calcium was added for the final 90 s to quantitate store-operated calcium entry (SOCE) into cells. Changes in the 340/380 nm fluorescence excitation ratio (505 nm emission) were detected using an InCyt Dual-Wavelength Fluorescence Imaging System (Intracellular Imaging, Cincinnati, OH, United States). Resulting fluorescence ratios were converted to [Ca 2+ ] i (nM) using a standard curve created with solutions containing known concentrations of Ca 2+ .

Histology
Salivary glands were harvested from mice at indicated timepoints post-IR and submerged in 10% neutral buffered formalin overnight. Tissues were sent to IDEXX Bioanalytics (Columbia, MO, United States) for processing where they were dehydrated with ethanol and xylene, embedded in paraffin and sectioned at 4 µm.

Immunofluorescence
Salivary gland sections were incubated at 37 • C for 20 min, submerged in Histoclear (National Diagnostics, for 10 min and rehydrated in ethanol gradations (100, 95, 70, and 50%) and deionized water for 10 min each. Sections were permeabilized for 15 min in 0.2% (v/v) Triton X-100 in PBS, then washed with PBS three times for 5 min each. Next, antigen retrieval was completed by microwaving sections in citric acid (pH 6.0) for 10 min and then cooling for 20 min. Sections were washed with PBS three times for 5 min each, blocked in 0.5% New England nuclear blocking agent (Perkin Elmer, no. 2346249) for 1 h at room temperature and then incubated in anti-amylase antibody (1:1,000 in 1% BSA; Sigma, no. A8273) or anti-Ki67 antibody [1:400 in 1% BSA, Cell Signaling, no. 9129] overnight at 4 • C. Slides were washed three times for 10 min each, incubated in fluorophore-conjugated goat anti-rabbit IgG antibody (Alexa Fluor 488, 1:500 in 1% BSA; Thermo, no. A-11008) for 1 h at room temperature, washed again three times for 10 min each and then rinsed in water for 5 min. Cell nuclei were stained with 4' ,6-diamidino-2-phenylindole (DAPI, 1 µg/mL; Invitrogen, no. D1306) for 3 min then washed with water for 10 min. Amylasestained sections were mounted with 50% glycerol in 10 mM Tris-HCl (pH 8.0), and Ki67-stained sections were mounted with ProLong Gold Antifade mounting agent (Invitrogen, no. P36934) and imaged the following day. Images were captured on a Leica DM5500 with 4-megapixel Pursuit camera at 400× magnification with identical camera settings used for all images. Quantification was done using ImageJ software with images from 20 fields of view per mouse with 5 mice per treatment for amylase area, or 5 fields of view per mouse with 4 mice per treatment for Ki67. Graphs depict the mean percentage of amylase positive area or the average percentage of Ki67 positive cells out of the total cell number. Each point represents an independent mouse.

Statistical Analysis
Statistical tests were run using GraphPad Prism 9 software (San Diego, CA, United States). Normally distributed data was assessed by Brown-Forsythe test. To determine significance between groups, a student's t-test, or a one-way analysis of variance (ANOVA) was used, followed by Dunnett's post hoc comparisons when comparing to a control group, or Bonferroni's post hoc comparisons when comparing all groups. A p-value of less than 0.05 is considered statistically significant. Specific p-values are indicated by the number of asterisks above groups ( * p < 0.05, * * p < 0.01, * * * p < 0.001, * * * * p < 0.0001).

Post-radiation Indomethacin Treatment Restores Salivary Gland Function in Mice
FVB/NJ (Figures 1A-D) or C57BL/6J (Figures 1E,F) mice were exposed to 5 Gy of radiation and were injected with the NSAID, indomethacin (1 mg/kg body weight), 1 h prior to IR only (Figures 1A,B, IR+1 Indo) or again at days 1 and 2 post-IR (Figures 1A,B, IR+3 Indo). Alternatively, mice received injections of indomethacin at days 3, 5, and 7 post-IR (Figures 1C-F). Stimulated salivary flow rates were measured on day 3, representative of an acute timepoint post-IR damage, and day 30, representative of a chronic timepoint post-IR damage (Grundmann et al., 2010). The results indicate that a single injection of indomethacin (1 Indo) prior to IR exposure does not preserve salivary gland function post-IR (Figures 1A,B). Interestingly, mice receiving both pre-and post-IR indomethacin (3 Indo) injections have reduced stimulated saliva output on day 3 that is only modestly improved by day 30 (Figures 1A,B). In contrast, FVB/NJ (Figures 1C,D) and C57BL/6 ( Figures 1E,F) mice receiving indomethacin injections on days 3, 5, and 7 post-IR (Figures 1C-F) show salivary flow rates at days 10 and 30 post-IR similar to untreated, vehicle injected (UT+Veh) mice (Figures 1C,D). In addition, indomethacin treatment alone does not alter salivary gland function in mice not treated with IR (Figures 1C,D). Combined, these data suggest that inhibiting eicosanoid synthesis with indomethacin following radiation exposure leads to restoration of salivary gland function.

Indomethacin Treatment of Primary Parotid Gland Cells Reduces COX-1 Activity and Prostaglandin E 2 Secretion Induced by Radiation
Indomethacin inhibits both COX-1 and COX-2 functions (Tanaka et al., 2002), but has greater selectivity for COX-1 (Warner et al., 1999;Brown et al., 2001). To evaluate whether indomethacin is COX-1-or COX-2-selective in salivary glands, parotid acinar cells were treated with vehicle or indomethacin (25 µM) for 24 h and COX-1 and COX-2 activities were measured. Results indicate that indomethacin preferentially inhibits COX-1 in primary parotid acinar cells (Figure 2A). We have previously shown that there are elevated levels of PGE 2 secreted from primary parotid gland cells following IR exposure, with lower PGE 2 levels detected in IR-exposed P2X7R −/− cells that correlate with improved salivary gland function (Gilman et al., 2019). To determine if indomethacin treatment modulates PGE 2 secretion, primary parotid gland cells were exposed to 24-72 h post-IR, whereas PGE 2 secretion is also inhibited by indomethacin in cells not treated with IR ( Figure 2B). These data show that indomethacin functions by inhibiting COX-1 activity and reduces constitutive and radiation-induced production of PGE 2 in primary parotid gland cells.

Prostaglandin E 2 Treatment Activates JNK Signaling in Primary Parotid Gland Cells
We have previously shown that the JNK pathway is activated at day 5 post-IR and promotes radiation-induced compensatory cell proliferation that correlates with the loss of saliva secretion (Grundmann et al., 2010;Hill et al., 2014;Wong et al., 2019). PGE 2 has been shown to stimulate JNK and c-Jun phosphorylation in human endometrial stromal cells (Zeng et al., 2015) and primary human skin fibroblasts (Arasa et al., 2019). To evaluate whether PGE 2 activates JNK signaling in salivary glands, primary parotid cell aggregates were treated with varying doses of PGE 2 for 10 min and phosphorylation of the JNK p54 and p46 isoforms and their downstream target, c-Jun, were assessed Immunoblots were generated from cell lysates using anti-phospho-JNK/SAPK (p54/p46) T183/Y185 or anti-phospho-c-Jun S73 antibodies, which were stripped and re-probed for JNK/SAPK (p54/p46) or c-Jun and then ERK1/2 as a loading control. Densitometry was performed using ImageJ software and protein content was normalized to the average of DMSO-treated cells (A-D) or cells treated for 0 min (E-H). Graphs represent the mean ± SEM. Each closed circle represents an independent sample. Significant differences from DMSO-treated cells or cells treated for 0 min were determined via a one-way ANOVA followed by Dunnett's post hoc comparisons (*p < 0.05, **p < 0.01, ***p < 0.001).
via Western analysis. Phosphorylation of both p54 and p46 JNK are enhanced following treatment with all doses of PGE 2 tested (Figures 3A-C), whereas c-Jun phosphorylation is modestly increased following 100 nM PGE 2 treatment (Figures 3A,D). To determine kinetic changes in JNK and c-Jun phosphorylation induced by PGE 2 , primary parotid cell aggregates were treated with 1 µM PGE 2 and phosphorylation of JNK p54 and p46 and c-Jun were assessed after 0, 5, 10, 20, and 30 min. Results indicate that phosphorylation of p54 and p46 JNK are elevated at 10 min following 1 µM PGE 2 treatment (Figures 3E-G). Interestingly, c-Jun phosphorylation is increased at all timepoints measured, with significantly higher levels observed at 5 and 20 min following PGE 2 treatment as compared to the 0 timepoint (Figures 3E,H). These data show that PGE 2 activates the JNK/c-Jun pathway in parotid glands.

Prostaglandin E 2 Decreases Amylase Levels and Store-Operated Calcium Entry in Primary Parotid Cells
Amylase is a marker for differentiated acinar cells and, following radiation damage, amylase levels are reduced in parotid glands, which correlates with reduced gland function (Grundmann et al., 2010;Hill et al., 2014;Morgan-Bathke et al., 2014). To evaluate whether PGE 2 influences amylase production in parotid glands, primary parotid cells were treated with 10 µM PGE 2 for 24 h and amylase levels were determined via Western analysis. Amylase levels are reduced in PGE 2 -treated cells when compared to vehicle-treated cells (Figures 5A,B). To identify other mechanisms by which PGE 2 mediates alterations in salivary gland function, primary parotid cells were treated with PGE 2 for 24 h and loaded with the calcium indicating dye, fura-2. Intracellular calcium concentration ([Ca 2+ ] i ) was measured for 6 min following carbachol (100 µM) stimulation in media lacking calcium, and 3 mM calcium was added 4.5 min after carbachol stimulation. The peak [Ca 2+ ] i in parotid gland cells following carbachol stimulation is not different between vehicle and PGE 2treated cells, indicating normal functionality of muscarinic type 3 receptors on parotid cells and no effect of PGE 2 on the [Ca 2+ ] in intracellular stores (Figures 5C,D). Following addition of 3 mM calcium into the media, there is reduced re-entry of calcium in PGE 2 -treated cells indicated by a reduction in [Ca 2+ ] i as compared to vehicle-treated cells, suggesting that PGE 2 signaling blocks SOCE in parotid glands (Figures 5C,D). Taken together, these data illustrate that PGE 2 reduces amylase levels and inhibits SOCE in primary parotid cells.

Indomethacin Treatment Reduces JNK Signaling in vivo at Day 8 Post-IR
We have previously shown that IR activates JNK signaling 5 days post-IR, which partially regulates the compensatory proliferation response in parotid glands (Wong et al., 2019). It is well described that JNK activation induces c-Jun phosphorylation (Johnson and Nakamura, 2007). To determine if indomethacin treatment modulates JNK signaling in vivo, mice with or without 5 Gy IR exposure were injected with vehicle or indomethacin (1 mg/kg body weight) at days 3, 5 and 7 post-IR. Parotid glands were harvested at day 8 and used for immunoblots. Phosphorylation of JNK p54 and p46 are increased in irradiated parotid tissues at day 8 post-IR, a response reduced with indomethacin treatment (Figures 6A-C). Further, phosphorylation of c-Jun that is significantly increased at day 8 post-IR is reversed with indomethacin treatment (Figures 6A,D). These data illustrate that post-IR indomethacin treatment blocks JNK activation in parotid glands, supporting a role for indomethacin in protection from IR-induced salivary gland damage.

Post-radiation Indomethacin Treatment Enhances Parotid Gland Amylase Levels and Reduces Compensatory Proliferation at Day 30
To confirm the in vitro findings that PGE 2 decreases amylase secretion, mice were untreated or exposed to 5 Gy radiation with vehicle or indomethacin injections at days 3, 5, and 7 post-IR and parotid glands were collected at day 30, representing a chronic timepoint following damage (Grundmann et al., 2010). Protein was extracted for immunoblot analysis, which shows that total amylase levels are reduced in irradiated, vehicleinjected mice (Figures 7A,B), whereas indomethacin treatment Immunoblots were prepared from cell lysates with an anti-amylase antibody which were stripped and re-probed for ERK1/2 as a loading control. Densitometry was done using ImageJ software and protein content was normalized to the average of DMSO-treated cells. Graphs represent the mean ± SEM. Each closed circle represents an independent sample. Significant differences from DMSO-treated cells were determined via a student's t-test (*p < 0.05). (C) Representative tracings of time-dependent fura-2 fluorescence in mouse primary parotid acinar cells pretreated with or without PGE 2 (10 µM) for 24 h, stimulated with carbachol (100 µM) in calcium-free media followed by addition of calcium (3 mM) where indicated. (D) Peak changes in [Ca 2+ ] i induced by carbachol and the indicated addition of 3 mM calcium. Values are mean ± SEM of results from 23 to 33 cells from two separate cell preparations as shown with closed circles. Significant differences from untreated cells were determined via a student's t-test (***p < 0.001).
increases amylase levels in irradiated mice to similar levels as unirradiated, vehicle-injected mice (Figures 7A,B). To confirm these findings and evaluate the proportion of amylase producing acinar cells in vivo, parotid glands were collected at day 30 for histological analysis and immunohistochemistry was used to determine the area of salivary gland tissue staining positive for amylase (graphed as percent of total area). The proportion of amylase secreting acinar cells is reduced at day 30 in irradiated, vehicle-injected mice, but indomethacin restored amylase levels to those in unirradiated, vehicle-injected mice (Figures 7C,D). To evaluate whether or not indomethacin exerts modulatory effects on other proteins that influence salivary gland secretion, whole tissue homogenates were used for immunoblots to assess levels of aquaporin 5 (AQP5) and muscle, intestine, and stomach expression 1 (MIST1). Irradiated, vehicle-injected mice have reduced levels of AQP5 in parotid gland homogenates when compared to untreated glands, with indomethacin-treated glands having levels that were not statistically different from either group ( Supplementary Figures 2A-C). Unexpectedly, MIST1 levels were not significantly different across treatment groups ( Supplementary Figures 2A-C). PGE 2 is able to induce proliferation following EP2-4 activation (Figure 4 and Supplementary Figure 1) and compensatory proliferation that occurs in irradiated salivary glands correlates with loss of function (Grundmann et al., 2010;Hill et al., 2014;Morgan-Bathke et al., 2014). However, the evaluation of indomethacin as a compensatory proliferation modulator post-radiation has not been previously explored. Parotid gland tissue sections were stained for the proliferation marker Ki67 to evaluate differences in compensatory proliferation following radiation in vehicle and indomethacin-treated mice. Ki67 positive cells are significantly increased in irradiated, vehicle-injected mice and reduced with post-radiation indomethacin treatment (Figures 7E,F). These data demonstrate that indomethacin treatment reduces the compensatory proliferation response and increases the concentration of the major enzyme amylase in irradiated parotid glands indicating improved differentiation of acinar cells, which suggests that this pharmaceutical should be further tested for Immunoblots were generated from tissue lysates using antibodies against phospho-JNK/SAPK (p54/p46) T183/Y185, or phospho-c-Jun S73 , which were stripped and re-probed for JNK/SAPK (p54/p46) or c-Jun and then ERK1/2 as a loading control. Densitometry was performed using ImageJ software and protein content was normalized to the average of UT + Veh-treated mice. Graphs represent the mean ± SEM. Each closed circle represents an independent sample. Significant differences were determined via a one-way ANOVA followed by Bonferroni's post hoc comparisons (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns = not significant).
restoration of the normal protein composition of saliva in patients undergoing radiotherapy for HNC.

DISCUSSION
Wound healing is a highly complex process that requires delicately coordinated signaling for adequate repair to occur. In irradiated salivary glands, there is conflicting data regarding induction of inflammatory responses (Marmary et al., 2016;Lombaert et al., 2020;Zhao et al., 2020). However, it is well characterized that excessive cell proliferation occurs, but without adequate production of functional differentiated cells (Grundmann et al., 2010;Hill et al., 2014;Morgan-Bathke et al., 2014), suggesting that inhibition of the pathways responsible for compensatory proliferation may improve salivary gland function post-IR. Indomethacin is an NSAID that inhibits both COX-1 and COX-2 functions, but COX-1 has a 10-fold lower IC 50 value for indomethacin than COX-2 (COX-1 IC 50 : 13 nM; COX-2 IC 50 : 130 nM) (Warner et al., 1999). These findings are consistent with our data from primary parotid gland cells, where 25 µM indomethacin caused a significant reduction in COX-1 activity after 24 h but had no effect on COX-2 activity (Figure 2A). Our previous studies demonstrated reduced PGE 2 secretion from primary parotid gland cells of P2X7R −/− compared to wild type mice that correlated with the preservation of salivary gland function post-IR, although the differences observed in expression levels or activity of COX isoforms did not correlate with changes in PGE 2 secretion (Gilman et al., 2019). Here, we show that indomethacin reduces PGE 2 production and secretion in both untreated and irradiated primary parotid cells, which correlates with reduced COX-1 activity (Figures 2A,B). PGE 2 has many physiological roles under homeostatic and inflammatory conditions. However, in the context of radiation damage to salivary glands, the generation of PGE 2 appears to be detrimental to the cell regenerative response, suggesting that NSAIDs are a possible therapeutic modality for restoring salivary gland function post-IR (Goldberg, 1986;Dennis and Norris, 2015;Gilman and Limesand, 2020). While suppressing PGE 2 production appears to be protective for irradiated salivary glands (Figures 1, 2) (Gilman et al., 2019), how PGE 2 influences salivary gland dysfunction has not been clearly delineated. Previous work on pancreatitis, an inflammatory disorder that is associated with high serum amylase content, found that PGE 2 treatment blocks amylase secretion from pancreatic acini (Mössner et al., 1991). Furthermore, canines with allograft pancreatic transplants that were treated with dimethyl-PGE 2 had reduced amylase levels in urine compared to untreated controls, indicative of improved pancreatic function with PGE 2 (Garvin et al., 1989). Indomethacin treatment has been shown to increase isoproterenol (ISO)-induced amylase secretion from rat parotid gland tissue, where enhanced amylase secretion observed following ISO (0.1 µM) stimulation was abrogated by cotreatment with PGE 2 (Hata et al., 1990). Here, we show that PGE 2 treatment of primary parotid gland cells reduces amylase levels (Figures 5A,B) and that post-IR indomethacin treatment of mice increased amylase levels in whole parotid tissue (Figure 7). We also show that although carbachol-induced Ca 2+ release from intracellular stores was unaffected, PGE 2 decreases SOCE in primary parotid gland cells (Figures 5C,D). Research on the mechanism whereby PGE 2 inhibits SOCE is limited, although studies have linked production of PGE 2 to activation of SOCE (Kuda et al., 2011;Jairaman et al., 2015). One previous study found that PGE 2 treatment (2 h) modulated FIGURE 7 | Indomethacin treatment enhances amylase levels and reduces compensatory proliferation in whole parotid glands at day 30 post-IR. (A,B) C57BL/6J or (C-F) FVB mice were untreated (UT) or received 5 Gy ionizing radiation (IR) with intraperitoneal injections of vehicle (Veh, saline with 10% ethanol) or indomethacin (Indo, 1 mg/kg body weight) at days 3, 5, and 7 post-IR. Parotid glands were extracted at day 30 post-IR. (A,B) Immunoblots were generated from tissue lysates and an anti-amylase antibody, which were stripped and re-probed for ERK1/2 as a loading control. (A) Representative Western blot images of changes in amylase levels following radiation and indomethacin treatments. (B) Densitometry was performed using ImageJ software and protein content was normalized to the average of the UT + Veh group. (C-F) Parotid glands were fixed, sectioned and immunohistochemistry was performed with (C,D) an anti-amylase antibody or (E,F) an anti-Ki67 antibody, as described in section "Materials and Methods." (C) Representative images of amylase positive area (40× magnification, scale bar: 50 µm). (D) Percent positive amylase area was determined using ImageJ software. The graph represents the amylase positive area as a percentage of the total area. (E) Representative images of Ki67 positive area (white arrows indicate Ki67+, DAPI+ cells; 40× magnification, scale bar: 50 µm). (F) Ki67 positive and total cell numbers were manually counted from 5 fields of view/mouse. The graph represents the Ki67 positive cell number as a percentage of the total cell number. (B,D,F) Graphs represent the mean ± SEM. Each closed circle represents an independent sample. Significant differences were determined via a one-way ANOVA followed by Bonferroni's post hoc comparisons (**p < 0.01, ****p < 0.0001; ns = not significant).
T cell receptor activation-induced calcium mobilization and SOCE in T cells (Choudhry et al., 1999), which express all EPR isoforms (EP1-4Rs) (Nataraj et al., 2001). In the current study, we show that PGE 2 treatment (24 h) of primary mouse parotid cells that express EP1, EP2, EP3, and EP4 receptors (Supplementary Figure 3) had no effect on carbachol-induced release of intracellular stores but did reduce subsequent SOCE (Figures 5C,D). Radiation has been shown to cause a loss of SOCE in salivary gland cells due to STIM1 cleavage through activation of the TRPM2 pathway (Liu et al., 2017), although there have been no reports on interactions between PGE 2 and its receptors (EP1-EP4) and this pathway. Further investigation of how PGE 2 modulates amylase production and SOCE in salivary glands is warranted and this information may be applicable to other exocrine organs, such as the pancreas and lacrimal glands. PGE 2 has been shown to activate the JNK/c-Jun pathway (Zeng et al., 2015;Arasa et al., 2019) and induces proliferation of pulmonary tumors (Zhong et al., 2015) and human mesenchymal stem cells (Yun et al., 2011). Previous work from our lab indicated that JNK/c-Jun activation leads to compensatory proliferation at day 5 post-IR (Wong et al., 2019) that remains elevated through day 90 (Grundmann et al., 2010) and reducing this dysregulated compensatory proliferation correlates with improved salivary gland function post-IR (Grundmann et al., 2010;Hill et al., 2014;Wong et al., 2019). The mediators responsible for radiationinduced cell proliferation in salivary glands are not well defined. Here, we show that PGE 2 treatment of primary parotid gland cells activates the JNK/c-Jun pathway (Figure 3) and PGE 2 or selective EP2, EP3, or EP4 receptor agonists induce cell proliferation (Figure 4), suggesting that PGE 2 mediates the compensatory proliferation response observed in parotid glands in vivo. Further supporting this observation, post-radiation indomethacin treatment significantly reduced JNK pathway activation in parotid glands of mice ( Figure 6) and Ki67 positive staining at day 30 (Figures 7E,F), which correlates with the ability of indomethacin to improve salivary gland function in irradiated salivary glands (Figures 1C-F). Taken together, our data suggest that indomethacin treatment is a viable pharmacotherapeutic approach to protect salivary glands from IR-induced damage by blocking PGE 2 production and its activation of the JNK/c-Jun pathway leading to dysregulated proliferation. Relevantly, excessive PGE 2 production has been implicated in the progression of cancer (Nakanishi and Rosenberg, 2013). In multiple HNC models there is evidence that increased production of PGE 2 increases tumor cell proliferation that can be inhibited with NSAID treatment (Zweifel et al., 2002;Pelzmann et al., 2004;Ye et al., 2004). In HNC xenografts with head and neck squamous cell carcinoma-1483 (SCC-1483) cells, increased COX-2-mediated PGE 2 production was observed and treatment with a COX-2 selective inhibitor or a PGE 2 -neutralizing antibody reduced tumor growth (Zweifel et al., 2002). Interestingly, treatment of oral SCC-25 cells with genistein, celecoxib or indomethacin reduced PGE 2 production and cell proliferation levels compared to controls (Ye et al., 2004). Lastly, indomethacin treatment of SCC-25 or SCC-9 cell lines reduced cell growth and induced apoptosis in vitro (Pelzmann et al., 2004). Therefore, the use of NSAIDs, such as indomethacin, may be a general means to restrict the growth of cells that become dedifferentiated, as is the case with irradiated salivary gland acinar cells and metastatic tumor cells, both of epithelial origin.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, and further inquiries can be directed to the corresponding author/s.

ETHICS STATEMENT
The animal study was reviewed and approved by the University of Arizona Institutional Animal Care and Use Committee.

AUTHOR CONTRIBUTIONS
KG curated and visualized data and drafted the manuscript. KG and JC designed and conducted the experiments. KG, KL, and GW obtained funding for completion of this work. All authors conceptualized experiments, edited the manuscript and approved the final version of the manuscript.

FUNDING
This work was supported by the NIH R01 grant DE023342 to KL and GW, and the F31 grant DE028737 to KG. The funding agency had no role in study design, data collection, data analysis, the decision to publish this study or the preparation of this manuscript.