Therapeutic Significance of Elevated Tissue Transglutaminase Expression in Pancreatic Cancer

Materials and Methods

Materials. Rabbit polyclonal antibodies against phosphorylated Akt (pAkt; Ser⁴⁷³) were obtained from Cell Signaling Technology. The anti-TG2 monoclonal antibody CUB7401 was purchased from NeoMarkers. A monoclonal antibody against vascular endothelial growth factor (VEGF) was obtained from Santa Cruz Biotechnology. An anti–β-actin antibody was purchased from Sigma Chemical Co., and horseradish peroxidase–conjugated goat anti-rabbit and sheep anti-mouse antibodies were purchased from Amersham Biosciences. DMEM/F12, fetal bovine serum, and the antibiotic (Normocin) were purchased from Invivogen. The liquid DAB+ Substrate Chromogen System-horseradish peroxidase used for immunohistochemistry was obtained from DakoCytomation. Gemzar, which was supplied by Eli Lilly and Company, was stored at 4°C and dissolved in sterile PBS on the day of use. d-Luciferin potassium salt (Xenogen Corp.) was dissolved in sterile PBS at a concentration of 40 mg/mL.

Cell lines. The Panc-28 cell line was kindly provided by Dr. Shrikanth Reddy (The University of Texas M. D. Anderson Cancer Center, Houston, TX). The cells were cultured in DMEM/F12 supplemented with FCS (10%, v/v), Normocin (0.1 mg/mL), l-glutamine (2 mmol/L), and HEPES (10 mmol/L; U.S. Biochemical) at 37°C in a CO₂ incubator.

TG2 down-regulation by siRNA and liposomal siRNA preparation. TG2-targeted siRNA sequence (5′-AAGGGCGAACCACCTGAACAA-3′) was synthesized and then purified as described previously (22). A sequence (5′-AATTCTCCGAACGTGTCACGT-3′) that did not have homology to any human mRNA as determined by blast search was used as a control. For in vivo delivery, siRNA was incorporated into neutral liposomes composed of 1,2-dioleoyl-sn-glycero-3-phosphatidylcholine (DOPC) as described previously (23). Briefly, the siRNA and DOPC were mixed in the presence of excess t-butanol at a ratio of 1:10 (w/w). Tween 20 was added to the mixture, which was then vortexed, frozen in an acetone/dry ice bath, and lyophilized. Before administration into mice, the preliposomal formulation powder (lyophilizate) containing the lipid and siRNA was hydrated with normal saline at a concentration of 15 μg/mL. Once reconstituted, the DOPC-siRNA preparation was used within 12 h.

PDAC orthotopic tumor model. Male athymic nu/nu mice (4 wk old) were obtained from the Department of Experimental Radiation Oncology at M. D. Anderson Cancer Center. The animals were housed four per cage in standard acrylic glass cages in a room maintained at constant temperature and humidity with a 12-h light and darkness cycle and fed a regular autoclaved chow diet with water ad libitum. Before initiating the experiment, animals were acclimatized to a pulverized diet for 3 d. All studies were conducted according to an experimental protocol reviewed and approved by the M. D. Anderson Institutional Animal Care and Use Committee.

At 70% confluence, Panc-28 cells were stably transduced with the firefly luciferase gene as described previously (24). Cells were harvested from subconfluent cultures after brief exposure to 0.25% trypsin and 0.2% EDTA. Trypsinization was stopped with the medium containing 10% fetal bovine serum. The cells were washed once in serum-free medium and resuspended in PBS. Single-cell suspensions with >90% cell viability were used for implantation into the mice. Animals were anesthetized using i.p. injection of 200 μL ketamine (100 mg/kg) and xylazine (5-10 mg/kg). For orthotopic implantation of tumor cells into the mice, a small incision (2 cm long) was made on the left flank, the pancreas was gently pulled out using blunt forceps, and a 100 μL suspension of luciferase-infected Panc-28 cells (1 × 10⁶) was injected directly into the pancreas using a 27-gauge needle (1-mL disposable tuberculin syringe). The incision was closed in a single layer using sterile surgical Autoclips (Braintree Scientific), antibacterial mycotic cream was applied to the wound, and the animal was placed on a heating pad to maintain body temperature and closely monitored until conscious.

Luminescence imaging. Tumor growth in the mice after tumor cell implantation was monitored weekly using a noninvasive bioimaging technique. Before imaging, animals were anesthetized using 1.5% isoflurane/air inhalation and injected i.p. with d-luciferin potassium salt in PBS (150 mg/kg). Ten minutes after injection, each mouse was placed in the right lateral decubitus position, and bioluminescence was measured using a cryogenically cooled IVIS imaging system coupled with a data acquisition computer running Living Image software (Xenogen). A digital grayscale image of each animal was acquired followed by acquisition and overlay of a pseudocolor image of the spatial distribution of photons emerging from active luciferase within the tumor. The signal intensity was quantified as the sum of all detected photons within the region of interest per second per cm² per steridian.

Before evaluation of the therapeutic efficacy of TG2 siRNA, the ability of liposomal-encapsulated siRNA (Alexa Fluor 546 labeled) to down-regulate TG2 expression in Panc-28 tumors was examined by giving the liposomal formulation in the tail vein or i.p. in four mice each (150 μg/kg/dose/route, three doses total, each 3 d apart). Based on the extent of down-regulation of TG2 expression in the two groups, the tail vein route was apparently superior to the i.p. route for the delivery of TG2 siRNA (Fig. 1 ).

Therapeutic efficacy of liposomal TG2 siRNA. Panc-28 cells (1 × 10⁶) were implanted orthotopically in the pancreases of nude (nu/nu) mice. The resulting tumor growth was monitored weekly for 3 wk using the bioluminescence IVIS Imaging System 200 (Xenogen). After the 3-wk monitoring period, mice were randomly distributed into four treatment groups (five mice per group): group I received control liposomal siRNA (control siRNA; 150 μg/kg) thrice weekly by i.v. injection, group II received gemcitabine alone (25 mg/kg) twice weekly by i.p. injection, group III received liposomal TG2 siRNA alone (150 μg/kg) thrice weekly by i.v. injection, and group IV received both gemcitabine (25 mg/kg) twice weekly by i.p. injection and liposomal TG2 siRNA (150 μg/kg) thrice weekly by i.v. injection.

Mice were imaged for tumor growth on days 0, 2, 9, 16, and 23 during the treatment. After delivery of a total of 10 doses and 3 d after delivery of the last dose, all of the animals were killed. Their primary pancreatic tumors were excised, and the final tumor volumes were measured using a digital clipper and calculated using the formula V = 2/3πr³, in which r is the mean radius in three different dimensions (length, width, and depth). One third of the tumor tissue was formalin fixed and paraffin embedded for immunohistochemistry and routine H&E staining. The other one third of the tumor tissue was fixed in OCT for fluorescence and terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) assay, and the remaining tumor tissue was snap frozen and stored at −80°C until use.

Western blot analysis. Pancreatic tumor tissue samples obtained from control and experimental mice were minced and incubated on ice for 30 min in 0.5 mL of ice-cold lysate buffer [20 mmol/L Tris-HCl (pH 7.4), containing 10% NP40, 5 mol/L NaCl, 1 mol/L HEPES, 0.5 mol/L EDTA, 0.1 mol/L phenylmethylsulfonyl fluoride, 0.2 mol/L sodium orthovanadate, 1 mol/L NaF, 2 μg/mL aprotinin, and 2 μg/mL leupeptin]. The minced tissue was homogenized using a Dounce homogenizer and centrifuged at 16,000 × g at 4°C for 10 min. Fifty micrograms of total protein from each sample were then fractionated using SDS-polyacrylamide gel, electrotransferred onto nitrocellulose membranes, and blotted with appropriate antibodies, and antigen-antibody reaction was detected using enhanced chemiluminescence (Amersham Pharmacia Biotech). Some of the membranes were stripped using Restore stripping buffer (Pierce) and reprobed with another antibody. The protein bands obtained were quantified using the AlphaEase FC (FluorChem 8900) software program (Alpha Innotech).

Immunohistochemistry. Tissue microarray slides constructed using 65 formalin-fixed patient tumor samples and 9 normal pancreatic tissue samples were used for this study. Samples were obtained from patients with primary PDAC who underwent initial pancreaticoduodenectomy without prior chemotherapy or radiation therapy. TG2 protein expression in these tumor samples and xenograft tumors from siRNA-treated mice was evaluated immunohistochemically using the Vectastain Elite avidin-biotin complex method immunostaining kit (Vector Laboratories) as described previously (20). In brief, antigen retrieval was done by treating tissue samples in a steamer for 30 min. Mouse anti-TG2, rabbit anti-pAkt (Ser⁴⁷³; Cell Signaling Technology), rabbit anti–Ki-67 (clone SP6; NeoMarkers), or mouse anti-CD31 (PharMingen) antibodies overlaying the tissue sections were incubated at 4°C for 16 h. Secondary antibody incubation was done at ambient temperature for 1 h. Mayer's hematoxylin nuclear stain was used as a counterstain. Immunostaining results were evaluated and scored independently by a pathologist and the laboratory personnel in a double-blinded manner. TG2 expression in tumor cells was categorized as negative (weak or negative cytoplasmic staining) or positive (diffuse moderate to strong cytoplasmic staining). pAkt (Ser⁴⁷³) expression in tumor cells was categorized as negative (no or very weak cytoplasmic staining) or positive (diffuse moderate or strong cytoplasmic staining).

Proliferation. Ki-67 staining was used to evaluate the proliferation index of in vivo growing xenografts and expressed as the number of Ki-67⁺ cells ± SE as determined by counting 10 random microscopic fields for three different tumor samples per treatment group.

Microvessel density. A rat anti-mouse CD31 monoclonal antibody was used to determine the microvessel density in xenografts, as previously described (24). The CD31-stained slides were observed under a Leica DM4000B fluorescence microscope (Leica Microsystems, Inc.) equipped with SPOT-RTKE digital camera (Diagnostic Instruments), and the images were acquired and stored using SPOT advanced software (Diagnostic Instruments). The stored images were processed using NIH Image software. The vessel density in each image was estimated by setting a minimum threshold of 4 pixel². Setting a minimum threshold enables to avoid the background nonspecific staining and to eliminate the fluorescence signal from single endothelial cell. Results were expressed as the mean number of vessels ± SE per high-power field (×100 magnification). A total of 20 random high-power fields were examined and counted in four different tumor samples from each treatment group.

TUNEL assay. TUNEL assay (Roche Molecular Biochemicals) of frozen sections to detect apoptotic cells was done according to the manufacturer's protocol. The apoptotic fluorescence-positive cells were counted under a microscope and expressed as the percentage of total cells ± SE. Ten fields in three tumors per treatment group were examined.

Electrophoretic mobility shift assay. To determine the level of NF-κB activation in xenografts, electrophoretic mobility shift assay was used as described previously (21). Briefly, nuclear extracts from tumor homogenates were incubated with a ³²P-end–labeled 45-mer double-stranded NF-κB oligonucleotide (15 μg protein with 16 fmol DNA) for 30 min at 37°C, and the DNA-protein complex that formed was separated from free oligonucleotide on 6.6% native polyacrylamide gels. The dried gels were autoradiographed, and radioactive bands were quantitated using a STORM 220 PhosphorImager (Amersham Biosciences) with the ImageQuant software program (Molecular Dynamics).

Statistical analysis. We used SD of 100 and mean difference between control and modified/treated groups as 200 for tumor volume to do the power calculation for in vivo studies based on the two-sided, two-sample Student's t test (nQuery 5.0). Using P < 0.05, we observed that n = 5 mice per group is adequate to achieve a power (1-β) of 80%. For tumor volume calculation, values were initially subjected to one-way ANOVA and then compared among groups using unpaired Student's t test. For TG2 expression in patient samples, values were calculated using Fisher's exact test with P < 0.001 set as statistically significant. For Ki-67 staining, microvessel density, and in situ TUNEL assay, values were initially subjected to one-way ANOVA and then compared among groups using unpaired Student's t test.

Results

TG2 expression in PDAC tumors. We previously reported that TG2 protein is aberrantly expressed in most PDAC tumors and cell lines (20). In the present study, we examined additional 65 PDAC tumor samples and 9 normal pancreatic tissue samples for TG2 expression and confirmed that TG2 expression was significantly higher in PDAC tumor samples (78%) than in normal ducts (11%; P < 0.0001). Although the sample size was too small to comment on the correlation between TG2 expression and histologic grades (1-3), node status (N stage), or metastasis (M stage), we observed a trend toward an increasing TG2 expression as the tumor stage increased [T1, 3 of 5 (60%); T2, 32 of 44 (73%); T3, 13 of 16 (81%)].

We also previously reported that TG2 expression confers resistance to gemcitabine and promotes invasiveness in PDAC cells (20). Therefore, in the present study, we chose to evaluate the potential of TG2 as a therapeutic target for chemoresistant and metastatic PDAC. For this purpose, we used highly aggressive, gemcitabine-resistant, and high TG2-expressing Panc-28 cells. We stably infected cells with luciferase gene and injected them into the pancreases of nude mice to establish orthotopic pancreatic tumors. Resulting Panc-28 tumors in the mice were highly aggressive and advanced rapidly to locally disseminated disease over ∼6 weeks after tumor cell implantation.

In vivo silencing by TG2-specific siRNA-DOPC liposomes. Based on the evidence that elevated expression of TG2 in PDAC is closely related to aggressive tumor behavior, we investigated its potential as a therapeutic target. In a preliminary experiment, we first determined the optimal route for delivery of DOPC-encapsulated TG2 siRNA to down-regulate TG2 expression in tumors. Immunohistochemical and Western blot analysis consistently revealed >80% reduction in TG2 expression in the tumors obtained from mice that received TG2 siRNA i.v. (Fig. 1A). The reduction in TG2 expression was also evident in tumors obtained from mice that received TG2 siRNA i.p., but the degree of inhibition from mouse to mouse was less consistent than that in i.v. group. Furthermore, fluorescence microscopy revealed a significant uptake of fluorescent-labeled TG2 siRNA in tumors treated i.v. (Fig. 1B). Based on these results, we used the i.v. route for administration of siRNA-DOPC in subsequent therapeutic experiments.

In vivo therapy of PDAC using liposomal siRNA targeting TG2. Three weeks after tumor cell implantation, mice were treated according to protocol described in Materials and Methods. Three days after the last treatment, mice were killed and examined carefully for tumor size, metastasis, and other gross pathologic changes. Luciferase bioluminescence data derived from weekly IVIS images revealed that the tumor growth in mice given TG2 siRNA-DOPC in combination with gemcitabine was significantly retarded (P = 0.019) when compared with that in the mice given control siRNA-DOPC (Fig. 2A and B ). Treatment with either TG2 siRNA-DOPC or gemcitabine alone resulted in ∼50% reduction in the final tumor volumes (Fig. 2C). Thus, the combination of TG2 siRNA-DOPC and gemcitabine had a much superior efficacy than did TG2 siRNA-DOPC or gemcitabine alone. Mice tolerated siRNA-DOPC treatment extremely well as there was no change in eating habits or motility and mouse weights were not significantly different among the four groups of animals.

The Panc-28–induced tumors rapidly progressed into locally advanced disease in nude mice. Therefore, we also examined the effect of various treatments on the number of metastatic foci in organs adjacent to pancreas, such as the spleen, mesentery/omentum, and liver, in these mice. The results shown in Fig. 3 clearly show that treatment with TG2 siRNA-DOPC alone was highly effective in inhibiting the progression of PDAC tumors to metastatic disease. Interestingly, treatment with gemcitabine alone, which caused noticeable reduction in tumor growth, was completely ineffective in inhibiting metastasis. Mice given TG2 siRNA-DOPC in combination with gemcitabine also had significant decreases in their metastasis scores similar to those in mice given TG2 siRNA-DOPC (Fig. 3; Table 1 ). These results suggested that treatment with gemcitabine alone, although capable of inhibiting the growth of or inducing cell death in some Panc-28 clones, does not affect the survival or growth of metastatic clones.

Effect of targeting TG2 on survival pathways. Based on our previous results that TG2 expression results in constitutive activation of the FAK/Akt and NF-κB survival pathways, we next examined the effect of TG2 siRNA treatment on these pathways. As shown in Fig. 4 , TG2 siRNA-DOPC–mediated down-regulation of TG2 in xenograft tumors was associated with inhibition of NF-κB activity and expression of its downstream target gene, VEGF. Moreover, tumors obtained from TG2 siRNA-DOPC–treated mice showed inhibition of pAkt (Ser⁴⁷³; Fig. 5 ). Tumors obtained from mice that received gemcitabine or control siRNA-DOPC alone showed no significant changes in NF-κB activity or pAkt levels.

Effect of TG2 inhibition on cell proliferation, angiogenesis, and apoptosis. To determine potential mechanisms underlying the antitumor effect of TG2 siRNA-DOPC, we examined its effect on several biological end points, including cell proliferation (Ki-67), angiogenesis (CD31), and apoptosis (TUNEL). More than 90% reduction in Ki-67 expression was evident (P < 0.001) in tumors obtained from mice given TG2 siRNA-DOPC alone or in combination with gemcitabine (Fig. 5). Tumors obtained from mice given gemcitabine alone or control siRNA-DOPC consistently showed high levels of Ki-67 staining (Fig. 5).

Next, we evaluated the blood vessel density in tumors obtained from mice in the different treatment groups. The representative images shown in Fig. 5 reveal a significant decrease in the mean blood vessel density in tumors obtained from mice that received TG2 siRNA-DOPC or gemcitabine alone. Densitometric analysis showed a significant difference in the mean microvascular density (P < 0.0013) in the mice that received TG2 siRNA-DOPC plus gemcitabine compared with the mice that received control siRNA-DOPC. Taken together, these results suggested that down-regulation of TG2 expression in combination with gemcitabine can result in decreased VEGF expression and decreased mean blood vessel density. Thus, treatment with TG2 siRNA-DOPC plus gemcitabine can affect the neovascularization process in a paracrine manner by affecting VEGF levels rather than due to down-regulation of TG2 in endothelial cells.

Finally, we evaluated apoptosis in orthotopic tumors using TUNEL assay. PDAC tumors treated with gemcitabine alone or in combination with TG2 siRNA-DOPC showed significant increases in the numbers of apoptotic cells. No differences in the apoptotic index in TG2 siRNA-DOPC–treated tumors were evident, and the extent of apoptosis in this group was similar to that in the control siRNA-DOPC group (data not shown).

Discussion

In this study, we examined the potential of TG2 as a therapeutic target for treating PDAC in an orthotopic nude mouse model. Using DOPC liposomes as the delivery system, we showed that TG2 siRNA could effectively down-regulate TG2 expression in tumors and inhibit their growth and metastatic spread. Previous studies convincingly documented that overexpression of TG2 in various cancer types is closely related to drug resistance and metastasis (9–17). In addition, the majority of PDAC tumors and cell lines have increased basal levels of TG2 expression (20). A comprehensive analysis of 33,000 genes using three different methods identified TG2 as one of the most differentially expressed genes in PDAC tumors (16). Similarly, in an attempt to identify metastasis-associated proteins using proteomic analysis, Jiang et al. (17) identified TG2 as one of 11 proteins selectively amplified in metastatic human lung carcinoma. Based on these observations, we reasoned that elevated expression of TG2 in pancreatic cancer cells could serve as a novel target in treating PDAC tumors, which exhibit intrinsic resistance to chemotherapeutic drugs and high metastatic potential. The data presented here clearly show the usefulness of down-regulating TG2 expression in inhibiting PDAC tumor growth and metastatic spread of the disease. Notably, down-regulation of TG2 expression by siRNA-DOPC rendered PDAC cells sensitive to gemcitabine, the drug most commonly used to treat this tumor, both alone and in combination with other therapeutic modalities, without much effect on survival.

Several recent reports have documented a role for TG2 in promoting cell migration and invasion. For example, TG2 was implicated to play a role in transmigration of T lymphocytes (25); migration of retinal pigment epithelial (26), monocytic (27), and neuroblastoma cells (28); and invasion of breast (22) and pancreatic cancer cells (20). Many types of cancer cells, especially when rendered resistant to drugs or isolated from metastatic sites, have high basal levels of TG2 expression (9–17, 29). Indeed, inhibition of TG2 expression using siRNA, antisense, or ribozyme approach resulted in reversal of drug resistance in breast and lung cancer cells (11, 14, 15). Similarly, down-regulation of TG2 by siRNA inhibited the invasiveness whereas ectopic expression promoted the invasiveness in breast cancer cells (22). These findings were further supported by a recent report showing diminished dissemination of tumors on the peritoneal surface and mesentery following TG2 knockdown in an i.p. ovarian xenograft model (30).

Despite all of these findings, the mechanism by which TG2 contributes to the development of chemoresistance and metastatic phenotypes remains unknown. A recent report by Akimov et al. (31) suggested that TG2 could interact and form stable complexes with β₁ and β₃ integrins. This finding led to the delineation of significance of increased TG2 expression in cancer cells. We previously showed that, in addition to β₁ integrin, TG2 could bind to and form complexes with β₄ and β₅ integrin in breast cancer (11), malignant melanoma (12), and pancreatic cancer⁷ cells. TG2 is known to associate with integrins primarily in the extracellular domains of integrins and to promote their interaction with the extracellular matrix ligands, such as fibronectin, collagen, and vitronectin (22, 31). Therefore, it is possible that TG2 contributes to the aggressive behavior of PDAC by constitutively activating the integrin-mediated cell survival signaling pathways.

Indeed, our efforts to understand the biological significance of increased TG2 expression in PDAC cells led us to the discovery that TG2 expression results in constitutive activation of FAK, Akt, and NF-κB (20, 21). Thus, TG2 expression activated and autophosphorylated FAK (pY397), and down-regulation of TG2 expression inhibited this activation. Moreover, TG2-induced activation of FAK results in activation of the phosphoinositide 3-kinase/Akt cell survival signaling pathway (20). Our more recent studies have provided some interesting leads and suggest that expression of TG2 in PDAC cells is inversely correlated with PTEN expression (32). TG2 inhibited the phosphorylation of PTEN at Ser³⁸⁰ and promoted its degradation via the ubiquitin/proteasome pathway. Thus, TG2 plays a dual role in the FAK/phosphoinositide 3-kinase/Akt signaling pathway by promoting direct phosphorylation of FAK and inhibiting its upstream phosphatase, PTEN, resulting in constitutive activation of the phosphoinositide 3-kinase/Akt pathway.

Moreover, TG2 expression results in constitutive activation of NF-κB (14, 21, 33). Under normal conditions, the inflammation-induced NF-κB activation is transient, which is critical for the controlled growth and survival of normal cells. Thus, TG2-mediated constitutive activation of NF-κB in cancer cells may be critical to conferring chemoresistance and metastatic phenotypes. Indeed, our recent data supported this contention and suggested that down-regulation of TG2 expression by siRNA in drug-resistant breast cancer and PDAC cells results in inhibition of NF-κB and increases their sensitivity to chemotherapeutic drugs (11, 20, 21). Kim et al. (14) observed a similar correlation between TG2 expression, drug resistance, and NF-κB activation in breast cancer cells. Taken together, these results suggest that TG2 plays an important role in promoting cell survival signaling and thus protects PDAC cells against drug-induced cell death.

Importantly, the present study provided the proof of concept that therapeutic delivery of gene-specific siRNA to PDAC tumors can be achieved using DOPC liposomes. Two recent reports documented a similar effect of DOPC liposomal delivery of siRNA to silence the expression of the oncoprotein EphA2 (23) and FAK (34). Using this approach, both studies documented encouraging responses to the treatment of ovarian tumors in mice. Thus, the significance of the results of this study can be rapidly translated into the clinical setting for treatment of aggressive forms of PDAC.