Expression of the PAX8/PPARγ Fusion Protein Is Associated with Decreased Neovascularization In Vivo

Abstract

ThePAX8/PPARγ fusion protein (PPFP) occurs in 36% of human follicular thyroid carcinoma (FTC) and is associated with favorable prognosis. To elucidate the function of PPFP in FTC, we analyzed the consequences of PPFP expression in immortalized thyrocytesin vitroandin vivovia xenograft tumorigenesis. While PPFP-expressing cells exhibited oncogenic hallmarks, including increased growth and decreased apoptosis,in vitro, xenograft tumors were initiated but not sustainedin vivo. PPFP xenograft tumors exhibited reduced CD31 staining and VEGF expression, suggesting that PPFP modulates neovascularization. Microarray analysis demonstrated increased expression of tissue inhibitor of metalloproteinase (TIMP-3), an inhibitor of angiogenesis, in PPFP cells and tumors, a finding confirmed by quantitative PCR and immunohistochemistry. Immunohistochemical staining of archival human thyroid tumors demonstrates a significant decrease in CD31 staining in all adenomas and carcinomas containing the PAX8/PPARγ rearrangement. Decreased angiogenesis in PPFP-containing tumors is directly correlated with our observations in the xenograft model and provides evidence for the first time that PPFP may impact FTC tumorigenesis by modulating angiogenesisin vivo.

Keywords

follicular thyroid carcinoma PAX8/PPARγ fusion protein xenograft tumors neovascularization angiogenesis tumor suppressor

Introduction

Follicular thyroid carcinoma (FTC) accounts for 20% of all thyroid cancers but is associated with 40% of the deaths. ThePAX8/PPARγ fusion protein (PPFP) is detected in about 40% of FTC¹and arises through a balanced chromosomal translocation, t(2;3)(q13;p25), fusing the 5′ portion of the thyroid-specific transcription factor genePAX8in frame with the entire translated reading frame of the peroxisome proliferator-activated receptor gamma (PPARγ) nuclear receptor gene.^2,3Chromosomal aberrations (translocations, inversions, deletions) are usually thought to be associated with malignant potential.⁴Papillary thyroid carcinoma (PTC), the most common thyroid cancer morphotype, is associated with 3 different rearrangements: RET/PTC,NTRK1, andBRAF.⁵However, unlike these rearrangements that are shown to drive PTC initiation and progression, it is not yet known whether PPFP expression in FTC is sufficient to cause malignant transformationin vivo. Most studies to date have dealt with PPFP effectsin vitro, demonstrating that transient expression of PPFP is oncogenicin vitro, leading to increased cell proliferation, decreased apoptosis, and decreased anchorage dependency.^2,6-8However, studiesin vivohave been restricted to the clinicopathological evaluation of FTC in the context of PPFP expression. Although some of these studies correlate the presence of PPFP with more aggressive cancers,^9-11the majority indicate that PPFP expression is correlated with mildly invasive FTC tumors with a more favorable prognosis,^12-17indicating a potential role for PPFP in the modulation of FTC tumorigenesis.

There are currently no cell culture models available that express PPFP for use in mechanistic studies. The present study sought to address this lacuna by generating a biological model of immortalized thyrocytes, permanently expressing PPFP. This model was then used to study the consequences of PPFP expression in terms of oncogenic potentialin vitroand tumorigenic potential using a xenograft modelin vivo. The observations from the cell culture and mouse xenograft studies were then verified in archival human thyroid tumors to elucidate PPFP function in the physiological context. Our study demonstrates for the first time that PPFP expressionin vivois associated with decreased vascularization and provides insight into the potential pathways involved in this PPFP function in FTC tumorigenesis.

Results

PPFP Expression in Immortalized Thyrocytes Enhances Oncogenic PotentialIn Vitro

We transfected the pPPFP construct (Fig. 1A) into immortalized thyrocytes (NT) to generate NT-PPFP stable cells that express PPFP (Fig. 1B) and compared them with NT-vector cells. We compared cell growth, apoptosis, colony formation in soft agar, and matrix invasion for NT-PPFP and NT-vector cells. The NT-PPFP cells showed a 1.2-, 1.4-, and 1.6-fold (P<0.005) increase in cell growth at 48, 72, and 96 hours with a 2.6- to 2.3-fold (P<0.005) lower rate of apoptosis at 72 and 96 hours, respectively, compared to NT-vector cells (Fig. 1 CandD). The NT-PPFP cells formed a significantly (P<0.005) higher number of colonies in soft agar (Fig. 1E), which was increased by 3-fold over vector cells. In the matrigel assay, NT-PPFP cells exhibited 3.5-fold (P<0.005) increased invasion indices compared to the vector control cells (Fig. 1F). The results show that permanently transfected NT-PPFP cells exhibit oncogenic characteristics very similar to those of transiently transfected NT cells.⁸

Figure 1.

PPFP expression enhances oncogenic activityin vitro, initiates, but does not form sustained tumorsin vivo. (A) Scheme for PPFP expression construct. (B) Immunoprecipitation (IP)/Western blot analysis of PPFP expression in NT cells stably generated with either vector alone or pPPFP as described in the Materials & Methods section. Cell growth (C) and apoptosis (D) in NT-vector (open circles) and NT-PPFP (black circles). Colony formation in soft agar (E) and matrigel invasion (F) by NT-vector (white bars) and NT-PPFP (black bars) cells. Subcutaneous injection of 10⁷cells of either NT-vector (open circles) or NT-PPFP cells (black circles) resuspended in media alone (G) or in 50% matrigel (H) into the right and left flanks of athymic nude mice was monitored over a period of 5 to 8 weeks, and tumor volume was measured and plotted as a function of time. (I) Ki67-stained images of NT-vector NT-PPFP xenograft tumors harvested at peak (week 1) at a magnification of 200x. (J) Quantitation of the Ki67 staining intensity showed a significant (P<0.05) increase in the NT-PPFP versus NT-vector tumors. (K) Reverse transcriptase–PCR of total RNA from NT-vector and NT-PPFP xenograft tumors harvested at peak growth (week 1) and during tumor regression (week 2) confirmed expression of PPFP, PAX8, and TTF-1 in the NT-PPFP tumors. The lowest panel shows expression of GAPDH mRNA, the housekeeping gene. Data inCare an average of 6 independent wells. Data inD,E, andFrepresent the average data of 3 independent experiments, each conducted in triplicate. The values inGandHrepresent the mean ± standard error from 8 (NT-vector and NT-PPFP) independent tumors. Error bars indicate standard deviation.

Enhanced Oncogenic Potential of PPFPIn VitroDoes Not Result in Sustained TumorigenesisIn Vivo

Subcutaneous injection of athymic nude mice with NT-PPFP cells led to the development of tumors that were 4 times larger than NT-vector cell-induced tumors by week 1 (Fig. 1G). However, tumors were not sustained, with complete involution occurring by week 4. Involution of the NT-PPFP–associated tumors was slower than that of the control group. Resuspension of NT-PPFP cells in 50% matrigel prior to tumor induction did not alter tumor initiation or growth (Fig. 1H). Tumor volume data were subjected to analysis of variance for repeated measures, which demonstrated differences at theP<0.05 level forpost hocBonferroni tests at individual time points. Tests for the time versus interaction of NT-PPFP and NT-vector were highly significant with and without matrigel (P= 0.0006 and 0.0001, respectively). Ki67 staining of the xenografts (Fig. 1I) clearly demonstrated a 1.8-fold (P<0.05) increase in the NT-PPFP tumors (Fig. 1J), indicating that cell proliferation was responsible for the difference in xenograft size. Expression of PPFP, PAX8, and TTF-1 mRNA was also evaluated in tumors at weeks 1 and 2 (Fig. 1K), demonstrating that lack of tumorigenesis was not due to loss of PPFP expression in the xenografts as the tumor progressed and that the thyroid phenotype of the tumors was not changed after implantation (Fig. 1K). Overall, the data suggest that in contrast to the oncogenic potential observedin vitro, PPFP is able to initiate tumor formation but is not able to sustain tumorigenesis in a xenograft mouse modelin vivo.

Initiation of Xenograft Tumorigenesis Is Dependent on Inhibition of PPARγ Function

Because our earlier observations demonstrated that PPFP acted at least in part through inhibition of PPARγ,⁸we evaluated NT cells expressing a dominant negative (DN) mutant of PPARγ (Fig. 2A) for its tumorigenic potential in the xenograft model. DN-PPARγ was efficiently expressed in the stable cells and readily detected by Western blot analysis (Fig. 2B). Reporter assays confirmed that activation of the PPARγ response element (PPRE) was significantly inhibited in both NT-PPFP (29%) and NT-DN-PPARγ (41%) cells (Fig. 2C), demonstrating that PPFP retains its dominant negative PPARγ function in the NT-PPFP cells and that DN-PPARγ is expressed in a functional manner in NT-DN-PPARγ cells. Although the NT-DN-PPARγ cells induced somewhat larger tumors compared to the NT-PPFP cells (Fig. 2D), the DN-PPARγ tumors also underwent spontaneous involution, with complete regression by week 4. Thus, NT cells expressing either PPFP or DN-PPARγ exhibit identical phenotypes, providing evidence that inhibition of PPARγ activity in these cells results in tumor initiation but not sustained tumorigenesis. However, because endogenous PPARγ activity was not completely suppressed in either NT-PPFP or NT-DN-PPARγ cells (Fig. 2B), these data raise the question whether complete inhibition of PPARγ would result in sustained tumorigenesis or whether other factors are responsible for the observed effects of PPFP expressionin vivo.

Figure 2.

Dominant negative inhibition of PPARγ does not induce sustained tumorigenesisin vivo. (A) Scheme for the pDN-PPARγ expression construct. (B) Expression of PPARγ, DN-PPARγ, GFP (hrGFPII, vector), and β-actin proteins by Western blotting in NT-vector, NT-PPFP, or NT-DN-PPARγ cells. (C) Relative PPARγ activity in NT-vector, NT-PPFP, and NT-DN-PPARγ cells. The respective cells were transiently transfected with a PPRE-Luc reporter and assayed for promoter activity as described in the Materials & Methods section. The fold basal luciferase/renilla ratios are expressed in each case compared to the vector control. Asterisks denote significant difference (P<0.05) from the control cells after analysis of variance andpost hocBonferronittests. (D) Subcutaneous injection of 10⁷cells of NT-vector, NT-PPFP, or NT-DN-PPARγ cells into the right and left flanks of athymic nude mice was monitored over a period of 4 weeks, and tumor volume was measured. Compared to the NT-vector cells (open square), NT-PPFP (black square) and NT-DN-PPARγ (black circles) cells demonstrated a significant (P<0.05) spurt in tumor growth by day 3 followed by regression of the tumors by week 2 before ultimate tumor involution by week 4. Data inBrepresent a mean of 2 experiments each done in triplicate. Data inDrepresent a mean of 6 tumors per group. Error bars inCandDindicate standard deviation and standard error, respectively.

PPFP-Expressing Tumors Have Decreased Angiogenesis and Neovascularization

Because tumor growth and progression are dependent on neovascularization,¹⁸we analyzed PPFP-expressing xenograft tumors at their peak size, for vasculature by immunohistochemical analysis of the endothelial cell–specific marker CD31 (platelet endothelial cell adhesion marker 1 [PECAM]) (Fig. 3A). Quantification of intratumoral microvessel density showed a significant reduction of 2.4-fold (P<0.02) in the PPFP tumors compared to the controls (Fig. 3B). Because vasculature is primarily controlled by expression of the vascular endothelial growth factor (VEGF), an angiogenesis-specific factor, we also quantified VEGF expression by immunohistochemical staining (Fig. 3C), which demonstrated a 1.4-fold reduction in VEGF expression in the NT-PPFP tumors (P<0.05) versus NT-vector tumors (Fig. 3D). RT-PCR confirmed a reduction in mRNA levels of the different VEGF isoforms in the NT-PPFP tumors at peak growth (Fig. 3 EandF).

Figure 3.

CD31 and VEGF expression is significantly reduced in PPFP xenograft tumors. (A) Xenograft tumors induced with either NT-vector or NT-PPFP cells were harvested at a peak of 1 week and evaluated for intratumor microvessel density by CD31 staining. (C) Tumors were also stained for the expression of the angiogenic-specific marker, VEGF. Quantitation of stained slides (Materials & Methods) demonstrated a significant decrease in the staining intensity of CD31 (B) and VEGF (D), respectively. RT-PCR of RNA from xenograft tumors was evaluated (E) and quantitated using densitometry (F) for VEGF transcript levels (Materials & Methods). (F) Graph shows that the 4 different VEGF isoforms are significantly reduced (*P<0.05, **P<0.005) compared to GAPDH controls in NT-PPFP (black bars) compared to NT-vector (white bars) tumors. Data represent the average data of 6 tumors per cell line (error bars indicate standard error).

To understand the possible mechanisms underlying the lack of sustained tumorigenesis and reduced neovascularization, we performed gene expression analyses of RNA from NT-PPFP and NT-vector cells. Differential regulation was observed in the expression of 67 genes, 8 of which are involved in angiogenesis (Supp. Table S1). These genes include proangiogenic PFKFB3, PDK1, NDRG1, BNIP3, and MALAT1, which were decreased 2-, 2.3-, 2.5-, 3.3-, and 4-fold, respectively, in NT-PPFP cells. Expression of antiangiogenic TIMP-3 was increased 2.6-fold, and 2 antiangiogenic genes, ADAMTS2 and PGK1, were down-regulated 2- and 4.5-fold, respectively, in the PPFP cells. TIMP-3 is known to inhibit expression of ADAMTS2,¹⁹which could explain its decreased expression in NT-PPFP cells, but the reason for lower expression of PGK1 is unknown. Because TIMP-3 is tightly bound to the extracellular matrix by interaction with heparan sulfate²⁰and is an inhibitor of angiogenesis through blocking VEGF binding to the VEGF receptor-2,^21,22it was studied further.

TIMP-3 transcript and protein expression were validated in NT-PPFP cells and tumors using real-time PCR and immunohistochemistry, which demonstrated a 6.5-fold increase (P<0.05) in mRNA (Fig. 4 AandB) and 1.6-fold increase (P<0.05) in protein (Fig. 4 CandD), respectively. TIMP-3 staining appeared to be predominantly cytosolic in the presence of PPFP compared to the vector-only controls.

Figure 4.

TIMP-3 expression is increased in NT-PPFP cells and xenograft tumors. Total RNA from PPFP-expressing cell lines was extracted and TIMP-3 transcript levels were detected by real-time PCR. (A) Representative amplification plot of TIMP-3 transcript in the NT-vector and NT-PPFP cells. (B) Quantitation of transcript levels normalized to GAPDH showed a significant (P<0.05) increase in TIMP-3 expression in NT-PPFP versus NT-vector cells. Graph represents the average data of 3 independent experiments, each conducted in triplicate. (C) Xenograft tumors induced with NT-vector or NT-PPFP cells were harvested at peak growth and immunohistochemically stained for TIMP-3 expression. Representative areas of tumors stained for TIMP-3 expression, top panel (5x) and lower panel (20x) magnification. (D) Quantitation of stained slides (Materials & Methods) demonstrated a significant (P<0.05) increase in the staining intensity of TIMP-3 in the presence of PPFP. Data represent the average data of 6 tumors per cell line (error bars indicate standard error).

Follicular Thyroid Tumors Expressing PPFP Have Decreased Microvessel Density

To begin to understand PPFP’s pathophysiological role in thyrocyte transformation and tumor initiation/progression, we evaluated a cohort of follicular adenomas (FA) and carcinomas (FTC) with and without PPFP for microvessel density (Fig. 5) and VEGF expression (Fig. 6). While both PPFP-containing FA and FTC demonstrated a 2.3- and 2.5-fold (P<0.05) decrease in microvessel density, respectively (Fig. 5F), only PPFP-FA showed a 1.3-fold (P<0.03) decrease in VEGF expression (Fig. 6F) similar to NT-PPFP cells. No significant differences were observed in VEGF expression in the FTC tumors (Fig. 6F). To understand the clinical implications of TIMP-3 expression and its regulation by PPFP in terms of follicular thyroid tumorigenesis, we evaluated the same thyroid tumors for TIMP-3 expression using immunohistochemistry (Fig. 7). While TIMP-3 expression was down-regulated 1.3-fold (P<0.005) in FTC containing PPFP, in concordance with other studies,²³FAs containing PPFP had a 1.3-fold increase (P<0.005) in TIMP-3 expression compared to tumors lacking PPFP (Fig. 7F), similar to the NT-PPFP cells. The increase in TIMP-3 expression accompanied by decreased CD-31 and VEGF in the PPFP-FA suggests that the NT-PPFP cells could be representative of the PPFP adenomas and not carcinomasin vivo.

Figure 5.

Microvessel density is reduced in thyroid tumors containing PPFP. Immunohistochemical analysis of FA (A,B) and FTC (C,D) demonstrated a 2.3- and 2.5-fold decrease (P<0.005 andP<0.05, respectively) in microvessel density as measured by CD31 staining in adenomas and carcinomas containing PPFP, respectively (B). Normal thyroid tissue (E) had vascularization equivalent to FA that did not express PPFP. (F) Graph represents quantitation of antibody-positive signal measured from 3 normal thyroids, 6 FA (3 with PPFP), and 18 FTC (8 with PPFP), as described in the Materials & Methods section (error bars indicate standard error).

Figure 6.

VEGF expression is reduced in follicular adenomas containing PPFP. Immunohistochemical analysis of FA (A,B) and FTC (C,D) demonstrated a 1.3-fold decrease (P<0.003) in VEGF protein expression in the FA containing PPFP (B). Normal thyroid tissue (E) expressed significantly lower levels of VEGF protein compared to FA. (F) Graph represents quantitation (Materials & Methods) of antibody-positive signal measured from 3 normal thyroids, 6 FA (3 with PPFP), and 18 FTC (8 with PPFP) (error bars indicate standard error).

Figure 7.

Follicular adenomas containing PPFP exhibit higher TIMP-3 expression. Immunohistochemical analysis of FA (A, B) and FTC (C, D) demonstrated a 1.5-fold increase (P<0.001) and a 1.4-fold decrease (P<0.001), respectively, in TIMP-3 protein expression in the presence of PPFP (B, D). Normal thyroid tissue (E) expressed significantly lower levels of TIMP-3 protein compared to tumor tissues. (F) Graph represents quantitation (Materials & Methods) of antibody-positive signal measured from 3 normal thyroids, 6 FA (3 with PPFP), and 18 FTC (8 with PPFP) (error bars indicate standard error).

Discussion

Since the discovery of PPFP in FTC in 2000,²and the numerous studies that have demonstrated its presence in approximately 36% of FTC (reviewed¹), there has been considerable interest in understanding its role in thyroid tumorigenesis. Subsequently, it has also been found to be present in FAs (11%) and to a much lesser extent in Hürthle cell cancer (<2%), indicating that PPFP may be an initiating oncogene that requires additional secondary hits for complete transformation. Although there are multiplein vitrostudies that support its role as an oncogene,^2,6-8the function of PPFP and its mechanism of actionin vivohave yet to be elucidated. Because there are currently no human thyroid tumor-derived cell lines available that express PPFP, we used immortalized thyrocytes to overexpress PPFP, thereby generating a biological model to assess PPFP functionin vivo. We evaluated the PPFP-expressing cells for their oncogenic potentialin vitroand tumorigenic potentialin vivoin a mouse xenograft model. Although the PPFP-expressing cells demonstrated enhanced oncogenic potentialin vitro, they initiated but failed to support sustained tumorigenesisin vivo, even in the presence of matrigel.

The NT cell line we have used as an experimental model is an immortalized cell line and therefore fulfills the criterion of unlimited replicative potential.²⁴The large T antigen transformation used to immortalize these cells adds further growth regulatory dysfunction to the PPFP-related effects in the form of impaired RB and TP53 signaling.²⁵Furthermore, thein vitroand, at least the initial,in vivodifferences in cell growth and behavior between PPFP-expressing and control NT cells suggest that the growth-promoting effects of PPFP do not overlap substantially with those due to the inherent RB and TP53 abnormalities of these cells.

Without angiogenesis, most tumors cannot sustain significant growth because of diffusion limitations of oxygen and essential nutrients.²⁶Analyses of the NT-PPFP xenografts confirm that both neovascularization (as evidenced by CD-31 staining) and VEGF protein expression levels are significantly reduced. Moreover, gene expression analysis of the NT-PPFP cells demonstrated an increase in the expression of the tissue inhibitor of metalloproteinase, TIMP-3, a potent inhibitor of angiogenesis and tumor suppressor. Of the 8 differentially regulated genes involved in angiogenesis, we focused on TIMP-3 for further study because V189 and V165, the 2 extracellular matrix (ECM)– associated isoforms of VEGF, were down-regulated in the presence of PPFP, and TIMP-3 is known to inhibit VEGF-mediated angiogenesis by blocking the binding of VEGF to its receptor, VEGFR-2.^22,27,28Chettyet al. showed that MMP-2 inhibition in lung cancer cells resulted in decreased induction of VEGF in endothelial cells that culminated in TIMP-3 induction, resulting in inhibition of p-AKT and induction of p-ERK1/2 leading to inhibition of angiogenesis and endothelial apoptosis.²¹Although we did not observe any significant change in MMP-2 activity in our zymogram studies (data not shown), we do see a significant decrease in VEGF expression in the PPFP xenografts. Because TIMP-3 is a known inhibitor of angiogenesis,^29,30both the down-regulation of VEGF and up-regulation of TIMP-3 provide a plausible mechanism for the observed decrease in angiogenesis and neovascularization.

Early studies demonstrated that PPFP functions by dominant negative inhibition of PPARγ.^2,8While more recent studies have confirmed the dominant negative mechanism, they have also provided evidence that PPFP can function as a partial agonist of both PAX8 and PPARγ pathways that is cell and promoter context dependent.^6,7Because the NT-DN-PPARγ cells exhibit essentially identical behavior in nude mouse xenografts as the NT-PPFP cells, it is likely that the observed effects on initial tumor cell growth and subsequent involution are mediated at least in part via inhibition of PPARγ-dependent pathways. However, the possibility that PPFP directly affects selected PPARγ pathways cannot be excluded because PPARγ exhibits both positive and negative actions on both tumorigenesis³¹and angiogenesis.³²Interestingly, both VEGF³³and TIMP-3³⁴contain PPARγ response elements (PPREs) in their promoter regions; however, how PPFP regulates these genes is unknown. While activated PPARγ inhibits VEGF protein secretion,³³whether this is mediated at the transcriptional level is unknown, and the functional significance of the TIMP-3 PPRE remains to be established. The factors regulating TIMP-3 expression are incompletely understood. There is growing evidence that methylation of its promoter is frequent in a variety of cancers, including colorectal³⁵and esophageal squamous cell carcinoma.³⁶In thyroid cancer, methylation and subsequent silencing of TIMP-3 are significantly associated with aggressive papillary thyroid cancer (PTC), including extrathyroidal invasion and lymph node metastasis.³⁷Thus, the increase in TIMP-3 expression in PPFP-FA is consistent with a tumor inhibitory role for PPFP.

To extrapolate our observations in the PPFP xenograft model to human thyroid tumors in a physiological context, we screened FAs and FTCs with and without PPFP for their relative expression of VEGF, CD31, and TIMP-3. As summarized inTable 1, CD-31 staining, a marker for microvessel density,³⁸clearly shows that PPFP expression in both adenomas and carcinomas is associated with reduced vascularization. Also, PPFP-FA exhibited decreased VEGF with increased TIMP-3 expression similar to the NT-PPFP cells, suggesting that the NT-PPFP cells appear to be more closely related to PPFP-FA than PPFP-FTCin vivo(Table 1). Although there was no discernible relationship of the NT-PPFP cell microarray expression data when compared with the FA (unknown PPFP status) or FTC microarray data of Giordanoet al.³⁹or Luiet al.,⁴⁰TIMP-3 down-regulation in FTC containing PPFP³⁴correlates with the observation in the current study. Further evaluation of the clinical history of the tumor specimen used in this study did not demonstrate any differences in tumor size in either FA (2.5 ± 0.7 cm v. 3.5 ± 2.6 cm) or FTC (3.8 ± 2.8 cm v. 3.8 ± 1.5 cm), with or without PPFP, respectively, suggesting that although tumor size was not affected, the reduced neovascularization observed in PPFP-expressing FTC provides a plausible explanation for the evidence of favorable prognostic impact of PPFP,^12-17in terms of better tumor differentiation and lesser invasiveness. Our findings that FA containing PPFP exhibit increased TIMP-3 expression and decreased neovascularization and angiogenesis suggest that the NT-PPFP cells mimic FA-PPFP and might be useful to identify the additional factors required for the transition of an adenoma to carcinoma, given that PPFP is also detected in 11% of follicular adenomas.¹

Table 1.

Effects of PPFPIn Vivoon Neovascularization Markers as Evaluated by Immunohistochemistry

	NT-PPFP^a	FA-PPFP^b	FTC-PPFP^b
CD-31	−2.4	−2.3	−2.5
VEGF	−1.4	−1.3	No difference
TIMP-3	1.6	1.5	−1.3
Phenotypein vivo	Tumors initiated but not sustained	Benign tumors	Minimally invasive tumors

Fold expression in PPFP xenograft tumors compared to vector-only tumors.

Fold expression in human FA or FTC PPFP-containing tumors compared to tumors without PPFP.

In summary, by using a combination ofin vitroandin vivoPPFP models, followed by validation of the data in human tumors, we demonstrate for the first time that PPFP expression is associated with decreased vascularization and therefore may have potential implications on the prognosis and management of follicular thyroid cancer.

Materials and Methods

PPFP Expression Constructs

pPPFP (Fig. 1A) was constructed by subcloning thePAX8/PPARγ NotI fragment from pCR3.1/PPFP²into hrGFPII vector (Stratagene, La Jolla, CA). The adenovirus tripartite leader-intron (TI) sequence from pMT2⁴¹was removed with PvuII-EcoRI digestion, blunt ended with T4 DNA polymerase, and inserted into the EcoRV site of pPPFP. The dominant negative pDN-PPARγ lacking 16 C-terminal amino acids was generated from mouse PPARγ2 template by PCR using 5′-attgcggccgcttaatgggtgaaactctggga-3′ and 5′-ctgcggccgcctgtctctgtcttcttgatcacatg-3′ primers, producing a 1350-bp fragment with NotI ends. This fragment was cloned into hrGFPII vector, generating DN-PPARγ fused to FLAG. Plasmids were verified by DNA sequencing.

Cell Culture and Transfection

Human Nthy-ori-3-1 (NT) cells (#90011609; European Collection of Cell Cultures, Porton Down, UK)⁴²were grown in RPMI-1640 medium (Invitrogen, Carlsbad, CA) supplemented with 10% FBS (Hyclone, Logan, UT) and 1X penicillin-streptomycin (Invitrogen) at 37 C in humidified conditions with 5% CO₂. NT cells were seeded at 1 × 10⁶cells per 100-mm Petri dish and transfected 24 hours later in multiples of 6, with 10 µg each of vector, pPPFP, or pDN-PPARγ, using Fugene 6 (Roche, Indianapolis, IN). After 72 hours, NT-vector, NT-PPFP, or NT-DN-PPARγ cells were obtained by selection in media containing 400 µg/mL of G418 (Invitrogen) with media changes every 72 hours. Pools of cells were harvested and analyzed for mRNA expression by RT-PCR and protein by immunoprecipitation and/or Western blotting.

RNA Isolation, RT-PCR, and Quantitative PCR

RNA from each group of cells was isolated using TriZol Reagent (Invitrogen). RNA (1 µg) was reverse transcribed with oligo-dT primers using the Message Sensor RT kit (Ambion, Austin, TX).PAX8/PPARγ fusion gene was detected by PCR using 5 µL of cDNA and a sense primer from exon 7 ofPAX8, 5′-CGCGGATCCGCATTGACTCACAGAGCA-3′ and antisense primer from exon 1 ofPPARγ1, 5′-CCGGAATTCGAAGTCAACAGTAGTGAA-3′. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was amplified using kit primers. Expression of TIMP-3 mRNA was evaluated using 5′-GCAGCAGCGGCAATGACC-3′ and 5′-CCGGATGCAGGCGTAGTGTTT-3′ primers. Xenografts were harvested at weeks 1 and 2 postinduction, homogenized in TriZol Reagent (Pierce, Rockford, IL), RNA extracted, and RT-PCR performed for PPFP, TIMP-3, and GAPDH as described above. Endogenous PAX8 expression was detected using a forward primer 5′-TTTCAATGGCAGAACTTCCC-3′ and reverse primer 5′-GGATGCTGTGCTCAAAGTA-3′ derived from the 3′-UTR of the PAX8 mRNA sequences not present in PPFP. Annealing temperature for PPFP, TIMP-3, PAX8, and GAPDH was 55 C for 30 seconds. PCR for TTF-1 and VEGF was carried out as described by Bhattacharyyaet al.⁴³and Munautet al.,⁴⁴respectively. Real-time PCR was performed using 1 µL of cDNA and SYBR green reagent mix from ABI (Foster City, CA).

Protein Expression Analysis

Stable cells (2 × 10⁶per 100-cm Petri dish) were harvested at 24 hours postseeding in M-PER reagent (Pierce), and protein was quantified (Coomassie; Pierce). Protein expression was assayed by Western blotting.⁸PPFP and PPARγ proteins were detected using anti-PPARγ antibody (sc-7273; Santa Cruz Biotechnology Inc., Santa Cruz, CA). DN-PPARγ and actin were detected with anti-Flag M2 (Stratagene) and antihuman β-actin (Sigma-Aldrich, St. Louis, MO) antibodies, respectively. Presence of vector, pPPFP, and pDN-PPARγ were confirmed with anti-hrGFP antibody (Stratagene). Detection and densitometric analysis utilized the Kodak ISO 4000 instrument (Rochester, NY) with normalization to β-actin. For immunoprecipitation (IP), 1 mg of cell lysate was used and processed using protein A/G beads (Santa Cruz Biotechnology Inc.) as per the manufacturer’s specifications. IP products were resolved on a 4% to 20% SDS-polyacrylamide gel followed by Western blot analysis.

Cell Growth Assay

Cells were plated (5 × 10⁴per well) in 24-well plates. After 24, 48, 72, and 96 hours, cell viability was quantified by MTT (3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyl tetrasodium bromide) (Sigma-Aldrich) per Denizotet al.⁴⁵

Apoptosis Assay

After 24, 48, and 72 hours, adherent and floating cells were harvested. Cells (10⁶) were washed twice in binding buffer, biotin-annexin V (Trevigen, Gaithersburg, MD) was added (1:200), and cells were incubated for 30 minutes on ice and washed once before labeling with phycoerythrin-conjugated streptavidin (1:200) for 15 minutes on ice. Cells were washed and resuspended in 400 µL of binding buffer and analyzed by FACS.

Colony Formation Assay

Cells (10⁴) were trypsinized, resuspended in 1 mL of 0.3% agarose/RPMI-1640, and layered onto a solidified bottom layer of 0.5% agarose/RPMI-1640 in 6-well plates. Normal growth media (1 mL) was added onto the gel, and media was changed twice a week. Colonies visible to the naked eye were identified, circled, verified by microscopic examination, and counted.

Invasion Assay

Matrigel inserts (#354480; BD Bioscience, Sparks, MD) were thawed and rehydrated along with an equal number of control inserts (#354578; BD Bioscience). Complete growth medium was added to the bottom of the wells and the chambers and control inserts were transferred to the wells. Cell suspensions (5 × 10⁴cells/mL) in serum-free medium were added to the chambers and incubated for 24 hours. Noninvading cells were scrubbed from the upper membrane surface, and cells on the lower membrane surface were stained with Diff-Quik stain (VWR Scientific, West Chester, PA) and air dried. Invading cells were counted and data expressed as percentage invasion compared to control cells.

Luciferase Reporter Assays

Cells (0.3 × 10⁶cells/well) in 6-well plates were transfected using LT-1 reagent (Mirus, Madison, WI) with 10 ng of Renilla (Promega, Madison, WI) and 100 µg of PPRE-Aco-Luc reporter²in triplicate. After 24 hours, cells were lysed and assayed using the dual luciferase kit (Promega). Data were calculated as a ratio of the luciferase to renilla expression and plotted as fold-change compared to vector.

In VivoXenograft Tumor Formation Assays

NT-vector, NT-PPFP, or NT-DN-PPARγ cells (10⁷) were resuspended in 100 µL of growth media and injected subcutaneously into both flanks of 5-week-old athymic nude mice (Fox1^nu/nu; Harlan, Indianapolis, IN). At least 3 mice were injected per group. Tumor size was determined weekly by measuring tumor length (L), width (W), and height (H) using vernier calipers (V_tumor= 0.5236LWH). For the matrigel experiment, 10⁷cells (100 µL) were mixed with matrigel (#356230; BD Bioscience) in sterile growth media (100 µL). Animal studies were approved by the Mayo Clinic Institutional Animal Care and Use Committee.

Tumor Histology and Immunohistochemistry

Three mice per group (6 tumors) were formalin-fixed, paraffin-embedded, and individual sections stained with VEGF- (RB-222-P1, 1:100 dilution; Pierce), CD31- (sc-1506-R, 1:200 dilution; Santa Cruz), TIMP-3- (prediluted PA1-21136; Pierce), or Ki67- (MIB-1, 1:100 dilution; DAKO, Carpenteria, CA) specific antibodies for 30 to 60 minutes and counterstained with hematoxylin. Antigen retrieval was done in EDTA at 37 C for 30 minutes after deparaffinization. Sections were blocked with peroxidase for 5 minutes and protein for 5 minutes. Antibody staining was visualized using the DAKO EnVision plus System for 15 minutes followed by DAB chromogen for 10 minutes in the case of VEGF and CD31 staining. TIMP-3 and Ki67 antibody binding was visualized using the DAKO EnVision plus Dual Link System for 15 minutes followed by DAB and HRP chromogen for 5 and 15 minutes, respectively. Separate tumor sections were stained with hematoxylin and eosin to visualize tumor histology.

Stained slides were scanned using the Hamamatsu NanoZoomer (Bacus Laboratories Inc., Chicago, IL). Digital images composed of multiple tiles (4096 × 64 pixels) were captured at 20x magnification. Image analysis was conducted according to Monahanet al.⁴⁶For each tumor section, the relative signal intensity from 10 randomly chosen fields (205 × 300 pixels) was quantified by histogram analysis (Adobe Photoshop, San Jose, CA). Data were expressed as signal intensity (signal pixel count divided by total image pixel count).

Six FAs (3 expressing PPFP) and 18 FTCs (8 expressing PPFP) were stained for TIMP-3, CD-31, and VEGF expression. Ten fields per FA and 4 fields per FTC were quantified using pixel color intensity as described above. Studies with the human specimen were approved by the Institutional Review Board.

Gene Expression Arrays

RNA (5 µg) from 3 different passages of NT-vector and NT-PPFP cells was submitted to the Mayo Clinic Microarray Core Facility for quality control and gene expression analysis on human U133plus2.0 chips (Affymetrix, Santa Clara, CA). Data were analyzed with Partek software (St. Louis, MO) using a Bonferroni correction for multiple comparisons with a false discovery rate of <0.05. Only genes exhibiting a ≥2-fold change were considered.

Statistics

Quantitative variables, including cell viability, apoptosis, soft agar colony formation, matrigel invasion, mouse tumor volumes, and immunohistochemical staining intensities were analyzed with a Student 2-sample pairedttest. Xenograft tumors were compared by analysis of variance for repeated measures (SAS 9.1.3, Cary, NC) to assess the effect of PPFP expression. All values withP<0.05 were considered significant.

Footnotes

Acknowledgements

The authors express their appreciation to Dr Shankar and Mr Rakhshan Rohakhtar at the Mayo Clinic Microarray Core Facility. They are grateful to Ms Bridget Hoesley at the Mayo Clinic TACMA Core Facility for her technical assistance with the immunohistochemistry studies. They thank Dr Henry J. Hiddinga for assistance with the animals.

The author(s) declared no potential conflicts of interest with respect to the authorship and/or publication of this article.

This work was supported by a grant from the American Thyroid Association (H.V.R.) and NIH grant CA80117 (N.L.E.).

Figures 1, 3-7 appear in color in the online version of this article.

References

Eberhardt

Grebe

McIver

Reddi

.The role of the PAX8/PPARgamma fusion oncogene in the pathogenesis of follicular thyroid cancer.Mol Cell Endocrinol 2010;321:50-6.

Kroll

Sarraf

Pecciarini

.PAX8-PPARgamma1 fusion oncogene in human thyroid carcinoma.Science 2000;289:1357-60.

DeLellis

.Pathology and genetics of thyroid carcinoma.J Surg Oncol 2006;94:662-9.

Albertson

Collins

McCormick

Gray

.Chromosome aberrations in solid tumors.Nature Genet 2003;34:369-76.

Gandhi

Evdokimova

Nikiforov

.Mechanisms of chromosomal rearrangements in solid tumors: the model of papillary thyroid carcinoma.Mol Cell Endocrinol 2010;321:36-43.

McBride

Wilhelm

Jr. .PAX8-peroxisome proliferator-activated receptor gamma (PPARgamma) disrupts normal PAX8 or PPARgamma transcriptional function and stimulates follicular thyroid cell growth.Endocrinology 2006;147:367-76.

Espadinha

Cavaco

Leite

.PAX8PPARgamma stimulates cell viability and modulates expression of thyroid-specific genes in a human thyroid cell line.Thyroid 2007;17:497-509.

Gregory Powell

Wang

Allard

.The PAX8/PPARgamma fusion oncoprotein transforms immortalized human thyrocytes through a mechanism probably involving wild-type PPARgamma inhibition.Oncogene 2004;23:3634-41.

Lacroix

Mian

Barrier

.PAX8 and peroxisome proliferator-activated receptor gamma 1 gene expression status in benign and malignant thyroid tissues.Eur J Endocrinol 2004;151:367-74.

10.

Nikiforova

Lynch

Biddinger

.RAS point mutations and PAX8-PPAR gamma rearrangement in thyroid tumors: evidence for distinct molecular pathways in thyroid follicular carcinoma.J Clin Endocrinol Metab 2003;88:2318-26.

11.

French

Alexander

Cibas

.Genetic and biological subgroups of low-stage follicular thyroid cancer.Am J Pathol 2003;162:1053-60.

12.

Sahin

Allard

Yates

.PPARgamma staining as a surrogate for PAX8/PPARgamma fusion oncogene expression in follicular neoplasms: clinicopathological correlation and histopathological diagnostic value.J Clin Endocrinol Metab 2005;90:463-8.

13.

Marques

Espadinha

Catarino

.Expression of PAX8-PPAR gamma 1 rearrangements in both follicular thyroid carcinomas and adenomas.J Clin Endocrinol Metab 2002;87:3947-52.

14.

Dwight

Thoppe

Foukakis

.Involvement of the PAX8/peroxisome proliferator-activated receptor gamma rearrangement in follicular thyroid tumors.J Clin Endocrinol Metab 2003;88:4440-5.

15.

Cheung

Messina

Gill

.Detection of the PAX8-PPAR gamma fusion oncogene in both follicular thyroid carcinomas and adenomas.J Clin Endocrinol Metab 2003;88:354-7.

16.

Marques

Espadinha

Frias

.Underexpression of peroxisome proliferator-activated receptor (PPAR)gamma in PAX8/PPARgamma-negative thyroid tumours.Br J Cancer 2004;91:732-8.

17.

Banito

Pinto

Espadinha

Marques

Leite

.Aneuploidy and RAS mutations are mutually exclusive events in the development of well-differentiated thyroid follicular tumours.Clin Endocrinol 2007;67:706-11.

18.

Folkman

.The role of angiogenesis in tumor growth.Sem Cancer Biol 1992;3:65-71.

19.

Wang

Lim

Nagase

Greenspan

.TIMP-3 inhibits the procollagen N-proteinase ADAMTS-2.Biochem J 2006;398:515-9.

20.

Meng

Brew

Woessner

Jr .TIMP-3 binds to sulfated glycosaminoglycans of the extracellular matrix.J Biol Chem 2000;275:31226-32.

21.

Chetty

Lakka

Bhoopathi

Kunigal

Geiss

Rao

.Tissue inhibitor of metalloproteinase 3 suppresses tumor angiogenesis in matrix metalloproteinase 2-down-regulated lung cancer.Cancer Res 2008;68:4736-45.

22.

Ebrahem

Moore

.A novel function for tissue inhibitor of metalloproteinases-3 (TIMP3): inhibition of angiogenesis by blockage of VEGF binding to VEGF receptor-2.Nature Med 2003;9:407-15.

23.

Lacroix

Soria

Bidart

Schlumberger

.Oncogenes et tumeurs de la thyroide.Bull Cancer 2005;92:37-43.

24.

Hanahan

Weinberg

.The hallmarks of cancer.Cell 2000;100:57-70.

25.

Ahuja

Saenz-Robles

Pipas

.SV40 large T antigen targets multiple cellular pathways to elicit cellular transformation.Oncogene 2005;24:7729-45.

26.

Gimbrone

Jr. Leapman

Cotran

Folkman

.Tumor dormancy in vivo by prevention of neovascularization.J Exp Med 1972;136:261-76.

27.

Yang

Liang

.Inhibition of human leukemia xenograft in nude mice by adenovirus-mediated tissue inhibitor of metalloproteinase-3.Leukemia 2006;20:1-8.

28.

Drynda

Quax

Neumann

.Gene transfer of tissue inhibitor of metalloproteinases-3 reverses the inhibitory effects of TNF-alpha on Fas-induced apoptosis in rheumatoid arthritis synovial fibroblasts.J Immunol 2005;174:6524-31.

29.

Anand-Apte

Bao

Smith

.A review of tissue inhibitor of metalloproteinases-3 (TIMP-3) and experimental analysis of its effect on primary tumor growth.Biochem Cell Biol 1996;74:853-62.

30.

Anand-Apte

Pepper

Voest

.Inhibition of angiogenesis by tissue inhibitor of metalloproteinase-3.Invest Ophthalmol Vis Sci 1997;38:817-23.

31.

Burgermeister

Seger

.PPARgamma and MEK interactions in cancer.PPAR Res 2008;2008:309-469.

32.

Kim

Surh

.The role of 15-deoxy-delta(12,14)-prostaglandin J(2), an endogenous ligand of peroxisome proliferator-activated receptor gamma, in tumor angiogenesis.Biochem Pharmacol 2008;76:1544-53.

33.

Peeters

Vigne

Tee

Zhao

Waite

Taylor

.PPAR gamma represses VEGF expression in human endometrial cells: implications for uterine angiogenesis.Angiogenesis 2005;8:373-9.

34.

Lacroix

Lazar

Michiels

.Follicular thyroid tumors with the PAX8-PPARgamma1 rearrangement display characteristic genetic alterations.Am J Pathol 2005;167:223-31.

35.

Brueckl

Grombach

Wein

.Alterations in the tissue inhibitor of metalloproteinase-3 (TIMP-3) are found frequently in human colorectal tumours displaying either microsatellite stability (MSS) or instability (MSI).Cancer Lett 2005;223:137-42.

36.

Ninomiya

Kawakami

Fushida

.Quantitative detection of TIMP-3 promoter hypermethylation and its prognostic significance in esophageal squamous cell carcinoma.Oncol Reports 2008;20:1489-95.

37.

Liu

Tufano

.Association of aberrant methylation of tumor suppressor genes with tumor aggressiveness and BRAF mutation in papillary thyroid cancer.Int J Cancer 2006;119:2322-9.

38.

Ferrara

.VEGF and the quest for tumour angiogenesis factors.Nature Rev Cancer 2002;2:795-803.

39.

Giordano

Kuick

Thomas

Rhodes

Wilhelm

Jr. .Delineation, functional validation, and bioinformatic evaluation of gene expression in thyroid follicular carcinomas with the PAX8-PPARG translocation.Clin Cancer Res 2006;12:1983-93.

40.

Lui

Foukakis

Liden

.Expression profiling reveals a distinct transcription signature in follicular thyroid carcinomas with a PAX8-PPAR(gamma) fusion oncogene.Oncogene 2005;24:1467-76.

41.

Jang

Davies

Kaufman

Wimmer

.Initiation of protein synthesis by internal entry of ribosomes into the 5′ nontranslated region of encephalomyocarditis virus RNA in vivo.J Virol 1989;63:1651-60.

42.

Lemoine

Mayall

Jones

.Characterisation of human thyroid epithelial cells immortalised in vitro by simian virus 40 DNA transfection.Br J Cancer 1989;60:897-903.

43.

Bhattacharyya

Coenen

Bahn

.Thyroid transcription factor-1 in orbital adipose tissues: potential role in orbital thyrotropin receptor expression.Thyroid 2005;15:422-6.

44.

Munaut

Noel

Hougrand

Foidart

Boniver

Deprez

.Vascular endothelial growth factor expression correlates with matrix metalloproteinases MT1-MMP, MMP-2 and MMP-9 in human glioblastomas.Int J Cancer 2003;106:848-55.

45.

Denizot

Lang

.Rapid colorimetric assay for cell growth and survival: modifications to the tetrazolium dye procedure giving improved sensitivity and reliability.J Immunol Meth 1986;89:271-7.

46.

Monahan

Rybak

Raetzman

.The notch target gene Hes1 regulates cell cycle inhibitor expression in the developing pituitary.Endocrinology 2009;150:4386-94.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.10 MB

0.00 MB