| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Radiation Oncology,
Department of Biostatistics, and
Animal Care, Vanderbilt University School of Medicine/Vanderbilt-Ingram Cancer Center, Nashville, Tennessee, USA;
Department of Medicinal Chemistry and Molecular Pharmacology and
|| Department of Foods and Nutrition, Purdue University, West Lafayette, Indiana, USA;
¶ Department of Pharmaceutical Sciences, College of Pharmacy, University of Kentucky, Lexington, Kentucky, USA; and
# Department of Radiation Oncology and
** Division of Comparative Medicine, Washington University School of Medicine, St. Louis, Missouri, USA
2 Correspondence: B902 TVC Radiation Oncology, Vanderbilt University School of Medicine, Nashville, TN 37232. E-mail: K.R.S., raja.konjeti{at}vanderbilt.edu; M.L.F., michael.freeman{at}vanderbilt.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
70%. The ability to undergo neoangiogenesis, as measured in a window-chamber model, was also inhibited by 70%. Screening of biochemical pathways revealed that analog II inhibited the enzyme ENOX1 (EC50 = 10 µM). Retroviral-mediated shRNA suppression of endothelial ENOX1 expression inhibited cell migration and tubule formation, recapitulating the effects observed with the small-molecule analogs. Genetic or chemical suppression of ENOX1 significantly increased radiation-mediated Caspase3-activated apoptosis, coincident with suppression of p70S6K1 phosphorylation. Administration of analog II prior to fractionated X-irradiation significantly diminished the number and density of tumor microvessels, as well as delayed syngeneic and xenograft tumor growth compared to results obtained with radiation alone. Analysis of necropsies suggests that the analog was well tolerated. These results suggest that targeting ENOX1 activity represents a novel therapeutic strategy for enhancing the radiation response of tumors.—Geng, L., Rachakonda, G., Morré, D. J., Morré, D. M., Crooks, P. A., Sonar, V. N., Roti Roti, J. L., Rogers, B. E., Greco, S., Ye, F., Salleng, K. J., Sasi, S., Freeman, M. L., Sekhar, K. R. Indolyl-quinuclidinols inhibit ENOX activity and endothelial cell morphogenesis while enhancing radiation-mediated control of tumor vasculature.
Key Words: angiogenesis • (Z)-(±)-2-(1-benzylindol-3-ylmethylene)-1-azabicyclo[2.2.2]octan-3-ol
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. Development of the tumor microvasculature represents a very early event in tumor growth: it has been shown to occur in tumors as small as 60 to 80 cells (1)
. Inhibition of angiogenesis has emerged as a new therapeutic approach to control tumor progression (2
, 3)
. However, there is a need for new strategies that target tumor vasculature, specifically those that synergize with cytotoxic therapy, in order to overcome resistance that can develop with current therapeutics (4
, 5)
.
The nonsteroidal antiinflammatory compound indomethacin [1-(4-chlorobenzoyl)-5-methoxy-2-methyl-3-indoleacetic acid] has been reported to target tumor vasculature by inhibiting ERK2 activation (6)
. Indomethacin inhibits the proliferation of rat endothelial cells, human dermal microvascular cells, and human umbilical vein cells (HUVECs) (6
, 7)
. Indomethacin can inhibit formation of capillary-like structures formed in Matrigel (6)
, as well as neocapillary formation driven by tumor cell/Matrigel mixtures implanted subcutaneously in mice (8)
. In preclinical rodent models, oral administration of indomethacin has been shown to decrease tumor microvasculature density (9)
. Unfortunately, indomethacin concentrations on the order of 500 µM or more are required for efficacy. Chronic exposure to such high indomethacin concentrations can produce significant toxicity (10)
, thus limiting its clinical utility.
We used the indole moiety of indomethacin as an initial structural platform in a chemistry-driven drug discovery program adapted for identification of gene product function (11)
. Cell-based functional phenotype screening that employed synthetic chemical libraries was used to identify compounds that produced a desired activity. These screens were not designed to provide insight into mechanisms of action (12)
. Mechanistic insight was acquired from use of approaches such as annotated compound suppressor screening followed by genetic and biochemical verification (11
, 12)
.
Novel compounds consisting of combinations of indole, benzyl, and quinuclidine moieties were synthesized. Cell-based phenotypic screening revealed that noncytotoxic concentrations of (Z)-(±)-2-(1-benzenesulfonylindol-3-ylmethylene)-1-azabicyclo[2.2.2]octan-3-ol (analog I) and (Z)-2-(1-benzylindol-3-ylmethylene)-1-azabicyclo[2.2.2] octan-3-ol (analog II) inhibited HUVEC and 3B11 endothelial cell migration, and the ability to form capillary-like structures in Matrigel. The ability to undergo neoangiogenesis, as measured in a window-chamber model, was also inhibited by these compounds.
Screening of biochemical pathways identified the ECTO-NOX family of cell surface enzymes as potential targets for these indole analogs. ENOX plasma membrane-associated enzymes (ENOX1, ENOX2, and arNOX) are located on the external surface of the plasma membrane and exhibit a protein disulfide-thiol interchange activity (13
, 14)
. Cytosolic NADH serves as a source of reducing electrons for the disulfide-thiol interchange activity. A quinone reductase located at the cytosolic surface of the plasma membrane oxidizes NADH and transfers 2 electrons to a plasma membrane- embedded coenzyme Q10 (ubiquinone), which in turn is oxidized by an ENOX enzyme (14)
. The disulfide-thiol interchange activity is unique in that ENOX proteins do not exhibit a conserved C-X-X-C motif characteristic of thioredoxin or protein disulfide isomerase. ENOX1 (CNOX) was originally isolated from liver (15)
and is not inhibited by capsaicin, and we, as well as others (16)
, have found that it is also expressed in endothelial cells. ENOX2 (tNOX) is a cancer-specific enzyme whose activity is inhibited by capsaicin (14)
. arNOX is an age-specific enzyme (14)
. ENOX activity is required for appropriate activation of Rho and Rac (17)
.
Inhibitor studies conducted with purified ENOX1 revealed that analog II inhibited ECTO-NOX activity. Retroviral-mediated shRNA suppression of HUVEC ENOX1 expression inhibited cell migration and tubule formation, recapitulating the effects observed with the indole analogs.
Studies have shown that the addition of ionizing radiation to a strategy that targets tumor vasculature has the potential to synergistically increase tumor response (18
19
20)
. Therefore, we administered analog II 30 min prior to fractionated radiation therapy. This therapeutic regime significantly diminished the number and density of tumor microvessels, as well as delayed syngeneic and xenograft tumor growth compared to results obtained with radiation alone. Necropsies were performed, and the analysis suggests that the analogs were well tolerated. These results support a rationale for targeting ENOX activity as a novel molecular mechanism for enhancing the radiation response of tumors.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
22
23)
. The geometry of the double bond in these molecules was established from X-ray crystallographic data (21
, 23)
.
Two-dimensional migration assay
The assay was performed as described by Osusky et al. (24)
. Slides were either photographed directly or fixed with 100% cold methanol and 1% of methylene blue. Cell migration into the scraped area was quantitated using TotalLabr software (Nonlinear Dynamics, Newcastle, UK).
Formation of tubule-like structures in Matrigel
The assay was conducted into 24-well plates coated with Matrigel as described by Yazlovitskaya et al. (25)
. HUVECs were incubated in the absence or presence of analog for 7 h at 37°C. Cells were either photographed directly or fixed with cold 100% methanol and then stained with Mayer’s hematoxylin solution. Assays were performed in triplicate. Image J software (NIH, Bethesda, MD, USA) was used to quantify the tubule formation. Tubule formation was quantitated as described by Yazlovitskaya et al. (25)
. In brief, cells were stained with 1% methylene blue and photographed (4 times), and the average number of tubules was quantified from 3 individual fields/dish, 3 dishes/point.
Quantitation of tetrazolium salt reduction
Quantitation of WST1 or MTT was performed according to the manufacturer’s instructions (Cayman Chemical Co., Ann Arbor, MI, USA).
Measurement of ECTO-NOX activity
This was accomplished as described in Yagiz et al. (26)
. In brief, ENOX1 purified from plasma membranes was added to 150 µM of NADH in 25 mM Tris-MES buffer (pH 7.5) supplemented with 1 mM KCN, and the indicated concentrations of analog. Samples were incubated at 37°C. Absorbance at 340 nm was monitored.
Dorsal skinfold vascular window chamber
A dorsal skinfold vascular window chamber was placed onto C57BL/6J mice, as described by Edwards et al. (27)
. To quantify vascular length, vascular windows were photographed using an x4 objective to obtain a 40-fold magnification. The photographs were scanned into Adobe Photoshop (Adobe Systems, San Jose, CA, USA) and converted into tif files. The images were analyzed using Image Pro Plus software (Media Cybernetics, Bethesda, MD, USA). The Image Pro Plus software measuring tool was used to quantitate vessel length (27)
.
RNA interference
Six retroviral plasmids expressing ENOX1 shRNA (Open Biosystems, Huntsville, AL, USA) were pooled and transfected into LinX cells using Arrestin transfection reagent as per the manufacturer’s protocol. Similarly, retrovirus expressing control GFP shRNA (Open Biosystems) was transfected into LinX cells. Twenty-four hours after transfection, medium was collected and filtered through a 40-µm filter, and the filtrate was used as virus stock. Passage 3 HUVEC cells were transduced with virus containing medium along with 2 µM of Polybren. Cells were reinfected 24 h and 48 h after the primary infection. shRNA-mediated suppression of ENOX1 was verified by immunoblotting.
Immunohistochemistry
Immunohistochemical analysis was accomplished using 1) goat CD34 (Santa Cruz Biotechnology, Santa Cruz, CA, USA) followed by donkey anti-goat Alexa Fluor 568 (Invitrogen, Carlsbad, CA, USA) and counterstained with DAPI mounting medium or 2) anti-ACTIVE caspase-3 antibody (Promega, Madison, WI, USA) followed by horseradish peroxidase (HRP)-conjugated secondary antibody (DakoCytomation, Copenhagen, Denmark)/3,3-diaminobenzidine (DAB). Slides were counterstained with Mayer’s hematoxylin. Other slides were stained with hematoxylin and eosin (H&E).
CD34 staining was analyzed using the program Image J and the methodology described by Wild et al. (28)
. The analysis was performed using 4 tumors/treatment and 4 representative sections/tumor. Images were converted to grayscale, inverted, and then converted to binary scale, allowing quantitation of the number of CD34-positive vessels, as well as vessel size.
The extent of activated caspase 3 staining or the degree of necrosis was determined from low-power microscopic magnification, 10 random fields/slide, 3 slides/point. Slides were coded. Necrosis was quantitated as follows: fields with <20% necrosis in the optical field were assigned a value of 1; if necrosis was observed in 20 to 50% of the optical fields, that slide was assigned a value of 2. If necrosis was observed in >50% of the optical fields, the slide was assigned a value of 3. The extent of caspase 3 activation was determined in a manner similar to quantitation of necrosis.
Tumor growth inhibition
These studies were performed under the Guidelines for the Care and Use of Research Animals, Vanderbilt University Animal Studies Committee or Washington University Animal Studies Committee. Hindlimbs of C57BL/6J mice, 
6–8 wk of age, were subcutaneously implanted with Lewis lung carcinoma (LLC) cells. Hindlimbs of homozygous nu/nu athymic nude mice, 6–8 wk of age, were subcutaneously implanted with HT-29 human colorectal cancer cells. LLC tumors were grown until they reached a volume of 250 mm3; HT-29 tumors were allowed to reach a volume of 125 mm3. For LLC tumors, mice were divided into groups of 5 animals each. For HT-29 tumors, 10 animals/group were used. Mice received daily i.p. injections of DMSO (25 µl) or 40 mg/kg of analog II in DMSO (25 µl) for 5 or 8 consecutive days, followed 30 min later by 0 or 3 Gy of X-rays (300 kVp/10 mA). During irradiation, mice were shielded such that only the tumors were irradiated. Digital calipers were used to obtain the length and width of each tumor. Tumor volumes were calculated 3x/wk according to the formula: length x (width)2/2. The starting volume of each mouse was normalized to 1.0.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The structures of 2 efficacious nontoxic analogs (analogs I and II) are shown in Fig. 1A
. The structure of analog III, an inactive compound, is also provided. HUVECs, murine 3B11 tumor vascular endothelial cells (29)
, and human microvascular endothelial cells were exposed to analog I or II (25 or 50 µM) for 7 h, and the ability of endothelial cells to form capillary-like structures was then quantiated as described by Yazlovitskaya et al. (25)
. A representative experiment (Fig. 1B
) illustrates the inability of HUVECs to form capillary-like structures in the presence of either analog I or analog II. Data provided in Supplemental Fig. 1 illustrate the inability of murine vascular 3B11 endothelial or human microvascular endothelial cells to form capillary-like structures in the presence of analog I or II. Exposure to either analog resulted in a statistically significant decrease in tubule formation. For example, exposure of HUVECs to 50 µM of analog I or analog II inhibited tubule formation by 80% compared to DMSO control (P=0.001; Student’s t test).
|
Utilization of a 2-dimensional in vitro migration assay demonstrated that exposure to analog I inhibited cell migration by a factor of 5 (P=003; Student’s t test; Supplemental Fig. 2), whereas exposure to analog II inhibited migration 8-fold compared to DMSO (P=0.005, Student’s t test). In contrast, analog III, the inactive compound, did not inhibit migration.
3H-TdR incorporation was used to measure HUVEC DNA synthesis. A concentration- and time-dependent decrease in tracer uptake relative to solvent control was observed (Supplemental Fig. 3A; P=0.025; ANOVA). We also tested the ability of analog I to inhibit uptake of 3H-TdR into HT-29 human colon carcinoma cells. Exposure to 25 µM of analog I for 24 h did not affect uptake (P>0.05; Student’s t test; data not shown).
Inhibition of ECTO-NOX activity
It has been reported that the NADH oxidoreductase activity of ENOX proteins is responsible for reduction of cell-impermeable WST1 (14
, 30)
. Reduction of WST1 was inhibited in a concentration-dependent manner on exposure to analog I (P<0.05; ANOVA; Fig. 2A
). Because we did not observe analog-mediated inhibition of the cell-permeable MTT, we hypothesized that these novel analogs target ENOX activity. To test this hypothesis, we first determined whether endothelial cells express ENOX proteins. The immunoblot shown in Fig. 2B
illustrates ENOX1 expression in HUVECs. HUVECs do not express capsaicin-inhibitable ENOX2 activity (data not shown). ENOX activity was measured using a well-characterized assay (14)
. ENOX1 was partially purified from HeLa cells (14)
and yielded an activity of 2.2 ± 0.13 nmol/min/mg enzyme (data not shown). Exposure to either analog I or analog II inhibited enzyme activity; EC50 = 50 µM (n=5; Supplemental Fig. 4). The ability of analogs I and II to inhibit ENOX2 isolated from HeLa cells was also tested. The EC50 for inhibition of ENOX2 by analog I was 50 µM, while the EC50 for inhibition by analog II was 10 nM (data not shown). The data presented in Fig. 2C
illustrate inhibition of ENOX1, partially purified from HUVEC cells. The EC50 for inhibition of HUVEC ENOX1 by analog II was 10 µM.
|
ENOX activity is involved in regulation of cell enlargement (31)
. The kinase p70S6 K1 has been shown to integrate cell enlargement with proliferation (32)
. Loss of p70S6K1 activity results in suppression of DNA synthesis and proliferation (33)
. In HUVECs, loss of p70S6K1 activity results in inhibition of tubule formation (34)
. Phosphorylation of p70S6K1 at T389 occurs in a PI3K-dependent manner and disrupts an inhibitory interaction between the amino and carboxyl termini (33)
. The immunoblot shown in Fig. 2D
indicates that phosphorylation of p70S6K1 at T389 was inhibited in HUVECs on exposure to analog II for 2 h, as was phosphorylation of the downstream target ribosomal protein S6. These results are consistent with the work of Shinohara et al. (34)
.
We used a retrovirus-based shRNA approach to determine whether loss of ENOX1 affected the ability of HUVECs to form capillary-like structures or to migrate. The immunoblot shown in Fig. 3A
demonstrates that a pool composed of 6 retroviruses expressing unique ENOX1 shRNAs suppressed expression of ENOX1 by 73% relative to expression of GAPDH, as determined by quantitative imaging software. The ability of a single ENOX1 shRNA vector to suppress ENOX1 is also shown. This vector (ENOX1 shRNA1) suppressed ENOX expression by 55%. Because we do not passage the HUVECs more than 4 times, we did not select individual cell clones expressing retroviral-mediated shRNA. Rather, the immunoblot represents a polyclonal population of cells with respect to ENOX1 suppression. Suppression of ENOX1 by shRNA inhibited formation of capillary-like structures in Matrigel by 64% (P=0.003; Student’s t test). With the caveat that we are interrogating a heterogeneous population, the degree of inhibition that we observed supports the hypothesis that ENOX1 activity is required for the formation of capillary-like structures.
|
A 2-dimensional migration assay was used to assess the role of ENOX1 in migration (compare Fig. 3D, E
). Twenty-four hours after scraping, cells infected with nonsilencing shRNA control vector have essentially completed migration: the leading edges have essentially coalesced (Fig. 3D
). In contrast, migration of the retroviral shRNA population was significantly inhibited (3-fold, P<0.05; Student’s t test; Fig. 3E
). Thus, genetic manipulation of ENOX1 recapitulated two of the phenotypic effects observed following chemical inhibition of the enzyme.
Pharmacokinetics, bioavailability, and necropsy
The in vivo half-life of analogs I and II was determined in plasma following either intravenous administration or oral gavage of rats (unpublished results and ref. 35
). Analogs I and II exhibit a short half-life in plasma of 30 min. With a short plasma half-life, it was important to ensure that the analog rapidly accumulated in tumor. We found that 15 min after i.p. injection of 120 mg/kg, HT-29 tumors contained 4.4 ± 0.6 mg analog II/g tissue. As both analogs exhibited approximately the same ability to inhibit endothelial cell morphogenesis, the same plasma half-life and the same ability to inhibit ENOX activity, we arbitrarily chose analog II for the remainder of the studies. Female nu/nu mice were injected i.p. with either 100 (n=9) or 120 (n=5) mg/kg of analog II. Fifteen nu/nu mice were injected i.p. with solvent control (30 µl DMSO). Thirty days after injection, mice were subjected to necropsy. Gross and histological examination did not reveal evidence of toxicity that could be attributed to injection of analog II. Both a complete blood count and a white blood cell differential count were performed. No significant differences were noted. This analysis indicates that analog II did not produce toxicity up to a dose of 120 mg/kg. For further details of these studies, see Supplemental Data.
Inhibition of tumor vasculature and tumor growth
We used a dorsal skinfold vascular window model to directly visualize tumor neoangiogenesis induced by injection of LLC cells (Fig. 4
). Mice were given daily i.p. injections of vehicle control (DMSO) or analog II (40 mg/kg for 5 d) beginning 72 h after tumor cells were injected. Neovascularization was monitored daily. In control animals, blood extravasation was observed soon after implanting LLC cells. This was followed by vascular remodeling. Vessel length was quantitated at 240 and 312 h (Supplemental Fig. 5). Analog II inhibited vessel growth by 70% compared to the time-matched control (P=0.007, Student’s t test).
|
We used immunofluorescent staining of CD34-positive capillary vessels present in LLC tumors growing in the hindlimbs of C57BL/6J mice to estimate tumor microvessel density following irradiation (36
, 37)
. Tumor-bearing mice were injected with either solvent control or analog II (i.p.; 40 mg/kg). Thirty minutes later, tumors were irradiated (0 or 3 Gy). This protocol was repeated daily for 4 additional days (q.d.x5). Twenty-four hours after the last irradiation, tumors were harvested, fixed, and sectioned. CD34 and DAPI immunofluorescent staining in representative tumor sections is shown in Fig. 5
. Daily injections of analog II alone did not produce a statistically significant change in either the number of CD34-positive vessels (P=0.294; Student’s t test) or their size (P=0.771; Student’s t test). Compared to control, 5 daily 3-Gy fractions produced a statistically significant decrease in vessel size (P<0.001; Student’s t test) but did not decrease the number of vessels (P>0.05; Student’s t test). However, a combination of analog II plus irradiation produced a statistically significant decrease in vessel size (P<0.001; Student’s t test) and a decrease in vessel number (P<0.001; Student’s t test) compared to control. Vessel size was significantly smaller in tumors exposed to analog II plus radiation compared to radiation alone (P=0.002; Student’s t test). The analysis indicates that injection of analog II significantly affected the response of tumor vessels to irradiation.
|
Tumor sections were also subjected to immunocytochemistry staining with antibody to activated caspase 3. We quantitated the percentage of caspase staining per random optical field inspected under low-power microscopic magnification. Staining for activated caspase 3 was independent of treatment with analog II, irradiation, or analog plus irradiation (P>0.05; ANOVA; data not shown). Tumor sections were also subjected to H&E staining in order to assess necrosis. Although the degree of necrosis appeared to increase in sections representing tumors treated with analog II and/or irradiation, the change was not statistically significant compared to control (P>0.05; ANOVA; data not shown).
We assessed the ability of fractionated IR to provide local control of LLC cells implanted in the hindlimbs of mice. LLC tumor-bearing mice were administered 8 daily fractions of the following treatments: 1) i.p. injection of solvent control DMSO alone; 2) i.p. injection of 40 mg/kg of analog II (q.d.x8); 3) i.p. injection of solvent followed 30 min later with 3 Gy (q.d.x8); 4) i.p. injection of 40 mg/kg of analog II followed 30 min later by 3 Gy (q.d.x8). The fold increase in tumor volume (±SE; n=5) is shown in Fig. 6A
. The data are expressed using a linear mixed-effects model fit by REML. This analysis indicated that tumors treated with solvent control increased 11-fold in size by d 13. In contrast, tumors treated with 8 daily injections of 40 mg/kg of analog II (q.d.x8) or treated with 8 daily 3-Gy fractions required 16 d to increase 11-fold in size. Thus, 8 daily injections of analog II by itself produced the same statistically significant tumor growth delay as 8 daily 3-Gy fractions. The observation that 5 daily injections of analog II did not affect tumor-associated vasculature (Fig. 5
) but 8 daily injections of analog II inhibited tumor growth was interpreted to indicate that 5 daily fractions represented a subtherapeutic schedule.
|
Tumors that underwent a combined treatment of 8 daily i.p. injections of analog II followed 30 min later with 3 Gy required 25 d to increase 11-fold in size. Whereas radiation alone produced a 3-d tumor growth delay, the addition of analog II to the irradiation protocol increased tumor growth delay by 11 d (Fig. 6A
).
HT-29 human colon carcinoma cells were grown as xenograft tumors in the hindlimbs of nu/nu mice. Tumor-bearing mice were administered 5 daily fractions of the following treatments: 1) i.p. injection of solvent control DMSO alone; 2) i.p. injection of 40 mg/kg of analog II (q.d.x5); 3) i.p. injection of solvent followed 30 min later with 3 Gy (q.d.x5); 4) i.p. injection of 40 mg/kg of analog II followed 30 min later by 3 Gy (q.d.x5). The fold increase in tumor volume is shown in Fig. 6B
. Tumors treated with solvent control or with analog II alone increased 7-fold in size by d 50. We hypothesize that 5 daily treatments with analog II were insufficient to affect tumor growth. Tumors treated with 5 daily 3-Gy fractions required 75 d to increase their size by 7-fold. However, tumors treated with 5 daily i.p. injections of analog II followed 30 min later with 3 Gy stopped growing and remained dormant. Thus, injection of 40 mg/kg of analog II 30 min prior to irradiation (q.d.x5) provided excellent local control compared to irradiation alone.
Radiation sensitivity
We assessed the radiation sensitivity of proliferating and nonproliferating HUVECs by measuring apoptosis and colony formation (Fig. 7
). Analog II increased radiation-mediated apoptosis only in proliferating cells (Fig. 7B
). Subconfluent proliferating (6.7x104/cm2) or confluent nonproliferating (1.7x105/cm2) HUVECs were irradiated (3 Gy) in the presence of DMSO control or analog II (50 µM). Cells were then assayed for caspase 3 activity using a fluorogenic DNA-binding dye and counterstained with DAPI. Numbers of caspase 3-positive and -negative cells were quantitated by microscopy (Fig. 7A
, B). Degree of proliferation was measured by Ki67 and BrdU staining (data not shown). Confluency inhibited proliferation by 70%. In proliferating cells, analog II increased radiation-mediated apoptosis 10-fold (P<0.001; Fisher’s exact test; Fig. 7B
). Irradiation did not alter the ability of analog to inhibit phosphorylation of p70S6K1 at T389 (data not shown). In nonproliferating cells, radiation-mediated apoptosis was <5% and was independent of analog exposure (P>0.05; Fisher’s exact test).
|
We assessed radiation-mediated apoptosis in HUVECs expressing control shRNA or ENOX1 shRNA. shRNA-mediated loss of ENOX1 produced a statistically significant increase in radiation-induced apoptosis (Fig. 7C
). Analog II (50 µM) inhibited ENOX1 activity by >90%, whereas ENOX1 activity was decreased 65% in ENOX1 shRNA-expressing cells (data not shown); these decreases in ENOX1 activity are reflected in the increased radiation sensitivity that was measured by colony formation assays (Fig. 7D
).
Neither analog I or II affected the radiation sensitivity of human HT-29 colon carcinoma cells (P>0.05; Student’s t test; Supplemental Fig. 6). These data support the hypothesis that exposure to analog II increased the radiation sensitivity of tumor-associated vasculature. Further, the data are not consistent with an effect on tumor epithelial cells.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, but very high concentrations are required. Small-molecule libraries, based on an indomethacin structure, were synthesized and screened for their ability to inhibit endothelial cell tubule formation. We identified several nontoxic active compounds and subjected them to further characterization with regard to migration and neovascularization. With respect to either vehicle control or an inactive compound (analog III), analogs were found to inhibit DNA synthesis, migration, and neovascularization.
The ability of endothelial cells to proliferate, migrate, and form capillary-like structures is regulated, in part, by vascular endothelial growth factor (VEGF) (38)
. The analogs illustrated in Fig. 1
did not inhibit VEGF-mediated phosphorylation of ERK1/2.
We found that the analogs inhibited purified ECTO-NOX activity. Neither analog I or II inhibited purified arNOX (data not shown) but inhibited HUVEC ENOX1 activity. Endothelial cells do not express capsaicin-inhibited ENOX2. The ability of the analogs to inhibit tubule formation, proliferation (unpublished results), and DNA synthesis is consistent with the observation that analog II inhibited phosphorylation of p70S6K1. This kinase is integral for regulation of cell size, growth, and proliferation in a cell-autonomous manner (33)
. Further, the kinase is required for formation of tubules by endothelial cells (34)
. The data suggest that ENOX1 activity is an upstream regulator of p70S6K1, but formal proof of this hypothesis is beyond the scope of this investigation.
Whereas analog-mediated inhibition of DNA synthesis was observed in HUVECs, it was not observed in HT29 colon carcinoma cells. Although this is not well understood, we hypothesize that this result is a consequence of the fact that tumor cells such as HT29 are relatively insensitive to mTOR-p70S6K1 inhibition. Buck et al. (39)
found that HT29 cell proliferation was inhibited by only 40% following a 72-h rapamycin treatment. Given that we treated HT29 cells for only 24 h with analog, we may not expect to observe significant inhibition of proliferation.
Studies have shown that pharmaceutical targeting of tumor microvasculature has the potential to synergistically increase tumor response to cytotoxic therapy. Because of the short plasma half-life exhibited by our analogs, we hypothesized that X-irradiation, where energy is rapidly absorbed (1 µs) by an entire tumor volume, would be an appropriate regime. Therefore, we assessed the ability of fractionated doses of ionizing radiation to provide local control of well-vascularized LLC tumors growing in the hindlimbs of mice. Modest control of tumor growth was achieved by radiation exposure alone (q.d.x8) or by injection of analog II by itself (q.d.x8). With regard to tumor vasculature, fractionated irradiation provided a modest decrease. However, the combination of analog II with fractionated radiation produced a significant decrease in tumor vasculature and tumor growth delay. A similar observation was made using the human HT-29 colon carcinoma model: a combination of fractionated radiation and analog II (q.d.x5) also produced significant control of tumors compared to a subtherapeutic radiation prescription. These observations paralleled the experiments that demonstrated that analog II increased radiation-mediated caspase-3-dependent apoptosis in proliferating endothelial cells (Fig. 7
). Similarly, suppression of ENOX1 via shRNA resulted in radiation-mediated caspase-3-dependent apoptosis in endothelial cells. The observation that nonproliferating cells were not radiosensitized is consistent with the knowledge that ENOX1 is dispensable in nonproliferating cells.
Immunohistochemical staining for activation of caspase 3 and H&E staining to detect areas of necrosis did not provide statistically significant support for the hypothesis that analog II affected the radiation sensitivity of LLC tumor cells. We found that exposure to analog II did not affect radiation sensitivity of HT-29 cells in cell culture models. Solid-tumor epithelium does not utilize apoptotic pathways when lethally irradiated (40
, 41)
, whereas vascular endothelial cells do (27)
. At a biochemical level, it has been shown that inhibition of p70S6K1 activity induces apoptosis in irradiated endothelial cells (34)
, a result of an inability to phosphorylate the proapoptotic protein BAD (42)
. In contrast, tumor cell survival following irradiation is independent of p70S6K1 activity (34)
. Thus, the differences in the biochemical pathways used to express cell death can explain the lack of radiosensitization in HT29 cells. Taken together, these data support the hypothesis that analog II increases tumor vasculature radiation sensitivity. In summary, these novel results are the first we are aware of that support a radiation-sensitization strategy involving targeting of ENOX1.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication January 20, 2009. Accepted for publication March 26, 2009.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
