Prevention of Irradiation-induced Salivary Hypofunction by Microvessel Protection in Mouse Salivary Glands

original article

© The American Society of Gene Therapy

Prevention of Irradiation-induced Salivary Hypofunction by Microvessel Protection in Mouse Salivary Glands Ana P Cotrim1, Anastasia Sowers2, James B Mitchell2 and Bruce J Baum1 Gene Therapy and Therapeutics Branch, National Institute of Dental and Craniofacial Research, Bethesda, Maryland, USA; 2Radiation Biology Branch, National Cancer Institute, National Institutes of Health, Bethesda, Maryland, USA 1

Treatment of most head and neck cancers includes radiotherapy. Salivary glands (SGs) in the irradiation (IR) field are irreversibly damaged resulting in severe hyposalivation. We evaluated the importance of SG endothelial cells to this IR-induced injury, and whether serotype 5 adenoviral (Ad5) vector-mediated transfer of basic fibroblast growth factor (AdbFGF) or vascular endothelial growth factor (AdVEGF) complementary DNAs would afford radioprotection. Four hours after IR, microvessel density (MVD) in SGs decreased by ~45%. However, if mice were pretreated with either AdVEGF or AdbFGF 48 hours before IR the loss in MVD was significantly reduced. An irrelevant vector, AdLacZ, encoding Escherichia coli β-galactosidase, was without effect. After 8 weeks, IR reduced salivary flow by ~65% in untreated mice. Mice pre-treated (using 5 × 109 particles/gland 48 hours prior to IR) with AdLacZ exhibited a reduction in salivary flow similar to untreated mice receiving IR. However, irradiated mice pre-treated with AdbFGF or AdVEGF showed a significant improvement in their salivary flow, to ~70% (P < 0.01) and 80% (P < 0.01), respectively, compared to non-irradiated control mice. These results are consistent with the notion that injury to the adjacent microvasculature may play an important role in SG radiation damage. Furthermore, our results suggest that a local transient treatment directed at protecting SG endothelial cells may be beneficial for patients undergoing IR for head and neck cancer. Received 7 February 2007; accepted 28 July 2007; published online 28 August 2007. doi:10.1038/sj.mt.6300296

Introduction Head and neck cancers account for ~5% of all cancers in the United States.1 The American Cancer Society estimated that ~40,000 new cases of oral cavity and oropharyngeal cancer would be diagnosed in the United States in 2006.2 The 5-year survival rate of such patients can reach up to 61% (ref. 3) and today there are more than 500,000 survivors of oral, head, and neck cancer living in the United States.1,2

The treatment of head and neck cancer typically consists of surgery and irradiation (IR). Chemotherapy can be added to this regimen to decrease the possibility of developing distant metastasis.4 Salivary glands (SGs) in the IR field are severely damaged and, consequently, this results in marked salivary hypofunction in ~80% of patients.4–6 The underlying mechanism of the IRinduced injury to the SGs is still an enigma.7,8 Typically, radiosensitive tissues are composed of primitive, undifferentiated cells with a high mitotic rate, e.g., hematopietic stem cells. However, SG acinar cells, which are the sole site of fluid secretion in this tissue, although highly radiosensitive, are well-differentiated cells exhibiting a low mitotic rate. Patients experiencing reduced salivary flow suffer considerable morbidity, including dental caries, mucosal infections, dysphagia, and extensive discomfort. Current management approaches remain palliative and are generally considered unsatisfactory.7–9 While most studies aimed at preventing IR-induced salivary hypofunction have focused on acinar cells, in the current study we explore an alternative possibility: that microvascular endothelial cells within the SGs are the primary target for IR damage. The possibility that microvascular endothelial cells might be targeted by IR was first put forth by Paris et al. in 2001.10 in their studies on gastrointestinal radiation damage. This idea is also supported by previous observations that indicate that (i) the survival factors for endothelium, e.g., vascular endothelial growth factor (VEGF), acidic and basic fibroblast growth factors (bFGFs), and interleukin 11, can protect the gut from radiation injury,11,12 and (ii) endothelium is a recognized principal target for radiation injury to lung and brain.13,14 Our results in the present study are consistent with earlier ones and confirm this association between IR-induced changes in endothelial cells and the development of IR damage to SGs. Specifically we show that 4 hours after IR, microvessel density (MVD) in murine SGs is significantly reduced. Furthermore, we show that a single local administration of a modest dose (of 5 × 109 particles/gland) of a serotype 5 adenovirus (Ad5) vector encoding either bFGF or VEGF 48 hours prior to IR (15 Gy), prevents rapid MVD loss in SGs and reduces the loss in salivary flow as measured 8 weeks post-IR.

Correspondence: Ana P. Cotrim, Gene Therapy and Therapeutics Branch, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, Maryland 20892-1190, USA. E-mail: [email protected] Molecular Therapy vol. 15 no. 12, 2101–2106 dec. 2007

2101

© The American Society of Gene Therapy

Microvessels and IR-induced Salivary Hypofunction

Initially, we exposed the heads of C3H mice to an IR dose of 15 Gy, a dose previously shown to reduce salivary flow by ~60%,15 and compared MVD in the SGs of these mice with that in SGs of non-IR mice. We monitored MVD from 0 to 24 hours after IR. Maximum loss of MVD was observed 4 hours after IR. The MVD in the irradiated group was reduced by ~45% (Figure 1). The average (±SEM) of the total number of microvessels measured in the control SGs was 376 ± 45, while for the irradiated group this value was 211 ± 10 (P < 0.001; Mann–Whitney t-test). These findings are similar to those reported by Paris et al.10 for MVD adjacent to intestinal crypts after IR. An alternative explanation for these findings is that IR does not actually reduce MVD, but rather leads to a reduction in the expression of the CD31 marker protein used for determining MVD. To test this latter possibility, we used human umbilical vein endothelial cells (HUVECs) and

a

Average number of microvessels/field

b

16 14 12 10 8 6 4 2 0

**

p < 0.001 *

**

Control

c

4h Non-IR

Average % of viable cells per group

60

50 Non-IR unmarked 40

30 Non-IR PE-CD31 20

10

0

8 Gy IR PE-CD31

0u

0CD31

2u

2CD31

4u

4CD31

8u

8CD31

Figure 1 Detection of endothelial cells in mouse submandibular gland with and without irradiation (IR). (a) Immunohistochemical detection of endothelial cells. CD31+ (BD Pharmingen) endothelial cells (positive cells stained in blue) are seen in sections of a submandibular gland from untreated mice. (b) Microvessel density (MVD) based on CD31+ cells. The average number of positive cells in 40 fields per gland in the nonirradiated control group and in glands from a group of mice 4 hours after IR is shown. (n = 6 mice/group). CD31+ MVD in salivary glands with 95% confidence intervals (CIs) are shown in a box plot. The upper and lower boundaries of the box represent the 75th and 25th percentile, respectively, of the number of microvessels per field per mouse. The horizontal line within the box represents the median value, and the error bars represent the 95% CIs. Asterisks above and below the bars represent out-of-range values. A total of 12 glands from six mice in each treatment group were studied, and 40 fields were counted/gland. (c) Fluorescence-activated cell sorting (FACS) analyses of non-IR cells and cells (HUVECs) irradiated at 2, 4, and 8 Gy either unstained (u) or stained with CD31 antibody to evaluate if CD31 expression is affected by IR. The bar graph shows the mean ± SEM for the results of FACS analyses with HUVECs. Representative FACS analyses are shown to the right side of panel c. HUVECs, human umbilical vein endothelial cells.

2102

measured CD31 levels in cells after IR. As shown in Figure 1c, 4 hours after IR, at several radiation doses, there was no reduction in the level of CD31expressed by HUVECs, i.e., the percentage of viable cells and the percentage of CD31 antibody-positive cells was comparable. Based on this result, we hypothesized that damage to microvascular endothelial cells during IR of SGs may be important in the development of IR-induced salivary hypofunction. Further, we speculated that protection of SG microvessels from IR damage might be accomplished via local production of transgenic angiogenic proteins. For the current study we used Ad5 vectors encoding either bFGF or VEGF, two often-studied angiogenic proteins. Initially, we determined if the Ad5 vectors used (AdbFGF and AdVEGF) could direct the expression of both angiogenic proteins in vitro with A5 and 293 cells. For example, 48 hours following transduction with vectors, culture media from A5 cells showed high levels of bFGF (3,657 ± 655 pg/ml) and VEGF (1,478 ± 477 pg/ml). Similarly, following transduction with these vectors, media from 293 cells also contained high levels of these angiogenic proteins (Figure 2). We next examined the ability of these vectors to direct the expression of these angiogenic proteins in vivo in murine SGs. Initial in vivo time-course experiments (data not shown) ­indicated that the peak expression of the Ad5 vector–encoded transgenes in mouse SGs took place ~48 hours after vector ­delivery. Therefore, AdbFGF or AdVEGF were each delivered to the submandibular glands of C3H mice and after 48 hours, SGs, saliva, and serum were collected from all mice. We were unable to detect bFGF or VEGF in either saliva or serum. However, ­ aqueous extracts from almost all SGs showed high concentrations of both transgenic proteins that were well above background levels, i.e., up to 140 pmol/mg-protein of bFGF and 4.3 pmol/mg-­protein of VEGF (Figure 3). We next directly tested our hypothesis, by finding out whether pre-administration of AdbFGF or AdVEGF could decrease the observed IR-induced reduction in MVD in SGs. Different groups of mice received either the irrelevant control vector AdLacZ, or either one of the two vectors encoding the angiogenic proteins, AdbFGF or AdVEGF; administration was to both submandibular

bFGF or VEGF (pg/ml)

Results

9,000 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0

A5 293

AdbFGF

AdVEGF

Figure 2  Vector-induced basic fibroblast growth factor (bFGF) and vascular endothelial growth factor (VEGF) expression in vitro. Data are mean ± SEM of results from three experiments. Enzyme-linked immunosorbent assays were performed with media collected from 293 or A5 cells either not treated, used as control, or cells 48 hours after transduction with either AdbFGF or AdVEGF. Both bFGF and VEGF were undetectable in control cells (not transduced with vectors). Ad, adenovirus.

www.moleculartherapy.org vol. 15 no. 12 dec. 2007

© The American Society of Gene Therapy

160

Microvessels and IR-induced Salivary Hypofunction

bFGF

5

VEGF 100

4

120

# *

90

100

3

80 2

60 40

1

20 0 Control

AdbFGF

Control

AdVEGF

Average number of microvessels/field

Figure 3  Vector-induced basic fibroblast growth factor (bFGF) and vascular endothelial growth factor (VEGF) expression in vivo. Growth factor production was measured by enzyme-linked immunosorbent assay in aqueous extracts of salivary glands obtained 48 hours after delivery of either AdbFGF or AdVEGF as described in Material and Methods. Ad, adenovirus.

14 12

*

10 8

**

* *

6

*

*

*

*

*

4

*

2 0

Naïve

Control

AdLacZ

AdVEGF AdbFGF IR

Figure 4  Microvessel density in murine salivary glands with and ­without vector treatment. In these experiments microvascular endothelial cells were detected by measuring aquaporin-1 positive cells. Microvessel density in salivary glands with 95% confidence intervals (CIs) is shown in a box plot. The upper boundary of the box represents the 75th percentile of the number of microvessels per field per mouse. The lower boundary of the box represents the 25th percentile of the data distribution. The horizontal line within the box represents the median value, and the error bars represent the 95% CIs. The asterisks above and below the bars represent out-of-range values. A total of eight glands from four mice in each treatment group were studied, and 40 fields were counted/gland. See Materials and Methods for details. Ad, adenovirus.

glands 48 hours before IR and then MVD was assessed 4 hours after IR. As shown in Figure 4, a naïve group that did not receive any IR or vector treatment had an MVD of ~10. Control mice that received IR only had an MVD of 4.6 ± 0.7 (~50% reduction). The group that received AdLacZ showed an MVD of 5.5 ± 0.4, statistically identical to that seen in the SGs of mice that received IR but no vector. Conversely, mice that were pre-treated with AdbFGF or AdVEGF had an MVD of 7.9 ± 1.1 or 8.0 ± 0.6 (~20% reduction), respectively. When these data were analyzed by pairwise multiple comparisons (Bonferroni t-test), with α = 0.01, results with AdbFGF- and AdVEGF-treated mice were significantly different from those with both AdLacZ-treated mice and irradiated control mice (P ≤ 0.001). Lastly, we tested the hypothesis that this vector-induced protection of MVD in SGs would be associated with elevated levels of saliva production in mice long after IR. To do this we administered vectors to the submandibular glands of mice 48 hours Molecular Therapy vol. 15 no. 12 dec. 2007

Salivary flow (% non-irradiated control)

pmol/mg protein

n = 10/group *P = 0.01 **P = 0.002 #n = 5/group

**

140

#

#

AdbFGF

AdVEGF

80 70 60 50 40 30 20 10 0 No vector

AdLacZ

AdbFGF

AdVEGF

AdLacZ

+ 15 Gy

Figure 5 Effect of serotype 5 adenoviral (Ad5) vector administration before irradiation (IR) on salivary flow 8 weeks after IR. AdLacZ, an Ad5 vector–encoding Escherichia coli β-galactosidase; AdbFGF, Ad5 vector–encoding basic fibroblast growth factor (bFGF); AdVEGF, Ad5 vector–encoding vascular endothelial growth factor (VEGF). Mice (n = 10) received either no vector or 5 × 109 particles of the indicated ­adenoviral vector in each submandibular gland 48 hours prior to IR. After 8 weeks, salivary flow was measured as described in Materials and Methods.

before IR and evaluated salivary flow 8 weeks after IR. Control mice that did not receive IR treatment or vector administration had an average salivary flow of 300 ± 23 μl/10 minutes. As shown in Figure 5, mice that received 15 Gy of IR, but no vector, had a significant reduction in salivary flow after 8 weeks (~65%; 96 ± 19 μl). Mice that were administered 5 × 109 particles/gland of AdLacZ, AdbFGF, or AdVEGF and were not exposed to IR had an average salivary flow of 262 ± 22 μl, 255 ± 12 μl or 243 ± 8 μl, respectively. While these values were not statistically different from those of the control mice, on average they were ~15–20% lower than the mean salivary flow seen in the control mice (Figure 5). Mice that received AdLacZ plus IR treatment had a marked reduction in their salivary flow (54% reduction; to 140 ± 24 μl; P < 0.01). As also shown in Figure 5, mice pre-treated with either AdbFGF or AdVEGF 48 hours before IR showed only a modest reduction in salivary flow after 8 weeks (26 and 17%, respectively; similar to the salivary flow in non-irradiated mice administered these vectors). Both of these results were significantly higher than those obtained with irradiated mice administered AdLacZ (P < 0.01 and P < 0.002, respectively).

Discussion The primary objective of this study was to evaluate the hypothesis that microvascular endothelial cells in SGs are targets for IR damage. Our results are clearly consistent with this hypothesis. Within 4 hours following exposure to 15 Gy, we observed a ~45% loss in MVD in targeted SGs, as measured by the presence of two separate endothelial cell membrane markers, CD31 and aquaporin-1 (AQP1). To assess if these results reflected a rapid reduction in the expression of the CD31 marker in endothelial cells due to IR, we used HUVECs in vitro. HUVECs were either non-IR or IR in the range from 2 to 8 Gy and the percentage of viable cells and CD31 positive cells was analyzed by fluorescence-activated cell sorting. The percentage of viable cells in each group remained ~40%, and all of the viable cells were positive for the CD31 marker. 2103

Microvessels and IR-induced Salivary Hypofunction

Treatment with recombinant angiogenic proteins has previously been shown to protect microvascular endothelial cells from single doses of IR in both lung and brain.13,14 We therefore next proceeded to investigate whether localized, intraductal pre­treatment of SGs with an Ad5 vector encoding either of the two potent angiogenic proteins (bFGF and VEGF) could prevent IRinduced changes in the MVD observed. For these experiments we administered vectors 48 hours before IR, so that peak transgene expression would coincide with IR exposure,16 and at a dose that would lead to modest transgene expression yet minimize local immune reactivity.16,17 This maneuver proved successful and the rapid loss of MVD (i.e., endothelial cell marker) occurring in SGs after IR was greatly reduced following pre-treatment with either AdbFGF or AdVEGF. Interestingly, both VEGF and bFGF can be found in normal human adult SGs and it has been suggested that they help maintain epithelial cell homeostasis. VEGF is localized in SG serous epithelium18 while bFGF can be found in the parasympathetic nerve and ductal region.19,20 We next explored whether protection against the rapid loss of MVD in SGs could also prevent the reduction in salivary flow seen in irradiated mice long after the single IR dose was delivered. We again pre-administered AdbFGF or AdVEGF to murine SGs 48 hours before IR and then evaluated salivary secretion 8 weeks following IR. The results clearly showed that animals pre-treated with vectors encoding either of the angiogenic proteins displayed significantly greater levels of salivary secretion than animals that were untreated or were administered a control vector. Thus, the present studies have demonstrated an apparently close association between IR, endothelial cell integrity and SG function, and resulted in two quite novel findings: (i) microvascular endothelial cells in SGs could be early targets of single dose IR, and (ii) protection against loss of these cells through the localized administration of angiogenic transgenes appears to prevent much of the significant and persistent loss of salivary secretion seen following IR. The overall tissue response to IR involves a complex interaction between the different cell types present. These interactions start immediately after IR exposure and can continue for months or even years later.21 While many studies have tried to elucidate the mechanism of IR damage to SGs, it has remained unidentified and enigmatic.7 Generally, there have been two accepted hypotheses used to describe SG damage after IR.4–8 One suggests the occurrence of lethal radiation effects in cells, possibly due to cell membrane disruption,22–24 while the second suggests that hypofunction might be due to sublethal DNA damage in these cells with subsequent cell death by apoptosis.25–27 While it is likely that both of these hypotheses are to some degree partially correct, their focus has been entirely on the salivary epithelial cells. However, the absence of clear SG epithelial cell loss long after IR,28,29 as well as evidence gained from the studies of Paris et al. with intestinal epithelial cells,10 suggests that alternative possibilities may be involved in the mechanism of IR damage to SGs. While the present studies may help us understand the sequence of events that leads to SG damage after IR treatment, it is important to recognize that these studies were performed with a mouse model, using a single dose IR treatment unlike the fractionated IR doses used clinically. As our results indicate, 2104

© The American Society of Gene Therapy

microvascular ­endothelial cells appear to be early targets of IR in SGs, and they in fact may be the primary targets. However, much additional study is needed, particularly in larger animal models. For practical reasons, we irradiated the head and neck area of mice, and therefore not only microvascular endothelial cells were exposed to IR but multiple other cell types present in the glands. Ideally, in a larger model, using more clinically relevant fractionated IR doses, and more focused collimation, such generalized radiation exposure could be avoided. Presumably, the differences seen in microvascular endothelial cells by us, led to diminished nourishment and thus to long-term dysfunction of salivary epithelial cells. Interestingly, Schuller et al. recently21 showed that IR directed specifically to the microvasculature in murine intestine had no effect on the survival of murine intestinal crypt stem cells. In summary, our findings suggest an association between IR-induced changes in microvascular endothelial cells and IRinduced damage to SGs. Although much more study is required to understand if there is a direct cause and effect relationship between these two observations, and define the mechanism involved, the present findings suggest the potential of microvascular endothelial cells as targets for prophylactic gene therapy prior to IR of patients with head and neck cancers. While treatment with vectors encoding angiogenic proteins is not really ideal (due to their pro-tumorigenic potential), aiming to protect endothelial cells locally in the SGs and preventing long-term salivary hypofunction due to IR, presents itself as a unique, and promising, therapeutic approach. Our goal here is not to suggest a therapy with angiogenic growth factors per se (that could also promote tumor cell growth), but rather to address a possible alternative mechanism of IR damage in SGs and a novel target for prophylactic therapy. Pre-treatment of cancer patients with vectors encoding angiogenic proteins is certainly not desirable, but a local transient therapy directed specifically at SG microvascular endothelial cells may lead to improved SG function after IR, and thus diminish the significant morbidity suffered by patients.

Materials and Methods Ad5 vector production and characterization. Generation of the E1

deleted, replication deficient Ad5 vectors was performed as previously described.30 Briefly, 293 cells were co-transfected with the shuttle plasmid pAC-CMV-pLpA containing either the 18 kd isoform of human FGF2,31 a generous gift of Dr. Fred. Gage (Salk Institute), or the LacZ complementary DNA, and the Ad plasmid pJM17, using a calcium phosphate transfection system (Life Technologies, Gaithersburg, MD) to generate the recombinant vectors AdbFGF and AdLacZ. The cytomegalovirus promoter was employed to drive VEGF, bFGF, and LacZ expression. The Ad5 vector– encoding VEGF165 was generously provided by Dr. Karl Czaky (of the National Eye Institute, National Institutes of Health) and was amplified as described below. Recombinant vectors were plaque-purified, propagated in 293 cells, and purified by CsCl gradient centrifugation, as described.30,32 After purification, recombinant vectors were dialyzed against 4 l of dialysis buffer containing 10% glycerol, 0.1 mol/l Tris (pH 7·4), 5 mmol/l MgCl2 for 4 hours at 4 °C and stored in aliquots at −80 °C for later use. Vector titers were initially determined by measuring the optical density at 260 nm as described,33 and then by real-time quantitative polymerase chain reaction using transgene-specific primers. Vector titers used herein were based on the quantitative polymerase chain reaction assay using the ABI Prism 7700 Sequence Detector (Applied Biosystems, Foster City, CA). For testing www.moleculartherapy.org vol. 15 no. 12 dec. 2007

© The American Society of Gene Therapy

the vectors in vitro, 293 or A5 cells were transduced at a multiplicity of infection of 100 viral particles/cell and the culture supernatants collected 48 hours later. A5 cells are a well-studied epithelial cell line derived from methylcholanthrene-treated rat submandibular gland cells.34,35 Enzymelinked immunosorbent assays were performed to measure the expression of the two transgenic proteins [bFGF (DFB50); VEGF (DVE00), R&D ­systems, Minneapolis, MN]. Animal radiation. Female C3H mice, bred in the National Cancer Institute

Animal Production Area (Frederick, MD), were used for this study. The mice were 7–9 weeks of age at the time of experimentation and weighed between 20 and 30 g. All experiments were carried out under the aegis of a protocol approved by the National Cancer Institute Animal Care and Use Committee and were in compliance with the Guide for the Care and Use of Laboratory Animal (1996), National Research Council. SGs were irradiated by placing each animal in a specially built Lucite jig in such a way that the animal could be immobilized without the use of anesthetics.15 Additionally, the jig was fitted with a Lucite cone that surrounded the head and prevented head movement during the IR exposure. Single IR doses at 15 Gy were delivered only to the animal’s head by a Therapax DXT300 X-ray irradiator (Pantak, East Haven, CT) using 2.0 mm Al filtration (300 kVp) at a dose rate of 1.9 Gy/minutes. As we have previously shown, this IR dose leads to significant (~60%) loss of salivary flow after 8 weeks.15 Immediately after IR, the animals were removed from the Lucite jig and housed (five animals/cage) in a climate and light controlled environment, and allowed free access to food and water. Ad5 vector in vivo delivery. Groups of mice (n = 4 or 5/experiment)

received 5 × 109 particles/gland of AdVEGF, AdbFGF, or AdLacZ, suspended in 50 µl of 0.9% NaCl, by retrograde ductal delivery in their submandibular glands.16,17,36 Animals were killed either 4 hours or 8 weeks (two separate experiments were conducted for each time) after vector delivery, and saliva and SGs collected as described below. Animal groups received 15 Gy, or were sham irradiated, 48 hours after vector delivery. In the two experiments in which animals were killed 4 hours after IR, SGs were collected for MVD determination, while for the remaining studies saliva and glands were collected 8 weeks after IR. Saliva and SG collection. For saliva collections, mice were first weighed

and then mild anesthesia was induced with a solution of Ketamine (100 mg/ml; Fort Dodge Animal Health, Fort Dodge, IA) and Xylazine (20 mg/ml; Phoenix, St. Joseph, MO) in sterile water, given intraperitoneally (1 µl/g of body weight). Salivary secretion was stimulated using 1 μl/g body weight of a pilocarpine solution (50 mg/ml) subcutaneously. Saliva collection began within 2 minutes of pilocarpine administration. Animals were positioned with a 75-mm hematocrit tube (Drummond, Broomall, PA) placed in the oral cavity and whole saliva was collected into pre-weighed 0.75 ml Eppendorf tubes for 10 minutes. The amount of saliva collected was determined gravimetrically. Immediately afterward, anesthetized animals were killed by cervical dislocation and the submandibular glands were removed. Submandibular glands were first cleaned of extraneous tissue and then fixed in either 10% formalin (for AQP1 determination) or formalin free zinc fixative (for CD31 determination; BD Pharmingen, San Jose, CA), embedded in paraffin, and 5 µm sections obtained. For the experiments shown in Figure 3, glands were removed after euthanasia as above and aqueous extracts were prepared by homogenizing the glands in 0.1 mmol/l NaHCO3 and protease inhibitors (cat. no. 1697498, Complete, Roche, Indianapolis, IN). Immunohistochemistry and determination of MVD. Submandibular gland

sections were subjected to immunohistochemical staining specific to either one of two markers for microvascular endothelial cells: CD31 (ref. 37) and AQP1.32 Similar results were obtained with both markers. Sections were first incubated in a 1:100 dilution of either anti-CD 31 (BD Pharmingen, San Jose, CA) or anti-AQP1 (Alpha Diagnostics, San Antonio, TX). Molecular Therapy vol. 15 no. 12 dec. 2007

Microvessels and IR-induced Salivary Hypofunction

Primary antibody was detected with appropriate biotinylated second antibodies followed by incubation with streptavidin-conjugated horseradish peroxidase and colorimetric detection with true blue (KPL, Gaithersburg, MD) for CD 31 staining and 3,3′-diaminobenzidine (SK4100, Vector laboratories, Burlingame, CA) for AQP1 staining. MVD was determined by counting the number of CD31 or AQP1 stained cells per field at ×200 ­magnification. The entire submandibular section was counted, i.e., an average of 40 fields/gland. Flow cytometry analyses. HUVECs were plated in triplicate in 6-well

plates. After reaching confluency, cells were either left untreated or irradiated with 2, 4, or 8 Gy and incubated for 4 hours at 37 °C. Cells were then trypsinized, washed two times with phosphate-buffered saline and resuspended in 300 μl of endothelial basal medium. Thereafter, 3 μl of RPhycoerythrin-conjugated mouse anti-human CD31 monoclonal antibody (BD Pharmingen, San Jose, CA; cat no. 555446) were added and incubated with cells for 1 hour at 4 °C. Cells were centrifugated at 1,000 rpm for 4 ­minutes and washed three times with phosphate-buffered saline. Cells were next resuspended in 300 μl of phosphate-buffered saline and analyzed by fluorescence-activated cell sorting in a FACS Canto Flow Cytometer (BD Biosciences, San Jose, CA). The results shown in Figure 1c, represent the percentage of viable cells in 10,000 measured events. Statistical analysis. Descriptive summaries (means ± SEM) of the data

were determined. A general linear model (analysis of variance) was used to analyze different response variables. Putative explanatory variables included radiation and type of treatment (with a different Ad5 vector), together with their first order interaction terms. Within each analysis of variance model group, comparisons were made using the pair wise multiple comparison procedures (Bonferroni t-test). Due to the multiplicity of comparisons that were made, we used α = 0.01 to offset the potential inflation in type I errors, i.e., only effects having P < 0.01 were considered as significant in our analyses.

Acknowledgments We thank Marc Kok, Jianghua Wang, Fumi Mineshiba, Changyu Zheng, and Simon Tran for helpful comments and assistance on the SG cannulation, and Corinne Goldsmith for help in Ad preparation. The research of Ana Cotrim was partially supported by a grant from CAPES, Brazil. This research was supported by the Intramural Research Programs of the National Institute of Dental and Craniofacial Research and National Cancer Institute, National Institutes of Health.

References

1. Greenlee, RT, Hill-Harmon, MB, Murray, T and Thun, M (2001). Cancer statistics, 2001. CA Cancer J Clin 51: 15–36. 2. American Cancer Association. Cancer Statistics (2006). Presentation http://www. cancer.org/docroot/PRO/content/PRO_1_1_Cancer_Statistics_2006_Presentation.asp 3. Division of Cancer Control and Population Sciences, National Cancer Institute (2004). 5-year relative survival rates based on cancer patients diagnosed from 1995 to 2000 and followed through 2001. Source: Surveillance, Epidemiology, and End Results Program, 1975–2001. 4. Vissink, A, Burlage, FR, Spijkervet, FK, Jansma, J and Coppes, RP (2003). Prevention and treatment of the consequences of head and neck radiotherapy. Crit Rev Oral Biol Med 14: 213–225. 5. Vissink, A, Jansma, J, Spijkervet, FK, Burlage, FR and Coppes, RP (2003). Oral sequelae of head and neck radiotherapy. Crit Rev Oral Biol Med 14: 199–212. 6. Nagler, RM and Baum, BJ (2003). Prophylactic treatment reduces the severity of xerostomia following radiation therapy for oral cavity cancer. Arch Otolaryngol Head Neck Surg 129: 247–250. 7. Nagler, RM (2002). The enigmatic mechanism of irradiation-induced damage to the major salivary glands. Oral Dis 8: 141–146. 8. Konings, AW, Coppes, RP and Vissink, A (2005). On the mechanism of salivary gland radiosensitivity. Int J Radiat Oncol Biol Phys 62: 1187–1194. 9. Shiboski, CH, Hodgson, TA, Ship, JA and Schiodt, M (2007). Management of salivary hypofunction during and after radiotherapy. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 103 (Suppl. 66): e1–e19. 10. Paris, F, Fuks, Z, Kang, A, Capodieci, P, Juan, G, Ehleiter, D et al. (2001). Endothelial apoptosis as the primary lesion initiating intestinal radiation damage in mice. Science 293: 293–297. 11. Houchen, CW, George, RJ, Sturmoski, MA and Cohn, SM (1999). FGF-2 enhances intestinal stem cell survival and its expression is induced after radiation injury. Am J Physiol 276: 249–258.

2105

Microvessels and IR-induced Salivary Hypofunction

12. Okunieff, P, Mester, M, Wang, J, Maddox, T, Gong, X, Tang, D et al. (1998). In vivo radioprotective effects of angiogenic growth factors on the small bowel of C3H mice. Radiat Res 150: 204–211. 13. Santana, P, Pena, LA, Haimovitz-Friedman, A, Martin, S, Green, D, McLoughlin, M et al. (1996). Acid sphingomyelinase-deficient human lymphoblasts and mice are defective in radiation-induced apoptosis. Cell 86: 189–199. 14. Pena, LA, Fuks, Z and Kolesnick, RN (2000). Radiation-induced apoptosis of endothelial cells in the murine central nervous system: protection by fibroblast growth factor and sphingomyelinase deficiency. Cancer Res 60: 321–327. 15. Vitolo, JM, Cotrim, AP, Sowers, AL, Russo, A, Wellner, RB, Pillemer, SR et al. (2004). The stable nitroxide tempol facilitates salivary gland protection during head and neck irradiation in a mouse model. Clin Cancer Res 10: 1807–1812. 16. Wang, S, Baum, BJ, Yamano, S, Mankani, MH, Sun, D, Jonsson, M et al. (2000). Adenoviral-mediated gene transfer to mouse salivary glands. J Dent Res 79: 701–708. 17. Voutetakis, A, Bossis, I, Kok, MR, Zhang, W, Wang, J, Cotrim, AP et al. (2005). Salivary glands as a potential gene transfer target for gene therapeutics of some monogenetic endocrine disorders. J Endocrinol 185: 363–372. 18. Turley, H, Scott, PA, Watts, VM, Bicknell, R, Harris, AL and Gatter, KC (1998). Expression of VEGF in routinely fixed material using a new monoclonal antibody VG1. J Pathol 186: 313–318. 19. Maharaj, AS, Saint-Geniez M, Maldonado AE and D’Amore PA (2006). Vascular endothelial growth factor localization in the adult. Am J Pathol 168: 639–648. 20. Kagami, H, Hiramatsu, Y, Hishida, S, Okazaki, Y, Horie, K, Oda, Y et al. (2000). Salivary growth factors in health and disease. Adv Dent Res 14: 99–102. 21. Schuller, BW, Binns, PJ, Riley, KJ, Ma, L, Hawthorne, MF and Coderre, JA (2006). Selective irradiation of the vascular endothelium has no effect on the survival of murine intestinal crypt stem cells. Proc Natl Acad Sci USA 103: 3787–3792. 22. Kolesnick, R and Fuks, Z (2003). Radiation and ceramide-induced apoptosis. Oncogene 22: 5897–5906. 23. Michael, JM, Lavin, MF and Watters, DJ (1997). Resistance to radiation-induced apoptosis in Burkitt’s lymphoma cells is associated with defective ceramide signaling. Cancer Res 57: 3600–3605. 24. Shi, YQ, Wuergler, FE, Blattmann, H and Crompton, NE (2001). Distinct apoptotic phenotypes induced by radiation and ceramide in both p53-wild-type and p53-mutated lymphoblastoid cells. Radiat Environ Biophys 40: 301–308.

2106

© The American Society of Gene Therapy

25. Radford, IR (1999). Initiation of ionizing radiation-induced apoptosis: DNA damage-mediated or does ceramide have a role? Int J Radiat Biol 75: 521–528. 26. Fei, P and El-Deiry, WS (2003). P53 and radiation responses. Oncogene 22: 5774–5783. 27. Preston, RJ (2005). Radiation biology: concepts for radiation protection. Health Phys 88: 545–556. 28. O’Connell, AC, Lillibridge, CD, Zheng, C, Baum, BJ, O’Connell, BC and Ambudkar, IS (1998). Gamma-irradiation-induced cell cycle arrest and cell death in a human submandibular gland cell line: effect of E2F1 expression. J Cell Physiol 177: 264–273. 29. O’Connell, AC, Redman, RS, Evans, RL and Ambudkar, IS (1999). Radiation-induced progressive decrease in fluid secretion in rat submandibular glands is related to decreased acinar volume and not impaired calcium signaling. Radiat Res 151: 150–158. 30. He, X, Goldsmith, CM, Marmary, Y, Wellner, RB, Parlow, AF, Nieman, LK et al. (1998). Systemic action of human growth hormone following adenovirus-mediated gene transfer to rat submandibular glands. Gene Ther 5: 537–541. 31. Chandler, LA, Sosnowski, BA, Greenlees, L, Aukerman, SL, Baird, A and Pierce, GF (1999). Prevalent expression of fibroblast growth factor (FGF) receptors and FGF2 in human tumor cell lines. Int J Cancer 81: 451–458. 32. Delporte, C, O’Connell, BC, He, X, Lancaster, HE, O’Connell, AC, Agre, P et al. (1997). Increased fluid secretion after adenoviral-mediated transfer of the aquaporin-1 cDNA to irradiated rat salivary glands. Proc Natl Acad Sci USA 94: 3268–3273. 33. Mittereder, N, March, KL and Trapnell, BC (1996). Evaluation of the concentration and bioactivity of adenovirus vectors for gene therapy. J Virol 70: 7498–7509. 34. Brown, AM, Rusnock, EJ, Sciubba, JJ and Baum, BJ (1989). Establishment and characterization of an epithelial cell line from the rat submandibular gland. J Oral Pathol Med 18: 206–213. 35. Zheng, C, O’Connell, BC and Baum, BJ (2003). Inclusion of Moloney murine leukemia virus elements upstream of the transgene cassette in an E1-deleted adenovirus leads to an unusual genomic integration in epithelial cells. Virology 313: 460–472. 36. Yamano, S, Huang, LY, Ding, C, Chiorini, JA, Goldsmith, CM, Wellner, RB et al. (2002). Recombinant adeno-associated virus serotype 2 vectors mediate stable interleukin 10 secretion from salivary glands into the bloodstream. Hum Gene Ther 13: 287–298. 37. Pusztaszeri, MP, Seelentag, W and Bosman, FT (2006). Immunohistochemical expression of endothelial markers CD31, CD34, von Willebrand factor, and Fli-1 in normal human tissues. J Histochem Cytochem 54: 385–395.

www.moleculartherapy.org vol. 15 no. 12 dec. 2007